RE: WEBDAV Observations on Disconnected Operation, Distributed/Re plic ated Systems, and Caching

Damn, you caught me! =) I left that section intentionally vague because
I hadn't written all my position papers yet and wasn't ready to talk
specific turkey. However most of my papers are now out and I can be much
more specific. I believe that useful work means the ability to create a
script of commands and then have that script replayed on the network.
This does not prevent a smart client program from performing command
compression, for example, if the user doing a series of partial puts to
an entity, the client program can always combine that into a single put.
So a user can now use all the DAV command of-line and when they come
back on-line the client program can be able to replay that script and be
able to deal with changes that have occurred in the meantime. I believe
we need to look at all the methods we are creating, as well as the HTTP
1.1 methods, and determine if there are scenarios where the script would
unrecoverable fail.

If a user orders a file moved that has already been deleted, that is not
an unrecoverable failure. An intelligent client can deal with this
situation. However if the user orders a file altered, that has already
been altered, I would characterize the lack of a MERGE method as an
unrecoverable failure. This is all quite subjective and will need to be
argued out on the list. However my current feeling is that a combination
of DIFF, to help reconcile the members of a structure, and MERGE, to
help reconcile the contents of an item, should give us all the
capabilities we need in order to satisfy my litmus test.

		Yaron

> -----Original Message-----
> From:	Mike_Spreitzer.PARC@xerox.com
> [SMTP:Mike_Spreitzer.PARC@xerox.com]
> Sent:	Friday, March 21, 1997 9:21 AM
> To:	Yaron Goland
> Cc:	w3c-dist-auth@w3.org
> Subject:	Re: WEBDAV Observations on Disconnected Operation,
> Distributed/Replic ated Systems, and Caching
> 
> Let me ask for a clarification of your litmus test.  When you say
> "perform
> useful
> work", do you mean to include DAV-specific stuff (all methods? most?
> some?
> which?), or will just local editing of a copy of the document suffice
> to pass
> this test?

Received on Friday, 21 March 1997 15:19:22 UTC