Re: "If the Semantic Web were ever going to fail..."

> > The Schema layer, when applied to RDF, will be the downfall
> > of the Semantic Web, simply because it isn't viable to apply
> > XML Schemas to RDF.
>
> I have no idea what you mean by this. Why is this the downfall of
> the Semantic Web?

RDF Schemas don't convey structure, and XML Schemas don't convey semantics.
Therefore, when it comes to such a time as we need RDF systems wih
structure, I think it will be very difficult to achieve. Look at the two
greatest uses of RDF to data: RSS and DC. Neither have either RDF or XML
Schemas. Why? Because RDF and XML Schemas wouldn't sufficiently describe
them: for DC it would be pointless, and RSS next to impossble (you'd have
to use Schematron instead).

This equates to a downfall, because that means that the RDF model probably
won't be able to coform to the partial understnading model put forwards by
Tim Berners-Lee. One part of the model is defined in RDF M&S itself, i.e.
triples and the like, but the other part is mixing namespaces:-

[[[
"The XML and RDF schema languages are now (1999) available in draft form.
New namepsaces must be designed assuming the use of schemas, and being test
cases for the schema language drafts, and not relying on DTD functionality.
Where the functionality being introduced maps onto a logical assertion
model, then the mapping onto the RDF model below should be defined, and,
normally, RDF used." - http://www.w3.org/DesignIssues/Architecture.html
]]]

Tim seems to be implying that either logical assertion models mapped onto
the RDF model (i.e. RDF apps.) won't require any form of strict datatyping,
as provided by XML Schemas. That is surely incorrect...although RDF does
provide the basic framework for assertions and quotations, this appears to
me to be incompatable with the adjacent layer: Schemas. Quoting Tim again:-

[[[
"2. The basic assertion model provides the concepts of assertion (property)
and quotation. (This is provided by the RDF Model and Syntax
Specification). This allows an entity-relationship-like model to be made
for the data, giving it the semantics of assertions propositional logic.
See the Cambridge Communiqué about the XML-RDF relationship) The RDF syntax
was considered in need of a change.
3. The schema language provides data typing and allows document structure
to be constrained to allow predictable computable processing. See XML
schema's structure and datatypes drafts. " - ibid.
]]]

It seems to me that tihs "predictable computable processing" layer that he
talks about it (for the RDF M&S model) inherent in the draft, and not
explicitly defined anywhere. Thus, for outside and "unrecognizable"
extensions (those in different namespaces), we would need to provide schema
validation similar to that provided by the M&S model. This was mean to be
given by RDF Schemas, but they do not (IMO) allow that. Just look at all of
the non-standard tags already introduced into M&S and RDF Schemas already
(see Dave Beckett's list: sorry I've lost the URI). If there was to be a
langauge so complex in it's use that it required an XML Schema as well, I
believe that that would add so much complexity to a system as to simply
choke it. I'm pretty sure that everyone knows that this problem is coming
up, becasue I've seen it discussed before. THe bottom line is that as RDF
systems get more complicated, we are going to find out that prose
definitions such as the ones provided by RSS and DC aren't good enough, and
the Schemas that we have that are meant to be designed to cover this, RDF
and XML Schemas, are both either not-suitable for the job, or are suitable,
but simply too complex for anyone to apply.

Kindest Regards,
Sean B. Palmer
http://www.mysterylights.com/sbp/
http://www.w3.org/WAI/ [ERT/GL/PF]
"Perhaps, but let's not get bogged down in semantics."
   - Homer J. Simpson, BABF07.

Received on Tuesday, 19 December 2000 17:51:32 UTC