RE: FW: I-D ACTION:draft-dusseault-dav-quota-01.txt

> From: w3c-dist-auth-request@w3.org
> [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Eric Sedlar
> Sent: Wednesday, October 23, 2002 7:21 PM
> To: Julian Reschke; Lisa Dusseault; 'Stefan Eissing'; Webdav WG
> Subject: Re: FW: I-D ACTION:draft-dusseault-dav-quota-01.txt
>
>
>
> It seems like this draft is fine with a user-based rather than a
collection-based
> system--the server will report the same usage & quota for any collection,
but
> will return different results depending on who is authenticated, no?
> ...

No. IMHO, the draft describes a very specific quota model, which it
shouldn't if it's to be implemented generally. For instance

"   Quota is not additive.  A collection only has the quota assigned to
   it, not (in addition) the quota assigned to sub collections or any
   other collections.  Sub-collections can have different quota values
   than parent collections.  These ôsub-quotaös may act as additional
   constraints, or they may be under-constrained and have no effect.
   This can allow the delegation of quota administration from root
   administrators to collection owners. "

will not work with a quota system that is per user and filesystem/server.

To make this proposal work in the general case, it needs to be more generic.
For instance, the only questions I can easily answer with a Unix-FS-based
quota system are:

- the storage limit (for the authenticated user)
- the current storage (for the user)

and possibly

- the amount of storage taken by a particular set of resources (and that may
be expensive to compute).

Julian

--
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760

Received on Wednesday, 23 October 2002 14:42:23 UTC