Ian Bicking: the old part of his blog

Sun Comes Lately

Did anyone else see Sun's hiring of the JRuby developers as a pathetic attempt to catch up to Microsoft?

That sounds inflamatory. Sorry.

I mean:

  1. It's pathetic that Sun is trying to catch up to Microsoft on open source dynamic language environments. Catching up to Microsoft?
  2. It's also pathetic because they probably could have hired Jim Hugunin years ago when he was working on Jython, and both made Jython a well-supported tool and the JVM a more dynamic-language-friendly environment. They're playing catch-up now when they could have played themselves as a leader before.

So, apparently lacking any vision whatsoever, Sun waits until now to actually put real effort into the dynamic language community. Of course better late than never, and I have no criticisms for investing in JRuby (nor is this a criticism of JRuby or Ruby), but but the whole timing comes off as merely reactionary.

Created 08 Sep '06

Comments:

I think you're right, Ian. To me sounds like they happened to notice the virtues of dynamic programming and trying to fill the gap. What I despise most in this all stuff is their developers yelling "jruby is cool!", "ironpython is cool!" when the day before they were trowing mud onto Python or Ruby because "dynamic is slow" or whatever. Anyway, let things be :-)

# Lawrence Oluyede

I don't. Or at least not wholy.

Sun is a big company, requiring many people to work in concert to achieve anything. While it is possible that IronPython was part of the motivation for throwing support behind JRuby, I doubt that it was the largest motivation for the majority of people involved. Also, from Sun's point of view, getting behind JRuby also makes more sense than getting behind Jython, on account of all the Java programmers currently getting excited about Ruby and Rails. I'm hoping that much dynamic goodness will come of this for JDK 7 or 8.

Still, I agree that it is was short-sighted of Sun to pass on getting behind Jython back in 2001 or 2002.

# Alan Green

Honestly I don't know what Sun would do to get behind Jython now; the project doesn't have a champion anymore. So I don't fault them for getting behind JRuby now instead of Jython. But I think they were pretty dumb to let Jython wither in the way it has. And I don't know what it takes for Sun to get behind something, but all they're doing is spending a few bucks to hire some people so they can focus on a project that should be strategic for Sun. This is exactly the kind of thing that should be easy for Sun to do.

# Ian Bicking

You might want to look at Tim Bray's blog on the subject:

http://www.tbray.org/ongoing/When/200x/2006/09/11/Other-Languages

Sun is a big company. Tim was the guy who made it happen, and he has some really cogent comments on why it was JRuby, and what would be needed for Sun to hire another language guru.

As the Python FIT maintainer, I've had a request to make it run under Jython. I'm not going to do it for two reasons, one of which is that I'm not a Java guy. The whole ecosystem is a foreign language for me.

The other is that my current support base (minimum Python level) is 2.3. AFAIK, Jython is still stuck on 2.1. If someone wants to step up to the plate and make it run with Jython (or IronPython or PyPy for that matter) I'll certainly discuss applying the patches. However, I'm not going to go back on the minimum level of Python required.

The takehome is that if you want Sun to support a couple of Jython developers, talk to Tim. But I suspect that the people in question will have to show their interest in very concrete ways first.

John Roth

# John Roth

[Ian] > Honestly I don't know what Sun would do to get behind Jython now

Simple: Put some money into paying some of the existing jython developers. It's too late to get Jim Hugunin, it's too late to get Samuele Pedroni, and I don't know whatever happened to Finn Bock.

But it's not too late to talk to some of the newer guys, like Frank Wierzbicki, or Khalid Zuberi, or Charlie Groves, or ....

I've saying this on the jython lists for several years now, with the most recent version being

Sun will have snatched defeat from the jaws of victory. http://sourceforge.net/mailarchive/message.php?msg_id=36335251

It is somewhat pathetic to see Sun playing catchup with Microsoft: they should know better by now.

Alan.

# Alan Kennedy

Sun has a history of letting cool little language work go unnoticed.

15 years ago they had an entire windowing system with a built-in forth interpreter (NeWS) and let it die in favor of the whole X / open look / Motif mess. At the same time more or less completely ignored the cool perl and shell extensions that were going on (anyone remember windowed Ksh?).

They had a great scheme shell with all sorts of neat stuff but it languished and died at the same time MSFT was teaching the world to love VB and fostering an entire VB addon market. They did learn their marketing lession from that somewhat with Java while completely ignoring python. They had Tcl for a while but never did anything with it.

Now they are starting to notice python and ruby, which is good because at least the python and ruby developers have a good history of taking cool features from functional languages little by little. But where of all places is some of the best functional language work being done these days? Give you a hint (it's not Sun).

# anonymous

Unlike MS, Sun makes $$$ from selling hardware. Unlike IBM, Sun doesn't/can't provide the whole solution to the customer. For years Sun nerver gets the application software right. Otherwise there won't be weblogic and websphere. When the Sun management has no vision on how to make money on application software(they are a hardware vendor), what you can expect that they can get the programming language right. Tim Bray is exceptional. He would have done lots of good to Sun if he's joined in 1998. I love /usr/sbin/dtrace and /usr/sfw/bin/python.

# Victor

15 years ago they had an entire windowing system with a built-in forth interpreter (NeWS) and let it die in favor of the whole X / open look / Motif mess.
See the wikipedia entry on possible reasons for the demise of NeWS. In short 1) X was established before NeWS, 2) NeWS required licensing from Sun, and 3) it practice it proved difficult to use.
# Andrew Dalke

Sometimes you lead, sometimes you follow. Business as usual. ;)

# Sérgio