Consider the following "chain" of factors that influence the finding of
relevant information on the Web. These factors are all related to
"quality".
Relevant info from Web (a) <<== "quality" of user (b) <- "quality" of
query (c) <- "quality" of search engine (d) <- "quality" of search
engine indexes (e) <- "quality" of metadata (f) <- ...
Most of the discussions and results have uptil now been concerned with
(f) "quality" of metadata mainly:
* "quality" of content -> metaguide, templates, richness (number of
elements, for instance), etc.
* "quality" of data -> syntax, structure, templates, MD creators, etc.
* "quality" of data delivery -> correct HTML, correct SGML, mime,
Warwick framework, etc.
Since (e) depends on (f), this raises the question of what requirements
are needed for the metadata in order to enable the indexing algorithm of
(e) to make "quality" (better) indexes.
Therefore. Will all DC elements likely be indexed? Will qualifiers be
indexed, or what is their role in the indexing process? In general, what
is, or what could be, the (good) influence of DC to make better (for
instance, context/subject related or enhanced) indexes in phase (e) of
the "quality chain"? I think, this would be an important goal to bear in
mind all the time.
If we are concerned about finding relevant information on the Web (a),
we should be concerned now about (e), the "quality" of the indexes,
putting aside the intermediate stages in the "quality chain" for
simplicity.
Stu Weibel wrote:
>
>
> (subtitle: how Carl ended up with the middle seat all the way to oz...)
>
> As several others have noted, there will be multiple uses and
> applications of the Dublin Core. Yes this raises problems of
> consistency and interoperability. We'll have to deal with this. What
> a wonderful problem to have: sufficient deployment that this is a
> concern!
>
> The problem right now is to hammer out sufficient details to jump-start
> deployment and acceptance. With some luck and good sense along the
> way, we may succeed in converging on deployment details.
>
> In spite of the fact I intend to make Carl suffer all the way to
> Canberra, I agree that the time for serious engineering is at hand, and
> I am very encouraged by the movement in that direction (PICS, WebDAV,
> and others). One of the blessings of such efforts is that we might
> succeed in separating the siamese twins of Semantics and Syntax.
>
> stu
--
Frank A. Roos ([log in to unmask])
CWI - Centrum voor Wiskunde en Informatica / Bibliotheek
- Centre for Mathematics and Computer Science / Library
http://www.cwi.nl/cwi/departments/BIBL.html
|