On Wed, 25 Sep 1996, Lauren Wood wrote:
> I can see that using a scheme is a good idea for some of the metadata that may
> only be relevant to the scheme search engine or application, but surely we
> could use some common words without prepending them with a scheme? Trying to
> explain to people that they should write
>
> <META NAME = "DC.author" content = "Joe">
> <META NAME = "MS.author" content = "Joe">
> <META NAME = "XX.author" content = "Joe">
> <META NAME = "YY.author" content = "Joe">
>
> (assuming that the world of relevant schemes is made up of DC, MS, XX, YY)
> invites uncomprehending looks and a feeling that we must be missing something.
I guess we could extend the convention we've got at the moment to say
that if there isn't a dot in the NAME, we assume that element matches the
attribute set(s) that we're interested in. So then:
<META NAME="author" CONTENT=" Joe">
would be recognised by all the schemas as an author. Of course this
assumes that all schemas can interpret "raw" content values like "Joe" -
I guess if you particular schema can't then you just ignore this and
carry on looking for META elements that explicitly contain metadata from
your schema. And things like DC which can have sub-elements to qualify
the value of the element would still want to be marked with their own
schema as software for processing other schemas won't know that the
sub-elements are there or what they mean.
> What are the common words? Well, author of course, and title, and
> keywords. Not much more.
Yeah, I'd leave it at that. If implementors reckon that they can process
anything else then fair enough but lets be conservative on what we
encourage people to be able to handle.
Tatty bye,
Jim'll
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Jon "Jim'll" Knight, Researcher, Sysop and General Dogsbody, Dept. Computer
Studies, Loughborough University of Technology, Leics., ENGLAND. LE11 3TU.
* I've found I now dream in Perl. More worryingly, I enjoy those dreams. *
|