jdbc:postgresql://localhost/whatever?schema=foo

The Google indicates people have been asking for this for years!

I was just introduced to them - what a beautiful feature. Now I have to set it myself in the connection factory programmatically. Grrr

I noted that I've done this in the past and always got burned.

"Oh, I'm a good person. I'm a religious jew."

"So was Bernie Madoff."

Pause.

"Ok, maybe I shouldn't have mentioned that..."

"Restore my faith..."

I handed him the cash and a business card. We'll see what happens.

Update 2009-06-13 : Fool me twice, shame on me.

JavaOne has always been my favorite of the commercial conferences, but this JavaOne feels very different. Maybe it's the combination of the economy, the Oracle acquisition, and Java's maturity. Lots of my friends aren't here, and I feel very disconnected. Things feel broken, somehow. I hope it's just me.

This morning's keynote was interesting in a few ways. (I was there a little late due to late night, um, meetings, with Patrick and James....)

First, Sun 'announced' the Java Store, and what was interesting is that it's not only admittedly incomplete - the revenue collection parts aren't done - but the incompleteness was emphasized by Gosling and Schwartz. I was really surprised. Can you imagine Apple announcing 2/3-rds of a product or service? I don't fault them for not being complete (stuff happens), but that shouldn't have been a major talking point at the product announcement. They should have made it a sell to developers, explaining how this idea has legs (I'm doubtful) and getting people excited about getting involved. Instead, I'm thinking "Will these guys even get this done?" and I'm probably not the only one.

Second, it was great to have Scott McNealy on stage one last time. I thought his "JavaOne Japan" joke was a riot (I think there were 3 of us in the room that got it), and I was really surprised to see Larry Ellison there. Ellison said a few things that were interesting. First was the calling out OpenOffice and JavaFX. While I'm not a fan of OO simply due to the endless irritation over the years dealing with documents from people who mean well :) it's clear that there's still a huge potential for it in a cloud/SAAS context, and I'm honestly very interested to see what Oracle is able to do with it. Also, the logical implication of what Larry said is that JavaFX will have to be opened up (something that Simon Phipps has been pointing out for a while, AIUI), and continuing the move to make Java technology truly open (not just Free(tm)) is just great. Next, his discussion of Android and netbooks and such was puzzling. I look at Android (at least the Java runtime part) as the logical result of Sun's irrational perspective on Java ME. (Defending ME revenue is also the reason why Sun still wont' give the ASF a legitimate license to the Java SE TCK). I find the whole idea of Sun embracing Android as cognitively dissonant, unless ... I don't know. I have to think about it.

Finally, I thought that McNealy's speech touching - it must be hard to say goodbye like that. I'm going to miss the Sun that we know as it is now. I really will.

At 2.4M operations (iow, 2.4M documents inserted or updated), wham! CouchDB speed. ;) I clearly have no idea how to tune this.

People seemed to like the talk. Seems that we were fashionable as well (see the comment at the bottom of the page). Who would have guessed...

Whoops. I'm sure no one saw this coming.

(Omitting obvious snark about trying something independent like ASF or FSF next time...)

Do it. Now you have no excuse.

Simon Phipps suggested a "cloud profile" for Java on his blog, something that's been discussed in other places as well (e.g. CCIF).

I certainly think that managed runtime environments like the JVM or the .NET CLR are the way to go for at least a subset of cloud platforms. We've had a bit of this debate on CCIF - I think that standardizing on a higher-level runtime (like a JVM) versus the low-level "enabler" VMs (as in x86 virtualization) is the right way to go - developers have a platform that's consistent, and service providers have large degrees of freedom to innovate on implementation. I realize that this is a problem for people that want to run code that isn't in a JVM language.

I think that there's lots to do in this area. When 10gen was working on PaaS, I longed for a multi-tenant JVM where I could safely partition applications from one another (each w/ it's own heap), yet manage them collectively in the same JVM (rather than running one JVM per app). There's also good open questions on APIs - what APIs can be implemented safely in a shared/cloud environment?

Now, I disagree with Simon a little regarding creating a new "profile" for Java applications. First, I think it's too soon. We're just starting to figure out how this might work. Lets not rush to standardize something that is still so new and developing. The "cloud" space is overheated with hype and promise, and I think that the hard work is just beginning in terms of technology, both platform and application. It's become clear to me that applications for the cloud will be different (except in trivial cases), and we have a lot of learning to do both as developers and technology providers.

Second, if it's not too soon, I have some real worries about this happening at the JCP. Now, I'm not certain what Simon means by "profile". In strict JCP jargon, a "profile" is a "platform" (SE, EE, ME) plus additional specs, all of which must be included in their entirety. I assume that he really means adding a new "platform" - Java CE. So why worried about the JCP? The last thing we need is another spec controlled by Sun that no one is allowed to implement. It happened with ME (and now we have the Android "Java" SDK) and we're fighting over Java SE with Sun's refusal to provide a TCK for the Apache Harmony project. We had the *same* battle with Sun over Java EE for Apache Geronimo a few years ago, but back then Sun had people like Rob Gingell that had both a clue and power to get things done, so it all worked out and Java EE is flourishing.

So what to do? I can think of two things. First, if it has to be done now, and has to be done at the JCP, Sun can't lead it - handing Sun control over another platform would be just stupid. Sun is showing it's willingness to control markets for the benefit of it's own revenue with the SE fight, and I don't think anyone has the energy to fight them for the ability to implement another platform. The solution would be to have Google or the Apache Software Foundation lead it.

Alternatively, if we can avoid the rush to the JCP, lets do a "Cloud Profile" project at Apache Harmony. Harmony has a great codebase (the class library powers Android and the JVM is good w/ JIT, modern GC, etc), a solid and diverse community, works entirely out in the open, and is licensed to allow anyone to reuse as they see fit. Everyone participates as an equal - no "first among peers". The class library was modularized from the beginning, so not only is it already able to be "profiled", but the community has the experience with issues related to modularization.

I'd love to do this, and then bring back what we learn to the JCP at some future time. Or not.

How cool. And based on Apache Hadoop.

And much hilarity ensued...

Me: "Can you explain the cost and details of your monthly laptop broadband plan?"

CanYouHearMeNow: "Yes sir - it's very straightforward. The plan is $69 dollars a month for unlimited use, up to 5GB"

Me: "Er, then it's really not unlimited, is it?"

CanYouHearMeNow: "It is, up until the 5GB cap."

I can't hear him now.

You've got a serious problem when MSFT can call you out on openness and they - despite being MSFT - have a point.

Clearly, "interoperable" isn't the first word that comes to mind when I hear about Silverlight, IE8 and Azure, but it will be interesting to see where this goes.

I never thought of doing this to get the ancillary things a project needs, although to be fair, GitHub has everything but issue tracking, and if I were a betting person, I'd bet that's coming soon.

Love IBM or hate IBM, it will certainly shake up my little corner of the universe, and given the deadlock over the past few years, this is a good thing.

That said, this really saddens me on another dimension. I'm actually an secret admirer of Sun and was hoping to see it pull through. Sun played a significant role in shaping computing as we know it, and continue to provide some of the most colorful characters and interesting ideas in the industry. They still have some truly great engineers, and at least from the outside, a really great engineering culture. I can't see that surviving as part of IBM. (That's not judging IBM as "not great" - but it's different.) The world just won't be the same without the Sun we know today.

Anyway, I saw the thread late, and when I was traveling to London for QCon and jetlagged and tired, so I skipped over Dalibor's response (as I know he understands the facts) and tried to address the comments from others.

Yesterday, someone suggested I read his response, as it represents a really unique change in Sun's positioning of the situation. I think it boils down to the following (my interpretation):

Sun is withholding TCK licenses from those of us doing independent implementations for our own good. Passing the tests is hard, Sun is very concerned about us, and Sun don't want us to waste time and resources.

In the thread, Dalibor says :

So why does Sun restrict the OpenJDK Community TCK to 'substantially derived' implementations? I wasn't at Sun when the license was conceived, but I think the obvious answer is that TCK testing is a lot of (manual) work, even when one is starting from a feature-complete implementation, like OpenJDK 6, and just needs to focus on setup, testing and bug fixing.

I keep reading this over and over, thinking "he didn't really just say that, did he? He must mean something else..." Later on, he says :

So, in a world in which any compatible free software Java SE 6 implementation in the foreseeable future is likely going to be substantially derived from OpenJDK 6 for very simple, rational economic reasons (it works well, it can be done on a rational schedule, etc.) - do you just toss the TCK binaries over the wall, or do you nurture the community of free software, fully compatible Java SE 6 implementations and work on gradually having a sequence of doable success stories, while expanding the clout of the platform? The latter approach works actually much better in practice, as experience shows, and has already delivered several fully compatible, fully free software Java SE 6 implementations.

This is such utter, self-serving nonsense, it's hard to know where to start. Let me throw a few facts into this :

• The only logical reason why "any compatible free software Java SE 6 implementation in the foreseeable future is likely going to be substantially derived from OpenJDK 6" is because Sun won't allow any other free or open implementation to get a TCK license. If people can't get a license for the test kit unless they agree to add Sun's terms to the license under which the implementation is distributed, then of course you won't see any other open implementations that are compatible, because Sun's additional terms aren't open. Without the TCK, you can't test compatibility. If you can't test compatibility, you can't be compatible. Nice how that works out for Sun, doesn't it?
• Sun has made the mistake of underestimating what others are capable of doing before. When we started Apache Harmony, they were very gracious and welcoming publicly, as (so the inside baseball goes), they were utterly convinced internally we had no hope of succeeding. Their perspective changed radically a year later when we showed up at JavaOne with a functional implementation of AWT and Swing, and good progress with a virtual machine w/ a JIT and modern GC. (Ironically, the lore has it that Sun was running the TCK on Harmony to track our progress...) So implying that doing an independent implementation is economically irrational is either conveniently forgetting history, or willfully misdirecting the discussion.

Finally, I'm just astonished that Dalibor of all people - someone I respect, with a very commendable track record of open/free software contribution and leadership - would be directly suggesting that instead of a free ecosystem where people are able to spend their time and effort where they choose (independently implementing a spec and testing it), that it would be better instead to artificially limit work only to [the Sun controlled] OpenJDK codebase and derivatives (why do I keep thinking of Bush's limitations on stem-cell research here?) for the "clout of the platform" (for the good of the State?).

This stinks. The beauty of the Java ecosystem is that there can be multiple implementations of a spec. This has served us so well over the years, I'm surprised that I have to bring this up.

Think about Java EE for a second. (BTW, we had the same knuckle-headed resistance from Sun when the ASF licensed the Java EE TCK for the first time for Geronimo.) The incredible choice we have had over the years - Apache Geronimo, JBoss, Sun's Glassfish, IBM's WebSphere, BEA's WebLogic, Oracle's appserver, JOnAS, etc - gave users options in licensing, support, performance and features. Without sacrificing compatibility (which was Sun's public motivation for the resistance). We've also had beneficial variety in the Java SE space - Sun, IBM and BEA all had competitive, compatible SE implementations. While all licensing Sun's class library implementation, they had independent or semi-independent virtual machines, which helped drive benefits in performance, stability and manageability to a degree that a single-vendor ecosystem would never have realized.

Think of the positive, constructive disruption that Google's Android platform (which is powered in part by Harmony's class library) is bringing to the phone/embedded ecosystem. Smart phones are exciting again.

I hope I'm just misunderstanding what Dalibor is trying to say.

So, it turns out to be a 21st century version of 'carb icing'? Really? This is something that Lindberg and Earheart had to deal with...

My first thought, after growing up trying to keep farm water pumps and carwash systems working in freezing weather, is "wrap heat tape around that sucker", but given it's going to take Rolls Royce 12 months to fix it, that probably isn't the optimal solution.

I'm privileged to chair a track with Simon Wardley on "The Open Cloud".

Starts at 6:45pm. It's free. There will be beer and pizza. No excuses.

Update : this one seems to be a monster - 800 people registered....

MSFT is adding full RDBMS services to the Azure platform - aimed for "department-sized" databases (i.e. things that won't need to scale, since they can't...).

I think this is right - the "scalable" and "alternative" data stores (MongoDB, Voldemort, CouchDB ...) are all IMO an augmentation to the RDBMS, not a replacement. The RDBMS is here, it's very useful, and well understood. But new programming models and new scaling/size/deployment requirements means that we need additional data persistence technologies, and I think we'll see more architectures that synthesize a combination of one or more of these technologies with the 40-year-old workhorse, the RDBMS.

Why not just merge the communities? And they claim the ASF has too much process...

I think it would be quite interesting for both parties. MSFT would do well to have Sam around.

I wonder what they'd say if I brought drumsticks and a small set of cymbals and started making random noise...

For example, someone posted the following on MongoDB site as an anonymous comment :

My understanding (IANAL) of GNU AGPL, and probably that of the first commenter, is that GNU AGPL applies the GPL to applications accessing the server via network protocols, as well as to applications linking the libraries. That's the reason for the GNU GPL or GNU LGPL suggestion, so as to not force all applications that communicate with the server over the network to also be GNU GPL licensed (there by preventing most closed sourced or most commercial applications).

Clearly this is not the intent of the AGPL (although we shouldn't give Stallman and Moglin any ideas...). While the AGPL clearly exhibits the purest known form of FSF-style "Freedom(tm)" (more than the GPL or that fifth-column of FSF licenses, the LGPL), and it does capture the "All your modifications are belong to ... others" spirit of Freedom(tm), it clearly can't force a re-licencing or limitation upon code that touches it.

That would just be silly, and we all know that nothing silly ever happens with open source licensing.

Today, Sun sent me the same mail. Only difference is that the mails end with "Pthbbbbbbbbbb". I guess they wanted to rub it in.

Alas.

From my experience with our NIO-based Java drivers, I know that limiting memory churn when serializing the data to the wire format is critical, so that's the first place I looked. I didn't really like what I saw - a key array was allowed to resize on each write - but as I don't really know anything about Ruby, I wanted to prove that reusing a pre-allocated array was faster.

To make a long story short, I've run into the following mystery, and would appreciate any hint I can get. This "behavior" happens on Ruby 1.8.6 on both OS X and Ubuntu.

require 'benchmark'

class BufferContainer

 def initialize(initial_data)
   @buf=initial_data
 end

 def put_array(array)
   @buf[0, array.length] = array
 end
end

size = 1000
thearray = [0, 1, 2, 3, 4]

for i in (1..3)
 size *= 10
 a = Array.new(size)
# RUN WITH THIS COMMENTED OUT FIRST
# a[size] = 1
#

 puts "size = #{size} #{a.length}"
 puts  Benchmark.measure {

      10000.times {
        buf = BufferContainer.new(a)
        buf.put_array(thearray)
      }
  }
end

So what's the problem? For three different sizes of a pre-allocated destination array, I'm doing range copies, always of the same size, always at the same location.

   @buf[0, array.length] = array

Everything is pre-allocated, so there should be no reason to create garbage. However when I run this, I see :

size = 10000 10000
  0.050000   0.000000   0.050000 (  0.053745)
size = 100000 100000
  0.400000   0.000000   0.400000 (  0.409549)
size = 1000000 1000000
 17.770000   0.130000  17.900000 ( 18.338888)

What is this saying? For 10,000 iterations of the range copy, if the pre-allocated destination array is sized to 10,000, it takes, 0.05 seconds, if sized to 100000 it takes 0.4 seconds (an order of magnitude bigger), and if 1,000,000, then 18 seconds. (!)

My expectation was that it as constant in time - no allocations or memory moves should be being made - it should be a simple copy. I spent a lot of time trying to boil down this testcase, then play with it. I think this is entirely due to some GC issue (my primary theory), or lazy array creation, or something. I feel kind of stupid that I spent the time I did, but fixing this driver is important, I'm not a Rubyist so I don't know who to ask (The Google was no help), and I really hate a mystery.

On a hunch, I added the line in the code that is currently commented it out. (a[size] = 1), as I figured that if it was some weird lazy creation issue, this would force it to create the array.

Uncomment that line see what happens.

I'll dig up the Ruby source tomorrow, but if someone wants to save a Reluctant Rubyist the time....

Update : 2009-02-27 : Seems to be a bug in the implementation of Array, as it's reported gone in v1.9.1 of Ruby. It's also not in JRuby, but that's not a surprise since that's a re-write. I'm honestly surprised that no one has seen this before, or if they have, it wasn't fixed in 1.8.x

Alas.

This will force me to learn Erlang. Then I can hang with the cool kids. Maybe.

I've been working with PHP from the POV of making sure that this impl is tested in our Hudson CI farm, and while I find it somewhat baroque to use (e.g. I have to edit /etc/php.ini as root to add the module....), it's cool that we have this and I'm looking forward to see what PHP people think.

This gives them a way to leverage a different kind of data store either instead of, or more likely, along side of their RDBMS.

Whee....

This release is really a baseline release for us at 10gen after we re-focused the company on Jan 1 to the persistence layer of our appserver stack. The appserver will continue as an Apache Licensed project (http://www.babbleapp.org.

This release contains :

• The MongoDB database (of course)
• A new, slick, Google V8-based command-line client that lets you interact with the database in JavaScript.
• Basic tools like import/export, backup/restore.
• Drivers for Java, Ruby, Python and C++, with PHP probably going to be made available (although not yet releasable) today.
• A RoR ActiveRecord Connector.
• An implementation of the ActiveRecord pattern in Ruby (not to be confused with the RoR AR component).

