develooper Front page | perl.perl5.porters | Postings from July 2006

Re: Its time we set the score straight on Perl 5 and Perl 6 and debunk our own self generated FUD.

Thread Previous | Thread Next
Scott Walters
July 2, 2006 13:21
Re: Its time we set the score straight on Perl 5 and Perl 6 and debunk our own self generated FUD.
Message ID:
Use me as a case study.

my Foo::Bar $baz works now.  That'll make much, much more
usable.  Coro is broken on the 5.9 branch, which keeps me from using it
for quasi-production work, as I'm addicted to Coro.  Whether through
core support or a better userland-only implementation, autobox works
without a core patch now.  When I download a release candidate, usually
they work great, except for sometimes breaking the more intensive modules 
such as Data::Alias.  In these cases, I submit patches for the CPAN 
module -- or try to.  I run Perl on NetBSD Sparc, MacPPC, ARM, and x86, 
and have run Perl on NetBSD Alpha and other platforms in the past.  When
I run a release candidate, I like to think I'm doing a small service 
to other NetBSD Perl users on lesser common platforms.  

Perl is a victim of its success.  Releases work too well.  CPAN authors
have too much independence from what is being done in core.  True, 
problems are found later, but, testing is hard.  It's hard to do 
better than the t/ tests and the tinderboxes.

No one wants to walk in and say "code what I want rather than what you
as maintainers think is needed".  That is, no one reasonable does.   And
this is seldom needed.  

As long as I'm running off at the mouth, I'd like to comment on 
backwards compatability.  It seems every time I try to run a Python
program or Java program, it completely fails with a stacktrace for
reason of not having the exact right version installed.  If Perl were
really out of touch with users, it would have this problem too.  

Speaking as a CPAN author, all I can say is keep up the great work.  Oh.
And refactor core for me so I can jump in without getting confuzeled =)

Actually, this might be a good time to do that -- if Pugs and Parrot are
going heavy on the new big new features, that might mean less pressure for
new features from Perl 5 for a while, and cleaned up core is the thing
I'd like to see from Perl 5 more than anything else.  Especially if the
Pugs Perl 5 backend becomes as popular as I think it will.


On  0, Nicholas Clark <> wrote:
> On Wed, Jun 28, 2006 at 01:59:50PM -0700, Joshua Juran wrote:
> > But what is the problem, really?  Not enough perl maintainers?  Or  
> > are we just jealous that a weaker language has stolen our thunder?
> Well, if everyone subscribed to this list counts as a maintainer, there are
> a lot. It's just that very few are active.
> I think that there's a bigger problem - there's little feedback between the
> users of perl and the core developers. There's loads of activity on CPAN,
> yet very little traffic to perl5-porters from CPAN authors in relation to
> their modules. And very little external traffic apart from bug reports.
> Which gets demoralising. Everyone uses perl, and nearly all of them take it
> for granted.
> As a specific example, when I ask people to test a release candidate -
> F.A. feedback. And even feedback that late is not great - ideally many
> people actively using perl for their day jobs or fun tasks would be trying
> the scripts and modules they maintain or just use against the current(ish)
> development and maintenance branches. But that isn't happening. So
> "interesting" changes go unnoticed, until after releases. Which is bad.
> Hmm. I'm not sure if that actually answered your question.
> Nicholas Clark

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About