RE: Range locking

I am wondering if the issue here is over the locking of a portion of a
document or over the use of a range header to specify the portion of the
document to be locked.

However, first, to the justification. The reason we need to be able to
lock a portion of a document is because many people tend to share the
same document and the ability to specify a section of the document as
"locked", rather than locking the entire document, enhances the
interaction of users. If one can only lock the entire document then
sharing is hobbled. However experience with document management systems
have demonstrated that users often only need to lock a small portion of
a document, thus allowing other users full access to the remaining
sections of the document. Hence locking of portions of a document is a
necessary minimum feature for a real world distributed authoring and
versioning system.

To the question of implementation, the main argument against the current
spec is that our design principals require us to only operate on URIs
and we were using the range header to modify what the URI pointed to.
The now standard "Fielding Compromise" in such situations is to gain
permission from the server to munge the URI to point to a particular
section of the resource. I support this compromise as it allows us to
stick to our design principal but still provide a vital feature.

		Yaron

>-----Original Message-----
>From:	Larry Masinter [SMTP:masinter@parc.xerox.com]
>Sent:	Thursday, February 20, 1997 3:11 PM
>To:	Jim Whitehead
>Cc:	w3c-dist-auth@w3.org
>Subject:	Re: Range locking
>
>The overall requirement of WEBDAV is that we create a 
>way that (web) authoring tools can interact with
>(web) (versioning) servers in an interoperable way.
>The technical requirements in the requirements document
>need to be justified in terms of how they help in 
>accomplishing the overall requirement. In the case
>of range locking, this justification has not been
>made. That is, while range locking might help some
>web authoring tools interact with some versioning servers,
>it is not in support of any particular interoperable method.
>
>If there is consensus, it must be on the fact that
>the linkage is there.
>
>Larry
>

Received on Thursday, 20 February 1997 21:48:31 UTC