> Actually, filesystem metadata may not be quite as important as we were
> originally thinking ... I was just considering some ways to support generic
> "encapsulated object" file formats, and realized that using directories for
> that really isn't that bad an idea -- provided we provide other ways of
> doing it, too. Most such file formats basically try to be filesystems,
> anyway, so it's not that much of a reach.
>
> That got me thinking ... if directories could be used that way (although
> they aren't always that convienient), why not .tar files? Using a .tar
> format file does away with many of the problems of using a directory (i.e.
> difficulty copying/moving/transfering/hard linking), while giving about the
> same capabilities.
>
> Anyway, a container (directory, .tar, or whatever other file formats we
> support in addition) could just be browsable normally, although for
> containers that were intended to be encapsulated objects as such (say, a
> word processor document) the default action would be "Open" or "Edit" or
> something instead of "Browse". Be kinda cool to browse a word processor
> document like a directory, and see files for the encapsulated images and
> stuff, which you could manipulate just like they were in a directory. We'd
> also get "free" .tar browsing support out of it, and it'd completely
> generalize directories as containers.
>
> (for directories or .tar files, I suppose some special file in there
> indicating what it is -- maybe a file called "default" or "main" in there
> containing the 'primary' object, which the magic strings facility could pick
> up on to find the 'primary' type)
>
> Anyway, it'd be kinda neat to support encapsulated objects made out of .tar
> files and directories in addition to whatever main file format we adopt for
> that.
>
> -=MenTaLguY=-
>
Tar files? Ouch! Tar is _not_ a friendly format. I like fs-level
support much better.
>
> --
> To UNSUBSCRIBE, email to design-request@berlin-consortium.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@berlin-consortium.org
>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/