We should.
1: The minor issue of kernel archive size. I think the size
of the kernel archive is growing exponentialy. Graph
patchlevel against bytes yourself to see what I mean.
2: It confuses newbies. What's ext? Maybe I need it...
3: It has not been kept up to date with the latest VFS improvements.
4: Code is not really good for _explaining_ problems.
It would be better to document the design of ext2.
5: There are other simple filesystems to use as examples,
such as xia, minix, and sysv. Maybe minix is kept
up to date with VFS layer changes.
6: It's only good for 37 gotos anyway, so we still break 1000 :-)