So this think-of-it-not-as-a-database database has some interesting properties. It stores JSON-like documents. I say "JSON-like" because rather than just strings, numbers and booleans, it can store other types like dates, binary data, and distinguish between integer and floating point numbers. It's pretty quick - on my mac laptop, I can do 300k inserts/sec from a Java client (doing them in small blocks of 100 documents per network message), and random reads at about 30k/sec. (Awake readers will note that I'm not transactionally persisting that much data to disk at that rate... disks don't go that fast... a subject for another post). I can do fancy indexing on the "documents" - not just primary, but also index into sub-objects. E.G., f I have a document that in JSON would be structurally represented as :

{
   foo : {
       bar : ....,
       woogie : ....
   },
   x : ....
}

I can create indexes on things like foo.woogie. I can have multiple indexes per collection (think of a collection like a table).

It also has a rich query language that lets you do a lot of the things that you'd expect when coming from a SQL background, and lets you express those queries in a way that is compatible with thinking in the document structure you're working in (in JS notation with the "what I think about in SQL" above it in the comment):

 //  select * from mycollection where foo.bar == 10
  db.mycollection.find({ foo.bar : 10});
  //  select x from mycollection where foo.bar == 10 skip 10 limit 10 order by foo.woogie
  db.mycollection.find({foo.bar : 10}, {x:1}).skip(10).limit(10).order({foo.woogie:1});

Where the first example lets you find all documents in the mycollection collection where the value of bar of the foo element is 10. The second example goes further, skipping the first 10 elements, only returning 10 elements, ordering by the woogie subfield of foo, and limiting the return to partial documents that only contain the x field.

Also, you can do document updates - rather than replacing the whole document if you want to modify it (which is a horror show if you have large documents), you can just update elements of the document in-place :

// update mycollection set total = 10 where id = 12345
db.mycollection.update({id:12345}, {$set:{total:10}});

MongoDB also has some nice replication and semi-HA master pairing features, and sharding is on the way.

What's it good for? Well, as I argue when people give me the chance to speak about it, databases are changing - just look at what is available in the so-called "cloud" arena. It tends not to be a RDBMS if it's scalable. The storage engine under AppEngine, or Amazon's SimpleDB, or any of the Dynamo implementations, etc, all of which change your programming model to one that isn't "tables and joins". Or look at the excellent CouchDB, a JSON store. If the RDBMS isn't being replaced outright (like it has to be in "the cloud"), it can to be augmented with other persistence technologies that are better suited for a portion of the data requirements of a system.

So what's it good for? It works fine as a database, but you can't think relational. If you want to just replace MySQL with something else, but don't want to rethink your data model, MongoDB isn't for you. Because of it's pedigree and initial design requirements, it works very well as an "object" store for dynamic languages. JS objects, Python and Ruby hashes all go in and out very effortlessly :

db.mycollection.save({a:10, b:2});

We've had it supporting news-ish/blog-ish websites in production for a year now, and it does fine there. It does fine as a large object store - think big binary blobs here, like images and videos. We have a POC in progress where we leverage the server-side JS execution feature to provide transaction-like isolation for high-performance shopping-cart/inventory management. (4k a second at last check on a mac desktop). It has some interesting potential as a persistent cache - one where you aren't afraid to restart the cache for fear of the hammering the backend data store will receive.

I think that this DB has a lot of potential, and I look forward to seeing what other kinds of problems it can solve. Download it and try it. We have it available for OS X 32-bit and 64-bit, Linux 32-bit and 64-bit, Windows 32-bit, Solaris 32-bit. Let us know what you think.

http://www.mongodb.org

• Battery life isn't as bad as I saw that fateful day last week. Keeping GPS off seems to give me a full day of use, which is enough for me.
• The included mail IMAP client has lots of "opportunity for improvement". I've switched to K-9, a fork of the original Google code, which has some nice features. Still, after the iPhone, there's lots of work still to be done here.
• I can't find a way to read a PDF. The iPhone spoiled me, and now I think of this as table stakes.
• I miss my newsgator reader from the iPhone. Typing in feed URLs directly on the phone is no fun. Also, the RSS reader market is wide-open. Yes, I know I can use Google Reader, but the key for RSS on a phone is being able to read when offline - like on a plane or subway.

I'm really hoping to find some time soon to play w/ the SDK...

I spun up an instance of Ubuntu 8.10 32-bit (AMI from http://alestic.com/ - recommended...), installed the JDK, downloaded and installed MongoDB, ran it. Wrote email to user. Spun down instance.

Took about 20 minutes all told.

Definitely worth the $0.10 that AMZN will bill us.

...and seeing your Python, I know you wield a colon w/ the best of them...

Perfectly innocent. Nothing to see here. Move along.

I'm a long-time "Sun Kremlinologist", mostly driven by my relationship with Sun on behalf of the Apache Software Foundation. I'm also a Java Weenie(TM), and for the longest time, Sun (NASDAQ:---->JAVA<----) was the center of that universe. (The relatively recent ticker change seemed like a futile grasp for the old glories.) I've made a lot of really good friends at Sun over the years, and thus have been watching with admittedly hopeful eyes that Sun might pull itself out of the "nose down, engines off" flight path that it's on ("Saves energy! Quieter!"). Whether you love Sun or hate Sun, the industry just wouldn't be the same with out them. I mean, they've been responsible for some really great things (Java, for example) and recently, quite a bit of entertainment - who could you rely on for such constant hilarity like their broken IP policy or "Java FX". (Seriously? You want to take on Adobe *and* Microsoft at the same time in a mature market?)

Anyway, their recent round of layoffs was personally distressing. The title of this piece comes from an IM conversation I had with one of the victims - I was trying to figure out the thinking behind the layoffs, and the best I could come up with was "First, get rid of the good people", as it seems that they let go a lot of good older, senior people, people who knew the industry, had the relationships, got things done, etc.

So, like the "Stray Sunbeam" series being done by Tim Bray, I'd like to mention some of the people I know and suggest that if you are looking for solid people who get things done, you hire them. My POV is biased - these are friends. (Some I don't know, but mentioning anyway). Also, they don't know I'm doing this, and some will probably be mad at me. Now that Cheney's bunker is empty, maybe I can use that if they come after me.

• Onno Kluyt : My good friend and long-time nemesis on the JCP, Onno and I have developed our personal and professional relationships to the point where we could literally be screaming at each other over Sun's utterly foolish and destructive passive-aggressive patent policies (who ever guessed that the "Participation Age" would require a patent license?) and still sit down over dinner moments later. He's a seasoned manager and corporate politician, and I have no idea what he wants to do next. He's also brutally honest - when I first engaged with Joost, and was visiting the Netherlands, I asked him what to look for in the local food, as I like to eat local when I travel. His response? "Have you ever seen a Dutch restaurant outside of Holland?" (If you couldn't tell by the unpronounceable name, he's Dutch).
• Sara Dornsife : Another good friend, I met Sara during the open source Java wars - she was doing community and developer marketing for... ok, it never was clear - Sun's dev and community marketing strategery never really was that obvious ("Ensure that we do what we can to ostracize the most popular Java IDE - Eclipse - and fracture the market with our own because we can put our logo on it.") She's very versatile and execution oriented, is able to connect with people, understands developers and is able to get things done. I guess the best way to sum it up is that she understands that when trying to get some communities together, it sometimes requires a tequilla shot.
• Ray Gans : Yet another person I consider a friend, Ray was responsible for a lot of activity around the open-sourcing of their implementation of the Java SE spec. (i.e "Open"JDK). Despite the fact that I was seen as a "dangerous enemy" for helping put together Apache Harmony, an open source implementation of the same spec that gave Sun the push to do OpenJDK (see. "NIH, fear of, e.g. Eclipse"), he had the wisdom and the open mind to solicit my input into how they should OSS Java (Ray, you picked the wrong license...), graciously invited me to attend the OpenJDK launch, and since then has always ensure that I was welcome to hang with the Sun crowd at conferences and similar events. I think he has a wealth of experience in many areas, but his work in OpenJDK - which I do think is a milestone in open source history - is valuable to anyone either wanting to open a living codebase, or engage in a serious way with open source. He was a corporate guy - not an open source partisan - who jumped into an alien world, and he did very well.

The gPhone certainly does, or at least gobble it. Yesterday, the phone didn't make it to noon before nearly expiring with a low battery. Far worse than the iPhone, which I thought was pretty bad. I was out of the office, wandering about Brooklyn for a meeting, and maybe the WiFi searching was the problem. I also remember using the GPS in maps to be sure I was wandering the right way, and left it on that app. I wonder if that did it - I've done that before w/ iPhone when letting it track current position.

I've turned off the wireless this morning when leaving the house - I'll see if that helps.

Maybe I'm dumb and I just can't find it.

Also, AT&T is going to make this difficult. I put the iPhone SIM in and telephony and SMS work, but no data services. I went to AT&T store, and they said that the iPhone data plan is different than the PDA data plan. They cost the same. The services are the same. But they are different (I suppose so Steve J can get his tithe). I'll call AT&T Central Services and see if I can work something out. Until then, only get email on the thing when I'm at home.

Re the email client, I'm not convinced it works right wrt IMAP.... Things don't seem to be in sync.

I also forgot to turn comments on to yesterdays, so that's fixed now.

Update 1 : CrazyBob reminded me that ironically, I failed to turn on comments today. /me wanders off to see if I can change a default.

Update 2 : CrazyBob pointed me at http://groups.google.com/group/android-developers/browse_thread/thread/8fa165d1d2a5f1f0 to fix the data service problem. It works. :) Thanks Bob

First, thank you Chris.

Second, thank you Chris.

Third, it's amazingly fast (remember, I've been using an iPhone for the last year+). I'll need to ping Dan and tell him how impressive that is.

Fourth, it has this really nice "geek aesthetic" that you don't find on the very slick, very polished iPhone, and that's a compliment.

Fifth... this gave me pause - the phone doesn't appear to support any bluetooth profiles? I pine for the days of my T68i, where I could pair with my mac, effectively turning the mac into an extension of the phone - I could initiate and answer calls, send and receive SMS, etc. I really, really miss that integration, and wonder why I can't do it on the GPhone. Something to figure out.

The SDK is downloading as I write this - this should be fun.

(P.S. Chris, thanks again!)

Deep breath.

Sprint - yes sprint - from security to gate 36. In the new AA terminal at JFK it's probably close to 1/3 of mile. Sprinting. Roll-aboard in one hand, laptop bag in the other. Down the 4 story escalator. Under the tarmac. Up the other 4 story escalator. Gate 43, 42, 41..... I'm there! Run up to the door, ticket in hand.

"Sir, we're not boarding yet..."

Another deep breath...

Think semi-murderous thoughts. Ok, no prob. Eventually board. Sit. Do work. Plane loads. Announcement - we need to de-ice. Good plan. Announcement - ovens don't work, sending for mechanics. Good plan - need the warmed mixed nuts. Announcement - still delayed, waiting for rollback. Ok, why is the ground crew surprised that the plane they just loaded needs to leave? Announcement - waiting for de-ice. Still a good plan, but one that should get executed on rather than discussed....

8:04pm... heading towards the runway...

If that wasn't enough to get you excited, it apparently is production code from LinkedIn.

Have fun. I have a few plane rides in the next week, so that's one thing I plan to do...

Am I simply missing it? I don't think so - I just played with the new AWS management console (very nice!) - and there no UI slot or capability for adding information like that.

If someone wanted to parody the subject, this wouldn't have been a bad attempt.

It started out as a project just to prove I understood the wire protocol (I was trying to write the spec for something we wanted to do...). The protocol is fairly simple, so the driver comes out to be simple. I did add a few things, like a "dynamic" version to make it easy for people using dynamic or -ish languages on the JVM to use it (e.g. from Clojure, Scala, etc)

Tonight I decided to futz with it a little, and switched the I/O from Socket to NIO's SocketChannel, and then decided to see how much speed I could get if I switched to "direct" allocation of the ByteBuffers. (Basically, the storage for the buffer is from outside of the JVM heap, so that I/O can happen much faster since it doesn't have to be copied out of object on the JVM heap.)

Performance tripled. Instead of being able to write 8k objects per second to the database, I could write 28k objects per second to the database. Fantastic!

I thought, now what can I do - I now where I can get some further optimizations by avoid non-direct buffer allocations, and just....

The folly struck me. I'm guessing that a middle of the road server is going to let me do 250-500 flushes/second to disk..... maybe more, but how much more?

Other than bragging rights, for what little that's worth, since there's wasn't much effort required to do this, I now have to deal with added complexity of the software going forward. I may roll the change back :)

I had all sorts of Big Plans for what I was going to get done over the break. Not much of it happened, and in retrospect, that's just fine with me.

I did finally get some time to play with CouchDB as I'm interested in all things alternative database. Lots of interesting things came from that - I'll post this week.

Also, comments are currently broken. The kind folks at the haus promised to fix.

As I ponder my options for installing Java, it strikes me that in 24 hours, it will be 2009, and you *still* can't get Java 5 as open source or Free(tm) software.

And it's not for lack of trying - a lot of us put our money where our mouth is and tried to make it happen. And we did it.

If only NASDAQ:JAVA had a tenable business model, I wouldn't have to be pushing a big tarball over the wire, but could just do something like

$ apt-get install javase5-harmony

This is our first officially supported language after JavaScript and it's a lot of fun. Beside making "10gen style" apps but in ruby, you can also write or run CGI apps (not really sure why you'd want to, but you can...) as well as run RubyOnRails. Now, there's a catch to the last one - we don't have ActiveRecord support. ActiveRecord is SQL-tainted through and through, and our database Mongo doesn't do SQL. However, we're working on it, and since it's open source, help is welcome.

Go download and play with it and let us know what you think.

I used to cross 100k sometime in June or July.

This year, it will be in December. Latter half.

The difference between me and those other people is that my trip has a point - I'm out of See's Candy, so need to go to SFO to get some more :) 1lb Nuts and Chews. No need for giftwrap. (Actually, there is a business meeting, but still... it feels like I'm going to fly 5k miles for a box of candy...)

Anyway, I just ran into git4idea, a plugin for Intellij IDEA. It's an available plugin in the plugin system. Yay!

While I thought I understood how scary these things can be for "civilians", it wasn't until I got to participate in this first-hand that I really grokked the problem. The only bright spot is that my father doesn't use his computer for much - primarily web and email, and I think he'd do more with photos if his photo software wasn't so bad. (iPhoto to the rescue I hope...)

So on Black Friday, both because he was visiting for Thanksgiving and thus opportunistic as well as ... well, don't all geeks want to do this? ... we went to the Apple Store at 7am. (I wanted to go at 6am for full additional bonus geek points, but we both weren't ready...). My hope was that he'd get a laptop and a big screen and keyboard (his sight at 75 isn't what it once was), with the intention that he could take the laptop and all things digital with him when he went to the Apple Store for help or classes. However, he wasn't convinced that he wanted the laptop - he didn't think he'd be able to see it well, and figured the expense wasn't worth it.

Well, the LCD screen in the new 15" MacBook Pro is gorgeous and he can read it, and coupled with the zoom feature and easy font adjustment (open-apple-+), he has no problems. So we got that, an airport express, and a wireless printer. (It was $18 after rebate....). I took the pile home, got it setup so that all he had to do was just plug things in. Connect the Express to his DSL modem, power up the printer... what could possibly go wrong?

Well, what went wrong was that unlike the cable modem that I've had for what feels like 15 years now, where I just plug a router into it, configure it for DHCP, and away we go, Verizon's DSL does some seemingly awful things, requiring flipping the modem into ethernet bridge mode and configuring the AirPort express to use PPPoE.

Really? PPPoE? I faintly remember using that over dialup in the early 90's....

And as a bonus, the DSL modem that my father has isn't listed as supported by Verizon's own support site! Since I was guiding a neophyte over the phone, there was *no way* I as going to try this blind.

I really understand why people are afraid of new computer technology. It wasn't just this DSL issue - even the basics that I take for granted, like minor config changes and such via SystemPrefs or whatever are utterly out of the realm of experience for someone who isn't computer literate. I'd love to see a common UI for things like this, just like cars have a common UI for the basics. I understand the advanced stuff is harder, but just the simple things....

Anyway, he was a trouper and between schlepping between the local apple store and dealing with Verizon's phone service, it all got worked out, and he's now online. I'm sure there's going to be quite a bit of niggling things to solve, but I think the journey will be worth it. (E.g. This morning, Mail.app kept the Send button grayed out because there was no outbound SMTP server selected for the account even though there was one defined - this had me for a while, but eventually googled my way out of it and talked him through it... would be nice to see a hint in a tooltip when you hover...)

BTW, I like the new LED screen so much, I got a MBP for myself - I'm a little afraid of the reflectivity of the new glass screen, but so far, there's a lot to like. Plus, the kbd on the MBP is the same as my desktop, which helps as I switch back and forth.

Works great - shows me map of each step of the trip. It will even track position in real time using the GPS (merrily burning up the battery, but hey...)

So what I really need is this app to speak to me so I don't have to squint at it in the same way a regular car GPS does - "turn right in 500 feet at Hooha Street" or whatever.

With strategery like this, anyone still wondering why Sun's market cap is now probably darn close to or less than the value of their real estate and cash on hand?

Just do it. No excuse not to. Every vote is needed, and every vote counts.

Stanley Ho, the now-ex spec lead for JSR-277, is leaving Sun to go to RIM. (Good luck!).

