Re: Call for Comments: SKOS Primer: W3C Working Draft 21 February 2008

Hi Azamat,

On 27/02/2008, Azamat <> wrote:

>  I.  ''SKOS — Simple Knowledge Organisation System — provides a model for
> expressing the basic structure and content of concept schemes such as
>  thesauri, classification schemes, subject heading lists, taxonomies,
> folksonomies, and other types of controlled vocabulary.''
>  Here is a confusion of different entities: constructs and words (terms).

I don't know about confusion - perhaps a flattening, so that both
words in structured vocabularies as well as conceptual structures can
be represented in a coherent system.

>  Besides, it is the domain of global ontology to cover all sorts of
>  controlled vocabulary. No need to invent the SKOS.

As far as I am aware, there is no machine-processable expression of a
global ontology (assuming you mean the 'one true ontology'). In lieu
of such a thing - assuming it possible - it is the domain of
initiatives like SKOS to create things that are useful today.

>   II. ''The fundamental element of the SKOS vocabulary is the concept.
>  Concepts denote ideas or meanings that are the units of thought [Willpower
>  Glossary] which underly the KOSs used in a number of applications
>  [SKOS-UCR]. As such, concepts exist in the mind as abstract entities which
>  are independent of the terms used to label them.''
>  Here is a bad misunderstanding of the nature of semantics and its basic
>  notions; namely:
>  1. Vocabulary is a listing of words or the system of symbols and techniques,
>  and hardly it is a conceptual system.

Perhaps not, but there is at least a useful correspondence between
words and concepts (one that has well-served the human for millennia).

>  2. No ''concepts denote ideas or meanings''. There is general semantic rule:
>  some symbols (signs) designate, while all constructs mean something in the
>  world. Then, symbols (or signs) designate constructs (ideas, concepts,
>  propositions, or theories), whereas the constructs mean real things. So the
>  ideas are expressed by words, which signify the real entities to be named.
>  Thus the meanings of ideas are signified by words, not by concepts. We say
>  significance of the symbol and the meaning of the construct, what might be
>  equivalent.

I'm sorry, I can't parse that.

>  3. There are words (signs, symbols, terms, codes) which signify only things
>  in the world or only the ideas in the mind or both ideas and things.

Ok. Words are messy.

>  4. Concept is the unit construct, and the construct covers concept
>  (individual, class, relation), proposition, context, and theory.

I /think/ I get what you're saying there, and don't disagree.

>  If somebody is striving for semantic web, he must have a good learning about
>  the nature of meaning
>  ( ) and the modes of
>  signification ( by the kinds of
>  things signified and the kinds of things which signify (signs, symbols,
>  codes, terms, words).

There I disagree strongly, in the first place probably because I use a
different definition of "semantic web". My definition would involve
the approximate modeling of real-world/human-world constructs in a
form that lends itself to machine processing in the particular
distributed computing environment we call the Web.

This approximation, in terms of abstraction and sophistication, is
very close to that predominant in existing conceptual modeling on
computers - usually isomorphic to simple (named) entities & (named)
relationships, the stuff we generally call data (or structured data).
Apply the Web's naming scheme to this and take advantage of its
protocols, and you get the Semantic Web.

I think it's reasonable to say that a big part of the Semantic Web
initiative is about finding sweet spots between sophistication of
representation, ease of use with known computing systems, and
compatibility with the most successful distributed system to date.

So while "good learning" may be useful in the definition of the
languages, for the Semantic Web to be successful, something a lot more
elusive is needed. It needs the realisation that the system has to be
one in which anyone with any form of knowledge can contribute, even in
situations where there is disagreement.

SKOS allows the expression of a fairly broad range of information in a
useful form, and being built on basic Web technologies allows
distributed development of conceptual structures - however crude they
may be philosophically - even when those structures may not correspond
1:1 across communities. Because global agreement isn't a prerequisite
it has a good chance of widespread adoption.

If there was a single global ontology on which everyone could agree,
not only would there be no need for SKOS, our emails would be much
shorter :-)



Received on Thursday, 28 February 2008 23:24:29 UTC