<IMHO>
In support of the "Leave it at 15" argument:
1) "It is often not possible to distinguish between publisher and creator"
In the case of the Web, where a document is written and published by the same
person (the webmaster), then Publisher and Creator fields should just be the
same person. There will be times when this is different - for example, a press
release that is written by the PR department, which the webmaster only
publishes.
There are times when the creator is the publisher of "real" documents too - for
example, on a University campus, flyers distributed by a student political group
might be created by them and published by them (either by printing out a
thousand copies, or photocopying the document a thousand times). In cases where
the publisher and creator are the same person(s)/machine, then the simple
solution is to set both to the same value.
2) "For most sites, the publisher is the web master and not usually relevant for
retrieval"
I beg to differ. If I'm collecting material for an essay, research paper or
lawsuit, I don't care how "obvious" it is that a particular person is the
publisher. Sure, "everyone" may know that the "Unleashed" series of books are
published by SAMS, but does that give me the right to declare that "For the
Unleashed series of publications, the publisher is SAMS and not usually relevant
for retrieval"?
What if I want to find all the web sites run by a particular entity? Wouldn't it
make my life so much easier to just search for "Doe, John <IN> DC.Publisher"?
3) "For most sites, the creator is the web master and not usually relevant for
retrieval"
See point 1 for most of this argument. What about a site which serves a
telescope connected to the Internet? The Creator then is not the webmaster, but
the mechanical/software entity that is the telescope.
If DC is going to be used for more than just sticking fancy tags on HTML pages
on the Web, then we'll be best served by keeping this structure - 15 elements,
with the option of using or ignoring the "qualified" part. If you want to write
a search that combines three fields into one "Agent" or "Responsible", then do
that at the Search Engine interface, rather than butchering the DC data. The
search technology I'm most familiar with is Verity Search 97 - this software
allows you to "filter" a search query before it's matched against an index, so
if a user selected "Doe, John <IN> tSA.Responsible", the filter would convert
that to ("Doe, John" <IN> DC.Creator)<OR>("Doe, John" <IN>
DC.Publisher)<OR>("Doe, John" <IN> DC.Contributor), which achieves the desired
purpose.
</IMHO>
Regards,
Alex Satrapa
Weibel,Stu wrote (on behalf of Martin Dillon):
> At the OCLC Institute we have conducted many workshops which apply the
> Dublin Core to various sites and collections. In all I would say that
> different
> work groups have explored its usefulness across 20 or so applications
> ranging from a collection of maps to a university home page. In every
> single
> one of these applications confusion over the use of the three fields in
> question
> was the paramount concern of the participants. They had many objections;
> let me list a few.
>
> It is often not possible to distinguish between publisher and
> creator.
>
> For most sites, the "publisher" is the web master and not usually
> relevant for retrieval.
>
> For most sites, "creator" is also the web master and not relevant
> for retrieval.
|