Most of the time I just potter on the various projects I have running.
A combination of what I'd like to think is a "measure twice, cut once" coding attitude, and distractions from client projects means that although code does get done properly, it often takes me a while to get to things.
But after a year in which my business is finally making enough money to consider moving back to Sydney, and with my main contact on my current "big client project" quitting on Christmas eve, I've promised myself a week long hackfest to try and see how far I can push all these outstanding things and get some of them moving.
After the first three days I've been amazed by how fast you can move when clients aren't distracting you all the time.
The tally so far for the week.
A back-compatible but complete rewrite in the style of File::Spec and aware of the distinction between different types of data. File::HomeDir->my_documents Does What You Mean on both Windows and *nixes.
A largely cleaned up production release of the formerly badly broken module for murphing classes APIs without needing lex-wraping evil muck.
A unified driver-based API in the style of DBI for unifying all of the ways in which you can send SMSs. Super-happy with this module. Really clean code and excellent unit test coverage. Tons of room for expansion into MMS and various other things.
As an initial demo, I worked up a driver for the www.myvodafone.com.au website that lets me send SMSs to people and have it billed to my vodafone mobile account.
Actually, both of these are intended as bribes. The OpenOpenOffice project gains itself an experienced debian packager (and oddly, one of purl's original authors) in exchange.
The most serious attempt at "full blown" XML I've done yet, with full xmlns settings, namespaces, Schemas and validation, oh my! (although due to schema changes happening daily I've disabled validation for now).
PITA-XML (god I love that name) is at the core of the Perl Image Testing Architecture, and should provide a platform/language/format-neutral way for storing the results of Perl distribution testing (such as with CPAN Testers and such).
Also, I believe PITA::Report may have nudged me into the three figures mark for CPAN distributions. Even though mine are broken up into pieces a bit more than some others, I certainly understand how hard it is to keep them all bug-free and not stalling.
The more I get done the more I'm convinced is going to kick ass.