The nub of the argument here is essentially the question of what exactly is
meant by semantics in a computational sense. Until this question can be
answered formally and completely, we're just arguing about interpretations.

My own take, as a non-programmer:

1) If I define a schema or ontology, the tokens defined within that ontology
are not meaningless. While the language is an XML construct (or, for the
sake of completeness, an RDF construct), the specific tokens, their
relationship to one another, and the constraints acting upon their
associated property values or containment relationships are very much
directed by a clear sense of intent.

2) From the standpoint of a lexical parser the tokens are of course
meaningless - they generally describe hash table bindings of one sort or
another - or at a higher level, they develop an acyclic directional graph.
This is the strength of XML - the lexical parser acting upon languages that
are defined using it is, by its very nature, almost entirely devoid of
semantic interpretation.  The XML is syntactical, the language that the XML
expresses is not, necessarily.

3) Semantics would seem to me to be an emergent property of systems, and is
tied to the relationship between different levels of abstraction that a
given system is capable of achieving. Intent, similarly, emerges out of
multiple levels of abstraction, and is a measure to which cascading systems
provide an environment in which the number of interactions are high enough
that it becomes hard to tell what was a rule-following cascade and which was
"intent-ional". In that regard I would tend to concur with the notion that
semantics may in fact be a quantum property, because systems act
stochastically rather than deterministically.

4) This is one of the reasons why it is so hard to decisively state that a
given XML language is "declarative" vs "algorithmic". The XML itself is
fundamentally declarative, but this says nothing of the characteristics of
the language that the XML describes. XSLT is (mostly) declarative, because
it describes a rule based system of nodal bindings. Pure XQuery is
functional rather than fully declarative, and most XQuery extension sets
tend to introduce potentially significant imperative characteristics. For
the sake of argument, I'm assuming that an XML encoding of XQuery, such as
XQueryX, establishes the mapping between XQuery and an XML language. In
other words, once you get into a domain specific language, that domain
introduces a "semantic" that moves beyond the syntax.

Thus, I'd throw Roger's question back out to the list - what does
computational semantics "mean" in an XML context?

Kurt Cagle
Managing Editor
http://xmlToday.org


On Wed, Jun 3, 2009 at 7:47 PM, Peter Hunsberger <peter.hunsberger@g...
> wrote:

> On Wed, Jun 3, 2009 at 6:31 PM, Costello, Roger L. <costello@m...>
> wrote:
> >
> > Hi Folks,
> >
> > I'd like for us to nail down exactly what we mean by "descriptive markup"
> and "algorithmic markup."
>
> Ok, I'll bite, why?
>
> > Here's a stab at it:
> >
> >
> > 1. XML documents just contain markup and data. There are no semantics in
> XML documents. An XML document is just syntax.
> >
>
> XML documents describe a graph.  Computer programs describe a graph.
> The semantics are dependent on who / what is traversing the graph.
>
> >
> > 2. A software application can be written to perform specific, repeatable
> actions when it encounters markup in an XML instance document. Stated
> differently, applications can be written to add a layer of semantics on top
> of the XML syntax.
> >
>
> See above  :-)  But the applications aren't "adding" anything, rather
> they are expecting the syntax of the XML to map to the syntax expected
> by the application so that they can interpret the semantics in
> accordance to the vocabulary the application and the XML presumably
> share.
>
> > 3. When creating an XML vocabulary, it is common practice to write prose
> that specifies semantics for the XML vocabulary. Then software applications
> can be written to implement the semantics specified by the prose.
> >
>
> The semantics can be stored as meta data (a third graph), external to
> both the XML and the application if one wishes...
>
> >
> > 4. Sometimes the prose just assigns a definition to each element and
> attribute in the XML vocabulary. Software applications that implement these
> semantics "understands" the XML vocabulary. These applications don't execute
> any actions when given an XML instance document, beyond simply informing
> itself. For these applications the markup is purely informational or
> descriptive.
> >
> >    For applications faithful to the semantics
> >    specified by the prose, the XML vocabulary
> >    is descriptive markup.
> >
>
> How is this useful?  Configuration?  I believe you are grasping for
> something more...
>
> >
> > 5. Sometimes the prose associates actions to each element and attribute
> in the XML vocabulary. Software applications that implement these semantics
> will execute the actions when given an XML instance document. For these
> applications the markup represents instructions or prescriptions.
> >
> >    For applications faithful to the semantics
> >    specified by the prose, the XML vocabulary
> >    is algorithmic markup.
> >
> >
> > I'm sure that these can be expressed more clearly. I welcome your
> suggestions.
> >
>
> It's not clear what distinction you are trying to draw.  As such, it's
> not clear how to clarify it....
>
> --
> Peter Hunsberger
>
> _______________________________________________________________________
>
> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
> to support XML implementation and development. To minimize
> spam in the archives, you must subscribe before posting.
>
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Or unsubscribe: xml-dev-unsubscribe@l...
> subscribe: xml-dev-subscribe@l...
> List archive: http://lists.xml.org/archives/xml-dev/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>
>