Ed,
You make an excellent point. One of the problems that I have encountered
in making Dublin Core records is the lack of any rules for input. All we
have is a place holder for "author" (agent?) "publisher" "title" and so
on, but there are no guidelines for input.
As a result, I have seen "author" information in Dublin Core records on
the web in all sorts of forms. Forename-surname, surname-forename,
nicknames; and who knows what else that I haven't seen?
I return to my point: librarians will have no control over what goes
into the major web indexes. We can make our wonderful records,
completely controlled, etc., but if they go into Alta Vista, they will
be side by side with the Dublin Core that somebody's 5th grade class
made for their class website, along with all the business and porno
pages. And while we're at it, we'd better make sure we don't bump these
businesses out of first place on the results pages, because then they'll
haul us into court.
These are new issues for libraries and cataloging. We've never had to
compete like this before. We can make the most wonderful Dublin Core
records in the world (or even full AACR2) and they'll get buried in the
morass of Alta Vista pages. My concern is: if we promise too much to the
administrators, then implement Dublin Core with the same results of no
one can find anything, it could be a disaster.
The only way Dublin Core (or whatever metadata scheme) can work is to
begin to build our own search engines. Then we have the control we need.
Jim Weinheimer
Princeton University
[log in to unmask]
BTW, I have some DC metadata ready, but I'm doing a workshop for
Princeton's Catalog Division. I'm waiting for all of our catalogers to
create the metadata together.
Ed McNeeley wrote:
>
> I was out last week and without e-mail (by choice, and it was nice!)
> so I read all of this Thanksgiving thread on DC and search engines
> in one sitting. It looks to me like the thread runs the risk of coming
> back to the issue of what the major search engines will or won't do
> with our metadata before some really relevant (IMHO) items get
> some attention. I'm thinking of .........
>
> (1) Debbie Cambell's response to the original post.
> "As a consequence, in Australia, various organisations and portals
> have developed their own, or are currently developing them."
>
> (2) Jim Weimheimer
> "If our (librarians') dreams of metadata are going to work, we must
> give up on Alta Vista, Yahoo, etc. and develop search engines that
> serve our own needs (which include identifying, acquiring, selecting,
> describing, arranging, storing, and retrieving). "
>
> (3) Garry Forger
> "Part of our process is to divorce ourselves from the chaos that the
> entire web consists of, but use the technology to allow individuals
> to design and deliver html, sound, video etc over the web."
>
> (4) Alex Satrapa
> "So write indexers that handle the metadata for all the resources
> that *you* are responsible for. Then write a search engine that
> brokers your search to a bunch of known indices."
>
> (5) Dianne Hillman
> " In my experience, the people creating the web-based information
> resources, *even if they are working in libraries* are not thinking
> about metadata and what will be necessary to provide access to
> their web-empires once they reach a certain critical mass."
>
> I believe the comments of these five wise people hint at a
> fundamental problem; we want them (the major web search
> companies) to value DC before we've convinced our own Library
> institutions and organizations to value DC. I took a quick look at
> some of the sites of the folks that have posted to this thread, and a
> few pages from institutions that could be considered close to the
> issue and I think the problem is much closer to home than Alta
> Vista.
>
> - The only tag I found at the National Archives of Australia
> (http://www.naa.gov.au/index.htm) was......<META
> NAME=GENERATOR CONTENT="Claris Home Page 3.0 30 Day
> Trial">
>
> - None at Princeton....http://libweb.princeton.edu:2003/
>
> - At http://amol.org.au/ is see "NAME=title", "NAME=keywords",
> and "NAME=description"
>
> - At http://www.gloscc.gov.uk/circe/ I found "NAME=keywords",
> "NAME=description", and "NAME"="Microsoft Theme" (Now I'm
> really concerned!)
>
> - At http://www.ala.org/ there is no metadata, even on the
> documents that talk about metadata.
>
> - At http://www.w3.org/ there are a few tags of the "<meta HTTP-
> EQUIV=" type, and some PICS stuff, but I couldn't find much
> metadata on pages below 'home'. (to be fair I did find a document
> that used DC on the W3 site, but it was Stu's 1996 paper "A
> Proposed Convention for Embedding Metadata in HTML." )
>
> - At OCLC (http://www.oclc.org/oclc/menu/home1.htm) the 3 or 4
> tokens are also of the "<meta HTTP-EQUIV=" form, and I find the
> ever present <meta NAME="GENERATOR" CONTENT="Microsoft
> FrontPage 3.0">
>
> - And last, but certainly not least, at http://purl.org/DC/ there was
> no metadata. Shouldn't there be, and shouldn't that metadata be
> Dublin Core metadata?
>
> Only the Cornell WebGoddess (Dianne) and Gary (University of
> Arizona) get an A+ for using Dublin Core in live, public pages. A
> few servers didn't respond, so maybe there are a few others that
> use DC 'in house', and maybe (like at my library) its kinda 'in the
> works'......but......
>
> If only a few of the organizations we work for are using metadata on
> public pages in a fashion that promotes what we consider to be the
> best practice, then perhaps we should be more concerned about
> our inability to convince our colleagues and peers, supervisors,
> directors, trustees and funders that DC is useful than our inability
> to convince anybody else.
>
> I've got some more to say on this subject, but I'd like to 1st ask the
> group what way they want this search engine thread to go? And,
> for the record, I know how difficult it can be to convince colleagues
> and peers, supervisors, directors, trustees and funders that the sky
> is blue, but I still think thats the more important task.
>
> ed
>
> ******************************************************************************
> Edward McNeeley [log in to unmask]
> Delaware Division of Libraries
> 43 S. DuPont Highway VOICE (302) 739-4748 ext.116
> Dover, Delaware 19901 FAX (302) 739-6787
> *****************************************************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|