On Wed, 4 Dec 1996, Rachel Heery wrote:
> On Wed, 4 Dec 1996, Stu Weibel wrote:
>
> > Here all this time I thought that DESCRIPTION was BETTER than SUBJECT
> > because it was in use elsewhere for roughly the same purpose.
>
> I (and a few others I believe!) promoted inclusion of a description
> element, mainly because most internet search services try to include some
> sort of free text description/abstract; and certainly all subject services
> using ROADS include descriptions in their records.
>
> My proposal was that there would be a description as well as a subject
> element. As part of the last summary/compromise round it was suggested
> that Description would incorporate Subject which could be used as a
> qualifier. That seemed OK to me ...... but what I really wanted was a
> separate element, now you mention it (again).
I would withdraw My objections against "DESCRIPTION" if it would become
piece of prose describing some object.
<meta name="DC.description" content="(type=abstract) A_summary_of_something">
seems absurd to me. And
<meta name="DC.description" content="(scheme=DDC) Some_Dewey_classification">
could at least potentially confuse some spiders, but we could of
course choose to ignore that since it would most likely be a temporary
proplem. Note that these problems could affect any spider, not only
ones that are write by "amateurs". For instance, just a few months ago
it was revealed that someone had put about 20
<title>...</title>
in his/her document. Obviously AltaVista at that time concatenated
all these titles and as a result the document in question became a very
relevant one...
How long such transient problems will last is hard to tell. I have
myself records left from the time before I started to remove
<script>...</script> before making a text extraction, and I've seen
such records being presented by AltaVista just recently. (We are
presumably both using get-if-modified-since, and neither of us expires
the entire database just because we've made some minor change in the
indexer).
>
> I think the important thing is to be able to distinguish a 'hand-crafted'
> description if it exists, as opposed to keywords, just because its so
> useful for display purposes. For that reason it doesn't seem a good idea
> to bury it away within another element.
I agree, and the default behaviour of the old "SUBJECT" was less than
excellent in that respect:
"Without the scheme, the Subject element is a keyword and may
contain any word or phrase that describes the intellectual
content of the object."
Much too vague for me. The current praxis on the web is to embed
<meta name="keywords" content="comma separated list of keywords">
<meta name="description" content="piece of prose">
It is obvious from what is out there today that some authoring tools
enriches the <head>...</head> section with an awful lot of rubbish,
and it is very common with multiple KEYWORDS fields. Anyway, these
two functionalities must be easy to implement in DC.
The following would (e.g.) do the job (if we opt for ignoring some initial
difficulties for some spiders).
<meta name="DC.description"
content="(scheme=DDC) Some_Dewey_classification">
<meta name="DC.description"
content="(type=keywords) comma separated list of keywords">
<meta name="DC.description" content="piece of prose">
>
> Rachel
>
Sigge
________________
Docent Sigfrid Lundberg, Fil. Dr. [log in to unmask]
Lunds Universitets Bibliotek
Netlab, Box 3, 221 00 Lund 046-222 36 83
|