Re: Extreme Community (discussion)

[koolade]

That sounds to me too like it'd be a good project. The biggest problem would be find a good way to wade through the mess. I have a lot of perl code laying around too, but not all of it is "production quality". I think it's great that you have 80% of what someone's looking for, but will it take them more time to learn your existing code and finish it than it would for them to start from scratch?

I think the code would have to be organized very well, and put into well-defined titles. There's a lot of potential for forking code when you look at smaller programs. Deciding between a script that queries CDDB for mp3s named Artist_Album_Song.mp3 versus a script that looks for Album_Artist_Song.mp3 would be a nightmare. It might be better to focus on larger projects than those that are more on the side of under 1000 line throwaway scripts.

Also, there are hundreds of sites out there where you can download one of thousands of CGI scripts. It's really hard to find good stuff there, and something that fits your needs. This couldn't be one of those sites. Again, maybe it would have to focus more on consolidating duplicate efforts so that there's less noise, and less bad code.

But yeah, if we can solve some of the implementation problems, it's definitely something I'd be interested in.

[da]

I think this should not be a replica of CPAN... This is an opportunity to peer-review code that isn't polished enough to submit to CPAN to be saved for posterity. (Also it's an opportunity to review polished projects, but that shouldn't be the limit for inclusion.) If you catagorize by project, then yes, you'll have a lot of repeated projects that are impossible to choose between. If you catagorize by individual module or script, there will be many thousands of files to catagorize. I'm more in favor of an amorphous, dynamically organized collection, because who really wants to manually sort n-thousand individual scripts into catagories?

The code should definitely be indexed and searchable. Preferably NOT publically on google, because I don't want my grungy five-year-old mistakes being aired in public.

Code should be indexed on either comments and POD, or just on POD. That means the most descriptively documented code gets the best reviews. :-)

Is there any way to replicate google's "page rank" and further weight the result by a "code worthiness" and/or "coder XP rating"? That would help people who are looking for good code to learn from and use.

Finally- incentive to review. I think the XP system is a good one. Reviewers gain XP and grant code worthiness points. Code that's been deemed worthy should be offered upon the altar of CPAN. :-)

Implementation-wise, could this be done as E2, translating directories into nodes and files inside the directories as sub-nodes? (yes, I realize I said "finally" in the last paragraph. But implementation comes after design. :-)