that's why i've kept my coffee table: when we're already busy with the
coffee and cake, and the next guests show up with the champagne, i don't
need three arms.
Steven R. Newcomb wrote:
> [James Anderson <James.Anderson@mecom.mixx.de>:]
>
> > i'm concerned about the case were i want to get at [the base
> > architecture's parameter entities], not be isolated from them. i
> > would expect to have to do something like the following. suppose
> > that i have two base architectures, each with an element "name". i
> > would like to derive an architecture to comprise descriptions of
> > both sorts. nothing new.
>
> > <!-- base arch 1. -->
> > <!ELEMENT name ((first, last) | %name-content;) >
> > <!ATTLIST name
> > address-form (PSEUDONYM | LEGAL) "LEGAL">
> >
> > <!-- base arch 2. -->
> > <!ELEMENT name %name-content; >
> > <!ATTLIST name
> > address-form (ASSOCIATIVE | LITERAL) "LITERAL">
>
> > while the provisions of enabling architectures make it possible to
> > use architectural form attributes to disambiguate the "name" element
> > names by creating a 2-d namespace for element names, and to
> > disambiguate the "address-form" attribute names by using the
> > remapper attribute to, likewise, create a 2-d namespace for
> > attribute names, and then to map them onto unique positions within a
> > 1-d namespace, i have not found the provision to enable declaring
> > distinct values for the entities. how would one distinguish
> > %name-content from %name-content in a derived architecture in order
> > to specify the respective entity declarations?
>
> > maybe i don't need to. i thought i would.
>
> I don't think you need to. When you create an element subtype from an
> element type in a base architecture (a supertype), it doesn't matter
> whether the content model of the supertype was expressed by means of
> one or more parameter entities. For all purposes of subtyping, it
> only matters what the replacement text of those parameter entities
> was. I can see where it might sometimes be convenient to re-use the
> same parameter entities that were used in the base architecture, but
> there is no provision in the current enabling architectures syntax for
> that, and it's not necessary, anyway. I doubt it would be worth the
> added syntactic complexity. You would have to cause the names in the
> replacement texts of the supertype's architecture's parameter entities
> to be somehow automagically translated into the corresponding names in
> the subtyping architecture.
>
> [I'm trying to use the "subtype/supertype" vocabulary instead of the
> "inheriting/inherited" vocabulary here; does it work better?]
>
> -Steve
>
> --
> Steven R. Newcomb, President, TechnoTeacher, Inc.
> srn@techno.com http://www.techno.com ftp.techno.com
>
> voice: +1 972 231 4098 (at ISOGEN: +1 214 953 0004 x137)
> fax +1 972 994 0087 (at ISOGEN: +1 214 953 3152)
>
> 3615 Tanner Lane
> Richardson, Texas 75082-2618 USA