playlist label tagging

[6arms1leg]

i have a lot of playlists and i use them for all kinds of stuff:
besides making custom music compilations i often use them to remind me of new recommendations from friends or to save a certain "state" in time.
so i end up with varying kinds of playlists:

• "normal" playlists, just some mixed tracks
• temp playlists (e.g. tracks i added over the day)
• regular albums i plan to listen to in the near future (think of it as a note)
• audiobook playlists
• soundtracks
• different genres (like classical music, etc...)

it would be nice to be able to sort/filter them in some way. thats one feature i miss with all audio players. what i normally do until now as a workaround is giving them distinctive prefixes like "ost" for soundtracks or "ab" for audiobook, so they are sorted and i can easily browse (e.g.) only soundtracks. that sucks!
thats why i thought about label tagging, just like we have here in github. a search function for playlists is no replacement for this -- i dont want to search but (e.g.) only listen to classical music (or soundtracks) from my playlists right now. that makes sense, because often a user knows what kind of music he/she wants to hear, but dont know which artist, so whats better than listening to his/her own playlists (which certainly contains artists they like). 👍 ?

[tilboerner]

👍

i dont want to search but (e.g.) only listen to classical music (or soundtracks) from my playlists right now. that makes sense, because often a user knows what kind of music he/she wants to hear, but dont know which artist, so whats better than listening to his/her own playlists

I don't really get what you mean here. Tags are a flexible way to categorize things into meaningful subsets, and quickly access them. Kinda like that? Or is there a special use case you have in mind?

[devsnd]

I think github does it just right. There is no enforced structure, but still everything can be sorted out just fine. The question is: Should we make a collection of predefined tags, or should they be created on the fly (with autocompletion)?

I'd vote for on the fly creation (with autocomplete similar to like stackoverflow), maybe suggesting the most used tags, when not having entered anything.

This however needs the user to pay attention not to create two tags with the same meaning but different spelling/words etc. I don't want to implement a "batch-retagging" tool afterwards...

[tilboerner]

I'd vote for on the fly creation (with autocomplete similar to like stackoverflow), maybe suggesting the most used tags, when not having entered anything.

me 2. Predefined tags = communism 🐷 .

pay attention not to create two tags with the same meaning but different spelling/words

That's a situation we have to accept, I think. See genre ID3 tags.

don't want to implement a "batch-retagging" tool

Although that shouldn't be too hard in a normalized relational database. It would allow renaming tags, too, which might turn out to be a useful feature. But let's wait for that.

[6arms1leg]

@tilboerner:

I don't really get what you mean here. Tags are a flexible way to categorize things into meaningful subsets, and quickly access them. Kinda like that?

yes, thats what i meant.^^ sometimes i cant express myself, i normally strip naked then and dance to describe what i mean.

@devsnd:

I think github does it just right.

yep. just like github.

I'd vote for on the fly creation (with autocomplete similar to like stackoverflow), maybe suggesting the most used tags, when not having entered anything.

me 2, too. 👍

@tilboerner:

That's a situation we have to accept, I think. See genre ID3 tags.

i agree.