The Apache/Sun Relationship Worsens

d6y writes "Over on the O'Reilly weblogs there's an entry on the relationship between Sun's Java Community Process and Apache. Sun have been rubbing people up with wrong way (the problems of licensing open source J2EE containers; stuts v. JavaFaces; log4j v. JDK 1.4 logging....) and I hope this gets sorted out real soon. See also the original VNUNet article and Apache's position paper."

Comments

Apache/Sun

[Orville]

A couple of thoughts:

While it does matter in the aesthetic that Sun is restricting certification of open-source J2EE platforms, fortunately Sun has not taken drastic positions of 'shutting down' JBoss or anything like that. This letter from Marc Fleury seems to clarify the exact issue with JBoss.

This seeming 'rivalry' between Sun & Apache is not as clear-cut; Many of the Jakarta contributors are Sun employees and engineers. (Tomcat/Catalina is used as the 'reference implementation' for the Servlet/JSP specifications.) For more on this, check out the former 'open source guy' at Sun: James Duncan Davidson

Re:Apache/Sun

[acroyear]

Yes, many Jakarta contributors are from Sun, particularly with regard to Tomcat/Catalina. But many are not.

Outside of Tomcat, IBM has been the biggest commercial vendor contributing to Apache, including core items Xerces, SOAP, and Log4J, and in each case Sun has refused to let Apache be the de facto standard (inspite of that being what i believe most java developers want), and has instead put out alternative APIs which its embedded with the JDK, thus locking everybody w/ 1.4 down to a standard that now has little room to grow.

And in some cases the 1.4 bundled version is inferior to the Apache effort (yes, Crimson was/is better than Xerces 1.x because of IBM's redundant i18n overhead, but its NOT better than 2.0; and jdk1.4 logging is definitely not nearly as good as log4j).

But as M$ has shown, people will stick with whats bundled, even if inferior, 'cause "its just easier that way"...

...& all java developers will suffer.

Also, yes Tomcat is the "reference implementation" and all that, but that's because Sun has also released the code for watchdog, which is the standards-compliance testing tool for Servlets. What Apache and O'Reilly want to see is a similar effort to release the standards-compliance tests for other J2EE software specifications like EJB, and they consider it hypocritical for Sun to hold onto those tests in exchange for as much money as they charge for them.

This has been around for a while now...

[mmcshane]

here's a thread (J2EE considered harmful) on the jakarta-general list that precipitated the Apache statement.

Re:The Sun/OSS relationship

[acroyear]

The irony being that in their early days, when SunOS was only a minor-varient away from being pure BSD (thanks, B.Joy), they WERE actually giving away the OS and its features (like NFS), hoping on always being ahead of the competitors in speed to encourage the hardware sales that kept the company on top during the late 80s to mid-90s.

Things were slowly changing by 1991 with SunOS 4, then with 5/2 they had to definitely switch to a "buy it only" since they themselves paid so much for getting SystemV in the first place...

of course, just about every single one of us Sun users at the time were furious with the switch...Sun boxes to me are still crippled in speed because of SystemV's overhead compared to BSD, and the speed of BSD x86 boxes over SCO & other SystemV-based x86 releases just rubs our noses in it even more...

Re:This is most typical of Slashdot

[ChaseTec]

>> 2. They "adopted" the free and entirely non-sun code base for Java Servlets (Jakarta) and claimed it was the "Sun Reference Platform"

Yeah, they never did anything for Tomcat did they? (sarcasm) A few of the developers for Tomcat were Sun employees until recently. Did you bother to check any of your other rants?