Belated minutes of 55th IETF WebDAV meeting

WebDAV WG meeting, IETF 55, Atlanta, Nov 19, 2002

 

                

ACL Spec

Main point of discussion is the ACL spec, and the IESG review thereof.

·         Geoff points out major features of the ACL draft:

o        what privileges do I have

o        what is the relationship between privileges

o        what are the principles available

o        writing of the ACL

·         One issue raised by the IESG was regarding the privileges needed for DELETE.  The sense of the WG was that it would be ok to add <D:delete> permission is added to control access to the DELETE method

·         Major discussion on the complexity of the spec, especially of acl-semantics.  No conclusion other than that the IESG is very firm on this issue.

 

RFC2518bis issues

 

  a.. No objections in the room to If header simplification/clarification
  b.. Proposal for the Locktokens-Used header:
    a.. Geoff is concerned about the backwards compatability requirements of the Locktokens-Used
    b.. The assertion is that Locktokens-Used addresses no additional use cases beyond the If/Not-If technique that is supported (although not on many server) (Geoff)
    c.. Joel / Lisa believe that header length causes problems with the If/Not-If case
    d.. locktoken-used is for the use case where we want to write whether or not we have the lock, whereas If header is needed for the case where we only want to update if we have the lock
    e.. If header needs to be fixed to handle comma separation to avoid the header length problem regardless
    f.. Possibility of the (*) to match URL in the If token to unify the two headers
  c.. Consensus on required use of ETags, but only on resources that accept a PUT-i.e. not a generated resource
  d.. XML Namespaces-no new ideas other than some worry about control chars in XML tag names
  e.. Next version of RFC2518 should include warning about how horrible DAV: URI scheme is
  f.. Need to discuss whether the allprop include should be experimental on the list
  g.. Other drafts (allprop, quota, property datatypes, bindings)-should we postpone discussion on one of these?  No one seems to think we need to wait.
    a.. Quotas-clarify that the quota-free applies to current collection (don't need to say anything about subcollections)
  h.. Attribute values in property names?  Do these need to be persisted?  Why? (Joel)
    a.. Joel-no, this requires a real XML data store.  The attributes can be reserved for metadata.
 

WG business 

 

  a.. Interop report for last meeting?  Lisa promises to look into this