Re: joint occurrences?
Michael Bernstein <[email protected]>
| Newsgroups | gmane.comp.infodesign.facetedclassification |
|---|---|
| Message-ID | <[email protected]> |
Eric Scheid wrote: > On 1/11/03 9:51 AM, "Michael Bernstein" <[email protected]> wrote: > >>It would also let you search for "that episode that has a character >>occurrence of Evil Bob in leather pants and an occurrence of That Midget >>Guy in slinky nightwear". > > Not really ... you still wouldn't know that Evil Bob wears the pants in that > relationship, only that there are two characters there (Bob, Midge), and > someone is wearing pants, and someone is wearing slinky nightwear (possibly > the same someone, at that). > > I've thought some more on the problem, and "granularity" is almost the right > answer... what I really need to do is build a new database/collection of > "appearances", rather than "episodes" or "scenes", where each appearance can > be classified as to which character, dressage, episode, scene, role (lead, > minor, support, extra, etc). Each "appearance" entity would, by definition, > only contain at most *one* character classification (or possibly none, if > the character is unidentified). If you re-read my email, you'll see that that is more or less what I said, except the I used the phrase 'character occurrence' instead of appearance. I could have been a bit more clear that the episodes broke down to scenes, and the scenes broke down to one or more 'character occurrences', which were the subject of classification. > So, the fundamental question is "what are the *things* which we are > attempting to classify?" ... > > Thus, there would be one database of "episodes", and a separate database of > "appearances", where in one the episode is the entity being classified, and > in the other it is a quality of classification. Umm. Not every relational database task should be subject to reformulation as an exercise in classification. Episodes *contain* (or are composed of) scenes, which *contain* appearances (or, if you like, you can have appearances contained directly in episodes). This is a classic one-to-many relationship between two tables in a database. There is no need to go through additional contortions to make episodes a type of classification that appearances are subject to. > Marrying these two databases is left as an exercise for the student ;-) Make them two tables in the same database, not two separate databases. Cheers, - Michael Bernstein ------------------------ Yahoo! Groups Sponsor ---------------------~--> Rent DVDs Online - Over 14,500 titles. No Late Fees & Free Shipping. Try Netflix for FREE! http://us.click.yahoo.com/vhSowB/XP.FAA/3jkFAA/0bmwlB/TM ---------------------------------------------------------------------~-> ---------- Thanks for playing. To unsubscribe from this group, send an email to: [email protected] Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/