I took a quick look at that site a couple of days ago, but it did not
seem to have any working code yet, is this the case, or are the pages
outdated?
> It should work. However I think it would probably be better to
> write a new FS and you can then add other features at the same
> time.
I thought about that when I read the XFS paper. But, if we are going
to go trough the hassle of writing a new file system, we should at
least be able to compete with XFS, and competing with this file system
means that we need a lot of resources (disks :-) to make sure we even
come close to them.
According to the XFS paper, XFS/XVM was 50,000 lines of source code by
the time the article was published. Ext2fs is just 6,000 lines of
code.
> For example Ext2 has no support for logging and adding
> such support to it would be a significant re-write.
I disagree with you. Adding logging to ext2fs is much easier than
coming with a new file system. Adding logging just requires putting
the right code in the right position :-).
> While we're at it using B+ trees for directories as
> HPFS and NTFS do would be a great win for news servers and other
> systems with thousands of tiny files.
This would be nice to have. Somebody told me yesterday that there was
a long thread on linux-kernel at some point that discussed this very
thing. Anybody knows what the result from that discussion was?
Anyways, adding this B+ tree information for directories to ext2fs
falls also in the "possible without a massive rewrite" case.
Miguel.