He's transferred the spec lead mantle to Alex Buckley, who also is the spec lead for JSR-294 (Improved Modularity), JSR-308 (Annotations), JSR-908 (Language Spec), JSR-924 (VM Spec).

Alex probably deserves a raise. JSRs take a lot of work.

Makes for easy java <-> JS interop for the basics. Jython (as of 2.5) and JRuby can both do it...

I mean, these interfaces are standard workhorses, having been around since Java 1.2...

grumble...

"It's stupidity. It's worse than stupidity: it's a marketing hype campaign,"

I think he's actually right about the marketing hype - there's way too much of it. But he's wrong in claiming there's something fundamentally bad about cloud computing. According to the article :

The 55-year-old New Yorker said that computer users should be keen to keep their information in their own hands, rather than hand it over to a third party.

This is true, but it's a general statement about any computing system or technology. I think that notions of privacy and user control aren't intrinsically at odds with the big spectrum of technologies that are called 'cloud computing'. Rather, like any other computing technology used by humans, there are options, and we can choose to use and create tech that is secure and open, both in the implementation (as in open source), but maybe more importantly in terms of portability and data freedom, being able to move one's data to where one chooses.

I use a hosted IMAP provider for email - that's a kind of cloud computing. I have the freedom to move my data - my mail - to another service provider or simply run an IMAP server myself and host it myself. I have this freedom because IMAP is an open standard, one that anyone can implement. I'm no more interested in hosting it myself than I am in making my own butter. I'm pretty sure I can do it, but why bother when cheaper and more convenient alternatives are there for me?

We'll see the same with other cloud technology. At 10gen, where I work, we've open-sourced (is that a verb?) our platform. No lock in. Nada. Not so stupid.

It's using the Apache Harmony class library, so I'm happy to have played a small part in seeing this come to light.

Maybe Google will send me a phone :)

Every morning, I take the subway from GCT to 23rd St in Manhattan, which is near the NY Police Academy. Over the last year and a half that I've been working again in NYC, I've noticed the young men and women police cadets as they too ride the same subway, exiting at 23rd St. If you pay attention, you can see gentle change. First, the are just in their uniforms, maybe with their "New York Police Academy" duffel bags. Over time they eventually get a nightstick (or whatever they call them these days.. maybe a RMASIRCD ? "Rigid Mass for Alleged Suspect Influence via Rapid Contact Deceleration") which is neatly attached to the side of their NYPDA duffel with velcro straps clearly designed for such purpose. Very unthreatening. Then come the patrol pads, large, leather covered pads that they are required to carry in their right back pocket, which seems specifically tailored for this purpose.

After the pads come "the book", this *massive* binder which I can only assume is the penal code applicable to street policing. It's huge. Dragging these around, you'd think they were Marine recruits undergoing some sadistic boot-camp routine at Parris Island. Obvious jokes about "throwing the book at 'em" come to mind... Then comes the "utility belt" - gunless of course - that gets attached to the duffel bag strap. And so forth.

A few weeks ago, the cadets disappeared. Budget cuts? Shipped off to Iraq under some "It's not a draft.. really... they willingly sign their stop-loss orders..."-type Pentagon program? Some silly DHS drill harassing the aged and infirm at airport security checkpoints? Graduation? :)

This morning, coming down the stairs to the downtown subway platform, I noticed suits. Now, at 6:30am, there aren't a lot of suits riding the 6 downtown from GCT, just a lot of construction workers, software engineers, and other blue-collar types, so the density of suits were noticeable. And as I looked I noticed that they were somewhat I'll fitting and wrinkled - like that suit you had in high school that fit for the wedding you had to go to in your Junior year, but didn't work so well in your Senior year. And everyone had a duffel bag, but all different. And I caught them eyeing each other, but without real recognition. Well dressed terrorists, pretending to not know each other?

I stood next to one on the subway, and asked him :

"Excuse me..."
"Sir?"

"Sir?"? This is NYC. Who responds with "Sir?" in a NYC subway other than tourists? Did I run into a missionary group? Maybe the terrorist theory has some merit...

"Well, I couldn't help noticing the suits..."
"Yes sir"
"And it can't be a coincidence that a multi-cultural group of young people in somewhat ill-fitting and uncomfortable looking suits, each with a random duffel bag of similar size and contents - judging from the shape - happen to converge on the downtown 6 platform at 6:30 in the morning by sheer coincidence."
"No sir...."

This was the new academy class. They hadn't been issued their uniforms, and clearly they hadn't been instructed to dress to blend in. I spent a few more minutes talking to the young man - what his hopes are (working with the Transit Authority), if he's afraid ("Well, NYC is better than it used to be...") and how to cope with the danger ("You just have to always be aware"... "Isn't that tiring?" ... "I guess it will be..."). I couldn't imagine doing what he's doing.

I didn't get his name, and I doubt he'll ever read this. (I mean, even my friends don't read this...) But either way, to that young man and the rest like him on that train and trains to follow - thanks. You're brave and we're glad you're here. I'm sure you'll make a fine officer and I wish you the very best in your new career.

And yes, I'll be watching for the bag, the pad, the stick the belt and the book. And once again, I'll make the "throw the book at 'em" joke. And I suspect I'll hear a very polite "Yes sir."

(We managed to tie the standing FOO Werewolf record. End of sat was 6:15am. Never trust Jane after 5am.... Imagine you find yourself in a world separated by time and space from your own, where you are a humble villager, seer-less, healer-less, werewolf-less... gets confusing awfully fast....)

To be clear, Sun = FOSS. We have transformed our company and aligned it around the belief that giving away our technology and investing in related communities will create greater adoption of our intellectual property and ultimately redound to the benefit of our shareholders, customers and the open source community. When it comes to Sun's commitment to open source - "the horse is out of the barn". Not only that, it's also had foals.

Be careful where you step. The horse seems to have left other things around. As long as it can remain hidden behind Sun's NDA requirements and the FSF peeps continue to remain complacent, maybe you won't smell it either.

Hopefully they won't destroy Java in the process just to hit their numbers for a quarter or two.

RedHat just announced a program where you can get JBoss App server (free as in beer software) running on an amazon EC2 instance for

"$119/month per customer plus $1.21 per hour for every deployed server, plus additional bandwidth and storage fees."

Putting aside the $119 and looking at published Amazon EC2 pricing, the $1.21/hr is a 50% markup on the $0.80/hour "Extra large instance", and a 1100% markup on the $0.10/hour "Small Instance", which I'd argue is enough for plenty of deployments. The $119 adds an additional $0.16 an hour.

Granted, you do save the costs of hiring someone to make a JBoss AMI for you, but still... amazing.

I guess it's now "was the voice of". John, we'll miss you.

"...but for the grace...", I suppose.

Until now I naively believed that self-destructive doomsday machines were fictional devices found only in James Bond movies. I never believed that anyone would actually create and activate one in real life. I guess I never knew about Yang and the Yahoo Board.

I can never get time back. I can always make another $20.

I've been meaning to announce that I've joined a new startup, 10gen. I'll write more about the change later. The upshot of 10gen should be well explained on the site (and if not, let me know so I can fix it), but in brief, we're developing a "platform" for what is very loosely called "cloud computing". I say "loosely", because the term encompasses a very broad range of technology and services, from basic grid computing to AWS to AppEngine to SalesForce to EngineYard to ... I'm working with some very smart and accomplished people (ignore the guy at the bottom of the page) and I'm lucky - we appear to all share the same views on code, agility, process, technology, product, etc.

What we're building is ambitious - an application server and a database that are designed for infinite scalability. In addition, we're building developer tooling, application libraries, and management infrastructure to help people develop, debug and deploy applications to said scalable application platform. A similar example is Google's AppEngine - it's an app server with a kind of database to which you can write applications, and those apps will run across the Google's resource grid. The 10gen platform is an app server with an object database (keep reading :), and this platform will run across a variety of resource grids - you can run it on Amazon's AWS, across standard managed hosting providers like Rackspace etc (even multiple ones, for geo and vendor diversity), or even run on your internal computing resources.

A key feature of our platform is that we're multi-language. We're focused on Javascript and Ruby at the moment for writing applications, but we have no religious or political convictions here - we're interested in supporting languages that people want to work on. (We are also dedicated to making such languages and related frameworks run fast.) The constraint is that we think that to get internet scalability, standard programming models have to change a little - you can't just toss a LAMP stack or Java EE server onto AWS and expect infinite scalability. While the benefits of AWS are clear - zero capex for infrastructure, dynamic resource availability, reduced operations personnel costs - you still have whatever scalability limitations you started with in your LAMP stack or Java EE server. Don't get me wrong - this isn't a ding on LAMP, EE, or AWS but just a recognition of the challenges we're all facing with our standard tools.

On the database side, yes Virginia, it's an object database. When I first heard this, I had what I imagine is the standard reaction by people that have only used RDBMSs - "huh?". But after playing with it, and thinking about the problem space, I'm convinced that this kind of database architecture is not only nice, but required. Clearly Google and Amazon think so too as their data stores are either object stores or tuple stores. Sure, you can run a RDBMS on at Amazon, but that's just as scalable as your current config. It's clear that an ODBMS isn't going to be the right database for all applications, but I think that what we have is very "fit for purpose", and the RDBMS isn't the right database for all applications either. (Think of how much time we all spend as programmers trying to deal with the RDBMS in a sane way via JPA, Hibernate, iBatis, JDO, Linq, ActiveRecord, ActiveTable, Django, SQLAlchemy, Storm, DataMapper, DataXtend, etc)

Anyway, we're looking for really good people to join the team. We're focused on hiring in our New York office on 20th Street in what is colloquially known as "Silicon Alley", but I'll consider other arrangements. We have some really big and interesting problems to solve, and we need people of all experience levels and backgrounds. The app server is written in Java, and the database in C++. There are all sorts of scalability, grid and management issues to solve. Our application libraries and frameworks are written in Java, Javascript and Ruby. We want to build tooling for Eclipse. We're going to be open sourcing major parts of our codebase. I need core appserver and db engineers, application library and framework engineers, Ruby and Ruby on Rails engineers, Eclipse plugin engineers, QA engineers, QA technicians, technical writers, community leaders and managers, developer relationship managers, etc. The list goes one. I could formally list the job specs, but I think that if you've read this far, our time would be better spent talking. Just drop me a mail at geir at 10gen dot com. If you have my phone number, call me.

This is an exhilarating, terrifying space with exhilarating, terrifying problems to solve, and we have an opportunity to make a real impact. Life is too short to be bored :)

Divorce long running queries. Ruby on Rails typically serves one request at a time and that’s a detriment if there are long running queries that block other incoming requests, said Zygmuntowicz. The workaround is that you have to set up a queuing mechanism so your database isn’t overloaded.

Serving one request at a time is a detriment. Period.

One simple mechanism for queueing requests would be SMTP - have the Rails framework send request data via SMTP and then fetch the next request when the (!) thread is free. I'd recommend IMAP for that - better chance of not losing requests when the server crashes. Other benefits include being able to use Mail.app or Thunderbird in your ops center to gauge real time load - just check the size of your inbox. You could also ditch Google Analytics or Omniture and use something like MarkMail for traffic analytics. You could also let people mail your website directly - save them from having to fire up Firefox.

1. A horse
2. A very fast catamaran
3. Amtrak regional service with 2 breakdowns
4. A MD-88 when operated by Delta Shuttle

I'm guessing he won't mind - it will support his conference and T-shirt business.

However, I want to lock branches once released. We announce to the team whenever we release something, but people make mistakes and accidentally add things to released branches.

Linus mentioned adding a very practical solution - letting me mark a branch as "readonly" in the .git/config file - but I don't think that really works I'd want to protect it in our central repository.

Could I just do something like make gitroot/thing/.git/refs/heads/version read-only? Anyone have any "best known method" for this?

(Taken w/ Elmo tethered to one hand, iPhone in the other)

]]> Travel geir 2008-05-25T14:02:20-05:00 http://blogs.codehaus.org/people/geir/archives/001695_ecma_doc.html Erm. Why are the membership documents for ECMA in .doc format?

But developing and maintaining a good TCK is a huge task, and one few open-source projects are accustomed to.

I think that Eclipse has 30k+ tests. Apache Harmony has a massive pile. GNU Classpath has something like 20k. There are tons in Apache Commons. Clearly it can be done. A TCK is really just a thorough (and if you've used them, not always so thorough) pile of unit tests :)

While I'm confident that Sun still has enough remaining technical chops to deliver the core technology - some of the smartest people I know work there, and work on this - I think that Sun needs to modify it's DNA and get people that not only understand how to market to the development and design community, but also create tooling for designers as Adobe (clearly) and Microsoft (to some degree) - the two companies that Sun has decided to take on, head on - have at least a decade head start on them. Hint I - this will require investing heavily now, rather than trying to limp by on the cheap. Hint II - another walled-garden OSS community ain't gonna cut it because the best OSS tooling are tools that developers built for themselves. (Eclipse, NetBeans (sorta), gcc, ant, etc...). Designers don't build these kinds of tools for themselves.

This reminds me - given the sheer number and quality of Java engineering defections (I've lost track of the world-class rockstars that are just at Google, let alone Azul, etc)... I worry about the effect this will have on Sun's ability to deliver the next rev of Java SE...

Patrick, Danny, Calinel :

Given that fact that the statements contained in

    http://www.regdeveloper.co.uk/2008/05/14/jcp_individual_representation/

are given by a Sun employee identifying himself in his job role, can I assume that Sun is interested in taking this discussion public? I think that is a really healthy approach. I think there is confusion about the basic facts and I think clarification will be useful for the community as a whole.

I think I'll wait to see what Sun's intention is here before addressing some of the problems in the article. After all, it could be a just a huge misunderstanding. Why do I care? Because openness, transparency and the equitable "rule of law" is inherent in the ASF's struggle in securing an equitable Java SE TCK license from Sun.

Hopefully Sun will allow me to publish their answer. Not being able to would be supportive of "A culture favoring closed-door meetings" :)

The UI is half-screen of keyboard, and half-screen of terminal window. What's interesting is seeing how they are learning how to leverage the touch features of the screen. A terminal using the iPhone kbd is a little challenging, especially w/ the small screen for those of us where glasses are required more and more :) so finding ways of incorporating graphics and touch will make this tool all the more useful.

They are using single-finger touch to bring up a neat "grid" menu, short and long single finger swipe, and two finger swipe. I'm still figuring it out, but what I know is nice. For example, short swipes up and down gets you the up/down command history in the shell, just like an up/down arrow would. Short swipe up to the "northeast" is a ctrl-c, to the "southwest" is tab. "west" is backspace, and "east" is space. Two-finger swipe up ("north") is the conf page, down is hide/show keyboard, "west" and "east" flip between the multiple terminal sessions. When you touch and hold, a square "menu" of buttons comes up, and sliding to them either does the function (e.g. "clear"), or changes the "menu" to a set of variants. For example, sliding to the "ls" button - which is darker to indicate that there are options there - switches the rest of the squares to variants : "ls -a", "ls -al", "ls -s" etc.

The results are pretty nice - if you have experience working in a shell, you can go pretty fast. I've only used it for a few things so far - ssh-ing into a server at work, or setting up a tunnel so that I can control a Hudson instance running inside our firewall. The iPhone is an incredibly powerful little computer, and having a good command line makes it more so. I wonder when Android will run on it? :)

"Most successful open source projects are using GPL," Mårten Mickos, former CEO at MySQL and now of Sun.

While "most" is debatable, I think it's interesting that it successful OSS projects either don't use the GPL, or don't just use the GPL alone, but have to modify it in some way to get around the enforcement of Freedom(SM) in GPL so people can use the project.

OpenJDK has the Classpath Exception along with the strict requirement of having Sun have complete copyright (so they can relicense to something commercially useful), MySQL also does the complete copyright thing (how else could they be worth $1B) plus several exceptions including one which I think of as The "We want to use Apache's APR, and we can for our commercial licensees because the Apace License doesn't restrict what we can do, but it's a big problem for those taking MySQL under the GPL" Exception, which I interpret as saying the FSF's opinion on license compatibility should be ignored when it gets in the way. Linux supposedly has some sort of exception for modules (I can't find it), and of course the standard unix library under linux is offered under the LGPL.

So yes, there are a lot of successful open source projects under the GPL, but there are a few others (Apache Httpd (aka "the webserver running the inter-truck" [Apache License], Apache EverythingElse [Apache License], Eclipse [Eclipse Public License], Firefox [Mozilla Public License], etc) that seem to do ok despite their non-GPL handicap :)

Are there actually *any* major, successful open source / free software projects available under a pure GPL?

As a bonus, Max said he'd ship it so I don't have to figure out how to explain it to the TSA peeps at SFO.

Read here

Also here : http://java.dzone.com/news/glassfish-v3-major-win-osgi

Nice evening last night. Dinner w/ Simon and Danese in North Beach. Found out that SF natives store their car keys in the trunk keyhole when at dinner. That wouldn't work in NYC. Hard to imagine why they bother locking them in the first place. Saw friends at OpenSolaris launch party (part 1) at Jillian's and then evening of debauchery w/ friends at HoS courtesy of RedMonk. Thanks guys, and apologies again for my tardiness, James.

Interesting problem on Leopard. I had been having problems in the last few days looking at PDFs on my laptop, but was so busy - and I have a big honkin' 8-core Mac at work - I didn't think much of it. Until this morning, when I was trying to look at the CommunityOne pdf and then upload photos from my camera into iPhoto...

