[iDC] The three unsolved problems in RDF/OWL ontology

[Danny Ayers]

On 17/03/2008, Paul Prueitt <psp at ontologystream.com> wrote:
>
>
>  The four unsolved problems in RDF/OWL ontology

Without concrete examples, it's hard to see what's unsolved here.

>  ontology merge

combine the ontologies by adding them to the same model,
the usual constructs allow merge: owl:sameAs, owl:subPropertyOf etc.

>  orchestration of different viewpoints

not quite sure what you mean here - the same information can be
presented in many different ways; mapping across diverse vocabularies
is possible (e.g. using SPARQL CONSTRUCT)

>  a process model that (sufficiently) accounts for changes in context

I don't know what kind of a process model or context you have in mind
- there isn't any process model inherent in RDF/OWL, though
inferencing can be treated as a process and processes can be expressed
using dedicated vocabularies (e.g. OWL-S or WSMO). Named graphs can be
used to represent most kinds of context.

>  discovery of new interpretations of data

not sure exactly what you have in mind here either - though by merging
data from different sources, relationships may be discovered that
might not previously have been apparent (serendipity is an aim!)

>  all of these are often managed well by humans using cognitive
>  abilities and natural language.

How do you define "managed well"? Only yesterday I was struggling with
Italian instructions on a tin of varnish, couldn't find a definition
of "polissandro" anywhere (turns out it was a typo - should have been
"palissandro", which might be rosewood). The last warning, offered in
both English and Italian was pure poetry:

Non gettare i residue nella fognature / Do not empty into dreams

>  Topic maps is a pre-cursor to a semiotic system (a system of emerging
>  signs used to convey information), it also has difficulties with
>  these four unsolved problems.  However, the topic map standard was
>  originally developed to not fully constrain meaning, but rather to
>  allow an interpretative process.

RDF doesn't constrain meaning: http://dannyayers.com may be my
foaf:homepage, but the ultimate definition of foaf:homepage and what
it means to me are in the minds of humans.

>  Topic maps was not designed to do what the RDF/OWL folks are trying
>  to do, and RDF/OWL over constrains "formal semantics" so that there
>  is a limit to the usefulness where usefulness is measured by average
>  people working on everyday problems.

An example would help, I don't know what you mean.

>  The difference between RDF/OWL and topic maps is very close to the
>  difference between early Wittgenstein and later Wittgenstein.  The
>  difference is seen in many religious arguments, and uncovers one's
>  predisposition to scientific reductionism or to a post reductionist
>  philosophy of natural science.

I am not a philosopher, but -
[[
A topic map can represent information using topics (representing any
concept, from people, countries, and organizations to software
modules, individual files, and events), associations (which represent
the relationships between them), and occurrences (which represent
relationships between topics and information resources relevant to
them).
]]
http://en.wikipedia.org/wiki/Topic_Maps

making the appropriate substitutions:

An RDF graph can represent information using resources and classes
(representing any concept, from people, countries, and organizations
to software modules, individual files, and events), properties (which
represent the relationships between them), and class properties (which
represent relationships between topics and information resources
relevant to them).

Was the later Wittgenstein very similar to the early Wittgenstein?

>  What the bleep is a GRDDL?

[[
GRDDL is a mechanism for Gleaning Resource Descriptions from Dialects
of Languages. It is a technique for obtaining RDF data from XML
documents and in particular XHTML pages.
]]
http://www.w3.org/TR/grddl-primer/

Cheers,
Danny.

-- 
http://dannyayers.com
~
http://blogs.talis.com/nodalities/this_weeks_semantic_web/