ASF vs JBoss bitchfight
Looks like the JBoss cult got its collective mind together and decided that enough is enough, there can only be one JBoss, and by god it’ll be led by the fleurys no matter what. Anyone who dares do otherwise must be punished like the wayward ungrateful children they are.
First, let’s be entirely honest and admit that there pretty definitely is some sort of misbehaviour going on. The ASF guys are generally a bunch of wankers too, and opensource being opensource, with everyone molesting everyone else’s code it’s hard to know exactly who did what, who’s idea it was, and who wrote it and designed it (no, cvs logs, @author tags, and strategically executed diffs don’t give you the whole picture), it’s hard to determine who owns what. Throw in selective memory and sheer happenstance into the mixture and it’s a recipe for much family fun all round.
November 11th, 2003 at 12:32 am
Always respect copyright.
November 11th, 2003 at 12:35 am
Uhh, I can still do whatever I want with JBoss source as long as I follow the license.
Rants and bile is all good but could you at least stick with some basics.
November 11th, 2003 at 2:40 am
Anonymous,
The basics are simple: JBoss code is licensed under the terms of the LGPL. Allegedly, someone has taken JBoss code and relicensed it under the Apache license. You can’t do that, because the LGPL doesn’t allow it. If the allegations are true, the JBoss guys are right to enforce the terms of their license.
November 11th, 2003 at 2:59 am
LGPL is void under the law because GPL is. (If SCO succedes, then JBoss will also be fair game ;)
November 11th, 2003 at 4:08 am
Does this mean I am not allowed to use setter and getter anymore, because I will be sued by the JBoss gang ?
November 11th, 2003 at 4:27 am
setter and getter are indeed copyrighted in the US by JBoss Group. You cannot use those in the US. If you request a special license from Fleury then he may, just may let you use them, if you include special disclaimers highlighting the copyright ownership of your code by the JBoss Group.
November 11th, 2003 at 6:02 am
Mr Popcorn, the original authour of the code maintains copyright under the terms of the LGPL itself. As such they can decide to reuse that code in anyway they please. Someone who is not the original authour of the code however could not change the license of the code bu removing the LGPL and adding an ASF license. Given the number of JBoss developers who jumped ship (the entire CMP team for example) its well within the realms of possibility that the code in question was written in the first place by these people.
I for one am very worried that Hibernate is now part of the JBoss group. I wonder how long it will be before Hibernates wonderfully informative and well maintained website atrophies to the circa-middle 90′s look and circa 1890′s functionality of the current JBoss website – and don’t even get me started on the attitude of Flery et al.
November 11th, 2003 at 6:46 am
humph, the so called CMP “team” of one did not write the CMP engine from scratch either
thank you for playing, drive through
November 11th, 2003 at 6:51 am
It doesn’t matter if they wrote the CMP from scratch. The point I made remains valid – the original authour of the code retains the copyright (granted under the LGPL itself). As such the original authours can quite legally release their code into the Geronimo project. The question remains as to if the code mentioned in the JBoss lawyer’s letter (in a SCO stylee one might say) has been leaglly introdcued to geronimo or not.
November 11th, 2003 at 9:16 am
Naaaaah ..
November 11th, 2003 at 10:14 am
Did you guys look at the code in the letter? The naming is similar in the log wrappers, but it’s also the same as that of the log4j examples.
Also, is WebSphere Portal built on Jetspeed?
November 11th, 2003 at 3:49 pm
The problem is, if you’re not working from scratch but building on top of existing LGPL work then you’re virally affected by the LGPL license. I think the LGPL is very explicit in this: you extend LGPL work (rather than using it as a lib) your work must remain LGPL. That’s the restriction you place yourself under by the fact that you’re using someone else’s LGPL’d work.
So it seems it does indeed matter whether their work was written from scratch or not.
November 11th, 2003 at 7:04 pm
No, that’s GPL.
November 11th, 2003 at 7:11 pm
Sam:
The CMP team of one apparently was a problem at JBoss. Bill Burke went on record saying that the CMP engine was a piece of shit. My read is that the Hibernate move was planned all along to replace the original code.