Symptoms were Preview.app starting, but not opening anything. Same with jpegs - double click or "open file.jpg" from command line, and Preview.app opens or becomes active, but no windows open. In iPhoto, I'd connect the camera, it would do it's thing for a bit, and instead of showing me the low-res thumbs on the import screen, I just got little rectangles. When I did hit import, it opened a dialog and said that either the images were of an unknown format (.jpg) or the data was bad. I knew the apps were ok - I had created a new user and tried both - so it had to be something specific to my user. I snooped w/ opensnoop for a while, but wasn't looking for errors correctly. Anyway...

Problem? Well, it turns out that on Leopard (maybe earlier versions as well...) these apps use temp space in /var/folders/{2 char dir name}/{big ugly dir name}/.. so they need to write there. Somehow (I suspect via DiskUtil's fix permissions "utility" which I was using on wed trying to fix a MSFT Word problem) mine was owned by root, and thus as me, nothing worked. The solution was simply to sudo chown -R geir on the directory. Make sure you choose the right one - there is one for each user on the system, I think.

In retrospect, I might have gotten a hint faster if I had looked at the system.log, but that kind of "aha!" is always easier once you know everything :)

Also, there's a thread on this in Apple's forums.

After the first drop, 10% of the screen was broken - the left side. Got another and was going to do a transfer when I got home. On the way home, it fell out of my bag again. Now over 50% of the screen didn't work.

Solution? Carbon Copy Cloner. Attached the two laptops together w/ a FireWire 800 cable, held option on startup to let me boot off the drive in my broken laptop, used CCC to clone the old drive to the new laptop.

4 hours later, had a complete clone, and got back to work. Perfect. Recommended.

She didn't quite mean it that way...

I did the export of the schedule to .ics and imported to iCal. Worked like a charm. Only problem is that it only was my sessions, and not the general ones, like the general sessions, pavilion hours, lunch, coffee breaks, and after hours activities. Silly.

I also got to spend time with Ean Schuessler and his brother Eric, who also was here. When Bruno told me "Hey, Ean has a brother! He's bigger and crazier than Ean is!" I knew that things were going to be more interesting than usual. It was great meeting Eric. These two guys are insanely talented, and inspired at a level that is indescribable.

I have two photo's to share. The first comes from a restaurant we went to sat night, one that specialized in filet mignon, where "specialized" refers to the various ways they cook it. Start with "Chicken-Fried" :

Yes, there's a piece of steak in there - and not a bad one. (Forget the bed of pasta it's one...) That was followed by "filet mignon parmesian", then w/ garlic, then we a yellow-ish cheese, then a white-ish cheese, then madiera, and then something else I forgot. All good, but odd :)

Here's another pic from the gym at my current hotel - it's SP in the evening, clouds and rain :

Congrats, my friend!

Hey, can I get on the OpenJDK governing board? Would be good to have some loyal opposition on it :)

h/t to ColmMac for the suggestion

Why they are doing ME rather than SE? SE should run just fine. :) I mean, Google has Android running on phones, and Sun themselves had SE (or most of it - I don't remember) running on a phone last year? the year before?

They modified the login page, at least here in Manhattan, to explain. In doing so, they modified it so that it doesn't work in Firefox on a mac - the login form doesn't render. I actually had to boot Safari.

I probably won't miss you.

(Hey, Rich... these are the kind of people you need to keep...)

Tom and I have been doing hand-to-hand combat over 'free' Java, open source Java, Apache's relationship with Sun, the TCK license battle for years now. He's a worthy opponent and I really respect him. His new things sounds great, and I can't wait until he reveals more publicly.

(So it's clear - this isn't my photo. I just ran across it somewhere and find it wonderful and wanted to share...)

I think this is one of the downsides of Sun's "Feudal" open-source model. All project models have strengths and weaknesses - I think that Sun's strengths are that they have many really smart people who really mean well, and they are also willing to just pump cash into something a la Microsoft (e.g. Netbeans) - but when the business plan needs projects to be a "walled garden" (e.g. OpenJDK) with statutory control designed or intended, things will get rough at the intercept with the natural ebb and flow of contributors.

I thought you'd be interested to know that IBM has released the "IBM Java SDK version 6" for general availability today [1]. The Java SDK is released across various combinations of Windows, AIX, z/OS, and Linux on IA32, AMD64/EM64T, Power32 & 64, and zSeries 31 & 64 processors. Of particular note is that this version of the IBM Java SDK contains a significant amount of Apache Harmony class library code, as developed by a number of people on this list. So thank you, everyone, for the hard work in implementing, testing and delivering production quality code. Of course, IBM's Java SDK v6 has passed the Java SE JCK compliance test suite, is installed on numerous IBM systems, and forms the basis of literally hundreds of IBM products, many used in mission-critical environments. I think this is a very exciting achievement.

Third, we file patents defensively. Like MySQL or Red Hat, companies similarly competing in the free software marketplace, we file patents to protect the communities from which innovation and opportunity spring.

Mr. Schwartz - if you really believe this, could you please direct Rich Green to give the Apache Harmony project a license for the Java SE TCK. Right now, your IP is being used as a weapon against that community, from which innovation and opportunity spring (I won't bother listing all the Apache software your company uses in it's products...)

(BTW, Evil Geir thinks giving the proceeds of your NetApp lawsuit to the SFLC and ilk is a brilliant move..)]]> geir 2007-10-28T07:22:14-05:00 http://blogs.codehaus.org/people/geir/archives/001628_phone_call_for_rms_phone_call_for_rms.html The Java Trap

And while you are pondering the idea of not paying, don't forget Sun does aggressively defend their intellectual property in court...

(thx to http://www.upgradetravelbetter.com/)

I don't like the segmenting, but I'll take what I can get until it comes onto Joost :)

I found the following clip, his first monolog after 9/11, particularly moving. Not much else to add.

Now, while I agree with Sam's assessment, I don't think that the "C" needs to be removed from "JCP". Now, it simply stands for "Customer".

I think this is brilliant, and now I understand completely what Rich is doing. My alter-ego, Good Geir - who does Apache Harmony - is somewhat bewildered. Good Geir presumes that this is the logical follow-on to the secretly successful but sadly misunderstood "Java Desktop System" product strategy, logically extending it now to the whole company. However, I think he's wrong - it makes perfect sense in the spirit of "go big, or go home". I can't imagine what Sun's partners in the Java ecosystem are going to do, but clearly Rich doesn't care.

http://blogs.sun.com/jonathan/entry/java_is_everywhere

(I do wonder if they'll be spinning off their server and microelectronics businesses now that they've finally settled on a business plan.)

UPDATE : This blog was referenced on http://rollerweblogger.org/roller/entry/blogs_on_sun_s_new, but I think that the author missed my point - if the business objective is to unwind the notion of Java as a collaborative ecosystem, and turn it into a one-company product (like .NET), then this *is* a brilliant move. The whole fight with Apache - shafting a long-time Java supporter and valued community member by Sun ignoring their JSPA obligations and public promises for the stated purpose of defending revenue - makes perfect sense in this light. Community be damned, there's revenue here somewhere... As my good-twin Good Geir suggested, just change the "C" in JCP to mean "Customer".

So, by the power vested in my by the encouragement of Bruce Eckel, I hereby challenge Brian to... blog about it. Please? :)

I was looking forward to attending this year and hearing how OpenJDK is going, community-wise... it's been slow going, but now that that they've had over a year to get it together, I was curious where they were. I'm hoping that Mark still doesn't claim that Apache community governance is complicated :)

So today it was good to see that Intel and OLTP have made peace.

Anyway, today Boeing rolled out it's first 787. Here's what the head of Airbus had to say :

"Even if tomorrow Airbus will get back to the business of competing vigorously, today is Boeing's day -- a day to celebrate the 787," Airbus co-CEO Louis Gallois said in a letter to Boeing Chairman and CEO James McNerney.

"Today is a great day in aviation history. Whenever such a milestone is reached in our industry it is always a reflection of hard work by dedicated people inspired by the wonder of flight," the letter said.

(link)

Nice.

http://tellison.blogspot.com/2007/06/gnu-forks-openjdk-despite-claims-to.html

I was the one asking Tom about closing down the other projects, as Tom thought it a good thing to only have one implementation.

I can't wait to see how this works out. If IcedTea accepts contributions from others, how are they going to contribute to OpenJDK? Maybe they'll force their contributors to hand over copyright...

Apologies to all in my life and in my Apache family - I'm coming back :)

Joost has to be the strangest and most intense place I've ever worked (and after Bloomberg, that's saying something). But it's certainly the most fun I've had for a while...

Maybe they mean to say that the Final Draft makes the GPLv3 compatible with the Apache License v2? After all, the Apache License didn't need to change for this to happen :)

And that's the 4-star part of the hotel. Guy may or may not come and fix it tomorrow. I'm in the Holiday for the night...

I don't work for Sun, IBM, Oracle or any other company making their money with Java, so my influence on the business decisions made by any one of them is going to be limited by their willingness to listen to outside opinions.

I don't understand. OpenJDK is all about one company (Sun) and their business decisions to try and make money with Java. IBM contributed to an open source java implementation - Apache Harmony - long before Sun ever seriously thought about doing so, so I'm not really sure how IBM or Oracle fit into this. You're going to need to influence Sun. You then continue :

I'd like to see OpenJDK rapidly evolve into something larger, and getting that right will be easier to do with all the interested parties involved inside the project, rather than waiting outside it - which is why it is so important to me to get the runtime projects like the Classpath VMs, for example, and GNU Classpath, to become closely associated with OpenJDK (and eventually Apache Harmony & J9 as GPLv3 happens) as soon as possible,

I sorta got the impression that GCJ, Kaffe and GNU Classpath were just going to close up shop because OpenJDK is here ;) Or is that only Red Hat-led projects? Seriously, I'd like to see this too, but I'm not sure how "closely associated" Apache Harmony will get to Open JDK simply because of licensing and IP assignment problems. I guess one way is to try the path we wanted to go down with Harmony, namely look for architectural alignment with things like modularity.

Other ways? Well, a good first step would be to get rid of the requirement that all copyright is held (jointly or otherwise) by Sun, either by abandoning the requirement, or putting it in the hands of a trusted, independent third party. (Sun is having a few problems regarding IP management these days...) That may give you a big shot of credibility with the above mentioned IBMs and Oracles of the world, as that would help level the playing field - they could theoretically participate in the project as peers, rather than simply providing Sun with re-licenseable technology for free. An alternative would be to dual- or re-license the codebase under something commercially useful, like CDDL, CPL, EPL or such to give symmetrical freedom to users, rather than have Sun as the only participant with commercial freedom. (BTW, J9 is a VM owned by IBM, and not available under an open source license, so i don't quite know why it's included above).

Putting that aside, maybe a good first step of the Interim Government would be to explain exactly what the OpenJDK codebase is (it's not Java SE 6 and there is no such thing as Java SE 7) and who can use it and when as an implementation of Java. That should keep you occupied for a while ;)

However, the blog had a really confused perspective on corporate threats to open source, as well as a bit of confusion regarding their OpenJDK project. Re OpenJDK :

We've seen Java's acceptance made permanent, on servers and desktops and mobile phones and set tops, in no small part due to our decision to use the GPL license (to simplify the Linux/Java combination on consumer devices and industrial applications).

"in no small part"? How about, "in no part whatsoever"? Despite claims to the contrary, Sun actually hasn't released all the source to Java SE, and for the source they did release, it's not of any known version of Java, nor can it be called Java or used as Java. it's seed code for the not-yet-proposed Java SE 7 JSR, a JSR nowhere to be found yet. If the code was complete, you couldn't test it because there's no TCK out there that will let you ship the tested code under an open source license because of field of use restrictions on the shipped code placed there by Sun.

He does, IMO, get it right in the end :

All of which is to say - no amount of fear can stop the rise of free media, or free software (they are the same, after all). The community is vastly more innovative and powerful than a single company. And you will never turn back the clock on elementary school students and developing economies and aid agencies and fledgling universities - or the Fortune 500 - that have found value in the wisdom of the open source community. Open standards and open source software are literally changing the face of the planet - creating opportunity wherever the network can reach.

Perfect. Now can the ASF have a TCK license that's compatible with "open standards and open source software"? Please?

And the Apache folks were very angry at us for picking GPL. But we had to pick something. If we'd picked the Apache license, the GPL crowd would be upset with us.

No, Apache isn't "angry" - it's Sun's IP and they can license it as they choose. We believe that people should have the freedom to license their own IP as they see fit. (That's not what the GPL says, btw).

Also, lets be clear - Sun has already picked more than one license. Sun picked the GPL+ClasspathException for the OpenJDK project, but since Sun will maintain copyright over the whole codebase, Sun will be re-licensing the codebase under other licenses for their customers and other purposes. So, Sun could *also* license under the Apache License as well and not invoke the wrath of the "GPL crowd". I'm guessing, rather, that Sun simply doesn't want to because it loses control.

I called down this morning and asked for coffee. "Sir, we don't have room service in the Tulip Inn". I asked if they have it in the Golden Tulip. "Oh, yes sir". So I asked why they couldn't simply go left rather than straight when delivering the coffee. "Sir, we don't have room service in the Tulip Inn. That's the difference between a 3 and 4 star...."

It seems like room service is a cost center or somerhing. I wonder if they know they can charge for it?

I was wrong. The description is from here :

Sun is going to make it easy to certify the compatibility of implementations based on OpenJDK code by granting access to the Java SE TCKs under a special license, and that in turn will help safeguard the "Write Once, Run Anywhere" compatibility promise for everyone. The details are still being worked out, but we're committed to compatibility for Java technology as free software.

Good for OpenJDK, but I don't really know what this means. Since OpenJDK isn't actually Java SE 6, it's a future version of Java (like SE 7), I suppose we aren't going to see it for a while.

It's a strange thing, to mention the JCK in a keynote. It's software that 99.999% of all Java developers really don't care about. They just care about having compatible, performant Java. In the Harmony project, that's what we're trying to do, but the license terms under which Sun is offering the JCK are simply unacceptable.

The rumor is that Sun will offer the JCK as open source software. While that would be great, I suspect it won't actually address the problem - I don't believe it means what people are interpreting it to mean :

Current JCKs are available under an open source license, such that using the software as-is tests compatibility, and if passed, bestows all necessary IP as mandated by the JSPA.

Why don't I believe this? Because I can't believe that Sun would have forced the ASF into the open letter with all the negative consequences when they intended to not only comply with the terms of the JSPA as the ASF is requesting, and then go a step further and put the software under an open source license, which would be nice, but not actually a request from the ASF. Why would they do this now? If they had planned this, why wouldn't they have given the ASF a license without the FOU that violates the JSPA - just like every other TCK license the ASF has received from Sun - and then followed up with the open source announcement at JavaOne? They'd be heros to the FLOSS crowd. They'd be heros to me! Now, they find themselves in a public battle with a charity.

So, while the above would be great, I'm guessing it isn't going to happen. I could be totally wrong - it's happened before, more times than I care to admit. But suppose I'm right - what might happen? Here are a few guesses :

1. "It's not whether but when!" - In the same spirit of their "open source Java" announcement, they will announce that the JCK will be open source, in some future version, which may or may not be defined at the time of announcement. I'm guessing Java SE 7 is safe for them - that will be 2009 if history is any guide - which will be time enough for the organization to purge itself of the paranoid elements driving the current rift with the ASF, or if not, to prove the rational POV to be correct - that Harmony isn't any more a risk to Sun's business than OpenJDK is, and that more compatible, open source Java makes the Java ecosystem stronger, which makes Sun's value proposition to their customers even stronger. If I were Jonathan Schwartz, I'd be mighty worried about MSFT distributing a free, open source, cross platform .NET VM. But then again, that's the long view, and Sun seems to be rather quarter-to-quarter focused these days.
2. "The 'Look Ma! No IP!'" - This one is to release the source to the JCK under an open source license but not have it be the official JCK, in the same way that lots of Apache technologies like Tomcat and the XML stuff were the code that went into the RI of Sun's specs, but wasn't officially the RI. This will satisfy the FLOSS crowd, as they don't care about being compatible, and are living under the delusion that they have all necessary IP for an implementation. This will also confuse people quite a bit, as they won't realize that it doesn't solve the problem. Why doesn't it solve the problem? Because in this model, the JCK isn't actually open source - just some test suite that's part of the JCK. By definition, free software and open source software don't have field of use restrictions.
3. "The Flying Stallman" - This option is a bit weirder and also interesting. Imagine if Sun put the source of the JCK tests under the GPL, and then additionally had the JCK license allow FOU-less use on GPL-ed codebases. Note that this JCK license would violate the JSPA in the same way that the current JCK license does, as it doesn't remove the problematic element. It's also interesting as I'd hazard a guess that the FLOSS crowd would stand and salute even thought it's an ironic and clever example of a commercial entity leveraging the restrictions (aka "Freedoms"(TM)) of the GPL for commercial control or advantage. It's a tried and true business model (MySQL, OpenJPA, JBoss, OpenOffice) that works just fine, btw. Would they try something as nutty as this? Well, given Sun's confusion over their current compliance with the JSPA, who knows... it seems that the sky's the limit. Would it work? (With "work" defined as "placate the FLOSS crowd while calming Sun's misguided fears about Harmony") Maybe. I think it depends on the wording. Suppose someone had a project where one entity holds all copyright (like Sun does with OpenJPA). That entity could dual license under both the GPL and something else, but I suppose Sun could argue that their IP doesn't flow to the software under the other license (in violation of the JSPA).

