Hmmm, here's a *very* generic approach that is most likely too difficult to implement. I'll post it nevertheless.
Remove all the auxialary information from the anime table so that only title, type, number of episodes, year and whatever status flags you have/need are left.
Create a new table that lists allowed auxialary info-types (auxinfo), such as "Producer", "URL", "Animenfo URL", "Image" and whatnot, with three flags: singletons / multiple, primary / secondary, listed / unlisted.
Create a new table that contains generic information for all auxinfo bits, e.g. the auxinfo ID, associated anime ID (for singletons, -1 for multiples), title, text, URL and storage class.
Create a new table auxinfo-anime-rel that implements the relation between animes and multiple-auxinfo bits.
Primary singleton auxinfo bits are the Synonyms, Shortnames, URL, Animenfo URL, Prequels and Sequels and the Image(s). These are listed on the Anime page just where they are now. Note that an anime can have multiple singleton auxinfo bits.
Secondary singleton auxinfo bits are listed in a separate box on the anime page. For example, the MyList box can be moved to the right of the description and these auxinfo bits are listed below that in a "Other Information" box.
Now for the real beef...
Multiple auxinfo bits are what implements the collections that were mentioned earlier. They aren't associated with an anime directly but through the relations table. Primary multiples are shown in the same area where the primary singletons are listed. Typical entries for this would be "Producer" or "Distributor". Finally, secondary multiples are shown in a third box below the primary singletons (labeled "Crossreferences").
The "Collections" are created by creating a new secondary multiple auxinfo and assigning it to a number of animes. All animes in this collection will show a link to the collection in their "Crossreferences" section, and an additional Collections list can show all available multiple auxinfos that are of an auxinfo type that is flagged as "listed".
Finally, a note on the storage class thing mentioned above: If a text field, title and URL aren't enough (title and URL are used for things like the AniDB URL or images), the storage class can tell the script to use a specific Perl package to obtain more information on how to handle that auxinfo bit. This would include the name of another table that stores extra info for auxinfo bits of that type and a hash with info about the fields of that table.
Whew, I'm done.
As I said, it's a very long shot and I don't expect it to implemented, but hey, you might want to keep this in mind for the 2.0 version of AniDB.
