>>> [log in to unmask] 11-08-2005 4:04 >>>

I do not see the actual difference between

<dc:dateCreated xsi:type="ISO8601">2001-01-01</dc:date.created>
<element id="DC.Date.created" encoding="ISO8601">2001-01-01</element>

The first one is according to the DCMI guidelines for encoding dc in
XML which I think we should follow as much as possible.  However DCMI
defines the terms and not the structure of the record where these terms
are part of. It is also valid to add other elements from other
namespaces. Therefore DCMI allows a lot of flexibility and freedom. For
what you are trying to do I would advide you to use RDF according to the
DCMI guidelines.

The main problem is how we deal with schemas. An XML schema which is
useful at the time of creation of metadata doesn't have to be meaningful
at the time of search and retrieval. At the time of search and retrieval
it might be enough that the record contains for example a title and a
resolvable identifier, while at the time of creation one may validate
for example against controlled vocabulaires for specific subject
headings. Therefor I think that the schema used for local validation is
not relevant for the outside world. Besides that there can be 1000 s.
For specification of what a record actually means for the outside world
we need a common schema which defines common elements.
I think that the different opinions in this group relate mainly to the
amount of freedom that a schema allows and we need to find a reasonable
compromise between "ending up with thousands of different restrictive
schemas that all are doing about the same"  and  "having a single schema
that has no restriction at all".