W3C home > Mailing lists > Public > www-rdf-dspace@w3.org > December 2003


From: Butler, Mark <Mark_Butler@hplb.hpl.hp.com>
Date: Fri, 5 Dec 2003 17:17:32 -0000
Message-ID: <E864E95CB35C1C46B72FEA0626A2E808206327@0-mail-br1.hpl.hp.com>
To: SIMILE public list <www-rdf-dspace@w3.org>
Hi team,

Kevin and I were discussing the demo script yesterday. I'd like to share
some of these thoughts with the team. First we've spoken a bit about
browsing and searching, but I think there are at least three distinct
activities here:

i) drill down browsing "I'll know what I want when I find it"  
i.e. we start with a large number of resources and identify a subset of
resources. This is the type of browsing occurring in faceted search.

ii) related browsing "I want something like that"
i.e. we start with a single resource and identify a wider set of
resources, e.g. a see a picture I like, then click on the creator to
find other similiar pictures. David Karger notes the importance of this
type of browsing in the Haystack position paper.

iii) search / query "this is a description of what I want"
i.e. we specify a set of constraints, and retrieve the set of resources
that meet those constraints.

So I guess the novelty here is I'm proposing that there is subtle
difference between i) and ii), and when we write the demo script ideally
we want to demonstrate both. When Kevin and I talked, it became apparent
that I'd been thinking about demos that used i) whereas Kevin was
thinking more about ii) and iii). In addition we can combine these
approaches e.g.

iv) "I'll know what I want when I see it, but I want to be able to find
other related things"
this is i) then ii)

v) "I'm not quite sure what I want, so I'll get rid of things I
definitely don't want, then try to describe what I want"
this is i) then iii)

So we have at least five different approaches here.

The other point is that Kevin's script describes starting with OCW data
and moving to the Artstor data. This is because we have much more
Artstor data than OCW data, so it is easy to show value moving that way.
However, ideally we need to come up with a scenario that moves the other
way. This is because the Artstor data is easier to do i) on, so if we
want to demonstrate iv) then it might be easier if we start with the
Artstor data?

Does this make sense? Any comments?

Dr Mark H. Butler
Research Scientist                HP Labs Bristol
Internet: http://www-uk.hpl.hp.com/people/marbut/

Received on Friday, 5 December 2003 12:18:31 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:13:09 UTC