John,
Thanks for the reference to your valuable overview of Controlled Natural
Language approaches.
Reading that, I have a few questions: (01)
1. I question whether logic and reasoning and the bridge to natural language
interpretation should have such a strong emphasis.
In my view information exchange, information integration and unambiguous
interpretation should precede that.
Gellish Formal English (as well as Formal Dutch and language independent
Gellish) were developed from the data exchange, storage and integration
perspective.
I think it is a pity that you did not take Gellish Formal English into
account in your overview. (02)
2. In your slide about 'Multiple Word Senses' you don't distinguish between
the meaning of a concept and the various words and phrases that can denote
such a concept. I think that, instead of "to specify word senses" it would
be better to define concepts (represented by unique identifiers). Then
during creation of an expression the author should select the concept which
is meant (e.g. choose between the various concepts denoted by the same term
'give'). Then during interpretation that choice is not necessary anymore,
and the interpretation is unambiguous. (03)
3. In your conclusions about advantages and disadvantages of CNL's you
appear to assume that people write directly in a CNL. However, that is not
necessarily the case. System may continue to use conventional user
interfaces, whereas the CNL is the underlying information storage,
integrations and exchange mechanism. (04)
With kind regards,
Andries (05)
-----Oorspronkelijk bericht-----
Van: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] Namens John F Sowa
Verzonden: zaterdag 1 september 2012 19:37
Aan: ontolog-forum@xxxxxxxxxxxxxxxx
Onderwerp: Re: [ontolog-forum] Accommodating legacy software (06)
Denise, Michael B, Kingsley, and Andries, (07)
Before commenting on the details of your notes, I'd like to emphasize
an excerpt from the original DAML report by Tim Berners-Lee: (08)
TB-L
> The goal of interoperability between heterogeneous components that
> we build is one that will test the extent to which the Semantic Web
> is achieving its promise. The more diverse the systems interoperating,
> the greater the merit of the Semantic Web. (09)
The date of the proposal is February 2000:
http://www.w3.org/2000/01/sw/DevelopmentProposal (010)
That quotation comes from Section E on Evaluation. Twelve years later,
we would expect to see interoperating systems running all the languages
and tools that Tim mentions (plus a lot more). But instead of trying
to support maximum diversity, the SW developers narrowed their goals
to just RDF data with a deliberately restricted subset of logic. (011)
DB
> I'm including this post and the DAML proposal as required readings
> in my Semantic Analysis Methods course... (012)
I'm glad that you like the note. But you might add more progress
reports for the DAML project. The first is an "Intent of Work" from
March 2001, which is fairly close to the proposal, but the words
'diversity' and 'heterogeneous' are missing: (013)
http://www.daml.org/2001/04/iow/mit/ (014)
Following is a progress report from February 2002: (015)
http://www.w3.org/2002/02/iow2 (016)
An ominous sign is that the word 'interoperability' occurs only once.
The very expressive Semantic Web Logic Language (SWeLL) occurs 8 times.
A new language called DAML+OIL is mentioned 6 times. (017)
Another report from November 2002: (018)
http://www.w3.org/2002/11/DAML-IOW (019)
Interoperability is mentioned once. DAML+OIL is mentioned 6 times as
a combination of a query language plus a rule language. SWeLL is
mentioned 8 times, but a new language called OWL also occurs 8 times. (020)
Finally, the Final Technical Report from September 2006: (021)
http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA458366 (022)
The most disappointing part of this report is that SWeLL, which was
supposed to include propositional, first-order, and higher-order
logic in the original proposal has now become "Notation N3" for
representing RDF triples. N3 would allow some rules to operate
on RDF data, but most of SWeLL is gone. (023)
But the most disgusting part is that interoperability is now
defined as the ability to use N3 on independently developed
systems that all use RDF. (024)
No legacy systems, no relational databases, and no possibility
of interoperating with any data not represented in RDF. (025)
MB
> I recently wrote RDF2RDB for people who want to use RDF data but do
> not want to operate a triple store or learn SPARQL:
>
> http://www.netestate.de/De/Loesungen/RDF2RDB (026)
Yes. Translations in both directions are the minimum requirement
for systems to interoperate. (027)
But note that it is now 2012. In 2000, Tim B-L emphasized
interoperability among heterogeneous systems as a primary
requirement for the Semantic Web. (028)
KI
> Linked Data picks up where OLE-DB, ADO.NET, JDO and the like have failed
> woefully, due to myopic platform specificity. It also addresses to
challenging
> issue of platform agnostic structured data representation that
accommodates
> multiple syntaxes... (029)
That's fine. Tim's vision of 2000 could have addressed those issues.
Unfortunately, the SW became narrowed down to what the DAML project
was able to implement. I wouldn't expect one implementation project
to be able to implement the entire vision in the 2000 proposal. (030)
It's unfortunate that the SW became redefined to just what the DAML
implementers were able to finish by 2006 + some incremental refinements.
It's time to rethink the goals of the SW and redefine the layer cake
to incorporate a broader vision that is closer to original (plus the
new ideas that have been introduced since 2000). (031)
MB
> I recently stumbled over Pat's ISWC 2009 keynote proposing "surfaces" that
> would give RDF the full power of FOL
>
> http://videolectures.net/iswc09_hayes_blogic/
>
> and I wonder if the RDF WG is actually going a bit in that direction. For
> example, would it be feasible to realize "surfaces" with a simple semantic
> extension to the "RDF Spaces and Datasets" (032)
I have no idea what the RDF WG is going to do. But note that Pat was
suggesting that RDF evolve in the direction of the existential graphs
by Charles Sanders Peirce. The surfaces that Pat mentioned are the
"sheets of assertion" that Peirce mentioned. (033)
For more about Peirce's existential graphs, you can start with a talk
that I presented on the topic: (034)
http://www.jfsowa.com/talks/moving.pdf (035)
For a more detailed treatment, which includes my commentary on
a tutorial that Peirce wrote, see (036)
http://www.jfsowa.com/pubs/egtut.pdf (037)
AvR
> If we really want to achieve that systems talk a common language for
> interoperability and data integration, then we should develop and use
> a kind of Formal English (and other formal natural language variants) (038)
I agree. Following is a talk I presented about controlled NLs: (039)
http://www.jfsowa.com/talks/cnl4ss.pdf (040)
John (041)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (042)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (043)
|