[python-advocacy] [PSF-Members] Thanks for Community Efforts on Forrester Research Survey

[Guido van Rossum]

On 6/6/07, Jeff Rush <jeff at taupro.com> wrote:
> I would like to thank everyone that contributed to the recent effort to
> respond to the survey on Dynamic Languages by Forrester Research.  I really
> appreciate the many notes, advice and links I received in private mail and
> believe it gave our community a strong lead.  All material, including the four
> samples of source code provided by volunteers have been submitted to
> Forrester.  They will provide to us a first draft for fact checking on June 8th.
>
> They also later interviewed me about the responses, seeking general input on
> the strengths of Python not covered in the survey form and clarification on
> certain points.  Some of their questions were interesting.
>
> One was how is consistency maintained across the various implementations of
> Python.  What checks are in place to prevent forking of the language and
> feature creep?

The authors/maintainers of IronPython, Jython and PyPy all use
CPython's extensive suite of automated test ("unit tests" is really a
misnomer for the test suite) to check conformance. There are also
areas where the test suite is understood to verify CPython
implementation details which other implementations are free to ignore.

We don't really have a mechanism in place to prevent feature creep in
those versions; I trust their developers (even MS's IronPython, which
is spearheaded by a long-time Python developer). I also expect that
end users would give them a hard time for unnecessary deviations. Open
source helps here IMO.

> Another was about a security committee that watches for security weaknesses in
> implementations and privately fixes them before they become a problem.  I know
> there is such a committee but don't see it given much visibility.

Mail to security at python.org or Google for that exact term. The
committee mostly works behind the scenes, distributing patches to
vendors.

> And the last question, a bit odd, is whether multiple inheritance is unique to
> Python, in the class of dynamic languages that excludes Smalltalk and compiled
> languages.

Gee, I don't know. I always assumed Ruby and Perl would support it
too. Maybe it *is* unique though?

> Soon we'll see how we stack up to the other dynamic languages, but in any case
> it is good to see this class of languages getting improved visibility in the
> IT world.  I think we've also helped Forrester understand the world of open
> source better, and emphasized to them that there is more to dynamic languages
> than the web.

I find the purpose of educating Forrester a lot more important than
seeing "how we stack up". I'm getting tired of people proposing that
Python should grow feature X with no more justification than that Ruby
or Javascript supports it. Or threatening to learn Ruby if their
proposal isn't accepted.

-- 
--Guido van Rossum (home page: http://www.python.org/~guido/)