RE: MOVEs across file systems

Yes, one of the challenges of the binding protocol is to make sure
that we provide a simple and sensible model for the kind of system
that Kevin describes (since they are the systems that are most likely
to require and use a multiple-binding model), while still providing
sensible interoperation with file-system based implementations.

In particular, I believe it is likely that a repository such as the
one Kevin describes is likely to refuse to do anything other than
REBIND for MOVE and UNBIND for DELETE, both for performance and
semantic reasons.

Cheers,
Geoff

-----Original Message-----
From: Dyer, Kevin [mailto:kevin.dyer@matrixone.com]
Sent: Tuesday, March 11, 2003 8:49 AM
To: 'Clemm, Geoff'; 'WebDAV'
Subject: RE: MOVEs across file systems


Just putting in my 0.02 Galactic credits to make sure we keep in mind that
some implementations will not be limited by a file system paradigm as the
underlying engines will provide an abstraction to the actual source data.
These systems use meta-data as the pointer to the actual source data which
can exist on completely different servers and file systems. So moving data
across a resources is as easy as creating a DB relationship. The actual
source data is not moved and the system retains one version of the truth.
Kevin 

Received on Tuesday, 11 March 2003 09:26:01 UTC