-1, I really don't think anything needs to be said about this.
Bill de hÓra wrote:
Roy T. Fielding wrote:
On Oct 31, 2005, at 4:03 PM, Tim Bray wrote:
On Oct 31, 2005, at 2:03 PM, Luke Arno wrote:
* "Note: deleting a member also removes it from all the collections to
which it belongs."
I don't think we ought to legislate this.
I think that if an entry shows up in multiple collections, it's still
the same entry, and the semantics of DELETE are "make this resource go
away", so I think that it would be very surprising if we changed them
to "remove from one collection". -Tim
[...]
In message-based protocols, you should define the sender's
intent and expectations for a given message, not how the
recipient shall change its own state as a result.
<http://www.intertwingly.net/wiki/pie/PaceClarifyCollectionAndDeleteMethod>
== Abstract ==
make a necessary distinction between resource deletion and collection
deletion. State the client intent on sending DELETE, and ensure APP is
consistent with HTTP and WebDAV in that regard.
== Status ==
Open
== Rationale ==
APP up to 06 is overspecified; it conflates deletion of resources with
the DELETE method with collection removal.
== Proposal ==
remove from "5.3.3 Deleting a Resource",
{{{
Note: deleting a member also removes it from all the collections to
which it belongs.
}}}
and, add this text instead:
{{{
Note: deleting a member resource does not necessarily remove it from
the collections in which it appears. The intent of DELETE in the Atom
Protocol is to delete the resource identified by the Request-URI, not
to remove it from one or more collections.
}}}
== Impacts ==
draft06
== Notes ==
If we need client ability remove members from APP Collections (as
opposed to the web), that's a separate matter and requires another Pace.
This Pace is here to address overspecification only.
cheers
Bill
