Re: Controlled Vocabularies Aid Translation and Content Management

Hans, Frank --

Somebody commented at a conference a while ago that one of the words with
the slipperiest semantics is --  er  --  "semantics".

One of the problems of attempting to write a document to define terms like
'resource' is, as others have noted, that different readers will interpret
the document differently -- even if it is a masterpiece of clarity.

After all, if you look up a word in a dictionary, it will be defined in
terms of other words, and so on.  You may even complete the circle and find
the original word as part of its own definition.

So, can the situation be improved?

I'd suggest Yes.  But it means going beyond the current technical notion of
"semantics".

Here are some ideas for your consideration.

(a)  extensive use of examples in English*, preferably executable ones.
These tend to provide a sort of center of gravity for interpretations by
people.

(b)  make the document executable, so that you can 'run' it over previously
unseen data.  Then, meaning = input-output behavior.

(c)  tackle the different-people-interpret-the-document-differently problem
by agreeing on  a model theoretic semantics that defines what a machine
interpreter should do.  Prove that the interpreter does what the model
theoretic semantics says it should do.  Clone copies of the interpreter.
The interpretations will always be consistent with one another.  Actually,
they will be identical.

(d) tie together executable English semantics, model theoretic semantics,
and W3C-style data semantics, in one executable interpreter-compiler.

(e) ensure that proof traces from the interpreter-compiler are mapped
computationally to step-by-step explanations in everyday, non-techie
English.  This is a good source of the examples mentioned in (a) above.

I could mention that much of this is integrated into the system that is
online at the site below, and that shared use of the system is free.  But I
won't say that, because it would just look like too big a claim to be taken
seriously (:-).

                             Cheers,    -- Adrian


*  Or the natural language of your choice.

Internet Business Logic (R)
Executable open vocabulary English
Online at www.reengineeringllc.com
Shared use is free

Adrian Walker
Reengineering
Phone: USA 860 830 2085


On 10/4/06, Hans Teijgeler <hans.teijgeler@quicknet.nl> wrote:
>
> Hi Frank,
>
> Wouldn't this be a wonderful use case for RDF vocabularies, set up by the
> different W3C communities in accordance with a nice little common schema,
> and linked together by the Semantic Web? Why are the baker's children
> always
> eating old bread?
>
> Regards,
> Hans
>
> -----Original Message-----
> From: Frank Manola [mailto:fmanola@acm.org]
> Sent: Wednesday, October 04, 2006 17:39
> To: Hans Teijgeler
> Cc: 'Adrian Walker'; semantic-web@w3.org; 'Karl Dubost'; qa-chairs@w3.org
> Subject: Re: Controlled Vocabularies Aid Translation and Content
> Management
>
> Hans Teijgeler wrote:
> > Adrian,
> >
> >> So, how is this relevant to w3c?
> >
> > It is very relevant to W3C, because their Recommendations are, at
> > times, hard to understand for someone whose native language isn't
> > UK-English or US-English. Add to that the handicap of not belonging to
> > the happy incrowd of W3C, so not being conversant with much of the
> > W3C-specific slang and the abundantly used acronyms.
>
> Hans--
>
> I think you're making the situation simpler than it really is.  Surely by
> this time you realize that it isn't only people whose native language
> isn't
> English who find W3C Recommendations "at times, hard to understand"?  And
> as
> for "the happy incrowd of W3C", there certainly seems to be more than one
> such crowd, and they often don't appear very happy with each other :-)
>
> >
> > An all-inclusive and normative W3C glossary of terms and acronyms with
> > a crystal clear definitions (in understandable English) would help,
> > provided that all authors would normatively refer to that glossary. A
> > simple case of QA (meaning Quality Assurance - 'The process assuring
> > the quality of one organization's outcomes.' (according [1])).
>
> "An all-inclusive and normative W3C glossary of terms and acronyms with a
> crystal clear definitions (in understandable English)" would not only
> help,
> it would be a miracle!  (A great example of how simple this task would be
> is
> to consider the definition of "resource").  I think we might want to start
> with something much simpler, like an OWL ontology (nothing like eating
> your
> own dogfood!).
>
> I'd also note that it generally isn't "the-great-W3C-in-the-sky" that gets
> things done, it's *volunteers* who get things done, W3C itself not having
> an
> arbitrarily-large workforce.  Things like making W3C Recommendations more
> understandable to non-native-English-speakers (or even translating them
> into
> other languages) are examples.  That things don't get done doesn't
> necessarily mean the W3C doesn't see them as desirable.
>
> --Frank
>
> >
> > Regards,
> > Hans
> >
> > [1] http://www.w3.org/QA/glossary
>
> --
> No virus found in this incoming message.
> Checked by AVG Free Edition.
> Version: 7.0.407 / Virus Database: 268.12.12/462 - Release Date: 03-Oct-06
>
>
> --
> No virus found in this outgoing message.
> Checked by AVG Free Edition.
> Version: 7.0.407 / Virus Database: 268.12.12/462 - Release Date: 03-Oct-06
>
>
>