We often allude to DC "users" and "discovery" as if we all shared a common
conceptual model. We may in fact do that, but let's find out. Enunciating
a model will help us communicate our agreements beyond the meta2 group and
improve progress. Once we agree on the audience(s) and model, the goals and
priorities can be determined. Thereafter, if a proposed change is examined
as to its impact on the prioritized goals for our audiences, that will
streamline reaching closure on discussion of the change.
Resource discovery is really a huge area, most of which is out of scope
for the DC. For example, aspects of discovery that the Dublin Core is
not about are, by my reading,
- not search and retrieval protocols (eg, HTTP, Z39.50)
- not query languages (eg, SQL, RPN)
- not search diagnostics and results analysis
- not retrieved record syntaxes (eg, HTML, MARC)
- not search as a series of inter-related questions
- not goal of discovery (eg, def. of success, primary vs secondary info)
- not stateful, session-oriented search
- not multi-target, parallel search
So what do we actually include in our model of resource discovery? I think
we've assumed that all searching is composed of one or more stateless
request-response cycles (one request, one response), and that for the DC's
purposes, our focus is on metadata as it supports one such cycle, namely,
(a) the search access points and (b) distinguished search response elements.
Very roughly, I see several resource discovery phases and several audiences.
DC Resource Discovery Phases Main DC Audience
--------------------------------------- ---------------------------
Search Support:
(1) metadata provision phase PROVIDERS
(entering and updating, whether by (authors, secretaries,
hand or with automated assistance) site administrators, etc.)
(2) metadata collection phase INDEXERS
(gathering metadata, binding to (local programs, internet
a resource, analyzing, indexing) crawler robots, etc.)
Request-Response Search Cycle:
(3) search phase SEARCHERS
(formulation of request based (school kids, scholars,
on metadata access points) nurses, professional, etc.)
(4) selection phase SEARCHERS
(examination of returned records;
discard/select based on metadata)
The needs of Indexers will be different from those of Searchers and those
of the Providers. Whose needs come first? What are those needs? What are
the tradeoffs? These questions must be discussed before we can build on
meta2's past work. A truly useful proposal for a design change, whatever
it may be, will specify which part of discovery is being addressed and
which audience will be affected. (Of course we could still use to
articulate some goals, but that's for another message.)
-John
|