Hello,

in case it has not been mentioned here, and even in case it has been
discussed :)

I want to make aware / remind the developers about important
needs not historically connected to "file systems", that nevertheless
has lots in common with Coda.

It is backup.

No, backup clone is just a very useful hack, but it is not a "solution".

Then, I am not talking about disaster recovery backups.

But to be able to restore a file state for a given date (usually
very approximately) is an everyday activity for a computer support unit,
using very expensive hardware (tape robots) and software,
taking a long time (average transfer rate 35 bytes/sec is not a surprize
for a home directory restore - tape seeks are sloooow)

At our place we are looking into moving the whole backup archive to
harddisk arrays, a lot cheaper and more convenient... Still needing
suitable (probably expensive) software with database in the bottom.

If the Coda developers could teach Coda to be able to produce multiple
backup volumes of the same volume, then a set of "yesterday's files",
"last week's files", "last months' files" and "last year files"
mountpoints would eliminate that headache and the need of
administration/support for restoring files.
(while not using more disk space than absolutely needed!)

Then Coda would become *very* interesting as a filesystem with really
integrated backup function.

It looks like the functionality is already there, but is just not exported
in a proper way.

My 2c,
--
Ivan