Subject: Re: Free Software vs. Open Source
From: Ben_Tilly@trepp.com
Date: Wed, 13 Feb 2002 19:07:41 -0500


Simon Cozens wrote:
> Ben_Tilly@trepp.com:
> > And my understanding was that this episode was responsible for your
> > Sapphire project.  To my eyes that is one of the best demonstrations
> > that I have ever seen of the value of lowering licensing barriers to
> > code reuse and taking full advantage of open development techniques.
>
> No. Sapphire won't happen because GNU libraries just ain't portable
enough.
> Perl 5 ran on over 80 platforms; most of the libraries out there we want
to
> use struggle to run on anything that doesn't look like Unix. *That*'s why
we
> have to reinvent lots of wheels.

I hadn't realized that Sapphire was ever thought of as anything beyond
a one week proof of concept and code reuse.  It was a good demonstration
of what is possible in that time.  It is sad that the wheels you are
reusing were not flexible enough to be used on the full range of what
Perl supports, but there are many other open source projects which
would find where you got very acceptable.

I haven't been following the development of Perl 6.  I do know it is
being built on Parrot, which is intended to be a good platform for
building scripting languages.  However I think it would be very nice if
Parrot contained reusable libraries that others can use.  Perl 5 had
lots of useful stuff in it that others liked, but it was virtually
impossible for other projects to identify and borrow any smaller
useful chunk than an embedded Perl interpreter.  (OK, mini-Perl
interpreter.)  Which is why, for instance, even though everyone and
their dog advertises Perl-compatible regular expressions, only Perl
uses Perl's RE engine.

I could easily understand if that wasn't going to be feasible, but if
it was then it would be a good thing if the Parrot project began
advertising the availability of highly portable libraries for general
use long before the rest of Parrot (let alone Perl 6) was delivered.

Cheers,
Ben