- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Wed, 12 Nov 2003 17:08:32 +0100
- To: Lisa Dusseault <lisa@xythos.com>
- Cc: 'Brian Korver' <briank@xythos.com>, w3c-dist-auth@w3.org
Lisa Dusseault wrote: >>True. This is a simplification that works well as long as the >>quota is >>not authorable. If it becomes authorable (by means of this additional >>property), there's a big issue because the behavior of the server >>becomes completely unpredictable. > > > It should be simple to make things predictable by saying that the > server MUST (SHOULD?) support only one quota if it allows that quota > to be authored via the mechanism in this draft. That would work, but seems like a hack. First we sell it as a feature that the server selects a specific quota, and then a feature is added that is incompatible with that. > In practice, there is usually one important size limitation. On a > service like sharemation or Apple's iDisk, that's the quota. On a > hard drive that's the drive size. Even though in theory there may > be a size-limited hard drive behind a quota-limiting system, it won't > change how things work except in extremely rare cases. And let's not > design a system that is optimized for the extremely rare cases. We just reached agreement that physical disk limits should not be treated as quota, right Brian? Julian -- <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
Received on Wednesday, 12 November 2003 11:08:34 UTC