About two weeks ago, Max reported a bug with Email::MIME::Encoding related to its handling of QP and line breaks. QP and line breaks provide a totally insane problem that I talk about in my upcoming OSCON talk. It’s a mess.

I have been obsessing over how to number versions since I first started writing code for a living, around 2000. Nothing is satisfactory. I’d really like to make a decision and stick to it across all my code, and I’m pretty close to doing that.

I am an idiot.

I really wanted to do some new talks this year, but no really spectacular new ideas seemed like good ones to promise. I think “Email Still Hates the Living!” will have to wait for next year. Instead, I submitted a new talk on Rx, a talk on various new email libraries released in the last year, and my long-threatened intro to git, “Git is Easy!”

Rx is my schema system for validating data in a highly extensible and portable way. It is in the same problem space as JSON Schema and Kwalify (both of which, I think, Rx exceeds), and close to RELAX NG. I wrote it last summer, and I’ve used it quite a bit since then, and I have not been unhappy with it… except for its error reporting.

At both the Oslo and Birmingham QA Hackathons, my big project was the CPAN Metabase. I’m hoping that in 2010, I’ll be working on something new, but only because I like variety – I really like the product we’ve got and I believe it can be a great tool for a lot of problems. Here’s a simple overview of what it is and how we plan to use it.

The QA Hackathon was, just as it was last year, fantastic. It felt really productive, and left me feeling really professionally refreshed and ready to try to get more things done in general. How long until I am crushed back into grist beneath the grinding wheel? Only time will tell.

I’m getting ready to write up events of the Birmingham QA Hackathon. One of the outcomes was that I ended up really wanting the device I am about to describe. Bear with me.

Almost a year ago, I wrote to p5p asking about overloading the arrow operator. Specifically, I wanted to be able to change how methods were found and invoked. My most pressing motivation was to help me deal with my undying hatred of AUTOLOAD and UNIVERSAL (and, especially, their interactions), but it had other obvious benefits. Without needing to cater to the idea of “classes are packages” you’d be free to do all kinds of horrible awesome things.

When I left home for college, I planned to take my heavily, hackily upgraded 386 (actualy a Cyrix Cx486DLC) along with me. It had two full-height ESDI hard drives, and the weird, exposed ribbon cable on the back of one broke about two weeks before I left home. I quickly sold a bunch of things I owned, begged some cash from my dad, and purchased an AT&T mid-tower (a Globalyst, which later became an NCR line). This Pentium 90MHz, marvin, lasted through most of college, and was finally replaced by a Compaq that looked good in the store but turned out to be a real pain.