j2ee vs. python (and what our evil competitors are saying about python)

[Paul Boddie]

curt at journyx.com (curt finch) wrote in message news:<96c7f32.0207081518.52ea644a at posting.google.com>...
> 

[Quoting a competitor's "analysis" of Python]

I won't go through the points and produce a 300 line post like almost
everyone else has done since, as almost everyone has noted, most of
them are ridiculous and suggest either cluelessness, arrogance or a
vested interest in writing lots of low-level "old school" code,
preferably in C++ if only they had the chance.

It is true that Python is used to describe algorithms and to prototype
software, mainly because Java (for example) is relatively poor at
investigating concepts, getting things working, and so on. However, it
is absolutely untrue to say that Python developers don't tend to
practise code reuse - it's one of the best languages to "leverage your
investment in existing components" (in executive-speak), *especially*
in the Java scene. And it's laughable to start talking about a lack of
J2EE-compliance - Jython is "100% Pure Java (TM)" and works very well
with J2EE components... a lot better than various Java products on
most occasions, too.

> As a new relatively untested development platform in the licensed
> software community, Python has certain deficiencies with which all
> users and developers should be aware.

I suppose when the term "licensed" is used, they actually mean "closed
source, proprietary licensed", but this is untrue as well: the
Infoseek intranet search products, for example, were both closed
source and proprietary Python software before "enterprise Java" was
even at the "short trousers" stage.

Reading the full list, I suspect that the person who wrote it has
never used Python and may even be an inexperienced developer who has
been asked to do some "research" on behalf of the project leader at
your competitor. Things like...

> ·         Python does not provide multiple ways to perform tasks as do
> most other languages.

...sound like a bizarre rephrasing of "there's only one way to do it"
taken literally by someone who doesn't know their stuff.

And this...

> ·         Since Python is relatively a new language, there are only a
> few resources available for information on its latest developments.

...deserves special attention. For Python, there's a shedload of
decent free, quality material out there on the Internet, and should
things not work as expected with a package or module, there are
usually good support possibilities and the source code is most often
provided and is easy to inspect. That's typically two more things than
any given comparable Java technology, in my experience.

Firstly, take a look at any Java support forum and it's flooded with
questions on how to get things to work. Now, one might claim that the
majority of these are from "clueless newbies" who should have read the
instructions, but a brief inspection of the instructions frequently
suggests that certain classes of Java project just don't have decent
manuals. But anyway, such entry-level questions tend to obscure the
threads describing non-trivial problems, thus making it difficult for
the average developer to solve them and get on with their work.

Secondly, it's all very well that Java packages are shipped in binary
form for easy installation, but when things go wrong, the class name,
method name and line number details are pretty useless. On numerous
occasions, it has been necessary for me to investigate configuration
issues with WinZip, just to see which classes are there and what
bizarre properties have been set by default. Ironically, in the
context of this message, the best tool for debugging Java
configurations is called Jython.

I agree with Anthony Baxter - you should name and shame these people,
if only for us to discover if it's a big name consulting house or not.
That would be interesting...

Paul