W3C home > Mailing lists > Public > public-owl-dev@w3.org > October to December 2006

Re: OWL "Sydney Syntax", structured english

From: Dan Connolly <connolly@w3.org>
Date: Tue, 28 Nov 2006 13:47:49 -0600
To: Kaarel Kaljurand <kaljurand@gmail.com>
Cc: Anne Cregan <annec@cse.unsw.edu.au>, public-owl-dev@w3.org
Message-Id: <1164743269.3997.1047.camel@dirk>

On Mon, 2006-11-27 at 14:20 +0000, Bijan Parsia wrote:
> what you describe is also one of our goals in the Attempto project
> (see: http://www.ifi.unizh.ch/attempto/). We have recently implemented
> a program
> that converts OWL ontologies into Attempto Controlled English (ACE),  
> see:
> http://www.ifi.unizh.ch/attempto/documentation/OWL_to_ACE/

I tried that on an ontology that is in development these days:

An Ontology for vCards
        14 November 2006
        Harry Halpin
        Brian Suda
        Norman Walsh

The results are sorta reasonable...

Every VCard revs at most 1 things .
Every Name family-names at most 1 things .
Every Name given-names at most 1 things .
Every Name additional-names at most 1 things .
Every Name honorific-prefixs at most 1 things .
Every Name honorific-suffixs at most 1 things .

though they show that the OWL_to_ACE tool doesn't expect
use of the http://esw.w3.org/topic/RoleNoun pattern.

I'd suggest supporting a special kind of label for the purpose of
controlled english generation; i.e. a subproperty of rdfs:label.
maybe ace:englishVerb.

I know the tabulator supports rdfs:label; I think it expects
them to be role nouns. Maybe it should coin a purpose-specific
tab:roleNoun property in case somebody wants to use rdfs:label
for something else. Hm... it seems to grab all subproperties
of rdfs:label

I wonder if the folks in the semweb-ui list discuss this stuff.

> This should give an overview of what we think the controlled natural
> language for
> expressing OWL should be like. Our claim is that an average user can  
> grasp
> the contents of an ontology better if it was presented in ACE rather  
> than
> visualized by Protege or similar editors. (We haven't done any  
> experiments
> yet to back this up.)

I have heard this claim from other places, though.

I agree that rendering to a constrained dialect of English is an
interesting user-interface technique; I made a note about that
during the OWL-ED workshop

"(Tom? stanford guy) says the English format spit out by SWOOP is very

> There is also a description of how to convert ACE texts into OWL:
> http://www.ifi.unizh.ch/attempto/documentation/writing_owl_in_ace.html
> Comments welcome.
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E
Received on Tuesday, 28 November 2006 19:48:14 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:58:14 UTC