To avoid overwhelming Andreas with all my indexing suggestions, and in
order to gain some flexibility to try out New Things without screwing
up everyone's CPAN access, I'm going to try a little experiment.
Hold onto your knees--I'm going to publish alternative CPAN indexes.
This is *not*, repeat, *not* a fork. The new indexes would all point
at the existing CPAN servers and modules. They would simply publish
additional information, different views on the data and try out new
indexing models. The indexes will live on a seperate server and will
be accessable via a subclass of the CPAN shell that has the ability to
read an index from one machine but download from another.
Things I will try out...
- Alternative indexing schemes that I've proposed in the last few days
- A more interesting indexing format, probably YAML
- rsyncing rather than downloading a new index
- publishing index fragments in each modules/by-module/Foo/ directory
- new indexing programs and schemas
- faster indexing by trusting the version # on the tarball name
We'll play around, and merge what works in with the main indexer.
--
Michael G. Schwern <[EMAIL PROTECTED]> http://www.pobox.com/~schwern/
Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One
Schwern Unit: a positive but insignificant quantity