Re: metadata for the web

Ralph R. Swick wrote:
> 
> On 1/30/97, Larry Masinter wrote:
> >...
> >"Representing the Dublin Core within X.500, LDAP and CLDAP"
> >...
> >I think it would make a good addition to the WEBDAV home page.
> 
> Good pointer, Larry.  There is also a proposal in the works to
> extend PICS values (dropping the values are rational number
> restriction) specifically to be able to use PICS as a transfer
> syntax for Dublin Core -- and other -- metadata.  See
> 
> http://www.w3.org/pub/WWW/PICS/970113/DigiLib/pics970113.htm

Everyone should also be aware that the group that developed
the Dublin Core has extended the model to include some
transport concepts. The best discussion of this proposal
can be found in the Dlib article that discusses the entire approach:
<URL:http://www.ukoln.ac.uk/dlib/dlib/july96/07weibel.html>
The basic summary is that there should be things called
containers that hold metadata of different types. Then
there is some type of relationship entity that shows the 
relationships between different containers. One of the
approaches uses MIME's multipart/related.

In addition to the X.500/LDAP/CLDAP approach there is
also the application/directory MIME type that the directory services
folks are working on. For very flat and what I consider low overhead
metadata this approach gives you the ability to handle community
specific metadata to some degree as well as encodings and character
sets. The application/directory stuff is described in:
<URL:http://ds.internic.net/internet-drafts/draft-ietf-asid-mime-direct-03.txt>          

With all of this in mind I'm becoming increasingly worried that
with so many parallel metadata transport/encoding methods that we
may be fracturing a part of the infrastructure that 
desperately needs some commonality. While it is increasingly
obvious that there is no concensus for an end-all-be-all
metadata format there does seem to be concensus that they
should be transported as MIME objects and that we should
standardize on a few formats and schema (classes?)

As far as WebDAV is concerned we should take a look at
each of these current metadata proposals with an eye
for how WebDAV can act as the primary seed for filling
in each of these containers. Since WebDAV is currently
the part of the infrastructure that is closest to the
author then we should make the metadata formats as
flexible as possible.

Thanks!

-MM

Received on Friday, 31 January 1997 10:22:43 UTC