|
Task Description
If anyone considers this, or has questions, please also conside FS#7582 - used in concert, I think these two feature requests would result in a *lot* of flexibility as far as ad-hoc searching, sorting and filtering results is concerned
Currently specifying a hierarchy for identifying tracks involves combinations such as
genre → track
I’d like to suggest that the tags currently available for use in navigating in this way are augmented by pseudo-tags in the form keyword_artist, keyword_album etc.
The entries generated by these pseudo-tags are lists of words generated from the underlying tag in question (a “word” being space-delimited). Selecting one of these words presents of a list of any entries featuring that word.
As an example, take the following tracks/genres:
“Are Friends Electric” - “Electronic” “Waiting for Cousteau” - “Ambient Electronic” “Let Me Entertain You” - “Rock Pop” “The Honeydripper” - “Jazz” “Ace of Spades” - “Hard Rock” “Albedo 0.39” - “Ambient Electronic” “Main Sequence” - “Electronic Jazz” “Chameleon” - “Ambient Jazz”
Using genre → artist in the standard view results in this for the first level:
“Ambient Electronic” “Ambient Jazz” “Electronic” “Electronic Jazz” “Hard Rock” “Rock Pop”
Selecting “Ambient Electronic” results in:
“Albedo 0.39” “Waiting for Cousteau”
By comparison, using keyword_genre, the first list results in:
“Ambient” “Electronic” “Jazz” “Hard” “Pop” “Rock”
Selecting “Electronic” results in:
“Albedo 0.39” “Are Friends Electric” “Main Sequence” “Waiting for Cousteau”
and selecting “Jazz” results in:
“The Honeydripper” “Main Sequence” “Chameleon”
1) The requirement for manually entering fixed/dynamic search terms would be reduced dramatically - much easier to select from a list of words that already exist than to manually enter and “fish” with a search term.
2) Although I think this adds flexibility by itself, combining with FS#7582 would enable a very simple but powerful system for searching and categorising tracks. Selecting any “Ambient Jazz” becomes a simple as searching for “Ambient” and then “Jazz”.
Adding a suitable keyword to a chosen ID3 tag effectively adds an ad-hoc user-definable dimension to categorisation - instruments, moods, conductors etc.
It would be easy to envisage, for instance, adding keyword terms such as “60bpm”, “100bpm”, “150bpm” to a set of tracks and allowing the user to retreive music sutiable for the gym or relaxation etc.
3) At first blush this seems like it may result in very long lists and be demanding in terms of resources, but I don’t think this is the case - if anything extracting keywords should shorten the potential list for reasonable sample sizes.
I’ll try to extract my own tagging information and process the text with uniq to see what the balance is.
4) Clearly there may be issues in defining a “word”. As a starting point, I’d suggest space delimited, case insensitive, and with punctuation stripped out.
|