I am strongly in favour of being able to string together like
items belonging to the one scheme, e.g. subject headings. One can
have quite a few keywords and to have each with its own metatag
is going to increase band width unnecessarily.
For searching (which is what we are mainly doing this for), thesearch
engines searching for a string will find them just as easily (if not
more easily) if strung together in one line as they would if placed in
individual tags.
In Environment Australia's Online Service, we have decided that it
is the better way to go.
Arthur
__________________________________________________________________________
______
Arthur D. Chapman [Scientific Coordinator, Biogeographic Information,
ERIN]
Environmental Resources Information Network internet:
[log in to unmask]
GPO Box 787, Canberra, voice: +61-6-274 1066
ACT 2601, AUSTRALIA fax: +61-6-274 1333
>On Wed, 2 Jul 1997, Paul Miller wrote:
>
>>We add the ADS are currently most interested in your first option,
where
>>all of the values are given in a single string. We would only split
them
>>if the values were drawn from more than one SCHEME or TYPE, as each use
>>of
>>an element is only allowed ONE SCHEME and/or TYPE.
>
>I wasn't following this thread too closely until I realized it affected
>other elements besides "coverage." The "Hull, Yorkshire, England, etc."
>example is somewhat comparable to a resource with multiple controlled-
>vocabulary subject headings.
>
>Keywords are generally strung together in one meta tag, as Paul Miller
>suggests doing with geographic values, but in the couple examples I've
>seen using Library of Congress subject headings (LCSH) as the scheme,
each
>subject heading has had its own meta tag, e.g.
>
><META NAME="DC.subject" SCHEME="LCSH" CONTENT="xxx">
><META NAME="DC.subject" SCHEME="LCSH" CONTENT="yyy">
><META NAME="DC.subject" SCHEME="LCSH" CONTENT="zzz">
>
>I was about to implement something like the example above because that
>seemed to be what was out there already. But if geographic values will
be
>strung together, as long as they're from the same scheme, then it seems
>that all LCSH values should also be strung together (the approach I'd
>prefer, actually).
>
>I haven't yet seen a conclusive answer in the discussion of the
>"coverage" element -- is there anything in DC, as it currently stands,
>that dictates that one format is "right," or at least that one is
>preferred over the other?
>
>Diane Madrigal
>New York State Library
>State Education Department
>Albany, NY
>
>
|