[KLUG Members] New advances in filesystems.

[Adam Williams]

> Very interesting stuff!
>   http://cda.mrs.umn.edu/~mine0057/fs.pdf

Very interesting, although I'm puzzled they don't mention resource forks
which do much of the same things, and filesystems like XFS which offer
basically openended meta data (attributes, properties), etc...

The ability to put real constraints on file system objects is very cool.

But some of the things he suggests (enviroment in a filesystem, etc...)
seem a bit obtuse - I mean, why?

> Even makes LDAP obsolete! (see last page)  ;-)

Hardly.  The author looses a touch of credibility here.

1. "If the server goes down you *MIGHT* still be able to log in" - Oh,
thats just great.
2. LDAP is not a glorified version of /etc/passwd, LDAP addresses a much
larger problem domain.  If all your using LDAP for is to replace
/etc/passwd, why bother?  Just use NIS.
3. The whole issue of secure hot replication isn't addresses.  So far
every 'distributed filesystem' I've tried has been no where near
enterprise grade.  
4. And sharing data between organizations?  Via file system shares? Over
the Internet?  I'd loose sleep.