Do we need a central code repository?

[Joe]

Hi everyone,



I have always thought that http://www.cpan.org/">//http://www.cpan.org/ was a great idea and bespoke the maturity of the Perl language.



I had started to build a central code repository site for newLISP and then hurt my wrist so I'm trying to minimize by keyboard use.



In the meantime, I thought we might discuss the need for a repository and what features it might need, besides being able to browse and search for files?



Thoughts?



Joe

[cormullion]

I haven't given it much - or enough - thought, yet. Obviously there's no harm in having another source for useful code -  but there are some negatives as well.



A lot of newLISP projects are enthusiastically started but quickly become neglected; people's initial enthusiasm wanes, and they move on to new pastures and new languages. For example, somebody once suggested that the newLISP source code should be available in a public repository. It was duly done; but it was never updated, and now languishes unchanged since March 2008 (that's ages in newLISP years!) and obsolete. And consider the various wikis: after the hard work involved in setting them up, they quickly become neglected and unloved. Incipient newLISP projects rarely gather much of a crowd.



In the absence of a code repository up to now, people have made their own arrangements for distributing their work, either on static sites or places like google code. I'm not much of a module writer myself, but I can at least see the benefits of being custodian of my own code in this way; I keep it available online while I consider it useful or functional, in a form which pleases me, I can withdraw it from the public gaze whenever I choose, and I don't have to seek anyone's permission or approval. Of course, I'm happy to store code anywhere else as well - but it's too easy for multiple versions to get out of step.



I think you're almost right when you say that cpan.org reflects the maturity of the Perl 'language'. For 'language' read 'user base'; that's why newLISP doesn't have anything similar. The newLISP and Perl user communities have very little in common.



In other words - great idea, give it a try, do it because you want to, and don't expect to be swamped with contributions!