Re: [Banshee-List] [feature suggestion] library structure

[Nicholas Doyle <niskel gmail com>]

On Tue, 2009-02-10 at 15:23 +0100, Oliver Horn wrote:
> Hi there,
> 
> Am Dienstag, den 10.02.2009, 10:06 -0400 schrieb Nicholas Doyle:
> > For the record, these are my library structure patterns. In my mind they
> > are perfect. :)
> > 
> > file_pattern:   "%disc%-%track_number% %title%"
> > folder_pattern: "%album_artist%%path_sep%(%year%) %album%"
> 
> Is %album_artist% that what one can see as "Compilation artist" when
> editing the mp3-tags? Where is this information stored? It's not in the
> mp3 itself, is it?
> 
> > For my last nit pick on the subject, having excessive 'control
> > characters' to separate fields in a file name bother me when they are
> > not necessary. For example "%track%. %title%" or "%track% - %title%" is
> > annoying compared to "%track% %title%". "%track% %title%" remains
> > unambiguous and does not inflate my (often long) filenames by one or two
> > characters; it also looks cleaner in my opinion. 
> 
> I think separators are essential. As I mentioned programs like EasyTag
> can read the file names and fill the tags with it. But it needs
> separators to recognize where one information ends and the next begins.
> So if you have a file named
> 
> 	A Girl Like Me 01 SOS.mp3
> 
> you had to scan with the pattern %album% %track% %title%, but this would
> possibly end up in
> 
> 	Album = A
> 	Track = Girl (or empty because of not being numbers)
> 	Title = Like Me 01 SOS
> 
> Oliver

I agree that separators can be useful (and essential in some cases to
make unambiguous file names (see my separation between %disc% and %track
%), I just like it when you can have name a file name without a
separator and still have the fields be unambiguous. This is why
something like this should be optional.

For example the pattern "%track% %title%" is unambiguous and can be
matched and extracted with a regular expression like in this command:

grep -e '^\([[:digit:]]\+\) \(.*\)\..*$'

The first group would have the track number and the second group would
have the title. I am not arguing in any way that regular expressions are
easy, trivial or anything like that just that this kind of parsing isn't
so bad for people that know how to use them. I am just saying that I
would like to have the option to omit some separators because I don't
need them and can still parse my file names unambiguously. This
alleviates file name clutter and makes files easier to read (to me).
Filenames should be easily human readable first and foremost and machine
readable secondarily. Again, this is why I say it should be optional. If
the software you use has a less robust parser, then the option to have
separators might make sense.

This was a nit pick point anyways :)