[
Lists Home |
Date Index |
Thread Index
]
[Chiusano Joseph]
>
> I agree that by its nature, XML does not (and was never meant to)
> capture rich semantics and meaning. But I do not agree that XML is
> *completely* devoid of semantics.
>
> Consider the following XML schema snippet:
>
> <xsd:element name="ApplicantEstimatedAmount" type="xsd:decimal"/>
> <xsd:annotation>
> <xsd:documentation>This is the amount that the Applicant has
> requested for...[etc.]</xsd:documentation>
> </xsd:annotation>
>
> Can't one discern the meaning (at some level) of the element above,
> through a combination of a rich (ISO/IEC 11179-based) element name and a
> robust definition provided as documentation?
Well, no - to discover those "rich element names" you have to go beyond xml
and at a minimum digest part of the XML Schema Recs. Without that, you do
not even know that the string "ApplicantEstimatedAmount" is destined to end
up being the name of an element in an xml document, let alone what
"semantics" it might have once it is in there. XML by itself is bringing
__nothing__ to the semantic table here.
> The rest would be up to
> semantic registries such as ISO/IEC 11179 or the ISO Basic Semantic
> Register (BSR[1]), and technologies such as RDF and OWL.
>
They won't have any machine-processeable "semantics" either. Just take
WSDL - you can automatically work out how to connect to a service and how a
proper message for it must be constructed, but you cannot work out what
actual content to put into such a message. Likewise, with xml - by itself -
you can automatically work out how to convert a string of xml into a
tree-like structure, but you cannot know anything about how to populate it
with useful content.
No semantics there.
Cheers,
Tom P
|
