[tcs-lc] Names as Objects

[Richard Pyle]

Thanks, Bob.

> You can use the "union" type. There are examples in UBIF.EnumLib.xsd
> There is no real support for boolean combinations in the XML type
> system, so if your combinations get complex, the unions get complex too,
> but as the examples in UBIF show, for simple cases it works OK. However,
> If you are after structural decomposition, like "Person is a kind of
> Agent" you probably want to use the classical type derivations system,
> most likely type "extension". In this case, the really common stuff is
> usually put in an abstract type which is never instantiated. There are
> examples in UBIF, as I recall, and certainly in SDD

I'm not sure I speak XMLese fluently enough to understand how to represent
that.  I want to establish a set of elements within TaxonConcept that are
available only for a certain defined subset of TaxonConcept instances.  I'll
check the examples in UBIF.

> Since this topic has nothing obvious to do with "Names as Objects", am I
> supposed to change the Subject line in the email?

But it does have something to do with "Names as Objects" -- because I'm
specifically trying to understand how to represent in an XML schema the idea
of creating a set of nomenclatural elements that are only available for
instances of TaxonConcept where "Type='Nominal'".  This is exactly the sort
of question that is better asked on an email list, than on a Wiki page,
because it is an aside to the main topic of conversation.

Rich