by daved » Tue Sep 16, 2003 2:22 pm
I read with interest jt004c's proposal for "proportional representation" of genres. However, I just can't get my head round the need for such a scary feature. The only use I can think of where it would be of benefit is to create a sorted list, based on genre, with the hits matching the criteria the most closely at the top.
However, wouldn't a simpler, two or three genre/sub-type[/sub-type] model work just as well if it were possible to sort on any selection of the two or three fields in any priority? Granted, it would be a less effective solution by the time you reached the bottom of the sorted list, probably missing out dozens, hundreds or maybe even thousands of tracks that a weighted search/sort would have captured. Honestly, though, how many people want anything other than the few most relevant matches to their search?
Also, think of the time and effort in cataloguing thousands of tracks like that - and you can bet your last dollar that your view wouldn't match someone elses!
I read with interest jt004c's proposal for "proportional representation" of genres. However, I just can't get my head round the need for such a scary feature. The only use I can think of where it would be of benefit is to create a sorted list, based on genre, with the hits matching the criteria the most closely at the top.
However, wouldn't a simpler, two or three genre/sub-type[/sub-type] model work just as well if it were possible to sort on any selection of the two or three fields in any priority? Granted, it would be a less effective solution by the time you reached the bottom of the sorted list, probably missing out dozens, hundreds or maybe even thousands of tracks that a weighted search/sort would have captured. Honestly, though, how many people want anything other than the few most relevant matches to their search?
Also, think of the time and effort in cataloguing thousands of tracks like that - and you can bet your last dollar that your view wouldn't match someone elses!