A better part of last week was spent teasing apart the disgusting knots that I had been letting pass as tagging among my files. The very same knots that made Jinzora damn near unusable and made Amarok look like a dollar store threw up all over MySQL. It took about 24 man hours but I tackled the genre issue along with the problem of null artists and albums, however there still appears to be some lingering issues as Amarok is reporting dead or unplayable tracks. Permissions maybe?

I’m coming to the conclusion that huge collections can be a real nightmare if tagging and preventive maintenance is not addressed up front. Having left mine to evolve over some seven years I am now really paying the price as I try and get it into a functional whole.  Thankfully, there are some solid tools out there as I am leveraging Amarok with a MySQL backend for reporting (N to the E dropping in on the R and D), EasyTag, and Picard (Amarok has it baked in but it is nice to have discrete tools).

To get started I imported everything into amarok and then wrote a group of reports looking for malformed tags, particularly ones using reserved characters like “;” which would cause Jinzora to choke.  Combing through the report for the obvious issues like null genre I used the sort and mass edit feature of Amarok to fill them in which made for smooth sailing.  Tracks that were an absolute disaster I fixed either with EasyTag or Picard, though I found the later to offer up some less than desirable data due to taggers as lazy as I posting to Music Brainz.

At the moment I feel like I am about 80% finished, more than enough to listen to my collection the way that I like but with still enough bugs to get me hot under the collar.  The bonus with a cleaner collection is now we can play music off the server for our greyhound which means I can track his listening over at Last.fm.  Yes, I am that much of a dork.