Re: Element sets, schemas, and record syntaxes

On Thu, Feb 19, 2004 at 03:50:33PM +0000, Mike Taylor wrote:
> For XML records, there is no distinction between schema and
> element-set; or, to say the same thing a different way, the concept of
> element-set is subsumed into the broader concept of schema.  You would
> not ask for XML records in the Foo schema using "F" and "B" element
> sets; instead, you would ask for XML records in Foo-F and Foo-B
> schemas.  For the sake of simplicity, this information (the string
> indicating the schema name) is carried in the element-set name field,
> in according with an implementor agreement (or something) made last
> year.

I think its wrong to say element sets are part of schemas. Element
sets are defined based on elements of a schema, but are also defined
per record syntax - that is, you can define an element set to be
different for MARC, GRS-1, and XML. This makes sense because the
expressive power of the different record syntaxes is different.

But from what you are saying, there are pairs of schemas and element
set names that only appear with each other. That is, if you ask for
"Foo-B" as an element set name without specifying the schema in
the present request, you look around for an element set definition
that is defined in any of the available schemas (for the selected
record syntax) and use that. So the element set name *implies* the
schema to use. Same effect as what you are saying.

> ... though of course Explain Classic is at best moribund these days.

It may be, but we still use it every day in our product. We had to
add a new category to map CCL/CQL like field names to attribute lists
reliably, but other than that Explain Classic is useful. (We do not
support all of it mind you - there is a lot of stuff we do not need/use.)

Thanks
Alan

Received on Thursday, 19 February 2004 18:15:33 UTC