Brett McLaughlin's Blog

Open Source - A Divider?

So the other day, I posted about Java not seeming to be so cool anymore. And I obviously know cool -- I had enough sense to actually grab a mandolin when it came time to have my O'Reilly website picture taken (for the humor-impaired, that was a joke. Gotta' get out from behind the computer once in a while, people!). In any case, it caused quite a stir, it seems, with people agreeing, disagreeing, and coming down in the middle.

Rather than trying to post some sort of follow-up, though, I'd rather let that debate rage for a while. I actually think we desperately need some more discussion like this going on -- rather than this sort of boring "Is JSF going to properly incorporate Struts into the JSR?" I'd rather see some people be passionate about what they (presumably) spend 8, 9, or 10 hours doing. And yes, I get that you can be stable and mature, and not cool, although I would argue that this does not preclude there being cool aspects about a language. But, enough of that. I said I wasn't going to do that discussion today...

So I picked another topic, hopefully just as disagreeable (if you can't tell, some of my best experiences have come out of a really juicy argument or disagreement). What are you going to rant on today, you say? Well, I'm actually a little tired of the Java-open source dynamic. Here's another discussion that's raging along, with relatively little interaction between the really smart people on both sides. Instead, we seem to have lots of little skirmishes between worker bees, when nobody is looking.

I think about JavaOne, when there was this pseudo-panel on open sourcing Java. Personally, I thought it was a complete joke. First, when Tim O'Reilly asked who had ever used open source software before, there was a fraction of the audience that raised their hands -- I think calling it 5% would be generous. So it's obviously a discussion without context. And, surprise surprise, almost everyone (oh, let's say, the other 95%) was for Sun keeping control of Java. Now, pardon this next statement, but....

How can someone who admits they have never even used an open source software project say open-sourcing Java is a bad idea? How would they even know? Come on... let's at least have a sane, reasonable discussion, rather than just closing our eyes and arguing. That's not how people in a -- supposedly -- intelligent society communicate. If you hate open source, you better have a good reason, and it better involve at least some minimal experience with open source software.

On the other hand -- and here's where I hack off everyone who's still nodding their heads with me -- how about the idea that when people come to open source software, we leave them with a good impression? Look, I've started several projects (Turbine with Jon Stevens, JDOM with Jason Hunter, and Zeus) and been involved with many many more. And so often, those projects have this massively high barrier for entry.

• I mean, we're not going to comment our code, because we're real uber-hackers, right?
• And heaven forbid there is documentation. Come on -- O'Reilly is supposed to publish that for us, aren't they? Besides, nobody around here reads the instructions first, anyway.
• Mailing lists and CVS are plenty easy for code access. You don't know how to use those? Oh, get a life, and come back when you're smarter.
• I know it says it's version 1.0, but this is open source. Bugs are pretty normal around here. Get used to it.

Now, before you start throwing tomatoes, I realize that this isn't true -- in part or in whole -- about every open source project. But I'll bet you at least one of these applies, to some degree, in almost every open source project (I have a stronger opinion than that, really, but I'm feeling nice). So no wonder some of the folks check out this stuff, and say, "Thanks, but no thanks."

OK, I could go on, but this has been plenty for today. I think the point I'm making -- and I hope you're not so angry that you've stopped reading -- is that both sides have a lot of work to do. But, this sort of idealogical battle is a dumb way to approach things. I'd rather see the closed-source guys actually check out some open source projects before turning up their noses at the idea. And, those of us who are open source types (and I strongly consider myself one of those) -- well, we need to start taking this stuff seriously. I realize that most of us code late at night, after working a "real" job, but if we ever want this stuff in mainstream use, we've got to take the same stringent approaches to projects and releases that the big boys do.

BTW, if you hadn't noticed, the big open source players that are getting traction all have worked hard to deal with every point on my rather sarcastic list above. Might be something to that, huh? So bring it on -- intelligent responses very much welcomed!

Comments

• Survey bias
  I am pretty sure that the "only 5% of audience used Open Source" stat is due to survey bias. If he asked everyone who has NOT used Open Source to raise their hand, it will be 5% or so as well. :)

  Also, if the stat were true, it basically tells us that Open Source users do not even care when Java is Open sourced or not.

  Posted by: juntao on September 26, 2004 at 12:08 PM

• Correction
  That wasn't the question that Tim asked. He asked how many hacked on open source. It would have been surprising if 95% of the audience had never used Tomcat, Ant, Eclipse, NetBeans, or any of the other open source projects that Java developers use every day.

  Posted by: daniel on September 26, 2004 at 05:23 PM

• Correction
  Excellent point,

  There seems to be confusion in the terms free & open source users, (the large community), and the f/oss developers, (the small community).

