Re: [xml-dev] RE: Abstraction in Science, Mathematics, Software,and Mark

[Michael Kay <mike@saxonica.com>]

On 12/03/2011 10:34, Costello, Roger L. wrote:
> Hi Stephen,
>
>> Are you suggesting or even hinting that it should be possible
>> to define a type whose contents are types rather than elements?
> Funny you should ask! Yes!
>
> A couple weeks ago I pitched that very idea on the xmlschema-dev list:
>

Well, "complexType" and "simpleType" are the names of two types whose 
instances are types - the set of complexTypes and the set of simpleTypes 
respectively. I guess Roger's "cross-product" is another type whose 
instances are types, though he presented it more as a constructor for 
new types, akin to construction-by-list and construction-by-union.

The IRDS model from the 1980s had four layers: loosely, objects like 
"John Smith", types like "Person", meta-types like "Complex Type", and 
meta-meta-types like "Concept". The things in each layer are instances 
of the types in the layer above. The idea was that the meta-meta-types 
were fixed (they define the ontological framework), but the meta-types 
are completely extensible - the idea being that you can always 
incorporate new "data models" like the relational model, the XML model, etc.

I put "data model" in quotes, because one of the problems is that the 
term is used by some people to mean a set of types like "Person", and by 
others to mean a set of meta-types like "Complex Type".

Michael Kay
Saxonica