How would you like your Classical Albums sorted for display?

[Andre Gosselin]

(13-01-2014 20:29)winxi Wrote:  

(13-01-2014 11:58)DavidL Wrote:  As an example to illustrate the 2 alternatives I'll use Mozart's 27 piano concertos.

Would you prefer the sort order

'Best of the 80ies'
'Best of the 1960ies'
'Best of the 1970ies'

or

'Best of the 1960ies'
'Best of the 1970ies'
'Best of the 80ies'?

Your example is a valid one which illustrates a case where giving special treatment to numbers when sorting strings may produce unnatural results. But it misses the point if you mean by this example to demonstrate that "natural" sorting of numbers inside strings is to be considered a bad practice which should be condemned.

As always happens in this world of ours, there is no perfect solution to any issue, and the one raised here is no exception. The question is : how large is the number of cases where "natural" number sorting would be beneficial, vs the number those where it would be bad.

1. As the original poster initially wrote, caring for the special needs of classical music is the reason MinimServer was created in the first place.
2. In classical music, it is easy to give *thousands* of examples where numbers in composition titles should be correctly handled in order to be properly sorted.
- around 1020 or so works in the J.S. Bach BWV catalogue
- 555 keyboard sonatas in the Scarlatti Kirkpatrick catalogue
- 104 symphonies in the Haydn HWV catalogue
- more than 600 compositions in the Mozart "K." catalogue
- more than a thousand cantatas by Telemann in the "TWV 1" catalogue alone
- 32 piano sonatas by Beethoven
- etc...
All the above works share the common property that they are identified by numbers in their composer catalogue, and that reference to the work catalogue number is essential to precisely identify a work.
3. So, would one prefer to see this :
Beethoven: Piano Sonata No.1 ...
Beethoven: Piano Sonata No.2 ...
Beethoven: Piano Sonata No.3 ...
...
Beethoven: Piano Sonata No.10 ...
...
Beethoven: Piano Sonata No.20
...
Beethoven: Piano Sonata No.30
...

or this :
Beethoven: Piano Sonata No.1 ...
Beethoven: Piano Sonata No.10 ...
...
Beethoven: Piano Sonata No.2 ...
Beethoven: Piano Sonata No.20
...
Beethoven: Piano Sonata No.3 ....
Beethoven: Piano Sonata No.30
...
4. Of course, one can left-pad the composition number with a sufficient number of 0's (3 of them for Bach BWV 0001 to 0009, 2 of them for BWV 0010 to BWV 0099, and 1 for BWV 0100 to BWV 0999) to achieve a correct sort. But this is just not the way it is done in a music catalogue, and I have yet to find a single CD booklet where Bach works are numbered with those leading 0's.
5. If one does not mind imposing the use of those leading zeroes in the music tags to achieve proper sorting of Bach's more than one thousand works (and similarly for Scarlatti, Telemann, etc), then why by the same reasoning one could not be asked to rename the album "Best of the 80ies" to "Best of the 1980s" when tagging the album title, to achieve correct sort within the "Best of xxxxies" group ? If the answer is "No! This is not the exact title used by the disc publisher", then it can also be answered that there is no such work as "Scarlatti Keyboard Cantata K.007 ..." on CD anywhere. Just "K.7" .

To conclude, I do not mean to say that one should loose sleep over the inconveniencies caused by a pure lexicographic approach to sorting : those can be worked around (as I myself do, with the use of leading zeroes for ex.). There may also exist (I am almost sure) more important priorities in the MinimServer todo list than the implementation a new sorting algorithm. But it matters to me that the discussion proceeds on a correct basis. In my humble opinion, it is really unfair to ditch the possibility of "natural" number sorting inside titles as used everywhere in classical music just because of the supposedly "inexact" sorting of the "Best of the 80ies" album this would provoke. Unless there are thousands of cases similar to this one (which I doubt somewhat, but I may be wrong ...)

It would be very helpfull if MinimServer users could contribute other similar examples showing how bad, or how good, a eventual change in the sorting algorithm could be.

Regards to all