  To me, this was the tricky part:

  
  At Java One the f/oss user community were themselves also developers. However, they develop proprietary software, using free software; for which their employers pay their wages.

  
  To me, the surprising thing was that the user community did not yet seem to appreciate the ideals of the developer community.

  Posted by: cajo on September 26, 2004 at 07:00 PM

• As if commercial Software Releases are free of bugs
  I can remember the time when WNT 4.0 was released with the neat little bug that one could log-on as administrator without its password. And the documentation on WNT is not great either or do you need to know how to deal with a window. On the other hand I just downloaded an open-source DVD player for windows and even though it is a beta version it works and I can show my little son his favorite TV show on our vacation which is priceless.
  -Andy

  Posted by: schaefa on September 26, 2004 at 10:00 PM

• Correction
  I think most OS users appreciate the effort of the developers well but standing on the rim looking in (I've dabbled in OS development and am a user who's had contact with developers from time to time as well) I can see why users want to avoid the quagmire that are many OS projects where it comes to their compiler.

  When having trouble with an OS tool, the answer when asking help is more often than not "go fix it yourself", "did you pay for it, if not why complain" or something along those lines.
  As a corporate user, I can't have that.
  I've no time to go track down bugs in the source of my compiler and fix them, I've enough work tracking down my own bugs (and those of my coworkers).
  I don't WANT to have the knowhow to write compilers, IDEs, database engines and webservers in our corporation. It's not our area of expertise. That's precisely why we use 3rd party tools (whether OS or not doesn't matter in this) instead of writing everything ourselves.

  OS'ing Java, quite apart from dividing the platform in a hundred forked versions that won't work (well) together, will mean we will have to write our own compilers (or at least do our own bugfixing on them) as the community (from my experience at least) is unlikely to respond to bugreports from corporate users with an effort to fix the bug. They're too busy implementing nice new hot features to spend time looking to find a bug in some old code (heck, I like writing new stuff better than debugging old too, but it comes with the job I'm paid for).

  I'm not saying this WILL hapen if Sun releases Java to ESR and his cronies, but it CAN and that chance is not one we should look forward to.

  Posted by: jwenting on September 27, 2004 at 04:49 AM

• Stakeholders
  One can argue endlessly about the coolness and/or openess of Java. However it seems to me that whenever these discussions take place its usually in the context of deloper roles and their needs.
  If one actually considers the Java eco system, one will realize that its no longer just about developer mindshare. Java has also evolved the business-person mindshare. These folks do not care a hoots about the openess or blinkeredness of a system as long as its impact on their bottom line is minimal.
  The show of hands at Java One was not about how widespread the penetration of OSS is in the Java ecosystem (OSS is pervasive within the java ecosystem). It was more about the audience recognizing Tim's attempt at putting Sun on the defensive. The show of hands (or lack of) was a vote of confidence at the current predicament. If you take all the arguements made my Raymond in his "cathedral and the bazaar" document and apply it to Java, you will realize that a middle road is also possible, that closed and open software can co-exist (like apache and javasoft).
  In the end, this whole arguement about opening up java is not about freedom but about control, control not at the systems level but at the level of dollars and cents. Right now I trust Sun a lot more than I would IBM or FSF and Sun has done nothing to diminish my trust in them.

  Posted by: suhail on September 27, 2004 at 06:01 AM

• Correction
  1. I have never had an open-souce coder tell me "fix it yourself." I'm sure they're out there, but I've never encountered them. If you tell me that there's a problem in NoExceptions, I'll fix it. You might be able to fix it faster than me, but I'll fix it.

  2. Sun should still have people paid to work on Java. I will be very disappointed if they say "I don't want to fix that." (Of course, it's not like they're always speedy about fixing stuff as it is.) Open-sourcing will put more developers on the project, and that increases the odds that someone will find your bug "interesting" and fix it.

  Posted by: afishionado on September 27, 2004 at 10:36 AM

• five percent?
  Did they mean program open source or just use open source? Using IE you're using open source since it was based on the original mozilla. A huge amount of Unix software is based on open source and I'm sure guys at Sun have used Unix (15%? :) ).

  I'm against standardizing Java as they did with C++ due to all the different interperatations vs one format.

  Posted by: smartinumcp on September 27, 2004 at 11:48 AM

• Does it really matter if the product is open-source or not? Look at Linux. You can repackage it and "sell" it. It all leads to getting a business rolling out of the model. Open-source or not, the efforts all point towards making it a business. And in most cases, the people who are in the business side of the open-source movement gets the money.

  Posted by: e_mendz on September 28, 2004 at 02:57 AM

This work is licensed under a Creative Commons License.