On Fri, 13 Jun 1997, Bruno Kestemont wrote:
> I wonder why dublin core entities should be named "dc.entityname" and not
> simply "entityname", dc beeing supposed to be used by default. An entity
> name like "dc.description" is not natural: nobody will use it if he does not
> know DC.
The xx.yyyyyy form of names in META elements in HTML was a convention
devised a W3C working group. As DC is just one of an open ended range of
metadata standards that users may wish to use, making it the "default" so
that it didn't need to be preceeded with "dc" would turn it into a "first
amongst equals". Suddenly DC would have a competitive advantage over
other metadata formats, even ones not yet dreamed up. That sounds like
bad news to me; it would make it difficult for new metadata schemes to
gain a foothold outside of local niches.
> Or wouldn't it be easier to agree on the most used entity names
> ("description", "keywords", "author") as DC standard entity names (to be
> used by default on the WWW), giving the possibility to anybody to add or use
> other entities by specifying the scheme and the language if they want
> specific properties for the content.
We've been through this loop on the meta2 mailing list already (several
times I think) and its probably not wise to revisit it yet again so soon
after the 15 core elements were pressed into clay (I was going to say "set
in stone" but I doubt that's the case). I think we need a couple of
years of real usage of DC as it standards before its twiddled with again.
If we keep on changing element names, etc. Having said that, the DC 15
elements does include Description already and Keywords and Author are
synonyms for Subject and Creator (the latter two being the choosen name is
they are more general).
> Of course, one could argue that each robot could translate the entity names,
> recognising "object-type" as synonym of "dc.type", "function-type",
> "document_type", "objecttype", "resource-type", "vw96.objecttype",
> "index-type" etc. But can we ask them that?
At the moment precious few robots do anything with Dublin Core, especially
the "big" commercial services. We shouldn't need to ask for robots to do
anything; handling metadata schemes should be something that they just do
because it would add value to the indexing service that the robot forms
part of and make it more attractive to end users. But that won't happen
until there's quite a bit of metadata out there, which won't happen until
a large number of people realise that adding metadata is a good thing that
will affect how easily people can locate their resources. Its a "chicken
and egg" situation.
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. *
|