W3C home > Mailing lists > Public > www-webont-wg@w3.org > November 2002

RE: MT for imports (was: Re: Imports Proposal)

From: Jim Hendler <hendler@cs.umd.edu>
Date: Wed, 13 Nov 2002 22:17:44 -0500
Message-Id: <p05111706b9f8c471027b@[]>
To: Frank van Harmelen <Frank.van.Harmelen@cs.vu.nl>, www-webont-wg@w3.org

At 1:52 AM +0100 11/14/02, Frank van Harmelen wrote:
>I remain entirely baffled by all of this.
>As Jerome Euzenat wrote:
>>The timed web structure applies to ontologies exactly like it applies to
>>hyperlinks in html, xsl:include and xsl:imports in XSLT, etc.
>Of course, some advanced version of imports would depend on the 
>"timed web structure" (Massimo), but I cannot see how that should 
>stop us from providing something simple.
>Lots and lots of computer languages provide import-like things,
>varying from C to Scheme, and from LaTeX to XML. None of these have 
>any of the problems raised in the preceding discussion. Pat's 
>example of someone changing an imported file is common to all of 
>these, happily ignored by all of them, and rightly so, since it 
>doesn't seem to break any of them, the meaning and pragmatics of 
>these import constructions is entirely clear for all of them. If it 
>works for XML, why wouldn't it work for OWL?
>I can simply not imagine standing up in front of a crowd, proudly 
>explaining OWL, and having to admit that,... eh... no, well... 
>actually, in OWL you cannot import other people's ontologies...
>   ----

Frank - let me be clear, Dan and I are not arguing that we shouldn't 
have an imports mechanism, just that we shouldn't define it precisely 
at this time (and in a manner we may regret later).  The DAML 
solution was to provide imports, explain it in English, and 
essentially say that there was no formal semantics for it at this 
point.  I would be very happy for us to say the same.  What I believe 
is that our current proposed solutions are drastically different than 
the nice C, Fortran, etc. include statements, which DON'T have a 
formal semantics (and the new logic-based formal verification 
langauges currently don't allow imports as far as I can tell) - I'd 
like us to handle it like those languages do -- explain what we 
expect to happen, and leave it to the implementors to work out how -- 
once entailment tests entered the picture, I moved from in favor to 
against, and I remain there.
  I see POSTPONING this issue as being exactly the same as what 
DAML+OIL currently has - imports as magic syntax, that works just 
fine thank you and let's not screw around with it
p.s. If the WG votes in favor of this crazy semantic approach to 
imports, may I make the friendly amendment that we rename it 
owl:includes, which does strike me as a better name.
Professor James Hendler				  hendler@cs.umd.edu
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  240-731-3822 (Cell)
Received on Wednesday, 13 November 2002 22:17:53 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:04:38 UTC