I'll guess it's the first one. It is years away and therefore doesn't need to be sold internally very much....

I'm not sure if I can think of any other possibilities, except one - Sun figures out that moving forward with the ASF is more valuable over the long term than pointlessly trying to protect "make or break the quarter" revenue by taking a legal position that puts the legitimacy of the JCP in doubt. Lets call it a big miscommunication, shall we? Jonathan, I'd love to shake your hand on stage Tuesday (or offstage, or not shake hands, or.... Just send us an email if that's easier...) and move forward in the positive and productive manner that has characterized the relationship between Sun and the ASF for years now. Lets fix this.

Or not.

I tried to provide background and some detailed explanations of the situation surrounding this letter, including discussion about the "Field of Use" limitations, why the ASF thinks the offered license doesn't comply with the JSPA, and the implications for the JCP and Java as an open ecosystem.

I think it's worth the listen :)

For more information, there is also a FAQ available.

Update : Links :

1. Sam Ruby
2. Stephen Colebourne's Weblog
3. Jim Jageilski
4. Behrang Saeedzadeh
5. Tim O'Brien
6. Bryan Noll

His comment was interesting :

who would have thought Geir wanted to get rich quick? congratulations man.

Thanks James, but this made me think....

First, who wouldn't want to get rich quick? "No thanks... think of the tax bill...." I have no clue if I'm going to get rich quick at Joost - that's not why I did it - but I promise not to argue if someone decide to hand me a big check.

Second, why would James be surprised? This is my 4th real startup (the prior ones have either achieved positive cash flow and profitability or resulted in an acquisition), and my passion for open source and open standards isn't in conflict with whatever drives me to do things like Joost (disruptive business model? great people to work with? novel and innovative technology challenges? an opportunity for financial reward?)

After all, "open" is different than "Free"*

* "Free" is a service mark of the FSF and is should not be confused with the English word "free". After accepting Freedom, be aware that some pre-existing rights may not be available. Your mileage may vary. The FSF reserves the right to substitute existing rights with rights of equal or different value at any time. Any prior statements made about intended compatibility should be considered as advisory. Do not operate heavy machinery when under the influence of Freedom. Do not taunt Freedom in an enclosed area. Keep out of reach of children.

At Intel (for those just tuning in, that was where I was working), I learned an incredible amount about how things work on the mega-scale - it's an amazing organization, and I'm glad I was able to spend some time there. But this new opportunity was just too good, too exciting, too terrifying :) to pass up - I'm an entrepreneur and builder at heart. There are so many people I ought to thank at Intel, but I'd like to give special mention to a person assigned to me as my "Intel buddy" when I first arrived, a friend named Tony Baker that I had known from the JCP EC. We worked together as peers in first the Managed Runtime Division, and then (after The Small Reorg) the Middleware Products Division, and then (after The Big Reorg) stayed connected (almost daily) when he went off to bigger and better things in the Systems Software Division. His guidance and advice (learned after 25 years (!) at Intel) was always valuable and insightful, and I really am grateful for it.

This change in employment has no bearing on my dedication and passion for Apache Harmony - in fact, I think it becomes something more special for me, as there's no part of my job description now that has the word "Harmony" in it, and after some introspection, I like that. While the downside is that the time I can devote to Harmony is reduced, I think that in itself will make my contributions that much more focused, and hopefully as or even more effective. The upside, of course, is a fresh perspective on the community and ecosystem. I'm a big fan of true "open community" open source. It's clear that the old school "community OSS model" is giving way to the new school "supported contributor OSS model", but I still think these are aspects orthogonal to the basic control scheme of a project, and Apache is fundamentally about individuals working together as peers, no matter what their motivation or employer. When we started Harmony, it was a passion and a labor of love for me, not a part of my job, and I honestly never really imagined that I'd find it part of my job - and here I am, back full circle. Of course, being "back full circle" with Harmony at 96% percent completion in the class library with a capable and performant virtual machine is a great place to be.

Anyway, a new chapter with new challenges, opportunities, and growth. Wish me well :)

Now, I don't believe there's a problem there, and it's hard to believe that they think there's a problem either. The text where they explain it is really unconvincing, since they agree with the idea that upstream contributors have zero liability due to a downstream distributor's deal.

It will be fun to figure out where this came from.

This needs to be fixed, and urgently. I've met Kathy a few times and don't understand this at all.

This couldn't have happened to a nicer bunch of people.

Congrats Cam, Rob, Alex et al!

A free-as-in-beer tool that arguably frees users to choose ODF natively from Word is mentioned on page 3, just after "Alacos' Lettos Beta Online Conversion" (whatever that is)....

And people wonder why ODF is apparently such an uphill battle for the industry. I just don't understand ODF politics.

Update : 2006-03-10 : I got beaten about the head and neck with a stick by someone who felt that me pointing out that the newsletter wasn't available in ODF was somehow wrong, because PDF is an excellent format for "final documents", and having it in an editable format like ODF would be bad. I don't disagree, actually, but still - having the document also available in ODF is a good example of eating one's own dogfood, as well as a demonstration of the fidelity of the alternate format.

Story is here. I especially like the picture of it moving through Leopoldshafen.

The content is great, and I'm looking forward to more.

You can download directly from the link above, use the link to subscribe in iTunes, or even find him in the iTunes store. There are two there, and I think you want the FloatingSimonHead subscription.

You can also use the following to punch it right into iTunes : http://click.linksynergy.com/fs-bin/stat?id=kuK***sqbac&offerid=78941&type=3&subid=0&tmpid=1826&RD_PARM1=http%253A%252F%252Fphobos.apple.com%252FWebObjects%252FMZStore.woa%252Fwa%252FviewPodcast%253Fid%253D218534869%2526partnerId%253D30 Click it. It won't bite.

This puts Eclipse in the runtime business, and it will be interesting to see how this blossoms, if it blossoms. With the OSGi underpinnings continuing to gain traction and credibility, the popularity of the IDE and RCP ever growing, and what does appear to be a fairly large number of industry collaborators participating (163 member organizations?), there's a lot of potential, especially now that I imagine the RCP has world-class O/R mapping support. I do wonder how this will play out in the Age of Web2.0, but maybe if they had a runtime to ship with... Would that help in the SaaS arena?

Say what you will about the Java ecosystem, it's not boring :)

Latest on the list is NetBeans - we can already run Eclipse very well, and as we want to run everything, NetBeans is a good choice. (I love IDEA, but hey, I'm not doing the work, so I don't complain...)

Trying to work w/ NetBeans found two interesting problems :

• First, NetBeans uses private fields in Swing. Right - private fields. It uses reflection to get at JEditorPane.kitRegistryKey and JEditorPane.kitTypeRegistryKey. If it can't get them, it simply dumps the stack trace to the log, but it's not clear what harm comes form not supporting the private fields.
• Second, it also uses a sun private class "sun.awt.AppContext", so we can't actually compile it.

In Harmony, we have a policy that we want user's apps to Just Work, even if they do stupid^H^H^H^H contraindicated things like using private fields or implementation-specific non-API classes. We have a "Sun compatibility" module in Harmony where we collect those, and we will put in a "AppContext" class just so NetBeans can compile with Harmony. However, we can't figure out what those fields do as Sun's code is proprietary.

There is a bug report already filed against the use of AppContext, but I don't think anything is there for the swing fields.

We'd be happy to implement those in Harmony if we knew what they did - or maybe NetBeans can remove that code?

There's lots of coverage about what was said in lots of blogs, but I think what was most interesting was what wasn't said, which boil down to two words - "Harmony" and "OSGi". I was there focused on open source java, so attended mainly the OpenJDK/GNUClasspath track. I had asked to participate months ago to give an update on Apache Harmony, but wasn't allowed a timeslot by the organizers. That's cool - it's their track, and I guess there wasn't time. But, it would have been nice to give an update on our progress - I think we have the most complete open source Java class library, and our VM is nothing to sneeze at - solid JIT, good GC, and performance that seems to be getting darn close to Sun's Java SE 5. We still have lots of work to do, but we're not yet 2 years old, and we've come so far, so fast. We do have a minor license incompatibility problem with the GPL-ed Java projects, but the FSF has realized the error of their ways, and is fixing it ;) As for what was presented, I thought it a nice overview of a variety of projects, and nice to put faces on names that I only know from IRC, email and blogs.

What was really odd was in the 2 hour discussion on packaging and deployment, not one person uttered "OSGi". I can't figure out why other than some NIH problem - it's a proven, mature spec currently in v4, with tons of deployments out there, from everything from cars to enterprise app servers with multiple open source implementations. Even Eclipse is built on it. I heard a bit about JSR-277, a still-in-progess JSR with no implementations.

The packaging story for the linux distros will be interesting. The core problem is that OpenJDK won't be releasing binaries, so if the distros want to include "Java" rather than "software we built from the source that goes into Java" (SWBFTSTGIJ), they'll need to get a TCK from Sun and certify that their build passes. (That's a interesting little minefield for everyone involved - more on this later). Even if Sun was releasing binaries in OpenJDK, the distros still have their own needs, such as layout, libraries, etc, so again, to do Java vs SWBFTSTGIJ, they need the TCK. I imagine that RedHat's customers want Java - they are now selling a full stack solution ("100% Open Source!") - and won't be happy with SWBFTSTGIJ.

Anyway, despite my minor disappointment re Harmony, it was a good conference for me. It was nice to get the general "free software" perspective on the world in general, and I'm eager to see how the drama of OpenJDK plays out in my little corner of the world. My real learning experience is getting a better understanding of the perspective of the linux packagers in the case of redistributing Java. They aren't concerned with the same things that us java implementors worry about - they really just want a clean way of distributing the software. They don't really care about our religious beliefs regarding compatibility, and I understand better how the JCP hoops we ask people to jump through may be a big problem. I look forward to see how this plays out, because I think that it's important we do everything we can to ensure that distros ship Java, and not SWBFTSTGIJ.

I've been a week - Brussels, San Francisco, Portland and Stevenson WA (ish). 6 different airports, 5 different hotels. 3 train stations. 4 different in-room wireless providers. 2 airlines. 1 car rental company and a tram.

Now I can go home. I may have a whole week there :)

The first will be a talk on - surprise - Apache Harmony with my fellow Harmonizer Tim Ellison of IBM. The second is the "Java Technology Libre Panel", setup by my dangerous friend Tom Marble from Sun. I'm honored to be on this one with Mark Wielaard from GNU Classpath, Tom Tromey from GCJ, Dalibor "Dance to Feedback" Topic from Kaffe and of course Tom himself, the newly appointed OpenJDK Ambassador.

I want to say "thanks" to Sun here. For various reasons, I'm not Sun's favorite person these days, and I'm glad that the issues that make me unpopular were kept separate from JavaOne (one of my favorite conferences, truth be told...). I'm still having trouble grokking why Sun put both sessions I'm involved in on their first brochure, as noted by Simon, and as a physicist, I should understand the inner workings of the sun, but I don't :)

I figured that the obvious outlier, StarOffice, is simply an anachronism that will be brought over time in conformance with this strategy, as business considerations allow. Also makes sense.

But then I see things like the new, very cool ODF plug-in for MSFT Word. (Start here , click on the "Click here" to take you to it, and read more about the background here and other places.) It's not open source. Just to be sure, I read the terms of the license (but didn't agree, because as the license defines itself as Confidential Information, I couldn't write about it if I did...), and yes, this is a proprietary license.

To be clear, I think that it's fine for Sun to do this - they made the investment in creating the software, and it's theirs to license as they see fit (which appears philosophically orthogonal to their recent embrace of all things GPL, but hey, we know they had other reasons for that...). But as the world's leading purveyor of alternative-to-MSFT office suites via StarOffice and OpenOffice.org, I'd have thought that letting this beastie out in the wild would be in line with their OSS strategy. Curious...

It happened when the mac overheated - it closed it's lid, put it in the bag at the Starbucks before going to a meeting earlier that morning, and when I pulled it out, it was still on, very hot, fans madly clacking and screaming. It was still running (try that w/ the Thinkpad and see new and interesting BSODs). However, both CPU cores were pegged and the machine was sluggish. I did initiate a shutdown, and when it seemed to hang on the way down, I shut it down hard. Big mistake. Next time booting I saw the dreaded "file-folder w/ question mark" icon, which means both "system unable to find bootable operating system image" as well as "you're *beep*ed" (this is a family-friendly blog..)

So, beep-ed I was. I called Apple, made an appointment at the "Genius Bar" at the 5th Avenue Apple Store, and at noon, caught a cab and got up there. (Amazing how easy it is to catch a cab in NYC during snow storms....) The "genius" told me that yep, I was beep-ed. He was happy to reformat the drive for me. Given I've been a subscriber to the "faith-based computing" paradigm lately (aka no backups), that wasn't going to happen except as a last resort (after forensic data recovery, for example, although I hadn't quite worked out how to explain *that* on my expense report...). He was then happy to sell me DiskWarrior, set me up at the end of the bar with power, and wished me luck. He also let me take out my Thinkpad and use that for email while waiting for DiskWarrier to boot and do it's stuff (45 min!) but only on the condition I kept the mac in full view at all times.

Magically, DiskKeeper was able to rebuild the directory tree, with a loss of only a few OS files (resulting in an unbootable system, of course). I figured I was safe, and left the Temple of Steve, and went back to the SOA conference, and then home to continue the work. By the way, the 5th Ave Temple of Steve (59th and 5th) is simply stunning :

You walk in the front of the cube and down a circular staircase to the big store floor. If you are in NYC, it's worth a look if you find this sort of thing interesting. (The one in SOHO is nice too...)

Anyway, that night I bought a 260G LaCie firewire drive ("designed by F.A. Porsche" who I assume is a specialist in simple square metal boxes with a mirrored front), and backed up the entire laptop hard disk. I tried to fool the OS X installation disk to simply upgrade and archive OSX, saving my user prefs, but due to disk space limits and damage to the OS, it didn't work. Yesterday morning, I did a full wipe of the drive, fresh OS install, created a new personal account, and then replaced my home directory with that from the LaCie backup.

I'm amazed how much was restored. The only things missing that I've discovered so far is a driver for Parallels (I simply need to re-install), my license key for Keynote, and I'm guessing my license key for QT Pro, and my QT MPEG-2 playback component, all of which I can get back from the online Apple Store w/o a hassle. My ControlPanel enhancements (like Growl) all came over w/ that home directory copy. It's utterly amazing. As far as I can tell, there's been no loss. I hadn't saved two OmniOutliner documents (I use it for todo list management and time tracking), and it was able to recover from its own periodic background saves.

Oh, and one more thing :) This experience validated my decision last year to go "all in" with IMAP (hosted by Tuffmail). While doing the restore yesterday (which took a while), I was able to pick up another Thinkpad running Ubuntu, fire up Thunderbird, and see all my mail, and all my archives. No interruption. No loss. And full confidence that any email work I did on that thinkpad would be reflected on the Mac once I got back up. If you tend to be multi-client, or don't backup like you should :) consider IMAP. Good support on Thunderbird (so you can use the same client on Windows, Linux and OS X, and the OS X Mail.app has good support as well (that's what I use, for aesthetic reasons...)

And Tom, I'm only brusque because I care :)

Today, a friend of mine pointed out another funny aspect and ambiguity of the license, again, in the definition of "Program" :

"Programs" mean Java applets and applications intended to run on the Java Platform, Standard Edition (Java SE) on Java-enabled general purpose desktop computers and servers.

Who's intention? If I write a program that I intend to be run on Java ME - IOW, uses a subset of SE, but still would run on SE, can I run it on Java SE 6? What about if I group the object of intention to be :

"on the Java Platform, Standard Edition (Java SE) on Java-enabled general purpose desktop computers and servers"

So if I give you a program that I intended to be run on general purpose desktops running Java SE, could you therefore also run it on an embedded system (say you put that same desktop computer into a kiosk to give information at an airport) using Sun's software? Seems like you can.

I wonder what the intended (no pun) purpose of this phrase really is and why? Sun?

The Java SE 6 license says

Sun grants you a non-exclusive, non-transferable, limited license without license fees to reproduce and use internally Software complete and unmodified for the sole purpose of running Programs.

and then defines "Programs" as such :

"Programs" mean Java applets and applications intended to run on the Java Platform, Standard Edition (Java SE) on Java-enabled general purpose desktop computers and servers.

Maybe if you put your laptop down on the desk it's ok?

And to think, people wonder why we're working on Apache Harmony :)

Anyway, I was impressed to see that IDEA will have a RubyOnRails plug-in soon. I can't remember if I blogged about this before, but I've always thought that this was the basic strategy Sun had in mind for NetBeans when they hired the two JRuby guys - get good Ruby support in NetBeans, and then when/if RoR users need to integrate with existing Java systems, or need the features of Java EE, they are already working in an environment they are at home in - a page from MSFTs playbook.

Anyway, this is cool - another example of how JetBrains (the maker of IDEA) keeps a laser-sharp focus on the end developer. When I lived in IDEs 12 hours a day (I've been an IDEA user since v2.x) I always thought they were a leader in the space. Good to see they still have 'it'...

