Re: Comment on -01.1 document structure

Geoffrey M. Clemm (gclemm@tantalum.atria.com)
Tue, 25 May 1999 23:13:57 -0400 

Date: Tue, 25 May 1999 23:13:57 -0400
Message-Id: <9905260313.AA25469@tantalum>
From: "Geoffrey M. Clemm" <gclemm@tantalum.atria.com>
To: ietf-dav-versioning@w3.org
In-Reply-To: <004b01bea62c$41a8b5a0$d115c380@ics.uci.edu> (message from Jim
Subject: Re: Comment on -01.1 document structure


Both Edgar and Jim make good arguments.

For level-1 versioning, I think I'd keep it the way it is, for the
reasons Jim mentions.  On the other hand, for the level-2 versioning
constructs, we are trying to make the resource types as orthogonal as
possible, so that a server provider could just provide
"level-1 plus activities" or "level-1 plus configurations".
Separating each of the resource types defined in level-2 in the
way Edgar suggests might therefore be advantageous.

When I have a chance, I'll try making a pass through the level-2
part of the protocol, and see how it works out.  If it's not a total
abomination (:-), I'll post it to the web sites for comments.

Cheers,
Geoff

   From: Jim Whitehead <ejw@ics.uci.edu>

   Edgar Schwarz wrote:
   > I don't know whether there is some meta standard how a standard document
   > should be structured.
   > But nevertheless, when I read the draft I saw some definitions of terms
   > (data structures, objects) at the beginning. Followed by methods working
   > on these objects. This is a functional decomposition. Wouldn't it
   > be clearer to give the objects together with the methods it allows ?
   > This would make it easier to compare e.g. the subtle differences between
   > configuration and baseline.
   > In my experience OO isn't all hype (but also no silver bullet) so I would
   > try to start it already at this stage. Also there are e.g. OSI standards
   > which are also (sort of) OO.
   > Just a thought :-)

   Well, one rationale for the current document format is "tradition" -- we're
   trying to duplicate the style and organization of the HTTP/1.1 and WebDAV
   specifications.  Since there is a large body of people who have read these
   specifications, and are comfortable with their organization, warts and all,
   changing organization late in the game is not something to do lightly.

   Getting away from tradition, I think the current organization has some
   strengths.  Since the intent of HTTP/DAV/DELTA-V is to define operations
   which, as much as possible, behave the same across different types of
   objects (and by type I mean ordinary resource, collection resource, redirect
   resource, search arbiter, etc.), the emphasis on method definitions, rather
   than object definitions, is appropriate.