[preetbhinder]

I stopped using anything from the i* suite after my first Mac when I learned that it messed up my folder structure. All my carefully organised music and photos were imported into an opaque database that gave me no visibility into how things were stored.

[wlesieutre]

This part at least has been a configurable option as long as I can remember. http://www.usbdacs.com/Macintosh/files/page5_5.jpg

[Karunamon]

Did something change? Because the iTunes internal organization for music at least is quite easy to understand.

From the library folder:

    Music/Artist/Album/Track# Title.ext

If your files aren't tagged correctly (which is distressingly common), then the auto organization will make a pig's breakfast of your library.

A common example: "Artist" and "Album Artist" are two different ID3 fields - if you have an album with an Album Artist but a blank Artist, iTunes will catalog it as "Unknown Artist" when the consolidation happens.

This is slightly annoying to fix, but the mass tag editor in iTunes works well.

[wlesieutre]

A complaint I've heard before is that this works fine for most modern music, but the Artist/Album/Track layout doesn't apply to a lot of classical collections.

In which case you can just turn off the folder organization, but there's not an easy way to undo it once iTunes has gone and dicked it up.

[rsync]

"A complaint I've heard before is that this works fine for most modern music, but the Artist/Album/Track layout doesn't apply to a lot of classical collections."

More importantly, and more generally: what if my music organization format predates OSX/iOS/iTunes and I am not interested in altering it for todays fad ?

This is not apple-specific, either - many devices and services (audi MMI, Sonos, etc.) assume certain naming and organizational practices - and work very awkwardly with any other layout.

Granted, my Sonos system has never deleted anything or reorganized directories ... but my music is mounted read-only just in case.

[Karunamon]

"Today's fad" has been the iTunes standard since the application existed.

Also, take your pick. If the files aren't organized to a consistent standard, that means a database has to be stored somewhere with that info so the app can do its job. In iTunes' case, that's an XML file.

[rsync]

"Also, take your pick. If the files aren't organized to a consistent standard"

Actually, that's not true - one alternative to an organizational structure or a database is to actually name the files verbosely, which is what I decided on in 1996 when I ripped my first CD:

Artist Name - Album Name - ## - Track Name - time.wav

For instance:

Ferry, Bryan - Taxi - 03 - Answer Me - 2m46s.wav

This file can now be dropped into place anywhere without losing information, requires no DB and gives you full "metadata" for wav files that don't actually contain metadata.

[rsync]

"If your files aren't tagged correctly (which is distressingly common), then the auto organization will make a pig's breakfast of your library."

Pray tell, how should my WAV/PCM files be "tagged" ? What's the format for that ? Exactly.

[Karunamon]

You just described the reason for the existence of ID3. But, if you insist on using files that don't have a metadata standard, the only way to organize them is in the database of the library application. Which generally won't be portable to other applications.

[dangravell]

WAVs have had a metadata standard since they were invented, see RIFF INFO. And now more software is supporting ID3v2 inside WAVs too.

What you meant is a standard that was used. That's changing, a little.

[rsync]

"But, if you insist on using files that don't have a metadata standard, the only way to organize them is in the database of the library application."

Again, I disagree - see my response further up the thread for how to get around this problem with verbose naming:

Ferry, Bryan - Taxi - 03 - Answer Me - 2m46s.wav

[mistersquid]

Your proposal to store track info in with verbose naming isn't very robust and doesn't solve even the most basic problems that a simple metadata scheme does.

For one, even in your specific limited case of Bryan Ferry's album _Taxi_ there are 5 different versions of the album from 5 different countries[0].

Furthermore, your verbose naming proposal is not only unequipped to handle something as basic as alternate versions and international releases, it has no affordances for providing basic information such as year of release, publisher, and composer, let alone information about bitrates and compression schemes.

Even were we to restrict our attention to just the metadata your scheme does encode, your proposal would fail when cataloging an entry where the band/album/track name contains a hyphen surrounding by whitespace. There are also some albums which have more than 99 tracks (archival records, for example, which are distributed in multi-CD collections).

My response may seem a bit like hitting a fly with a sledgehammer but, having worked in a library and taken classes in information science, it's crucial to illustrate why hastily-conceived proposals to replace metadata with file naming conventions should never be taken seriously. In fact, I believe they should only be taken jokingly!

Metadata schemes are so important to information and library science that, in my opinion, any proposal to replace metadata schemes with "verbose naming" should be shown to be untenable unless the goal is to index fewer than 100 files in a restricted-access repository (and maybe even then).

From the standpoint of a librarian or digital archivist, you may as well have proposed storing the binary data in the file name, too, essentially eliminating the file name which itself is a piece of metadata.

[0] https://www.discogs.com/Bryan-Ferry-Taxi/release/1120442

EDIT: readability

[tripzilch]

> Your proposal to store track info in with verbose naming isn't very robust and doesn't solve even the most basic problems that a simple metadata scheme does.

But that's how anyone who is serious about collecting digital music has been ordering it since ages. Sure, I never considered putting track length in the filename as :) (but I can see why), and if I can help it I have them tagged correctly too. But the file/folder structure is how I keep it organised.

There may be better ways to go about this, but it at least needs to be an actual improvement. iTunes is not.

Currently I'm experimenting with a command line tool called 'beets', which from reading its docs, definitely has a philosophy that aligns with mine. Unfortunately I haven't quite figured out how to tell it where to get the metadata from, it defaults to MusicBrainz which seems to have quite a few inaccuracies in their data (spelling of "Kung-Fu", hyphenated or with spaces I don't care really but if you use both spellings on the same album, one of them is wrong). It's got plugins for discogs and figure-out-from-pathname so that's good, but then it still uses MusicBrainz too. All in all it's a bit fiddly, but I can't really imagine a way to do it better, if you want have accuracy, keep control, some assistance with automatic tagging, but without messing up weird edge-cases like the classical albums and bootleg recordings mentioned above (for which 'beets' tries real hard to do the right thing, that being what you decide it to be).