There are a couple of interesting bits in Ed Burnette's blog on ZDNet. First, of course Sun's reasons for embracing the GPL aren't "as pure as you think". Sun is a publicly-traded company, is allegedly ;) run for profit, and Jonathan Schwartz is a smart guy - there is an agenda, and it serves Sun's business interests. This is normal and expected. It's hard to imagine why anyone would ever think otherwise.

Ed gets to the heart of the matter :

The ultimate goal is nothing short of replacing GNU/Linux with "GNU/Solaris".

Ya think? :) This has been my theory for a while. But I think Ed misses an important point in the next paragraph. It wasn't "suspicion" that kept the "thought leaders" away from OpenSolaris. I think it was two things - first, the license was declared to be incompatible with the GPL by the FSF. Second, the OpenSolaris project doesn't have it's own distro. You can choose one of two quasi-open source distros from Sun (Solaris Express, Solaris Express Community Edition), and elsewhere there are other variants, such as BeleniX, marTux, NexentaOS and SchilliX. This is all well and good, but I think that given how closely Sun is holding onto OpenSolaris, they'd be well served with an open source distro maintained by the OpenSolaris community.

As I said before, I think that it's a reasonable strategy to move OpenSolaris to GPLv3 to try to take advantage of the rift that may happen if the Linux kernel stays on GPLv2, and let "Solaris/GNU" take on "Linux/GNU" on a level license playingfield. However, I think that Jonathan's and Rich's attempt to get this to happen backfired this week, and it will be interesting to see what happens going forward.

There's one more thing that deserves a comment. In the next paragraph, Ed says something that I think needs to be clarified :

By contrast, Sun's decision to release Java under the GPL (v2) was warmly received, except for those parties such as IBM and the Apache Foundation who would have preferred a more permissive license such as ASF (Apache), EPL (Eclipse), or BSD.

The Apache Software Foundation made no statements regarding the license Sun chose for their implementation of Java SE. The ASF has it's own open source Java SE project, Apache Harmony, that predates Sun's announcement of OpenJDK by a year, and will probably be 3 years old before Sun actually gets OpenJDK operationally off the ground. But the ASF doesn't really care what license people use for their own intellectual property. Now, it would have been nice if Sun chose a license that wasn't known to be incompatible with the Apache License that we use in Harmony (FD, I'm the PMC Chair of Harmony), but it was their choice to make, and they made it. Having another open/free Java implementation is good. Period.

Anyway, today I got a copy for the Mac. Naturally, it just worked. And it was really nice.

I'm Joost!

I like the slickness of the Word UI and prefer it to OO.org (and on a mac, there is no OO.org, just the NeoOffice/j port, IIRC), but I resent the file format. Kudos, Sun. I'm looking forward to trying it.

(h/t to James)

Update : the ad seems to be gone. It said "100% Open Source", advertising the RH stack....

"Wow!" I thought - as a co-founder of Apache Geronimo and Apache Harmony, this has been a dream of mine for a while - a 100% open source application stack. That would be at least an OS, a Java runtime, and a J2EE appserver. Since this is RedHat, it's going to be RedHat Linux (open source), JBoss J2EE app server (open source), but after that, I'm stumped. It needs Java to fit between them. There are no Java compatible open source runtimes, and the only way that JBoss could claim to be J2EE is if it was tested on a Java compatible runtime. We're working on one in Apache Harmony and Sun is as well with OpenJDK, and there are piece-parts in GNU Classpath, Kaffe, GCJ. But nothing is ready yet. What are they using?

I filled in the form to download, but it's been a few days, and I haven't heard back yet. I'm curious...

In the last two years, it's just gotten better for Mono :

• Runs on the Nokia 770/800
• Zing's Sansa Connect
• Second Life

Go, Mono, go! It's good to have something to chase :)

However, as I really don't know a lot about either spec, nor am interested in wading too deeply in what is very complicated political theatre (hey, Bob Sutor, when will you blog about the ODF Tools project at OpenOffice.org?), there was one thing that I'd like to talk about.

Miguel argues that in the area of formulas and functions for spreadsheets, the ODF spec is incomplete :

Depending on how you count, ODF has 4 to 10 pages devoted to it. There is no way you could build a spreadsheet software based on this specification.

I'm betting he's right. That in a vacuum, with just that spec, you don't have a hope of building a complete, inter-operative implementation. He reports MSFT took over 300 pages to describe that area, and if it contains silliness comparable to "autoSpaceLikeWord95", it's probably not enough either.

However, we don't work in vacuums. Good specs should have reference implementations (RI) that you can examine for correct behavior - for example, it's impossible to implement Java SE without testing what he RI does, because like all specs, there is vagueness and incompleteness. Even better, having an RI or other implementation in open source means you can not only see how the code should behave, but give some insight into how or why it behaves that way. As far as I can tell, there are 3 open source implementations to look at for ODF (OO.org, GNumeric, KSpread) and one proprietary one, Lotus Workplace. What are your choices for OOOOOOXML? A $400 closed-source product from MSFT that probably restricts how you can use the software in it's EULA?

Seems like others are starting to see the value of the model....

So now, the Harmony project has two production VMs to choose from, BEA's JRockit and IBM's J9, to use for testing in the project (as well as our own up-and-coming VM, DRLVM).

You can find JRockit here : http://dev2dev.bea.com/jrockit/jrockitVM

BEA has setup a forum : http://forums.bea.com

And of course, there's also the Harmony dev list for discussion on the Harmony-related aspects.

In the interest of clarity - this isn't JRockit going open source. It's a binary, evaluation-only license (and really, really useful to us).

Thanks BEA!

(w00t!)

I prefer simpler definitions. A standard is one that has multiple, inter-operable, independent implementations. An open standard, at least in the software world, is one where at least one of those implementations is open source.

I'd like to augment this by pointing out that the term "open source" implies the usual set of freedoms that accompany the copyright license, such as freedom to use in whatever context an end-user wishes without requirements to engage in a transaction or contract of any sort with any other party. Therefore, an open standard must allow such freedoms for users of open source implementations. If not, the implementation isn't open source, and therefore, according to Sam, not an open standard.

What caught my attention was the other thing it says :

"Solaris 10. Free & open source software from Sun"

What does that mean? Clearly Solaris 10 isn't free and open source software, because it's also proprietary. Does the ". " after the "10" and before "Free" mean something ? Not "is". Maybe "contains"? How about "will eventually completely be (as we position it strongly against Linux)"?

Don't get me wrong - I like the idea of Sun removing the religious and political barriers (aka "CDDL") to OpenSolaris adoption, and the much theorized prefixing w/ "GNU" (as in "GNU/OpenSolaris") is a very smart move as well. Is Solaris also was completely open source, that would be icing on the cake.

But they ain't there yet...

http://www.cafepress.com/meepzor.44654124

(And if you're looking for that union logo, it is made in the USA, apparently.)

Now, about a year after I joined Intel, this happened. Who would have guessed?

Clearly, this isn't my fault. And, honestly, in both cases, I think it's great.

Now, time to send that resume to SAP...

Anyway, Rich Green, EVP of Software at Sun, recently corrected the public record - they are still evaluating, which is wholly consistent with their position on GPLv3 for their implementation of Java SE, OpenJDK. Makes perfect sense.

So I tried moneybookers, the suggested alternate payment source. They don't like my mastercard either.

So I called mastercard. They never saw any attempted transactions. Could it be because I'm traveling and in Argentina, and they see the request coming from there? But wouldn't it make sense for a US-based traveler to... want to use skype when traveling?

I stumbled across WriteRoom via some new friends in the Boston area.

It's an editor that goes into full screen mode on the mac. No windows, no menu bar, no bouncing icons, nada. Just a clean place to think and write. It's nice. Try it. It's not a replacement for a word processor, but for the creative act of just creating words, it's compelling so far. I needed to write some important email today, the kind where you spend lots of time thinking and wordsmithing, and it honestly was refreshing to focus like that.

(I also like the model - use it for free. Pay for it if you want to. I'll certainly pay for it if I find myself using it.)

And I feel that I'd be remiss in what appears to be my newfound hobby if I didn't bring up the following. Note also the use of the BSD license. As he put in the blog entry :

"We ... settled on using BSD ... so that the widest possible community - Free software people, academics and commercial researchers - would be free to co-evolve the language and its implementation with us."

Good choice! (Btw, although a more modern license from the perspective of patents might have been good too...) However, think of what this is saying about Sun's choice to use GPL for their implementations of Java SE and Java ME. Walled garden? Replace "widest" with some other word, and "free" with some other word...

This will be really great to have another open source foundation in the JCP (the first being the Apache Software Foundation, of course)

I saw the term in a Cory Doctorow blog entry I was reading on a plane (using NewsFire). I woke up the business class cabin I laughed so hard...

I will note that for a while I was traveling with three laptops, and it really isn't that hard. Just takes practice, I guess.

Last year we had no beneficiary to donate the booth, and Sun wasn't quite so interested in open source then, so our hope to repeat it (with a conference-provided booth) didn't happen.

This year, Sun is clearly interested in open-source Java. They've placed a chunk of their JDK under the GPLv2, created a project called OpenJDK to eventually host it, promised the rest sometime in 2007, and if that wasn't startling enough, actually created a separate track on Open Source at the conference. (Woot!)

So Sun - how about creating a section of the expo floor where open source projects can participate, modeled after the standard "free software ghetto" you see these days at LinuxWorld? It shouldn't require much - a table per project, 2-3 chairs, standard curtain/backdrop where the sign is hung. Power and network would be great too (although I realize that does make it more expensive). It will certainly be an attraction for the open and free Java communities, which tend to stay away from JavaOne, and a good introduction to the variety strength of the open and free communities for the JavaOne attendees. How about it? I promise we'll behave :)

Update 2007-01-07 : Got a hint from someone who will remain anonymous since he mailed me privately... (I think because I had comments turned off...) that there's a site that offers a Solaris 10 VM for Parallels ready to go. I won't even bother asking under what license that OS is coming to me under. It runs too :) However, still want an open source operating system based on the OpenSolaris kernel (and Solaris 10 ain't it...)

]]> Misc geir 2007-01-03T08:48:25-05:00 http://blogs.codehaus.org/people/geir/archives/001475_rss_reader_for_os_x.html Time of the year when I challenge my work habits.... Anyone have a good suggestion for an RSS reader for OS X?

I disagree with the conclusion in that Sun will possibly marginalize Ruby. My theory is that working very hard to expand developer base for Java as well as try to keep NetBeans competing with Eclipse.

So, if they got good Ruby support in NetBeans, with the Ruby code running on the JVM for and debug and profiling support, it's a clear winner if you want to write in Ruby. Then, with all that lovely Java enterprise-ish stuff floating around, it's a no-brainer to start using that Java-as-in-language-based infrastructure in your ROR apps if you need it... :)

I think what they are doing is smart, and while I'm not sure if Sun has thought through a strategy like the above, it will be interesting to see what this will be a year from now.

Favorite part :

Executive Executive Summary
---------------------------
The Vista Content Protection specification could very well constitute the
longest suicide note in history.

Read it and pass it on...

(I have to admit I was a little disappointed that I couldn't run the YouOS demo in "WereWolf", the web browser for the YouOS desktop in the YouOS demo...)

For whatever reason, I changed my mind - maybe it was getting the Big Thick Disk for the TiVo. Doesn't matter. There are cool tools out there that I'm playing with now :

• tivodecode - takes a TiVo formatted file and converts to MPEG-2 (I think)
• Galleon - Home media server that works with TiVo, I'm using it to fetch content from the TiVo (better than using a web browser...)
• TiVo Decode Manager really slick front end for tivodecode - lets you fetch and convert. Problem is that for some reason, it has trouble downloading, and simply quits. Maybe I can help fix it.
• VideoLan - viewer for the MPEG-2's created by tivoconvert. For some reason, QuickTime can't grok them.

It's really not clear to me what I'm going to do with this (again, I really don't have time to watch TV...) but we'll see. Right now it's fun to fool around, something the Christmas break is good for...)

Update : Oh, I see - you can drag around in the right pane, but not the left. I wonder why?

prop.setProperty("mail.pop3.socketFactory.class", "javax.net.ssl.SSLSocketFactory");

and it just feels wrong...

So after it arrived, I pinged weaKnees via email. An hour later, they'd refunded the full shipping charges. No argument, no attempt to compromise on a ground rate - just a refund. Needless to say, I'll buy from them again.

Problem was, when I went back to TBird an hour later, I thought the same thing that Tim did - it does feel slower, and wasn't as visually appealing, etc. So I'm back on Mail.app for a while. It doesn't seem to have identities (which I think is a big flaw, but I can work around it), but otherwise I'm happy.

This is a validation of at least one reason why I chose IMAP - freedom from client. When I tried (and failed) to switch from Mac to Windows last year (because I was tired of carrying around multiple laptops...), I realized that Mail.app had it's own mbox format, and I vowed never to be beholden to one mail client again. IMAP lets me do this - I can use the best mail client for the platform, or even switch now and then to keep things fresh. Change is good.

So we have lots of code out that that claim to be implementations of the Java SE specification, but what actually makes those implementations "Java^TM", or as a Sun branding pedant would have it, "Java^TM compatible"? I'm actually a True Believer in the compatibility promise of Java - I think that it's very important - as a software developer, knowing that my runtime platform is Java (or "Java compatible") means that, in the memorable words of Rob Gingell, "my bytecodes won't be lied to". While this explicitly means that my programs will run as I expect them to, the implicit benefit is that I'm free of vendor lock-in. I can take my program and move it to another Java implementation, on another hardware platform, on another operating system. My choice, my freedom.

Traditionally, the way you know you have Java (or an implementation of one of the many JSRs that augment the base platforms) is if the implementation passes the "TCK" - the Technology Compatibility Kit - a test suite that any implementor can obtain from the spec lead to demonstrate that their implementation is compatible with whatever spec they've implemented. When you pass the TCK, you can claim compatibility with the spec and as a side benefit, you get all the "necessary IP" held by any members of the Expert Group that created the spec, a mechanism that keeps any member of the Expert Group - including the Spec Lead - from acting like a patent troll, being able to commercially control usage of the spec through patents. Java is an "open" ecosystem, right? ;).

So how is this related to the blossoming of open and Free implementations of Java? I was reading this recent blog post by Chris Adamson of OnJava. He asks if the DLJ - the Distribution License for Java - matters anymore, based on the premise that the linux distros will be doing their own packaging :

"Of course, with the GPL release of Java, the need for the DLJ would seemingly go away"

Does it? What exactly is the result when the distros compile and package the OpenJDK source? I'll argue it's by definition not "Java^TM" (or "Java compatible") because it hasn't passed the TCK. I'm sure that it's reasonably good software - after all, it's based on the same source code that Sun ships as tested binaries under their proprietary licenses. If a distro, especially a commercial distro that is feeding commercial enterprises, wants to deliver commercial-grade (whatever that means) infrastructure, I'm betting their customers will demand "Java^TM compatible", not "something based on the code that goes into Sun's implementation of Java".

One of the "success drivers" of open source is the notion of "good enough" - people are very happy to use solutions that are good enough to get the job done. They aren't necessarily the fastest, or the most stable, or full-featured, or pedigreed, but are fast enough, stable enough, functional enough, or "pedigreed enough". So whatever you get when you compile OpenJDK may be "good enough", but it won't be "Java^TM" or "Java^TM compatible". Now, I'll risk being accused of spreading FUD here, but I really wonder - is there an implicit "fork" coming in Java, with users choosing between "Java^TM compatible" runtimes and ubiquitous "Good enough" runtimes? (Maybe that's why I get winks and giggles from my Free Software friends when I ask them questions like this....) Last week, I spent some time traveling around Belgium and the Netherlands. Trains are two-class - first class means you probably get a seat, second class is good enough, but both tend to get the job done. (To really screw up this metaphor, I suppose I can conflate .NET/Mono and Amtrak as examples of the bad things that happen when you have a single supplier...)

I'm really not interested in seeing this. I've spent a lot of time in the last few years helping create compatible implementations of Java specifications in open source, including helping drive awareness of the value and utility in the Java compatibility promise. I'm hoping I'm wrong here - that this is a hypothetical, and that a solution will emerge where the Linux distros are shipping tested binaries. If I put on my tinfoil hat, I can see that this fits with my Walled Garden Theory about Sun's open source Java strategy, but the long term consequences are so disastrous, I can't believe it (even with the tinfoil hat...). Two obvious solutions - 1) OpenJDK releases binaries under the GPL, so that the distros that are happy re-packaging binaries from others can do so, and 2) those that need to build from source simply get the TCK from Sun and test what they ship. I suspect that the latter will be the way it will have to happen, given the need for the distros to build from source (and it will be fun to watch), but it leads to the outcome best for the ecosystem - preservation of the Java compatibility promise, and the long-term health of the platform.

Isn't having competition in the Java SE platform space (Sun, IBM, BEA, Apache Harmony, Kaffe...) grand!

Update - 2006-12-22 : David updated his weblog with actual numbers (probably required by SPEC). First thing to note - Sun needs to stop shipping AMD and start shipping Intel because, um, that Dell 2950 is simply hammering the tar out of the Sun Fire X4200 (FD - I work for Intel). Second, these are good numbers but if you go look at the latest posted SPEC JBB2005 results, nothing beats a good set of configuration parameters for your VM if you are interested in performance. Which makes me wonder - will we ever be able to get our VMs to self-tune optimally for the workload? I bet we will...

I found the following link on Simon's Blog, and think that the way CBR presented the material is extremely misleading.

