Consolidate some UPnP index nodes

[mcow]

1) Instead of a separate index node for every different punctuation mark (I have eleven different punctuation categories covering about fifty song titles), have a single [symbol] entry.

2) For names that begin with Roman letters with diacriticals, I'd like to see them included in the base-letter configuration: put Ö under O, for instance.

[nohitter151]

Agreed on all points.

http://www.ventismedia.com/mantis/view.php?id=10570

[Lowlander]

As per http://www.mediamonkey.com/forum/viewto ... =6&t=70352 this has been discussed in http://www.ventismedia.com/mantis/view. ... 740#c35129

[mcow]

This was implemented well in 1626 for indices of Albums, Album Artist, etc. But the All Tracks indices are still broken out on individual characters.

[nohitter151]

This was implemented well in 1626 for indices of Albums, Album Artist, etc. But the All Tracks indices are still broken out on individual characters.

From the notes in the tracker:

- there is usually a lot of files in library and therefore distinguishing between indexes À, A, 1, 2 makes sense for song titles (unlike artist, album).

[mcow]

- there is usually a lot of files in library and therefore distinguishing between indexes À, A, 1, 2 makes sense for song titles (unlike artist, album).

Well, I would argue that's just silliness. There are hundreds of songs beginning with "T" but they're not broken down further into Ta, Te, Th, Ti...

Also, separate indices for . # ( " and so on continues to be no use. Sorting by the first letter (or digit) in the title would almost always be OK, but barring that, a single all-symbol category would be helpful.

It's as important to make the list of indices manageable as it is to make the list of songs manageable. What's really needed is adaptive indexing, but that doesn't have a chance as along as simplistic advocates for "consistency" have the developers' ear, while people who in fact design UIs professionally and have studied information theory do not.

[Lowlander]

The MediaMonkey userbase is very diverse and has many use cases. Whereas for some merging title nodes will not be an issue for others it will. Especially as UPnP is very slow to start with, merging nodes will have a much bigger negative impact on users that prefer the separation than the separation has on those who seek merged title nodes. For example international users may require separation (for example ñ is an actual letter in the Spanish alphabet) and having to find tracks among possibly hundreds of tracks will be tedious at best over UPnP. Separation will prevent that problem.

[nohitter151]

- there is usually a lot of files in library and therefore distinguishing between indexes À, A, 1, 2 makes sense for song titles (unlike artist, album).

Well, I would argue that's just silliness. There are hundreds of songs beginning with "T" but they're not broken down further into Ta, Te, Th, Ti...

Also, separate indices for . # ( " and so on continues to be no use. Sorting by the first letter (or digit) in the title would almost always be OK, but barring that, a single all-symbol category would be helpful.

It's as important to make the list of indices manageable as it is to make the list of songs manageable. What's really needed is adaptive indexing, but that doesn't have a chance as along as simplistic advocates for "consistency" have the developers' ear, while people who in fact design UIs professionally and have studied information theory do not.

I can't really comment here, honestly I would never browse over UPnP by title, I think regardless of which nodes are available it is way too cumbersome to be useful. But that is beside the point I want to make, which is that you can use the MagicNodes addon and set up the nodes/subnodes however you wish - and they will also be available over upnp.