Besides they are not talking about CMP code, they are talking about invocation and log4j which seems to be identical pieces of code. I for one, think it is a good thing that an open source player stands up, has the resources (read lawyers) to avoid getting cluster fucked.
Good article hani, Apache being called “thieving twats” is the funniest comment I have seen in a long time.
November 11th, 2003 at 7:45 pm
Blah blah blah… I hate JBoss… Blah blah blah blah… I don’t understand what is going on, but JBoss is doing something and therefore it must be bad… Blah blah blah… I am obsessed with Marc Fleury… Blah blah blah… Why doesn’t Marc ever ask ME to … uh … never mind… JBoss is bad… Blah blah blah…
November 12th, 2003 at 3:05 am
Blah blah blah I love jboss blah must never leave the hive.. blah blah it’s good because .. blah blah.. everyone is just jealous blah blah.. they don’t understand to appreciate Fleury blah blah must use dumbass container blah.
You all Jsco lovers need to pick individual nicknames, or is ‘Anonymous’ HiveStandard(tm) ?
November 12th, 2003 at 4:38 am
Dark Shadow,
The reason I mentioned the CMP developers at all was merly to show that many of the people of jumped ship would of been the original authours of large parts of JBoss, and as such still retain the rights to release their code under alternative licenses. As to the CMP implementation being shit I cannot comment as I have limited experience with other J2EE containers. Hibernate itself is not of course a drop in replacement for the current CMP implementation – it isn’t itself an implementation of the CMP specification. There are obvious areas where parts of Hibernate could be used but it would require significant rework.
November 12th, 2003 at 6:02 am
CMP implementation is “shit” ? Is Bill Burke telling his customers this, or just Dark Shadow ?
Saying one piece of code is shit would always strike me as a bit harsh and not thought out. Things are written the way they are in particular circumstances, so blaming whoever wrote it is always lame. The guy who wrote it maybe wrote it for EJB 1.1, and then had to adapt it to 2.0 rapidly. You cant slag someone’s work off like that and expect to retain any respect people have for you. But then we are dealing with the “Fleury collective” here :-)
November 12th, 2003 at 6:12 am
It will be interesting to see if the comments made by Bill Burke (assumng thyey were made at all – I have no proof!) were done so before or after the CMP guys jumped ship firstly to the Core Developers Network and then to the Geronimo project.
November 13th, 2003 at 1:59 pm
Go read the geronimo-dev list [http://nagoya.apache.org/eyebrowse/BrowseList?listName=geronimo-dev@incubator.apache.org&by=date&from=2003-11-01&to=2003-11-30&first=1&count=244]
and you’ll find that not only did the Log4J code originate directly from Log4J, the invocation code that is in question was actually originally written by one of the Geronimo committers. Additionally with the invocation stuffs – some of these were originally contributed by one of the geronimo committers to the EJBoss (does that tell you how long ago this was?!) project in a somewhat different form. This is all just a bunch of FUD.
“First they ignore you, then they laugh at you, then they fight you, then you win.” – Mahatma Gandhi (1869-1948)
November 13th, 2003 at 8:14 pm
My understanding was that the code in question was still in incubation, and that nothing would be released under ASF until any leftovers of LGPL code that were being temporarily used as scaffolding were removed. If that’s the case, then the weren’t really trying to change the license, they were just using the code internally in predevelopment; the pieces of the server that they hadn’t worked on yet were being propped up with similar code that they were all familiar with.
Don’t know if that’s 100% ‘legal’ or anything, but the spirit of it still seems kosher to me.
I loved Burke’s code examples, though. Just about every book or article I’ve seen on design patterns used the exact same naming conventions for the items he pointed out, or they were straight from example code from the libraries they leveraged, like log4j. Who should be threatening who(m)?
July 13th, 2005 at 9:00 am
This page looks weird in firefox (the center content seems not to adapt to the page width, therefor the right column is overiding the center content…)
:S