Alphabetic Sort

[simoncn]

(18-03-2012 19:09)DaveWr Wrote:  A minor issue, but when I look at the artist list, k. d. lang appears at the very end (after the uppercase 'Y'). Note she prefers her name in lowercase. Is the alphabetic listing case independent?

MinimServer does case-independent sorting for some lists (I think the only one currently is the list of tag names). As you've discovered, the sort for tag values is lexicographical (case-sensitive). I agree that k.d. should have her name in the K section, and there are various ways that I could do that:

1) Change the artist sorting to be case-insensitive, but leave all the other sorting (Composer, album title, track title, etc.) as case-sensitive.

2) Change all the sorting to be case-insensitive.

3) Do something special case for lower-case initial letters in tag values (presumably there won't be many of these) and sort these values as if the first letter (only) were changed to upper case. This would mean that "k. d. lang" is sorted as "K. d. lang", and would therefore come after "K. E. Smith" (for example).

4) Leave the sorting as case-sensitive by default, but allow the user to specify a custom sort position for "exceptional" individual values. For example, the user could say that "k. d. lang" should be sorted as "K. D. Lang".

I don't like 1), as it seems arbitrary and inconsistent.

I do like 2), but I'm a bit concerned that there might be some situations where it isn't the right thing to do, though I can't think of any at the moment. I'm also a bit concerned that it might add overhead that would make MinimServer slower when sorting large lists, but having just looked at the code that does this sorting, I think it's likely to be OK.

I'm not keen on 3) because it could still give surprising results (with k.d appearing after K.E).

4) is convenient from a MinimServer implementation perspective because I could use the same internal mechanism that I'm planning to use to enable custom sorting in a future MinimServer release. The disadvantage is that is adds complexity and extra work for the user.

My current inclination is to try 2) in the next beta release (after I've done some performance testing on it) and see how people find it.