February is a short month and the last couple weeks have flown by. Since my last update, there have been a couple of significant milestones that bring us closer to CT2.0:
- I successfully created a test Metabase backed by Amazon S3+SimpleDB and have created, retrieved and deleted some user profile facts. This built on some of the earlier work by Leon, but reflected the work I’ve been doing to revise the guts of the Metabase libraries.
- I wrote a program that converts a given NNTP report into a CPAN::Testers::Report object (which is itself a Metabase Fact subclass.) This wound up being trickier than I expected due to potential for ambiguous mapping of distribution names to actual distribution files. (Thank you to Tokuhirom, Andreas, Takesako and Offer Kaye for reviewing and fixing my exception list.)
- I generated 900+ Metabase user profiles for all the known CPAN Testers, which will be used to linked with their reports during the conversion process. We still need to figure out a way to distribute these to testers so that new records will link up and other logistics, but this was a necessary step to prepare for conversion.
- I’ve started configuring Amazon EC2 instances – one for the parallel conversion of reports and one for the CT2.0 server itself. I’m still coming up the learning curve on EC2, but see no obstacles (except time)
My immediate goals are (1) to get a test CT2.0 server up and receiving reports so we can start testing it and (2) prep and run the parallel conversion process. I’ve already gotten some help on the #catalyst channel for #1 (thank you, hobbs) and #2 is now mostly a matter of programming it up now that all the foundational prep work is done.
Hitting the March 1 deadline is going to be tight. We’re still holding at about 2 weeks behind the original estimate (where we’ve been since mid-January). But I think progress is coming fast and furious and I hope to get a solid beta launch by March 1 and negotiate with the Perl NOC for an orderly transition as we ramp up the new service.