PDA

View Full Version : Unicode alphabetization



Corgi
06-19-2006, 11:11 AM
I hope I am stating this right. Currently, the Music Collection has a problem alphabetizing things that are outside of ASCII set. For example, I have Béla Fleck in my collection and it sorts it after all the other B's which seems a bit odd.

Also, would it be possible to add a Sort field for sorting by Artist. It would be nice to have it display Béla Fleck but sort by Fleck, Béla.

donny
06-20-2006, 06:32 PM
although these ideas are nice, and would be usefull, I don't think that they will be done soon...

1) sort is just the way unicode is... it would require a lot of work - manual work almost- to get it straight... and that work can be used on much more urgent things... sorry :(

2) the problem is to find out how to auto detect name/surname ? for many applications it would be easy to assume name"space"surname. But what about bands? "red hot chilli peppers" for instance can't be clasified at all... and even some artists would be problematic like that, i.e any artist with three names (Jerry Lee Lewis, Weird Al Yankovic, Jean Michele Jarre, just to name a few)

ChristinaS
06-20-2006, 07:35 PM
Normally just by the numeric value, É is way larger than E, so it sorts after Z.

But it looks like the collating sequence has been redfined to take this into account at least in some aplpications like Window Explorer and others.

Windows sorts file names alphabetically "properly" if they contain accented characters, so they come rigth after the unaccented character.

MS Word, Excel, also sort columns properly in the same way.

I have several languages defined in my Windows installation. Maybe that's why.

I don't have a music collection defined in dMC - so I cannot tell if it would sort properly or not.

But maybe if you test that windows Explorer can sort file names properly to start with, that could be a step towards a possible resolution.

Corgi
06-21-2006, 11:32 AM
although these ideas are nice, and would be usefull, I don't think that they will be done soon...

1) sort is just the way unicode is... it would require a lot of work - manual work almost- to get it straight... and that work can be used on much more urgent things... sorry :(

2) the problem is to find out how to auto detect name/surname ? for many applications it would be easy to assume name"space"surname. But what about bands? "red hot chilli peppers" for instance can't be clasified at all... and even some artists would be problematic like that, i.e any artist with three names (Jerry Lee Lewis, Weird Al Yankovic, Jean Michele Jarre, just to name a few)

1. I would bet that there are some commercially available algorithm to sort unicodes. A lot of applications seem to support this and I can't imagine that each one of them implements it individually. I know MS Access seems to get the sorting right. I know it isn't an urgent thing. Just an idea.

2) I wouldn't expect it to auto detect. What would be nice is another field so that the user can specify the sort order which could be different from the display name.