Having looked a little closer, I suspect that these may be possible using
the links section if we define some appropriate values for type (such as
logo, news, copyright info).
However, I also spotted a bug in the DTD:
I suspect that
<!ELEMENT links (subject+)>
<!ELEMENT link (#PCDATA)>
<!ATTLIST link
type CDATA #IMPLIED
>
Should really be
<!ELEMENT links (link+)>
<!ELEMENT link (#PCDATA)>
<!ATTLIST link
type CDATA #IMPLIED
>
Matthew
> > On a related matter, one of the other concerns from the
> > metasearch crew is about branding. One of the things I
> > focused on in a report I did for the JISC on service
> > descriptions is the wealth of bradning type information in
> > traditional Explain. For instance:
> >
> > An icon used to represent this database (in machine
> > presentable form) (today this would probably be termed a logo
> > rather than icon)
> > Whether there is charge to access this database.
> > A human-readable name or title for the database
> > A description of the database, in human readable text.
> > Any disclaimers concerning this database, in human readable text.
> > News about this database, in human readable text.
> > Hours of operation that this database is available.
> > Best time to access this database, in human readable text.
> > A description of copyright issues relating to this database,
> > in human readable text.
> > A notice concerning copyright which the target expects the
> > origin to display to the user if possible, in human readable text.
> > Description and contact information for the database
> > producer, database supplier, and for how to submit material
> > for inclusion in this database, in human readable text.
> >
> > At a quick glance many of these aren't yet in ZeeRex. Some
> > may need a little thought (e.g. News about database might be
> > better as a link to an RSS feed, or even embedded RSS, than a
> > string?) but I would have thought it not too difficult to add
> > these in ZeeRex 1.9/2.0 and hence tackle this issue. Also we
> > may need to consider branding information in our record
> > metadata discussions.
> >
> > Matthew
|