Distributed Authoring group meeting...

["Robert S. Thau" <rs...@ai.mit.edu>]

Last Monday, I attended a meeting of a (so far) ad hoc working group
on support for distributed authoring and versioning on the web.  This
group was hosted by the W3C (which also currently hosts its mailing
lists), but it does not have formal W3C sponsorship at this point;
it'll probably go that way, but there was also some talk of
materializing as an IETF working group.  Here's a brief report on some
of the high points of the meeting:

1) The current state of the effort is as follows:  there are a couple
   of documents circulating in draft form concerning requirements for
   distributed authoring support and versioning --- that is, what
   capabilities that the protocol must support.  There seems to be
   near consensus on these --- though it still isn't clear, to me at
   least, that the requirements document lists all the varieties of
   locks which servers might want to support.

   See http://www.ics.uci.edu/~ejw/authoring for drafts.

2) Members of the group are working on other foundational documents,
   including a "scenarios document" describing several interactions
   between clients and servers which the protocol would want to
   support, and most importantly, strawman drafts of actual protocol
   specs.  Unfortunately, none of those are up on the mailing list
   yet.  

3) There's a two-day meeting scheduled at Xerox PARC in November, to
   be hosted by http-wg chair Larry Masinter.  Hopefully, by this
   point, there will be actual discussion drafts of protocol specs.

4) Regarding the participation of "major industry leaders" ---
   Microsoft sent a very active representative, the program manager
   who is responsible for specing out the interfaces to the web and
   the internet at large from within Microsoft's products.  (A note on
   titles here: from what little I can see from outside, Microsoft's
   development teams are divided into "developers" and "program
   mangers", a/k/a "devs" and "PMs".  However, a single project can
   have several PMs --- their responsibility is less management per
   se than managing the interface between their own project and some
   other constituency, internal or external; furthermore, a dev who
   thinks that a particular PM's ideas are lousy will often just
   refuse to go along, and stands good odds of getting away with it).

   This particular guy is actually trying to coordinate the
   server-side needs of a whole bunch of Microsoft products --- he's
   interested in partial file locks (locks on a subrange of a file,
   for instance), in part because Excel wants to support them.
   However, it *seems* to me that he'll have to sell whatever the
   group produces back to those groups, and doesn't have the power to
   commit them on his own.  (Partial locking is not just a freebie for
   Redmond, fwiw; other people do want it as well).

   Netscape, true to form, didn't show up, though there was a
   representative from Mortice Kern systems, to whom they've
   subcontracted versioning support in some of their current products.
   However, things are looking up, a little --- they have at least
   been sighted on the mailing list.

   (FWIW, I mentioned our problems with Catapult to the Microsoft guy
   at the earliest opportunity... I'm not sure this had any effect).

5) There's actually a fair bit of concern within the group for staying
   open --- in particular, the main concern expressed about proceeding
   under the W3C was that a number of people didn't want the W3C's
   "members-only" nature to exclude people who had something useful to
   contribute.  Apparently, this is not too much of a problem, as W3C
   working groups can include non-members, at the working group's own
   discretion.   (Interestingly, the Microsoft guy was one of the most
   vocal about this --- make of that what you will).

rst