RE: three clarification questions on the If header.

At 12:24 AM 9/15/98 PDT, Yaron Goland wrote:
>I believe the spec is clear on the effects of the Depth 0 lock but I do
>admit that the term "Depth 0" gives the connotation that somehow the
>children aren't covered under the lock token. 

Yes, the spec is quite clear on this point.  As I understand it, since C is
a collection, its state includes among other things the set of resources
that it contains.  Adding to or deleting from this set affects  C's state,
and hence one needs the locktoken.  It is also clear that for
(pre-existing) resource C/R, doing a PUT (overwriting it) does not affect
C's state, so one does not need the lock token.

>... the write lock token for C in an untagged production on C/R is
>completely correct because the locktoken DOES describe C/R's state 

Sounds reasonable, but I would like this to be stated explicitly in the
spec, perhaps as an example, because it certainly is not obvious even after
careful study.  If not in the protocol spec, then please collect it for the
O'Rielly book (Please say you're planning to write one!.  What animal goes
on the cover?  Maybe a dove? or maybe some colonial insect such as an ant
or a bee, good examples of distributed cooperative authoring?)

Received on Tuesday, 15 September 1998 13:52:43 UTC