Jerome McDonough wrote:
> 
> but note that it is also a redesign of the schema to make it play 
> well with others.  If we were to take a completely independent design
>  approach, we wouldn't need to worry about interoperability, and we 
> wouldn't need  to make the change you suggest.

That would rather beg the question, why get mired down in XML Schema at
all, when a DTD, or even plain old English can do the job of specifying
the PREMIS standard?

To properly exploit the extensibility features of XML Schema (rather
than simply writing a DTD in XSDL, which is pretty much the case now) is
to enable PREMIS not only to have an independent definition of its own
ideal discrete record structure (mandatory elements and all), but also
to play well with /anything/ else that also has an extensible XML
implementation.

The redesign of the PREMIS XML Schema (not the PREMIS DD) is necessary
and pressing because it is currently a naïve implementation that /only/
supports the validation of standalone PREMIS records. The better
approach would be a separate library of PREMIS typedefs, that can be
used, refined and extended not only in the core PREMIS.xsd, but also by
developers of repository-specific PREMIS systems, and other hybrid APs.
This approach will also reduce the risk of the sort of type-clash that
invalidated the latest draft of 1.1.

Of course, it's not going to play nicely with anyone who wants to read
the simplest kind of raw schema, in a straight line from start to end - 
but that position was never likely to be sustainable.

Richard



-- 
/
\ Richard M Davis
/ Digital Archives Specialist
\ University of London Computer Centre
/ Tel: +44 (0) 20 7692 1350
\ mailto: [log in to unmask]
/