[topicmapmail] Making ontologies : RDF vs TM

[Murray Altheim]

Marcel Ferrante wrote:
> To maybe help in the ontology question:
> 
> Ontology [1]
>    For Philosophy
>      Ontology is a discipline of Philosophy that deals with what kinds
> of things exist -
>      what entities there are in the universe. It is a branch of
> metaphysics, the study of first
>      principles or the essence of things.

A rather simplistic and outmoded definition that seems to ignore most
of the past thirty years or so of philosophical investigation. This
particular definition also ignores the entire branch of philosophy
that springs from Wittgenstein, Dewey, Sellars, through to Rorty,
Derrida, etc., i.e., placing our understanding of reality under
metaphysics rather than social, psychological, political, linguistic
studies. Few philosophers today think of ontology as under metaphysics.
For myself, I tend to agree with people like Rorty and Robert Brandom,
who consider ontology as a way of people expressing themselves, and
their relationship with others. A social phenomenon. Richard Rorty
quotes and expands on some ideas of the physicist Arthur Fine in:

   A pragmatist view of contemporary analytic philosophy
   http://www.stanford.edu/~rrorty/pragmatistview.htm

basically suggesting that we stop considering some divide between
the "hard" and "soft" sciences based on some erroneous notion that
there is an objective reality on the one hand and not the other, and
look at all of scientific inquiry as a social and humanist endeavour.

One of the few benefits of postmodernism was to at least bring to
light the typical conflation of ideas in past approaches, instead
looking towards the realistic chaos of holism. But rather than
discuss what ontology *is* (since I'm on very thin ice anyway),
perhaps we should first ask what the problem is that ontology is
suggested as a solution for. [E.g., will creating a large structure
of fixed categories and then placing everything into those categories
work? Hasn't this actually turned into a problem in its own right?]

But before we proceed, one of the pieces of this puzzle that is
missing but always assumed is that we can simply "know", that there
is no filter between reality and how we decide to represent it. In
the past few years of research I've found hardly the slightest whiff
of any consideration for epistemological issues when discussing
"ontologies" within the field of AI/KR or IT. I've heard repeatedly
(from sources who I think should know better) that we can discover
the nature of reality by simply "trying harder", improving existing
techniques, that the platonic notions of such structures are merely
there to be grokked if we open our eyes and/or improve our tools.

>    For Information Science
>      Working model of entities and interactions in some particular
> domain of knowledge or
>      practices, such as electronic commerce.
>  
> So, we can see, from the ontology definition of Information Science,
> that is possible make a ontology with XTM. Also is possible construct:
> indexex, taxonomies, thesaurus, glossaries and semantics network.

Yes, this seems patently obvious, but so what? You can build a house
with wood. The graph structure of Topic Maps (or XTM by extension)
can be used as a building material for anything that can be built
with graphs. Same with GXL and RDF. This leads us back to the earlier
question: what exactly is it we're really trying to build?

> And "topic maps are not about the semantic web, but rather about
> seamless knowledge "
> Hum, interesting. So the scope is bigger than web. Let's move the
> focus instead the macro, as web, to the micro, our PC.

I'd suggest to stop centering on technology and think about what
problem one is trying to solve. "Seamless knowledge" is a solution
for problems where knowledge is not seamless. Like walking into a
library and not being able to find the resource, or worse yet, finding
a thousand resources. Now, if one's PC is a mess, perhaps something
like a Topic Map-based solution could help. Fair enough. This has
almost nothing to do with the Semantic Web.

> I read that next version of windows, longhorn, will has a new file
> system, winfs, that will be  "Search and Manage Files Based on
> Content" [1] as they said. They want use XML in the file system.
> That's interesting. I don't want thing now what they will do with this
> and web service in the future. But the analogy comes. Knowledge in own
> PC, organizing files and information with XTM ? Who knows...

Only if one reads marketing copy and believes it.

Knowledge is not stored within computers, data is. The only way that
data becomes information, becomes knowledge, is when it can be put to
good use in the process of learning and discovery. Having XML in a
file system is nothing new or particularly revolutionary, except from
the standpoint of developers who can benefit from it in creating more
seamless applications. It does nothing directly for end users, and
probably nothing that can't be done with flat files.

As both Pat Hayes and John Sowa have correctly been saying for a long
time, the mode of storage within the system is pretty irrelevant, it's
what is done with the data, and that hasn't changed much at all within
the past thirty years. The Semantic Web is doing a poorer version of
what AI researchers were doing in the early 1970's, just on the nice
big grand palette of the Web. That is progress in a sense. But our
understanding and use of research into computation linguistics and
related fields has not been altered significantly since even the 1960's
(just gradual, evolutionary improvements in various algorithms, no
major breakthroughs). This is why we still see Google ignoring metadata
and using word placement as its means of categorization documents, not
any sense of "ontology" or meaning. Google doesn't have any sense of
the meaning of a document at all, and according to computational
linguistics experts like Stanford's Geoff Nunberg, we're probably at
least a decade or more away from being able to effectively use anything
from CL. We still know comparatively little about human language or
thinking. This hasn't stopped anyone from bandying about the word
"knowledge" as if it were something you could put in a cup and drink
(and yes, the institute I'm studying at is just as guilty as anyone).

The Semantic Web is a solution in search of a problem. It probably
is more fruitful to develop tools to solve specific problems. Topic
Maps were designed as a tool to solve indexing and categorization
problems, and we're simply finding that the Swiss Army Knife we ended
up with had a few more blades than first thought. What that strange
one with the shiny bent hook is for, I dunno. It won't open a bottle
of wine, maybe it'll seamlessly integrate all of the world's knowledge
in a way that all human beings can suddenly grasp mutual understanding
and end all world conflict, and even uncurl Dick Cheney's lip.

Sorry. My caffeine ran out halfway through this message. Time to
head downstairs for another cup...

> References:
> [1] An evaluation of Topic Maps A Master's Thesis in Computational
> Linguistics Göteborg University May 2002 Anna Carlstedt Mats Nordborg
> [2] http://msdn.microsoft.com/library/en-us/dnwinfs/html/winfs03112004.asp

Murray

......................................................................
Murray Altheim                    http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK               .

   The Rise of Pseudo Fascism -- David Neiwert
   Part 1: The Morphing of the Conservative Movement
     http://dneiwert.blogspot.com/2004_09_19_dneiwert_archive.html#109028353137888956
   Part 2: The Architecture of Fascism
     http://dneiwert.blogspot.com/2004_09_26_dneiwert_archive.html#109563628314780505
   Part 3: The Pseudo-Fascist Campaign
     http://dneiwert.blogspot.com/2004_10_03_dneiwert_archive.html#109596147171278590