On 19 Aug 2008 at 7:35, Eric Dannewitz wrote: > A relational database probably is not the best thing to use for doing this > sort of thing. Unless you want to mine your library for stuff, or have > thousands and thousands of things in there > I'd say set up some sort of card system. That would the easiest thing to do.
I would strongly disagree. The simplest example I could give of why you'd want a relational structure is if you're searching for repertory for a specific ensemble. If you use a flat cardfile-type structure, you're forced to do some kind of tricky searching within a single text field, where you have no idea how the instrumentation has been entered (unless there is strict enforcement of how the instruments are entered in terms of order and spelling. On the other hand, the "spreadsheet" structure, where you have a column for each instrument, is just as bad, since you have to search across multiple columns. With a table representing a many-to-many join between your music table and your instruments table, you don't need to know the order of the instruments nor do you need to word about variant spellings. You'd query in just one place. This is both more convenient to use and easier to set up a user interface for others to use. And in the long run as your database grows, it will also be vastly more efficient than the other methods. That said, *something* is better than nothing for tracking your library. A friend of mine recently started cataloging his record collection by simply typing up a list in his word processor. This gives him a searchable document where he can find things much more easily than he could just scanning the shelves. But he still doesn't have 100% certainty that he can find everything that matches what he wants at a particular time, nor can he combine criteria for searching. But it's certainly much better. So, a flat file might be a decent place to start and then break out data into relational tables as you find that the flat file causes problems with providing needed functionality. But one warning: cataloging music is more complicated than just about any other data structure I've ever encountered. I would definitely *not* start out trying to create a model that would be a "perfect" data structure for cataloging music, because you may not need that kind of precision structure. The main point is that the way you structure your data will determine how easily and reliably you can find particular pieces, unless you're content to search on simple criteria like name and title. As soon as you need something more complicated than that, it gets tough (and even title searches can be problematic unless you enforce uniform title entry somewhere along the line). -- David W. Fenton http://dfenton.com David Fenton Associates http://dfenton.com/DFA/ _______________________________________________ Finale mailing list [email protected] http://lists.shsu.edu/mailman/listinfo/finale
