Linus Torvalds (torvalds@linux-foundation.org) on 1/27/09 wrote:
---------------------------
>slacker (s@lack.er) on 1/27/09 wrote:
>>
>>What if you have 5 applications wanting to write metadata
>>about a file? Keep 5 textfiles around in a directory?
>
>No, keep their own damn databases, and index it by the
>filename or something. Different programs should just
>keep their data separate. They shouldn't mix it up with
>files that they don't control.

1176 words, and you start with this. If the application is indexing the files by name, what happens when the name changes? You've already forgotten that the application doesn't control the file name, so now you're trying to rely on something which you don't control. Now you have a broken database, and zero utility to the user.

Even worse, a day ago, you were whinging about how FTP would break files in VMS. If you have metadata which gives relevance to a file, and that's stuck in some database, what are you going to do when you want to transfer the file? Dump the database and send it along with the file in a tarball? Well isn't that just a perfectly functional - yet completely idiotic - approach to things.

Beagle is a poor example for a number of reasons, including the fact that it uses xattrs to mark a file as indexed or not. That's Beagle's problem, not xattrs. I'm not going to further comment on disk access patterns, because that's an implementation issue, and your method of stuffing everything in to a common folder does absolutely nothing for improving these access patterns.

As to security, metadata as a whole is a complete pitfall in this regard. If your application saves your name as metadata inside of a document, your privacy could be compromised. Heck, an application writing it's own name inside of a file can be a problem (eg: when Microsoft marketing material was published on a Mac). What's your solution to this? Eliminating metadata as a whole? Or are you back to your folder argument? If you stuff metadata in to a directory, people will just stop looking at everything in the __METADATA__ folder because they've got better things to do, such as upload drunken photos of themselves to Facebook so the HR department at work can peruse the galleries at their leisure.

Extended attributes should have the file ACLs apply to them by default. I don't know what the implications of having xattrs having ACLs themselves would be. I suspect it could cause headaches for the user when they don't understand why they can do one thing, but not another, but ACLs themselves can lead to this problem. I don't know why xattrs need to have the versatility of a database to have any utility to you. I'm more than happy using it just to append simple metadata which can follow the file around.

Is it really your opinion that something like ID3 tags should be stored in separate databases rather than something associated with the file itself? I can't imagine anything more annoying.