The article in question talks about a response from Steve Harris of Oracle regarding Sun's OpenJDK project, and also references an earlier response by Rod Smith of IBM. In summary, they thought that contributing the code to an existing open source Java community under a different license would have been better than forming a new community. (I also think it would have been tremendously constructive for Sun to offer the code under a permissive license like the Apache License to the Apache Harmony project, but I also understand Sun's desire to retain full control.)

What bothered me is the position of the following quote from my friend Simon Phipps from earlier in the month (IOW, it doesn't appear that the quote was in response to either Steve or Rod) :

Speaking to Computer Business Review earlier this month Sun's chief open source officer, Simon Phipps, explained that using the GPL would overcome fears about license compatibility with Linux. "We've been working on the Java platform for a considerable time and we've got to the point where we're considering 'how do we grow the market'," he said. "The most important thing is that that the Java platform is not included in many GNU Linux distributions." Choosing the GPL, which is already used for Linux, avoids that issue, he explained. "Java now becomes the development platform of choice for enterprise GNU Linux users."

My complaint is that as presented, it makes it appear to the reader that Simon is saying there's a problem regarding license compatibility between the Apache License and Linux. He isn't (I checked) and there is none. Ever see a Linux distribution without the Apache HTTPD Web Server? It's under the Apache License, and runs 70% of the website in the world. (Answer : no). Money quote on this - when we first started Apache Harmony, one of the GNU Classpath people told me - "That's why we like the GNU Classpath exception to our license - we want to use Apache Tomcat". The problem was that Sun's licensing was neither open source or compatible with re-distribution requirements of the Linux community. They solved nicely - this should result in quick uptake of complete Java in Linux (the Novell - Microsoft pact will help too) - but lets be clear... there's no problem running Apache Licensed programs on Linux.

"Sun has every interest in encouraging innovation in the Java ME application space. The reason we have chosen not to add the classpath exception to Java ME is simply because Java ME applications are typically not shipped in a bundle together with the platform implementation and therefore an explicit classpath exception is not needed."

So... why not add the Classpath exception and prove your theory? As you are interested in encouraging innovation, you can't lose! If it's not needed, then no one will use it. If someone does take advantage of it, it's by definition an innovation.

As Bill Joy likes to say "the smartest people don't always work for you", and "innovation happens elsewhere". Let go. Let it happen.

(And these are not only some of the nicest people I know, but also some of the smartest - "wicked smaaht", actually, being in the Boston area - so if you have problems dealing with large amounts of transactional data in a cluster or grid, give them a call. They will fix it, and you will be happy.)

And the GPL acts like a patent shield between all parties.

You sure, Mark? It's possible to put yourself back in patent jeopardy through derivative works. Sun didn't open their Java-related patent portfolio - they just implicitly licensed it to a single, specific codebase, right?

However, there's a really bent and funny side to Sun marketing (remember the Duke condom?) as evidenced by these : #1 and #4. I had seen #1 before as an internal video, and especially as an Intel employee, I had to laugh.

Update - I see that Simon has a blog entry about them, with more links. Go see. Very funny.

If Linux remains under GPLv2, how does this help?

Putting that aside, it's nice to see Red Hat realize that the "Mono Trap" has always been worse than the "Java Trap" ever was. And now, coupled with Sun's action eliminating the "Java Trap" (if you believe there was one...) and Microsoft's action which put the world on notice, the difference is simply clearer.

I see NOVL's faustian pact with MSFT as simply great for Java - in one fell swoop, the battle lines have been redrawn, with Java as the rallying point. The Java ecosystem is a big tent - big enough and robust enough to have room for everyone, free software, open source software and proprietary software. Welcome, Red Hat :)

Well, it downloaded and installed ok. Looks good, v5.5, very pretty. There's a Sun logo at the bottom, which is weird, but maybe they forgot to take that out of the Forte contribution. I'll post a bug report. (Can you imagine the hyperventilation if there was an IBM logo in Eclipse?)

First thing I want to do is install the C++ support. I have no clue what to do. I find the Module manager under the Tools menu. Slick. An update button at the bottom takes me to a list of sites to check which takes me to a list of things to add. Very nice. Clean and clear. But I can't find C++/C. Maybe NetBeans can't install some things "hot". No biggie, I'll check the website.

Website. Wander around (I like the new site btw...), and find it easily on the products page. The section notes it's for "a variety of platforms including Microsoft Windows, Linux, and Solaris Operating Systems." Oh, please don't let it be that NetBeans doesn't understand the difference between "including" and "only"... yep. No OS X Support.

Why? Why? Why? Is C++ code harder to parse and syntax color in Java on a Mac than it is on Windows? A significant percentage of the Java developers I know work on Macs. I know that C/C++ is a niche, but why make OS X a second class citizen? It's no more niche that CLDC, CDC, or Glassfish. I wouldn't mind if I had limited features because is was a Mac (heck, I'm used to that), but not working at all? Hm. I wonder if this is just some sillyness on the website. I'm going to download for Linux and try that... Nope. The .bin for linux when run on OS X doesn't think a JDK is installed.

Never mind. I'll keep trying on Linux....

I hope that doesn't completely capture the spirit of the event. Did he forgot a few projects? I don't really remember... No matter. There's always opportunities to build bridges and collaborate.

(Maybe he can open the doors to the Walled Garden and check...)

Do not taunt SOA.

I hate to admit it. I love IDEA. IDEA is so elegant, so thought out, so polished, so coherent. Eclipse is so... 'extensible'. But I need C and C++, and Eclipse does deliver that. I know that Netbeans has it as well, but I just couldn't get Netbeans to do the simplest thing... build a project around an existing one. I'll have to try again, I guess.

Statements like this are confusing to people. It's really a cool thing that Sun did, releasing their implementation of Java under the GPL but lets be clear - the structure of the Java ecosystem is the same as it was before Sun announced OpenJDK. Java is still governed by the JCP. Maybe this will help us accelerate the opening of the JCP...

But what struck me as interesting was the package, which claims as a major selling point that it was "Created by a second grade school teacher". Huh? Maybe that school teacher uses teaching materials created by a subway conductor or fireman?

Thanks to Simon, Ray, Rich and Tom for making that happen.

More pics :

This is a good thing - it's more open source Java. I'm not a huge fan of the GPL myself, but that's ok. Sun has it's reasons and it's Sun's IP to license.

Sun was very gracious and has invited me to their announcement event today, and the press discussion afterwards. I'm very thankful that they did that. I just wish they held it in .... New York ... :) so that I didn't have to do JFK<->SFO twice in one week. Maybe next time :)

People are asking what this means for Apache Harmony. I don't think that it will change our day-to-day life much in the project. We have the same goals, the same problem to solve, the same work to get done. As I've noted elsewhere, Apache and Sun have different communities, with different licenses, different conditions for contribution and different governance models. Apache projects are a collection of peers, and each of us our own reasons for participating. I believe that this good news today from Sun doesn't change what we'll be doing - it just means even more open source Java, choices for users and contributors, and that's a good thing.

In my last post, I clearly didn't get it right, so I'm writing a new post rather than bury the "whoops, sorry" in the old one. (Given no one reads this anyway, it doesn't really matter, but I like to be honest with myself, anyway).

By placing under the GPL, a lawyer friend reminded me that Sun is implicitly granting patent rights to users of the code. Therefore, they do lose the ability to rattle the patent sabre for compatibility. However, because Sun will be creating code that is compatible with the specification - it is the reference implementation, after all - I'm betting that Sun isn't worried about that.

Further, as with any open source codebase under any license with explicit or implicit patent grant that I'm aware of, 'significant' modification of the work moves the modifier back into patent jeopardy if such exists, so Sun's putting the code under the GPL doesn't create compatibility risk any more than there is in the ecosystem today. (And I'll remind anyone reading that I believe the risk is very low - the market isn't interested in incompatible Java....)

Finally, I'll note as I have before, Sun placing this code under the GPL is really a NOOP if we want it to be - I'm guessing we'll still be able to get the tested binaries from Sun under the same licenses through which they are available now, and given their preservation of complete copyright, you can also license the full codebase under whatever terms you are willing to pay for. (I still think that's a mistake for community reasons, but Sun has different community goals than Apache Harmony does...)

In the end, this is simply creates more Java implementations under open source licenses. That's simply a Good Thing(tm).

Sun choosing the GPL for it's implementation of Java SE is interesting for a number of reasons, the most prominent is patents. GPL v2 has no explicit patent license - that means that you can release software under the GPL v2 for which you hold software patents, and there's no requirement that said patents are granted to the recipient of the software. More modern licenses, like the Apache License do have terms like this, which work to prevent patent owners from creating patent traps with source donations to open source.

One of the interesting aspects of the Java ecosystem is that when you create an independent implementation of a spec, you receive all "necessary IP" (whatever that means) from all members of the expert group via the spec lead when you pass the TCK, the test kit that certifies that your implementation is compatible with the spec. This is useful, as it takes away uncertainty about patent risk from the members of the expert group, so you can't use the JCP to create patent traps in the Java ecosystem.

So what happens if Sun uses the GPL for their implementation of Java SE? They have stated that they will use an OSI-approved license, which means GPLv2 rather than a one-off they create themselves. GPLv2 has no language that gives a patent license to recipients. There is language in the preamble that indicate it was the intent of the GPL authors to have patent licensing, but as far as I can tell, it's not there. Therefore, Sun maintains the ability to use patents to enforce compatibility for any implementation based on the GPL-ed source base. This in itself isn't very interesting, other than it's a supporting point to why they'll chose the GPL - they don't lose their patent weapon.

But what happens going forward? The GPL v2 states that unless the applicable version of the GPL is specified, you can choose any version of the GPL as defining the terms under which you receive the license. What happens then when GPL v3 is done? As of draft 2, it has explicit patent grant language in the form of a covenant not to assert :

You receive the Program with a covenant from each author and conveyor of the Program, and of any material, conveyed under this License, on which the Program is based, that the covenanting party will not assert (or cause others to assert) any of the party's essential patent claims in the material that the party conveyed, against you, arising from your exercise of rights under this License.

Will Sun restrict scope to v2?]]> Java geir 2006-11-09T19:28:26-05:00 http://blogs.codehaus.org/people/geir/archives/001432_open_source_java_iv_towards_a_walled_garden.html I muttered about Sun using the GPL for their OSS efforts a while ago, and there has been very little information about it in public, until now, and the whole industry is buzzing.

Jonathan Schwartz mentioned CDDL at the end of October, but that's been the only thing out there, and that was clearly an unscripted moment.

CRN posted this piece on Sun using the GPL, and - well - this supports what has been my bet for some time now.

Why? Because I think I understand now what Sun wants to achieve - basically, a 'walled garden' for its customers and partners, where they can create a stronger relationship, a participatory relationship, with those that still wish to be dependent on Sun for things like tested binaries, and such.

In retrospect, it's an easy conclusion to come to. Safe money bets that Sun will retain the same IP model they use for Glassfish and OpenSolaris, namely require that Sun is granted joint copyright ownership of any contribution, with the result that Sun will have a complete copyright and thus will be able to relicense to others under commercial terms. This creates a tilted "playing field" in Sun's favor, which I think is a mistake. It will discourage collaboration by Sun's licensee-competitors, which is a wonderful use for open source. There are countless examples of it, such as linux, httpd, tomcat, etc - but it relies on an equality in the ecosystem that Sun's model doesn't support. The GPL is the only logical license that supports this level of control. Other licenses, such as the BSD, MIT, CDDL, Mozilla, etc allow a recipient to freely augment or replace pieces and license those changes under their own license. That removes the control point that GPL gives Sun.

Sun's business model these days is 'open access and trial, come back for support and service'. So they will make things accessible for people to read, review, run, inspect, etc. The GPL is a great vehicle for this - it lets them retain control over who can make a proprietary derivative work, yet allow people to read and inspect it under a license that is "blessed" by the OSI.

So, it's a logical move in support of the "walled garden", one that fits what Sun appears to be doing as their global business model. It's a proven model, as demonstrated by MySQL, so it will probably work out for them. It does raise some interesting questions though, but that's for another blog....

And in Apache Harmony, we're getting close to being compatible... but we're still not.

Update : there was also this statement : "We expect that people who care about reliability and compatibility with the spec will use our version." Why? Anyone who passes the TCK is compatible, right?

Was Steve punk'd?

Graham was the Sun rep on the JCP EC, which is how I met him. A Sun Fellow and a VP, he was an uncompromising advocate of Java compatibility who simply didn't understand that my advocacy of open source was aligned with the "compatibility promise". Needless to say, we didn't see eye to eye.

That said, this is a significant loss for Sun, and the Java ecosystem as a whole. I'm proud to say that I've worked with Dr. Hamilton, and hope to do so again in the future.

I wonder what this means for Java, especially on Linux. Open source Java SE is a subject close to my heart, and I wonder what this "get out of jail free" card for Novell customers that run Mono as their managed runtime will mean in the long term. As Simon noted, there appears to be a bit of 'vendor lock-in' - not only will you have technology switching costs, but you may have a legal bill as well.

This certainly will keep things interesting for a while.

Hard to imagine what the joint offering would be.... increased uptime running Windows on top of LInux via virtualization?

Update : (thanks to Simon for the link) http://www.upi.com/NewsTrack/view.php?StoryID=20061102-041207-2995r

In the PR, they said they chose the Apache License and the Apache model of open project governance. This is good news - it's a good license and a good model that provides a level playing field for all involved. They didn't say they were doing it at Apache.

P.S. The Apache Harmony project was started by individuals of the Apache Software Foundation...

It's a shame about the name, because it looks like a really nice suite of stuff...

Could we create the "Apache Java Application Suite" over at the ASF? We've got good parts too - Apache Geronimo, Apache Tomcat, Apache MyFaces (JSF), etc...

Over half a mile, at 250 miles an hour...

Whatever, Russ - it's a good choice if you go that way. (...as I've said before...)

]]> Java geir 2006-10-26T17:36:34-05:00 http://blogs.codehaus.org/people/geir/archives/001417_unbreakable_leenooks_from_larry.html I think this press release from Oracle is a hoot. "We take Red Hat, de-brand it, and then fix it!"

Seriously, this is an interesting quandary for Red Hat - they have Oracle nibbling at them from the top, and Ubuntu (my favorite distro) from the bottom. Disruption in an ecosystem that was considered itself disruptive only a few years ago...

Now, the subject material is gone.

I'm sorry if anyone caught grief because of my comments. Going forward with that page was a Good Thing, a good decision, and a great step forward.

Here's a hint - you aren't going to get it right the first time, and that's ok. Openness is hard. It really is. And there are lots of us out here that want to help. Yeah, my comments were a bit snarky, and for that I apologize, but I really do want to see this succeed for you.

Let us help.

(tip o' the hat to Dalibor)

Question #1 : Do you believe that the JDK is a common term to describe the JRE + 'standard' tools?

Question #2 : Do you expect that you can get a JDK from any vendor, such as IBM, BEA, Sun, Apple?

I say yes to both...

First, maybe I'm just being picky, but I think that "Sun Microsystems will launch the open-source implementations" should be changed to "...our open-source implementations" in recognition that Sun's implementations of these technologies is one of many. Java is a multi-implementation ecosystem, which is one of it's best aspects, and there are other open source implementations in progress in Java SE, Java EE, and probably ME. I'm just saying... To be fair, they do get it right in the paragraph describing "OpenJDK" community, where they say "Sun's implementation of Java SE".

Next, consider adding links to stuff that does exist - like Glassfish. Also, why this weird statement about how "OpenJavaEE" will now be the place that

...governs the development of specifications, technologies, and products that work together to provide enterprise-level solutions under the Java EE paradigm.

Isn't the place that governs the development of specifications the JCP?

Now, as the name "OpenJDK" is so odd, I'm expecting that it's a placeholder until some internal battle is decided, and it will be renamed "OpenJavaSE" (pronounced "Oh Jeeze"?). Please don't do this. Please. There are several Open Java SE communities (and open JDK communities too) out there already, and we're doing just fine, thanks. Apache Harmony, Kaffe, JamVM, GNU Classpath, just to name a few. Yes, Java is your trademark, but it's now an ecosystem too, and that's what makes it valuable. Don't squeeze it too tight.

Anyway, I schlepped around the world with both my Mac and my Thinkpad for 6 months. I joined intel, and did the same with Intel's Thinkpad (although it did have Outlook, which at least is a normal mail client). But last Christmas I gave up. After trying a few mail clients on Windows, I settled on Thunderbird and moved. It was painful for bunches of reasons. First and foremost, at some point Apple's Mail.app switched from mbox to their own format. Second, Mail.app is slick, and mail clients on Windows are pretty bad by comparison. I suspect that MSFT killed the ISV market with Outlook and Outlook Express. Thunderbird worked, though, and is multi-platform, so I switched to that.

Recently, I got a MacBook Pro because clearly I need it for evaluation of the core duo and open source and (mumble) and... :) Seriously - via the Parallels and virtualization, I can run both windows and linux, so I can do all my Harmony build work on the road on one machine (yes, I was carrying 2 machines again, and actually now three - mac, ubuntu thinkpad and windows thinkpad...). So I wanted to solve my mail woes once and for all. Enter IMAP.

