[WEB4LIB] Re: Pointer Records for Metadata

[Jerry Kuntz]

    First, a question: is metadata supported in other file formats commonly
published on the Internet (and indexed by many of the major robot search
engines): .gif, .jpg, .doc, .zip, etc? If so, what tools allow the creators
of these files to enter the metadata elements?
    Secondly, at the risk of opening a can of worms or displaying my
ignorance (or both!), what's the point behind Dublin Core metadata? I
understand that a standard, thorough list of metadata elements would help
document authors identify the pieces of information they wish to offer as
access points. However, do the major search engines that now use metadata in
indexing and weighing hits treat Dublin Core metadata differently than
generic metadata?
    If the point of Dublin Core metadata is to support the indexing and hit
display of the major search engines, is there a plausible scenario that will
see enough document creators use it to make a difference? [I've heard the
scenario about all web publishing going through an authentication agency--it
doesn't sound plausible.]
    On the other hand, if Dublin Core is meant to support smaller indexing
projects, such as a site search index or a harvested index of targeted
sites, why do we want another inflexible standard that can't accommodate
differences in audience, scope, or topic? Or was Dublin Core designed with
an academic audience implied?
    As an example, the KidsClick! database of web sites for children that I
help manage uses metadata as the structure for the individual database
record [and the SWISH-E search engine can see the indexes produced from
those elements.] However, we needed elements not covered by Dublin Core:
reading level (with three possible values) and amount of illustrative
material (with three possible values.)

    ...apologies to for straying so far off the thread...
Jerry Kuntz
Ramapo Catskill Library System
jkuntz at rcls.org
-----Original Message-----
From: Brian Nielsen <b-nielsen at nwu.edu>
To: Multiple recipients of list <web4lib at webjunction.org>
Date: Thursday, November 05, 1998 12:45 PM
Subject: [WEB4LIB] Re: Pointer Records for Metadata


>David's question about PDF metadata leads me to suggest that the web
>publishing community approach Adobe to ask that they support IMS metadata
>and Dublin Core within the metadata framework that currently exists for
>PDF.  One can add title, author,  and keyword info into a PDF file now, and
>such an approach seems to me much more manageable than the dissociation of
>the metadata from the file that David suggests.
>
>I copy this note to technical folks at Adobe; maybe they have some
>suggestions.
>
>Brian
>
> Brian Nielsen, Ph.D.
> Manager, Learning Technologies Group
> Academic Technologies
> Northwestern University
> 2129 N. Campus Drive
> Evanston, IL  60208-2850
> (847)491-2170   fax:(847)491-3824
> email: b-nielsen at nwu.edu
> http://www.nwu.edu/people/b-nielsen
>
>
>At 09:24 AM 11/5/98 -0800, you wrote:
>>I'm embedding metadata in our HTML files.  However, more of our files
>>are going up in PDF format.  These are unable to have metadata embedded.
>>
>>It is possible to create a pointer record holding only metadata and a
>>redirect to the PDF file.  That way spiders could find a record, display
>>a title, description and URL and the user could be directed invisibly to
>>the PDF file.
>>
>>Sounds workable to me.  Is there something I'm overlooking?
>
>