I have to disagree…. kinda, with what Havoc says:
Filesystems are not the answer, what we need is a document store. None of the open source desktop projects will ever depend on ReiserFS or similar features. Because these projects have to be able to work with NFS, GFS, AFS, and across multiple UNIX flavors.
Basically what he’s saying is that to be nicely cross-(unix)-platform, you have to have userspace utils instead of relying on a certain filesystem being present or a certain kernel module present for a technology to work.
Currently, yes, this is true. I don’t see it moving elsewhere in the near future either. I somehow doubt that you’ll be able to convince people running their sunsparcs to redo their system with ReiserFS or FooFS or whatever the current FS is.
However, it also means that GNOME (and KDE) are restricted in what they can do because of not only cross-unix issues, but even problems with compatibility between different flavors or distros of Linux! This is something that both Microsoft and Apple got around long ago by simply not offering the user any choice at all. Microsoft gives you the option of NTFS and FAT, Apple gives you only one filesystem with journaling enabled or disabled. Under linux you have ext2, ext3, reiserfs, jfs, xfs and others that I can’t think of right now. Each has different capabilities and different uses, and none is the “standard” install on all distros.
A blessing and a curse. It allows you to put whatever you want on your computer, but it doesn’t let a desktop environment like GNOME to simply say “we’ll check for the metadata stored in the filesystem for our WinFS type system.”
I wonder if it might be worth it to work towards changing this, even just within GNOME. Have all the major vendors get together and say “this is the one way we will do it.” Users will of course still have choice, but each and every distro complies with the Freedesktop.org standards, the file system standard, and using a kernel with features X, Y, Z enabled with the FooFS filesystem running.
Of course, that was tried with United Linux, which didn’t go anywhere fast.
Of course, even within Linux development there is a disconnect between the graphics server (XFree/X.org)/Xserver), the Kernel (Linus et all), the desktop environment (GNOME, KDE), and the apps. Some of these parties work closely, the app developers and the DE developers come to mind, but (from what I’ve seen at least) the rest are for all intents and purposes, separate entities.
Which is again, a blessing and a curse. Choice is great, choice sucks.
Anyway, I think that in addition to creating a userspace model for the document indexing issue, start working with the FS developers to work towards the option of a filesystem based system similar to what Apple and Microsoft will be coming out with (in a year and a few years respectively).