Subject: Re: Successful FSBs
From: "Benjamin J. Tilly " <>
Date: Mon, 04 Nov 2002 21:07:37 +0500

"Stephen J. Turnbull" <> wrote:
> >>>>> "Tim" == Tim O'Reilly <> writes:
>     Tim> But if I can, say, just go to cpan and get the module I need
>     Tim> without writing any code, I've just accomplished a huge
>     Tim> amount of re-use.
> I think you're overestimating how much you've accomplished.  But be
> that as it may....

How much you have accomplished depends on the module
you are using.  There are some really stupid ones out
there.  But there is a list of really powerful ones as
well.  Try Template::Toolkit, DBI, Inline, or
Parse::RecDescent for a few examples.

> I may be underestimating the value of the reuse that's going on in
> free software.  But if so, I'm in good company.  Steve McConnell, for
> example, who concluded (IEEE Software, Jul/Aug 1999) that "open source
> methodology" wasn't ready for prime time.  Nor do I think our response
> (same issue) was particularly convincing to anybody who isn't already
> convinced.  Things have changed in OSS development, especially in the
> leading projects, it's true---but the outside world hasn't been
> standing still either.

I think that the amount of reuse that happens is
strongly dependent upon which community you are in.  I
am not sure how much of this is social and how much is
technical.  My suspicion is that there is a certain
"worse is better" aspect to Perl that discourages local
customization of the environment, which makes reuse
much easier.  Of course Larry Wall has a demonstrated
talent at shaping social communities as well.  (I would
say that he is better at writing social communities
than languages.)

For instance large C projects which wish to be portable
tend to define a slew of macros right off the bat.  But
now mixing two projects is hard because they have
different sets of macros that you have to reconcile and
map.  (And what do you do if either project changes
what some of *their* macros mean?)

> And that's the bottom line, isn't it?  You don't have to listen to me;
> you've been there, you know what can be done, you can make your own
> judgements.  But what you're talking about is convincing people who
> are far more likely to swear by McConnell's _Code Complete_ than by
> ESR's _Magic Cauldron_.  I don't think unsupported claims about "just
> going to CPAN" are very convincing.

I think that people like yourself who dislike Perl or do
not know it well tend to underestimate CPAN.  Here is a
data point for you.  Several prominent Perl hackers have
been heard sayingthat they would prefer to program in
other languages (particularly Ruby) but keep on coming
back to Perl because of CPAN.  IIRC Matt Sergeant is
among them.

That implies that there is some real value being
realized there...

Get your free email from

Powered by Outblaze