Sun explains open source at EclipseCon (NewsForge) [LWN.net]

Here is Joe Barr's latest report from EclipseCon. "Simon Phipps, formerly an IBM employee and for the last two years Chief Technology Evangelist for Sun Microsystems, gave his EclipseCon keynote address Thursday. As you might expect, he took exception to a couple of remarks from Wednesday morning's keynoter, Michael Tiemann of Red Hat. Phipps's talk was on "The Business of Open Source." It was interesting, almost a through-the-looking-glass experience, to hear a suit from Sun stand on a stage and try to explain open source to an audience of mostly proprietary developers. It should be noted that the ballroom where the keynote was given was not far from Disney's Fantasyland."

(Log in to post comments)

Sun explains open source at EclipseCon (NewsForge)

Posted Feb 7, 2004 8:04 UTC (Sat) by Duncan (guest, #6647) [Link]

> "It should be noted that the ballroom
> where the keynote was given was
> not far from Disney's Fantasyland."

Indeed. Take a look at these quotes from the article.

> Phipps said: "Open source was triggered by
> the release of Java in 1995."

Yeah, right, and Al Gore invented the internet, too!

> He also disputed Tiemann's claim that Java
> was a "religion," noting that both Perl and
> Python have "spiritual" sides to them.

Well, he DOES have a point, there, but at least their "Bible" isn't locked up
in some language only a privileged few have access to, but rather, it's "open"
and available to all that care to look at it. Talking about which..

During question time, he was asked

> why doesn't Sun open source Java?

>His answer [] was that he believes Sun has open sourced Java.

Of course, my response is if that's the case, where was the source when I
transfered to AMD64, and needed it to compile a 64-bit VM?

Fantasy land.. indeed!

Duncan

Sun explains open source at EclipseCon (NewsForge)

Posted Feb 7, 2004 11:52 UTC (Sat) by oak (subscriber, #2786) [Link]

> "Open source was triggered by the release of Java in 1995."

Heh, what a joke.

One of the strongest still going Open Source software systems is TeX. It's still widely used both by commercial and academic institutions, originally released at the start of 80's. I think its last bug was reported/fixed 1989 i.e. 6 years before 1995. I wonder when Java reaches the same stability level . :-)

Sun explains open source at EclipseCon (NewsForge)

Posted Feb 7, 2004 12:03 UTC (Sat) by oak (subscriber, #2786) [Link]

And here's a link telling a bit more about this:
http://truetex.com/knuthchk.htm

(1989 figure was not quite right, but anyway...)

Sun explains open source at EclipseCon (NewsForge)

Posted Feb 7, 2004 13:13 UTC (Sat) by khim (subscriber, #9252) [Link]

Obviously Java will never reach the same stability level - and that's Ok. The only reason for TeX stability level is two things:

Total lack of specifications (TeXbook is only translation of TeX code from machine-readable language to english, nothing more)
Deliberate refusal to solve new problems (color support, big character sets, right-to-left writing, etc, etc).

So instead of changing instrument which becomes easier and easier to use with time we have layers over layers of addons to solve bunch of problems TeX does not solve and all this mess becomes more and more awkward to use over time.

So the only thing to compare TeX with is JVM. And not just any JVM but JVM without JIT and so on (in TeX case such things are part of "system-dependent code" and bugs there are never declared as TeX bugs - correctly, of course). Java support libraries, compilers and so on are more or less like TeX macropackages.

And guess what ? JVM core is very stable today (may be not up to TeX level yet but close) and all problems are with java libraries and GC. Just like TeX is superstable but macropackages are full of quircks.

Think about it. If you try to do something really tricky with XyPic and TeX crashed with "out of memory" exception (yeah, yeah, it does not crash to OS but will just print error message and nicely exit - what's the difference though?) - it's not bug in TeX, right. When you call DNS bazillion times per second and resolver library crashes JVM - it's claimed as bug in JVM! Why so? It's relation to JVM is more or less the same as relation of XyPic to TeX...