Working on 3 machines, I wanted to have a consistent view of the Harmony mail list - what I had read, answered, etc, on whatever environment I happened to be working in. So I tried a few providers (pobox - who I adore for forwarding service, Runbox and Tuffmail) and settled on Tuffmail (www.tuffmail.com). The reasons were that they support server side scripting (necessary for having a consistent view of mail no matter what machine...) and it has the feel of a mail service provider that does retail only for beer money. IOW, they appear to be worrying about big clients rather than lots of people like me, and I figured that therefore I'd get great service. I wasn't wrong - I sent a note to support one saturday morning asking a question and I got an instant reply. I was able to chat with the support person for a few iterations asking about their machine room, reliability, the colo, and I got smart and useful information. If this was their 1st level support, I was darn impressed. (Ever call your cable provider about a mail issue???). So I started doing all my Harmony mail on tuffmail. Worked well.

Last week, the POP server that I have an account on died. As in "it's dead, Jim". It was the mail server of a friend's company, and he was letting me have an account there. Very convenient. However, there was no backup machine. Very inconvenient. So they switched over to Google Apps and wished me luck :), and I decide to go "all in" with IMAP. All my mail traffic (except Intel mail for security reasons) now goes to IMAP - personal, open source lists, everything. I've proven again to myself that Tuffmail's support is very good - I thought I had deleted a folder of important mail, and I got a response to my frantic email to support w/in minutes. They did inform me that it was my fault - I somehow did screw up - and then promptly gave me a new folder with a very recent backup of my whole tree. Within the hour of my first mail. And, as a bonus, they didn't call me a moron, which they had every right to. It turns out that I didn't actually delete the folder, I moved it accidentally to an adjacent one. I heart tuffmail.

Anyway, I'm now migrating all my old mail archives from the Thinkpad to the IMAP server. It's so convenient - add the IMAP account in thunderbird on the thinkpad, create the folders, and just drag. Simple. As a bonus, I can see my mail from Versmail on my Treo. One view of it all - what's read, what's answered, etc. Wish me luck :)

(And yes, I realize I'm a luddite...)

Simon pointed this out to me. What they really said was this. My read is that they seem to think that the GPL v2 is actually compatible with a whole bunch of OSS licenses that the FSF says are not, because nothing in the exception removes the issues for which the FSF claims the other licenses are incompatible. Fun.

So what does this mean? Does mean that the license that MySQL is under is now incompatible with the GPL? If not, can you use a snippet of it as a "de-Stallman-izer" to allow you to combine GPL code with Apache License code? (Answer : no )

If you read down the page, the list of packages that are covered by this exception includes just one, APR, the Apache Portable Runtime, under the Apache License. I guess MySQL wanted to use it...

Take this a step further. Suppose you did the same thing with a piece of code and then put under GPL + Exception and also remove 2.2, the requirement in the exception for source re-distribution. Does that work? I don't think so - they acknowledge that this isn't compatible with the GPL where they say "If you create or distribute a work which is a Derivative Work of both the Program and any other work licensed under the GPL, then this FLOSS Exception is not available for that work"

Of course, since it hasn't been certified by the OSI, this means that MySQL isn't open source any longer. Unless you combine it with GPL-ed code, and then it is. Magically. Isn't OSS licensing fun?

Naturally balanced, light and low in minerals, Evian water SUPPORTS YOUR EFFORTS.

I've noticed it in my speech - and consciously trying to stop using it - and hear it often in the speech of others. Listen for it.

• Packing for a 3 day trip to west coast, I left the iPod in the charger
• Driving to the train station, spilled small bit of coffee on shirt.
• Train timed to bring to to city in time for meeting was cancelled.
• Leaving train at GCT, handle to "rollaboard" suitcase broke. Now I have two posts coming out of top of bag, but no crossbar.
• U.N is in town, so person I was meeting was 30 min late, caught in traffic
• That made breakfast meeting go late, so had to rush to find cab.
• Awful traffic (see "U.N" above) to get to JFK

Ok, the flight to JFK went ok. If anything had to, that was it... 767 went up, 767 came down, and was even resuable - my definition of a good landing. I even got the upgrade at the last minute (I was back in row 33 when they came and got me to go forward...) Dear Business Class Passenger That Was a No-Show : thank you!

• Bag was pulled off of carousel when I was in Admiral's Club synching mail, leading to frantic search and debate with the AA baggage people
• AMEX screwed up the car rental, so it was not under Intel discount, and was a large "road sofa" Mercury Marquis. I kept having the urge to yell "helm's alee" and throw the helm over... Had to fix that.
• Hotel bar closes at 10:30.

Time to throw in the towel and just try again tomorrow :)

FOO is fun, and I get to talk more about Apache Harmony to anyone interested.

I'd like to thank the O'Reilly staff for helping me out. I'm very much in a "just in time" mode these days, and Sara, Vee and Nat really show great patience and tolerance. Thanks. I hope I can return the favors some day.

So how did they claim the GPL wasn't viral? They just simply ignored the conventional worry about the GPL - that when you link non-GPL-ed code with GPL-ed code, the whole work falls under the terms of the GPL, and went for the corner case - if you have a GPL-ed program on your disk drive, it has no effect on other programs on your disk drive.

Ironically, they note in section 6 of the article that the GPL also requires distribution of anything and everything short of OS and complier needed to get the program going, including build and installation scripts. So not only is it viral to your source code, but also to that source's build environment. C'mon, Dave...

I found something similar for OS X - Witch. By default it uses option-tab, which may not be so bad as it keeps the regular task switcher in place.

And I'm honestly surprised to note that it's not under an open source license.

Seriously - why not do Java SE 7 as open source? I know you couldn't start with the Codebase Formerly Known As Mustang* (CFKAM), but are we in that much of a hurry for Java 7? Start with a blank slate, and let the community watch the governance and code appear. Do the Java 7 EG like an open source project. That will take some time to get going, and by that time - according to Sun - the CFKAM will be available under an open source license. Make Java 7 open from the get-go!

* It appears the Mustang and the Dolphin were let go during Sun's last RIF

It will be interesting to consider their javac in Harmony, compared to the Eclipse one. Sun has a chance to one-up Eclipse by using a license with less restrictions, like a BSD license.

Update (2006-08-15) : http://news.com.com/2100-7252_3-6105601.html?part=rss&tag=6105601&subj=news

Like Glassfish and OpenSolaris, Sun won't do it at once. Some Sun people think they can, but they can't. There has been 10 years of very fast, very competitive growth in that codebase, and I'm betting that Sun doesn't have a real clue what's in there. The code will have to be sorted through and examined. I know from my experience at both IBM and Intel in OSS code contributions, it's prudent to go over the code with a fine-toothed lawyer, even when the code was specifically written to be clean-room and clear of external IP issues. In this case, there's going to be lots of encumbrances, and I'm really hoping that those with the encumbering IP help out and make this easy for Sun. This is a big and important step for the entire Java ecosystem.

What license will Sun use? This has been and will be a source of rampant speculation. Sun needs to balance two things - dealing with their darkest fears around compatibility with the need for a licensing regime in which all players can innovate and control their own IP. Clearly the GPL won't do. While I am a big fan of the Apache License and the full freedom it offers, I can live with the CDDL and other soft-copyleft licenses. At least then, people could invest in innovation, and choose how they wished to license the IP for the things they created that were truly new.

I've suggested to anyone who listens that if Sun does have to do the GPL, they at least dual license under something free-as-in-freedom like the Apache License, with a non-free/non-open-source addendum that licenses the code under the Apache License if and only if any derivative work passes the TCK - IOW, is compatible Java. While we couldn't use that code at the ASF, as it's not open source, they would at least make it easy for people to do innovative things with Java (port to new platforms, better performance, embed in new applications...) and retain the freedom over their own code.

Maybe the best solution is the CDDL for core stuff, and BSD-like license for the not-critical-to-compatibility parts. Maybe do all the tools and plugins under BSD (so we can use them in Harmony), and the rest under CDDL? (Except for the bytecode verifier that we also need in Harmony...)

The community that Sun creates around their open source implementation of Java will be the most important aspect - more important than the license. People tend to forget that 'open source' is really a combination of a license for code and a governance model for the community. Simply put, people want to participate on a level playing field where all activity is transparent, where everyone is a peer, and everyone shares in the same benefits of the collaboration. To that end I hope that Sun creates a community that allows anyone to be a committer after some amount of demonstrated competence, that the copyright model is symmetric (IOW, it's not like Glassfish where Sun's copyright contributions belong to Sun, and your copyright contribution belong to Sun...), and that the governance model is clearly defined, and open in it's execution.

Again, this project will only reap the collaborative benefits of open source if every possible participant - be it individual, academic or commercial - has the same chance to contribute and the same opportunity to benefit as every other participant.

This is the "Participation Age" after all, isn't it....

While it does seem like I'm shooting myself in the foot as I'm one of the founders of and a big proponent of Apache Harmony, I'm actually willing to help with this one. I just don't know who to ask....

So here's where the crazy talk comes... I'll bet, and it's only based on sheer optimism and that things happen for good in the universe, that the people from Sun listening to Jonathan at JavaOne got it wrong, and aren't going to just open source Java SE, but rather open source the whole kit and kaboodle. Java EE is almost done with Glassfish, so it's only Java SE and Java ME, and the JCP.

Now, that's just crazy talk, of course. But Sun has done some odd things, and if you understand Sun's business model, it actually makes sense. "Lift heavy or go home" as a good friend used to tell me. (I work at home now...)

Were they to do that, I have some thoughts here too :

1. First, do it in steps. You didn't get Glassfish (Java EE) right, so I don't believe you'll get Java SE right either. But it will be better, and the learning you do in Java SE can be both plowed backwards into Glassfish as well as forward into Java ME.

2. Second, engage the community on what to do with the JCP. A lot of us (speaking as the Apache EC rep) have poured a lot of love and energy into the JCP, and given the promise that it will become a truly open and neutral ground for governing the ecosystem, I think you'll get a lot of positive engagement on how to make it better.

3. Finally, be careful with the name "Java". Don't abuse and dilute it by branding this new endeavour as "Java" or "Something Java". Just as we at Apache Harmony are "Apache Harmony, an implementation of Java SE", so should you be "Sun 'Jonathan', an implementation of Java SE".

   You've done it in the past with "Java Enterprise System" and "Java Desktop System", and most recently with that awful idea, "Java DB". Yes, you own the trademark, but "Java" is bigger than a mere trademark now, and in a community sense, we each own a little piece of it. People are pouring in creativity and innovation, planning for the future on it, building their businesses, their educations. It's a part of their social fabric. It's a promise of things interconnecting, it's a way of thinking, of working, it's an important facet of our technologically-oriented lives. Let it continue to be the thing around which we all share, all collaborate. You've created a wonderful thing - let us all continue to build on it, rally around it and yes, as owners, help protect it.

Thanks for reading this far :)

Note : the above is my personal opinion, and doesn't reflect the view or policy of my current or past employers.]]> Java geir 2006-08-14T19:20:28-05:00 http://blogs.codehaus.org/people/geir/archives/001381_united_was_ok.html So I was away last week in Crested Butte for my favorite un-conference, and as I mentioned, I had to fly United. I have to report that it was pretty good. They upgraded me to 'first class'* on both legs between here and Denver, let me roam about in the Red Carpet club for a day, and generally were very nice at all times. Now, the 757 to Denver and the A319 coming back DIDN'T have power for my laptop which was annoying, especially since it was first class, but it's only 2100 miles so I survived (barely... near-dead battery walking off the plane...). I think American does have them beat on this, as they do power in first, business and a good part of economy on every real plane, and one my Laws of the Road is "Keep your Powder Dry (and Batteries Charged)".

* Means that the seats slightly recline :)

It turns out that United had a nice "status match" program - prove to them that you have some status level in another frequent flyer program, and they'll give you the next-to-equivalent level in their program. After asking for expedited handling, they've managed to get it done before my trip, and I'm now "Premier Exec". (Kinda like AA's Platinum, I suppose). It hasn't solved my seat problem yet - I'm still in the back in row 723, in the middle, near the head... - but maybe they can fix that near their airport. Maybe flying on United won't be so bad now.

I once asked Delta about this - they hadn't even heard of the idea, so it seems. That fits with their "we resent you being here" approach to customer service..

Anyway, on an Ubuntu 6 box I had an autoupdate that killed Firefox. It was an odd bug - you could show a page, but clicking on any link would crash Firefox. Today, the little 'update light' was lit, more code came, and it was fixed!

I'm not terribly comfortable with this (reminds me of Windows), but at least it works...

Please, try your favorite apps with it and let us know what's needed... We're now at the stage where user feedback is possible and very, very welcome. We have about 80% of the class library implemented, and the VM is reasonably useful. Has JIT and is not slow.

One recent development is that we are now producing JRE snapshots for Apache Harmony. While the stuff isn't ready for prime time, it is the core of a JRE - the VM and the classlibrary. All open source. All under the Apache License.

You can do a few things with it, like run Tomcat and Eclipse. But it's still early, so I don't want to hype it. ßWhat we want to see is users being able to take our snapshot JRE and test their apps, and tell us what's wrong, so we can fix it, and let this virtuous cycle continue.

So please try it. You can get it here.

See what you get on the right when you search for "Geronimo" on Google...

Put in "JBoss" and "Jonas" as well...

Last thursday, Thunderbird 1.5 was released. To best new featuers are real-time spell checking (i.e. as you type.... I was so used to this on the mac, systemwide, that any program that doesn't do it on Windows is IMO broken...) and auto save...

It seems to now boldface any folder whose subfolders have unread messages, so that's nice. It doesn't show how many, but maybe that's some obsure "advanced config" parameter. It also now does auto updates, and the default setting is horrific - it will by default just download and install any updates when they become available. Lots o' confidence there in Thunderbird developer land, I suppose.

(Aside : Hopefully a different lot than the Firefox people - I added an extension from Weather.com this morning for Firefox, and after that, Firefox wouldn't come up. No window, no dialog, nothing. I tried starting by holding down shift, ctrl, etc hoping that (like many mac apps) it would start in standard-no-extentions-or-other-stuff mode. No luck. Uninstall/reinstall? Nope. Finally I noticed that it kept the extentions after the uninstall (what a nice option that would be on uninstall...) so I blew those away and now it works again... I wonder how long it would have taken your average non-professional user to figure that one out, like my father....)

I'm starting to get used to the T43 and WinXP. One of the major irritants was its sloth - I'd sit there sometimes with the disk light pegged on for quite a while. It was like using an old mac.

I noticed that the standard drive, some Fujitsu 80G thing, was only 5400 RPM, and I knew how to fix that. I bought a Hitachi 0A25015, a 100G 7200 RPM unit, and a copy of "EZUpgrade", a USB 2.0 shell with software allegedly for laptop drive upgrades. I say "allegedly" because I just wasn't able to get it done. I would try, and it would fail about an hour in with "Failed to read from the sector 1 of the Source disk". Now, that sounded really scary. Sector 1 seems like an important one.

Tech support suggested that I run some WinXP sector checking program, which I did and hours later, it reported no problems. I tried to copy the drives again, and same error was reported.

I decided to try to find other software. I found Acronis who offered a free download of their DiskDirector package. It had a tool for making bootable images of disks, and amazingly, it looked exactly like the EZGig software. (It turns out they have an OEM program...). Anyway, I tried it anyway. sure enough, about an hour into it, "Failed to read from the Sector 10265332 of the Source disk". A ha! EZGig seems to truncate the message (assuming they OEM-ed it from Acronis). I told it to ignore. Several more times, too. 10 hours later, it was done.

I switched drives, powered up the T43 and after telling the BIOS to keep going even though the new drive wasn't supported (it made the stock drives look bad, I suppose), up it came. It feels so much faster. I have to do some perf tests to see, but the machine really snaps now. Eclipse comes up in 10s of seconds rather than over a minute. Maybe the old drive was broken....

On Wednesday, I went to an Intel Linux Strategy Summit at a lovely lodge outside of Portland, actually over the border in Washington. To get there, you need to cross the Colombia river at a dramatically named bridge, "Bridge of the Gods", which refers to an old Native American story of a landbridge that crossed there. I imagine that the area is just gorgeous, but it was hard to tell in the rain and fog that morning as I drove out.

Anyway, I left at about 10pm to get back to Hillsboro. It's dark there (out in the sticks) and I was worried about missing the bridge on the somewhat pastoral state road that winds along the Washington state side of the Colombia.

My fears weren't justified :

Or "Darklandia" as Danese calls it. They don't drive like they do on the east coast. I have to be careful. (It's slower here...)

I'm very used to - and attached to - the UNIX shell experience (It was very good on OS X). I've been trying to mimic it on WinXP using Cygwin, but the shell program use ctrl-c/ctrl-v for cut and paste. I guess if I think about it, it can't - I need ctrl-c to kill things, at least on the command line. However, I need to do this somehow with the keyboard - Right now, I have to go to that little menu in upper left, click on it, click edit, click paste.... I can't stand it. It really slows down work. In mac-land, I had cmd-c, cmd-x, cmd-v for this. (And those keys worked in every program... what's up with that? Windows had had 22 years of consistent Mac UI to copy that one single feature and still hasn't... )

So the solution that I'm trying is running Ubuntu in VMWare. So far, so good - VMWare is very slick, and Ubuntu installed w/o a hitch. I'm not sure if I like the UI, but we'll see. I need to figure out how to mount the windows file system in Ubuntu so I don't need to duplicate things like source trees and such.

Let OS/2 be free....

Anything

Today's early morning task was to get my address information over from the Mac addressbook to first Ou