From: <Gajdos_at_metatem.net>
Date: Mon, 2 Aug 2004 17:59:43 +0200 (CEST)

On Mon, 2 Aug 2004 at 14:28, Stuart Tedford wrote:

> > > There is nothing "natural" about it and I disagree with it at a
> > > fundamental level.
> >
> > Lol. We obviously went to very different schools. I was always taught
> > 12 comes after 2 not between 1 and 2. ;)
> >
>
> Not if we are dealing with strings, which is what filenames are. This
> "natural sort order" is actually breaking a second standard (ASCII), as
> well as breaking the leading zero standard. Maybe you think 2 wrongs
> make a right, I don't.

Even though a number is part of a filename, it is still a number, here
representing a position in a list.


> I didn't really want to start a flame session or anything (so I shouldn't
> have made the AOL comment, so sorry for any offence I caused), but I feel
> strongly that standards are usually there for a reason - i.e. because a lot
> of people think something is a good idea. So anything that breaks a
> standard should have a very good reason to do so, and I don't consider this
> to be one.

(Feel free to point to a standard that says something about MP3
numbering.)

Seriously, I admit that I myself always number numbered MP3's with a
leading zero, but there are occasions when I receive files that are not
numbered this way, and with the "natural sort" those files are presented
in order, the way I (a human) like them to be presented.


So, for a directory with 1500 files, the time spent sorting is:
strncmp: 84 ticks
strnatcmp: 119 ticks

If it is considered to be of major importance, the strnatcmp code can
surely be optimized to cut down the difference, so no new option would be
necessary.

/JG
_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox
Received on 2004-08-02