[
Lists Home |
Date Index |
Thread Index
]
James,
Please see comments below.
<Quote>
But there's no way to define in XSD to define what
an element or attribute means. XSD can be used to declare an attribute
as an ID, but this is just a constraint of uniqueness and not of
meaning. Why not use RDF to describe the semantics of the markup of the
document?
</Quote>
An approach that I have been using in the US federal govenment space
that is a sort of "middle ground" (i.e. not as robust as RDF, OWL, etc.)
for expressing the semantic meaning of elements is the ISO/IEC 11179
standard [1]. 11179 is a very mature standard for metadata registries,
and is comprised of 6 parts (I will refer you to [1] for these). Part 5
- Naming and Identification Principles - is the one I am referencing
here.
Part 5 identifies 3 main "components" that comprise an 11179-compliant
data element name:
(1) Object Class - represents an activity or object in a context
(essentially, "what is this item about")
(2) Property Term - serves to give the data element a unique
identification, given the other 2 components (i.e. what makes this
element unique among other data element that have the same Object Class
and Representatation Term)
(3) Representation Term - describes the form of the representation of
the data element (ex: Name, Amount, Number, Text, Quantity, etc.)
So an example would be (pipes serve only to separate components):
Employee|Last|Name
Where Employee=Object Class, etc. (in order from left to right)
Of course, we could have an element named EmployeeFirstName - this is a
case where the Property Term (First) distinguishes between all other
elements of the format "ElementXName". The specification goes further to
describe cases in which more qualification may be needed - that is, it
describes "Object Class Qualifier" and "Property Term Qualifier"
components in which one may need to differentiate between (for example)
a "CurrentEmployeeLastName" and "FormerEmployeeLastName" (where Current
and Former are Object Class Qualifiers).
Again, this approach does not in any way offer the same level of
semantic clarity that OWL, RDF, Topic Maps, etc. do (not even closely) -
but perhaps it is a step toward improved semantic clarity via the use of
a standard framework for naming data elements.
Kind Regards,
Joe Chiusano
Booz | Allen | Hamilton
[1] http://metadata-stds.org/11179
"James F. Cerra" wrote:
>
> A few days ago, I was reading about the silly HLink [1] and XLink [2]
> debate on XML.com. Along with nearly everyone else it seems [4], I
> wondered why XLink just hasn't caught on. This issue is a little scary
> to me, since the debate feels like the older one about ID in the
> XPointer scheme [5,6]. (Note: I still can't find any articles on the
> current general consensuses about the ID-debate.) Now that
> three-fourths of the XPointer Framework have been stabilized [7], I
> would *really* hate to see XPF to fall into oblivion like XPointer seems
> to have.
>
> It seems to me that both of these issues are about defining the
> semantics of xml tags in a machine-understandable way. HLink was
> invented to describe which markup represents a hyperlink [1], while most
> of the ID-issue solutions seem to try to declare what markup represents
> an ID for the processing machine [6]. Both of theses issues stem from a
> common problem of explicitly defining what an element's name means,
> whether an ID or a Link in these cases. In general, I suspect that this
> problem will block any attempt to define a standard that applies to a
> ***specific type of element*** rather that to the content or document
> itself.
>
> Note the emphasized text, 'specific type of element.' Defining a
> thing's type and meaning is an issue that has been of interest to the
> W3C and computer engineers for a while [8], and it has influenced the
> design of XML. Indeed, Tim Bray says that one of the main advantages
> that XML has over other languages is when it is used to describe the
> semantics of a document [9].
>
> However, there haven't been many ways to explicitly define this metadata
> until Topic Maps [10] and RDF [11] were invented. Although Topic Maps
> and RDF applications have been used mainly to describe document content,
> they haven't often been vectored for describing the semantics of the
> markup itself (in my limited experience).
>
> But why not? Xml Schema Definition Language (XSD) was created as a more
> flexible way to describe the structure of the markup of a document [12].
> It can be used to define what elements should be nested where. It can
> define the structure of text in an attribute or element (such as a
> string or integer). But there's no way to define in XSD to define what
> an element or attribute means. XSD can be used to declare an attribute
> as an ID, but this is just a constraint of uniqueness and not of
> meaning. Why not use RDF to describe the semantics of the markup of the
> document?
>
> So I tried it. Just as XSD describes the logical structure and
> organization of elements in an XML application, RDF can be used to
> define - in a machine-readable way - what different elements in an XML
> application mean. This allows the semantics of a particular markup
> language to be explicitly defined. Furthermore, RDFS allows one two
> apply the semantics of one markup language to another, as the example
> below demonstrates:
>
> <h:html
> xmlns:r="http://www.w3.org/1999/02/22-rdf-syntax-ns#";
> xmlns:s="http://www.w3.org/2000/01/rdf-schema#";
> xmlns:h="http://www.w3.org/1999/xhtml";>
> <h:head>
> <r:RDF>
>
> <s:Class r:about="http://www.w3.org/1999/xlink#href"; />
>
> <s:Class r:about="http://www.w3.org/2001/XMLSchema-datatypes#ID"; />
>
> <r:Description r:about="#xpointer(/*//*/@href)"
> r:type="http://www.w3.org/1999/xlink#href"; />
>
> <r:Description r:about="#xpointer(/*//*/@src)"
> r:type="http://www.w3.org/1999/xlink#href"; />
>
> <r:Description r:about="#xpointer(/*//*/@id)"
>
> r:type="http://www.w3.org/2001/XMLSchema-datatypes#ID"; />
>
> </r:RDF>
> <h:title>foo</h:title>
> </h:head>
> <!-- rest of document -->
> </h:html>
>
> In the example above, the all "href" and "src" attributes are declared
> to be a type of XLink "href" attributes. Also, the "id" attributes of
> all elements are declared to be a type of XML Schema datatype
> (specifically, "ID"). Thus, this mechanism may help when using several
> different markup languages where some of the semantics overlap between
> them.
>
> Of course, the above RDF document could be stored in a separate file and
> linked to html document as the semantic-analogue of an XML Schema
> (though the XPointer expressions would be different).
>
> But I'm definitely not an expert, so I leave this to yinz.
>
> --
> James F. Cerra
>
> [1] http://www.w3.org/TR/hlink
>
> [2] http://www.w3.org/TR/xlink
>
> [3] http://www.xml.com/search/index.ncsp?sp-q=hlink&search=search
>
> [4] http://www.xml.com/pub/a/2002/03/13/xlink.html
>
> [5] http://www.xml.com/pub/a/2001/11/07/id.html
>
> [6] http://weblogs.userland.com/eclectic/xml-identifiers.html
>
> [7] http://www.w3.org/2003/03/xpointer-pressrelease
>
> [8] http://www.w3.org/2001/sw
>
> [9] http://tbray.org/ongoing/When/200x/2003/04/09/SemanticMarkup
>
> [10] http://www.ontopia.net/topicmaps/materials/tao.html
>
> [11] http://www.w3.org/TR/rdf-primer
>
> [12] http://www.w3.org/XML/Schema
>
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
>
> The list archives are at http://lists.xml.org/archives/xml-dev/
>
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>
begin:vcard
n:Chiusano;Joseph
tel;work:(703) 902-6923
x-mozilla-html:FALSE
url:www.bah.com
org:Booz | Allen | Hamilton;IT Digital Strategies Team
adr:;;8283 Greensboro Drive;McLean;VA;22012;
version:2.1
email;internet:chiusano_joseph@bah.com
title:Senior Consultant
fn:Joseph M. Chiusano
end:vcard
|
