RE: [xml-dev] 'is-a' Relationships in XML?

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

My gut feel is "no", because genericode is suitable for expressing a 
fully- or sparsely-populated keyed table.  Given that both ontologies 
and schemas are hierarchical, and the table and its key are flat, it 
may take too much shoehorning to suitably express context when 
ontology and schema members have the same names in different contexts.

For schemas expressed using naming and design rules such as UBL's 
NDRs, it starts to make more sense to use genericode because the 
uniqueness for the key is found in the parent/child relationship of 
elements.  Context is uniquely found in parents in UBL.  I can 
imagine two columns: parent and child, and the genericode key being 
the combination of the two, and then supplemental information about 
that unique context in the columns.

Of course in UBL there already exists a single key value for 
parent/child relationships, that being the dictionary entry name, but 
that may not be true for other schemas that have uniqueness in the 
parent/child combination.

For the next issue to go two ways as you ask, it would then be 
necessary in genericode to express every member of the ontology 
uniquely.  So it would depend on properties of the ontology to know 
if such a key-able value exists.

I hope this helps.

. . . . . . . . . . . Ken

At 2010-05-08 10:42 +0000, stephengreenubl@gmail.com wrote:
>Maybe OASIS open's genericode could be used for mapping two ways 
>between ontology and schema for the markup.
>
>Stephen D Green
>
>
>-----Original Message-----
>From: Stephen Green
>Sent:  06/05/2010 4:52:07 pm
>Subject:  Re: [xml-dev] 'is-a' Relationships in XML?
>
>Maybe, as with character sets, what we need are three kinds of artefacts
>to more completely define the markup
>
>1. schema (to define the structure and structural constraints)
>2. ontology (to define the semantics - the real world things and
>concepts to which the markup relates)
>3. mapping table (to map between 1. and 2.)
>
>Again, as with my analogy to character sets and their mapping tables,
>we'd ideally need to use the table when 'understanding' the underlying
>markup (as a table is used to 'understand' ASCII in its binary/hex form).
>
>In some cases it might be possible to combine 2. and 3. (or even 1. and 3.).
>E.g. maybe RDF's 'about' can be used somehow with a combination of URL
>(or the like) and XPath (or the like) - or just XQuery??? - to denote
>the mappings.
>Or maybe the RDF would want it the other way round to map XML to the
>meaning (mapping XPaths/XQueries to RDF classes and properties via 'about'?)
>But the lookups need, I think, to be able to work both ways, which might be
>more than can be done with RDF alone.
>
>Best regards
>
>Steve
>---
>Stephen D Green


--
Principles of XSLT for XQuery Writers: San Francisco,CA 2010-05-03
XSLT/XQuery/UBL/Code List training: Trondheim,Norway 2010-06-02/11
Vote for your XML training:   http://www.CraneSoftwrights.com/x/i/
Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/x/
G. Ken Holman                 mailto:gkholman@CraneSoftwrights.com
Male Cancer Awareness Nov'07  http://www.CraneSoftwrights.com/x/bc
Legal business disclaimers:  http://www.CraneSoftwrights.com/legal