But I hate Python's debugger.

We need a better debug interface so that we can fire the pdb to intercept any running Python process just like the way gdb does on C/C++ process.

To me *this* article/discussion is short-sighted.

Jython (and JRuby, that at least has a non-alpha release in this last two years) are still ruby and python.
They have they'r own platform, libraries and so on.
Groovy integrates tightly with java, and groovy's library is Java's one.

Plus, if you ever took a look at the language it has some interesting features (notably, creative use of SmallTalkish/Rubyish blocks, the ability to have both ruby's function call without parenthesis and python's first class citizen functions, real lexical closures like lisp).

I believe many people, even in this discussion are reacting to groovy jsr in a bad way for some reasons:

1: this is a chance to grow *even* for jython, JCL, ObjectScript and so on. Dont' fall in the "they're stealing us momentum!" mood. If the platform wqich groovy is based upon (BSF) grows, and if agile languages are accpeted in the mainstream world (see the JSR) we all take advantage.

2: choice is good. If ruby never existed prolly python won't have iterators (even if the BDFL quotes CLU).
If python never existed probably perl won't have Objects. If AWK..

3: It's not SUN that is pushing groovy. The authors are indipendent devs, thus if the JSR is accepted, that will' a good demonstration that the JCP is really an open process, confirming successes like the GenericJava/Pizza integration.

Please, consider this a chance, not a problem. Try not to be insular. Look at the groovy wiki.

I agree with this article 100%. I'm a somewhat recovering Java junkie from CS who got interested in Python 2 years ago, but only started realizing how cool it can be recently. I've been toying with some ideas on how to work on Python to make it more competive like adding better support for streams.

Some ideas: http://www.blindmindseye.com/bmeblog/archives/000089.html

I'm curious why nobody's mentioned JPE:

http://jpe.sourceforge.net/

I too am a bit comfused why Jython wasn't mentioned in the article, although I would like an easier "CPython" integration route as well for certain applications. We successfully use Jython (and other languages that are translated into JVM bytecode) in our production work.

"""So, Groovy is a great first step. But for Java to really get past the awkward pause in its relationship with open source developers, those keeping the Java flame have to get over themselves and the whole "not invented here" mindset that has locked them in thus far. Architectural purity is great. But pragmatism is better."""

Rule 9 from the Zen of Python:

"Although practicality beats purity."

Sun should remember DoD's failure with Ada. The DoD was too heavy handed in trying to control the language and they ultimately failed in the market place as a consequence. Why should the open source community buy into Java when they have a language like Python That give them more than Java? Python has its own byte code interpreter, why should it envy the JVM? Python integrates functional programming and proecedural programming with a strongly, but dynamically typed OO language. It doesn't force the developer to live in an OO straight jacket. Python also has a very easy to use C API for integrating legacy software into Python modules. Java is hostile to other languages by comparison.

As someone who uses jython in a production application, I took a brief look at groovy to assess it. I find it ironic that it's similarity to java syntax is considered a strength. IMHO, it makes the groovy code look ugly compared to jython code. The ability to paste java code into a groovy code seems of little value since it is usually done the other way around--you tinker at a scripting console and get the algorithm working and then port the code into java if needed. I doubt if groovy would port much easier than anything else since the functionality is different (that's the whole point of a using an alternative language). I have helped several developers learn python sytax so they could start scripting in jython and they always pick it up in a day or two--and then they know python too (and can access it's vast libraries that come along with it) for when they aren't restricted to using a JVM. Functionality-wise, groovy looks like it has a decent blend of features and I didn't see anything I would miss except for python's terse syntax (and triple quote syntax for literals, useful for templating without escaping embedded single and double quotes). But that's based on just browsing the docs. Pedigree-wise, groovy is very immature compared to jython (and even more so compared to python). Hence the consternation on selecting groovy as the subject of a JSR. For me, the choice to stay with jython is obvious--even ignoring groovy's immaturity. I don't just want an "agile" language for the jvm. I want an another language for the jvm that can also bridge me into the python world loaded with tools, apps, libraries and books with a vibrant community. But if you work with developers who won't be comfortable working outside of java's syntax and you only want to run on a jvm and you don't mind the relative immaturity, then groovy is probably a good alternative.

A clarification: Yes, Jython is the fastest route to achieving most of this compatibility. Which is why I'm puzzled by the Groovy project being forwarded as a JSR first; why a whole new language just for the JRE when people are already coding in Python? What I'm saying above is that the JCP should formalize Jython support (and other support for Python executing outside the Java framework) as part of the Java plaftorm if they want to reach a wider audience of Linux developers.

Sorry I wasn't that explicit in the weblog entry; I'll add clarification on my website.

I briefly mentioned Jython in a previous post on the weblog this article was screen-scraped from. It's definitely something in the mix, but it seems, based on what the Java Community Process people are saying, that Jython is not "pure" enough for them, so Groovy is what they're pushing.

I think that's short-sighted.

I agree with the Jython comment above. Why even bother talking about a SOAP marriage when Jython has it all -- and even better, since Jython can use and sub-class any Java class.

Servlets in Python? Yep. SWING apps in Python? Yep. JDBC? Yep. Jython is the best thing Java has going for it.

Check out http://www.jython.org

Rather bizarre not to mention Jython in such an article.