I'm using in reference to a collection of data, however it's stored. It would really help to have a proper sampling of it, so I can see potential ways to migrate it to a more effective storage approach (and link other things together like episode commentary and the like).
As for the migration, I believe that was hardware/software version only, not really format or type. Having said that, the bugs that have been in the text since the migration would indicate it could have been a database rejig.
