W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > April to June 1998

Re: Advanced Collection Requirements

From: Judith Slein <slein@wrc.xerox.com>
Date: Thu, 16 Apr 1998 07:13:07 PDT
Message-Id: <3.0.3.32.19980416101307.0096f2b0@pop-server.wrc.xerox.com>
To: "Jukka-Pekka Numminen" <Jukka-Pekka.Numminen@poyry.fi>
Cc: slein@wrc.xerox.com, w3c-dist-auth@w3.org
Thank you for sending your scenario.

I am not quite sure whether we are using the word "ordering" in the same
way.  As it is used in the requirements draft, it refers to the sequencing
of resources within a single collection, where a collection is one level of
the hierarchy you describe.  

We do allow alternative hierarchies to be defined for the same resources.
Any resource can be referenced from multiple collections.  It would
certainly be possible to construct the two hierarchies you describe, which
reference all of the same resources but organize them in different ways.

The rationale for dropping the requirement for multiple orderings on a
collection was that you can accomplish the same thing by creating multiple
collections, each with a single ordering.

The ordering that is being addressed by the advanced collections
requirements is different from sorted search results.  Although we haven't
talked about implementations yet, I think a lot of us have assumed that the
ordering would be implemented as a property on the collection, whose value
is a list of the URLs of the collection members.  This ordering might be
equivalent to one you could get by sorting based on a metadata value, but
it would not have to be.  For example, I might create an ordering on a
collection that was "the best order for a beginner to read the documents in
order to develop an understanding of the field", even if there is no
property on the resources in the collection that could be sorted on to
arrive at this ordering.  

Search based on metadata is being addressed by another closely related
group, DASL.  The mailing list for that group is www-webdav-dasl@w3.org,
and their web site is at <http://www.ics.uci.edu/pub/ietf/dasl/>.  You will
find a requirements document there that includes ordering of result sets.
Although the DASL straw-man protocol does not address ordering, I assume
that this will be fixed as the specification evolves.

I think that the combination of the base DAV specification, the advanced
collections specification, and the DASL specification will eventually meet
your needs, but if not, we need to make some changes.

--Judy

At 11:38 PM 4/15/98 PDT, Jukka-Pekka Numminen wrote:
>Judith,
>
>I have tried to follow the WebDAV discussion because the results would be
very useful for executing industrial plant projects distributed among many
parties. If I have not completely misunderstood your work, the concept of
collections could be streched to cover an area of interest to industial
plant projects namely dynamic hierarchic collection. I hope that you taking
away the possibility to have multiple orderings of a collection does not
kill my scenario.
>
>The scenario (or an example):
>
>Engineering documents of an industrial plant can be ordered into
hierarchic collections in many ways. 
>
>To give you an example of two possibilities:
>
>Physical plant structure: plant -> department -> process -> equipment
>Document type structure: drawing -> electrical -> loop diagram -> department
>
>The first structure is the one a process or maintenence oriented person
uses and the second one is the one a design engineer uses. The user of the
system should be able to give the order (any order) of metadata elements
that form the hierarchy dynamically.
>
>There are tens (even hundreds) of thousands of documents in an industrial
plant project. To effectively navigate the mass a dynamic hierachy based on
the metadata would be very useful. In a sense this is an ordered collection
and the same collection has several ordering schemes. In addition to the
collection members it would be very important to get just the browsing tree
from the server (one could ask just for the tree that is the dynamic
hierarchy without any members)
>
>I hope that if this is not a part of the definition, the definition does
not prohibit (hopefully it supports) the creation of a system that works
like this.
>
>best regards
>
>Jukka-Pekka Numminen
>Jaakko Pöyry Group
>
>
>
>
Name:			Judith A. Slein
E-Mail:		slein@wrc.xerox.com
Internal Phone:  	8*222-5169
Fax:			(716) 422-2938
MailStop:		105-50C
Web Site:    http://www.nde.wrc.xerox.com/users/Slein/slein.htm
Received on Thursday, 16 April 1998 10:08:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:46 GMT