W3C home > Mailing lists > Public > www-webont-wg@w3.org > September 2003

Re: 600 too large? dl-904

From: Jim Hendler <hendler@cs.umd.edu>
Date: Wed, 10 Sep 2003 22:20:45 -0400
Message-Id: <p05200f0bbb858c1c62b5@[10.0.0.217]>
To: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "Sandro Hawke" <sandro@w3.org>, "Sean Bechhofer" <seanb@cs.man.ac.uk>
Cc: <www-webont-wg@w3.org>

At 5:18 PM +0200 9/10/03, Jeremy Carroll wrote:
>Concerning the dl test 904:
>
>http://www.w3.org/TR/2003/CR-owl-test-20030818/dl-900-arith#description-logi
>c-904
>
>
>Jim wrote:
>>  I would, however, like to suggest we make this test with the numbers
>>  200, 300, and 600 into an extra credit test -- now that my group has
>>  built Pellet, and run this problem with various values of i,j and k,
>>  I'm more sure then ever that we're just looking at a large
>>  combinatorics problem -- with enough space and time the system would
>>  prove it, and I think anything that doesn't need to enumerate
>>  examples will require some sort of heuristic -- we could pick a
>>  smaller number (20, 30, 60) to be the harder version of the problem
>>  if we think we need some values to push people a bit more...
>>
>
>
>My point of view in putting this test in was the following:
>
>1) there are implementation strategies, such as those investigate by the
>Racer team, that keep track of cardinalities as numbers and have a linear
>programming module that check such constraints. I do not know the state of
>that work, nor whether it is directly applicable to this test.
>
>2) 600 was intended to be big enough to blow a brute force approach.
>
>3) If the implementors cannot handle 600 then IMO it is a sufficiently small
>number that some OWL users may be surprised.
>
>4) Thus if the WG would choose to move this test to extra credit, I would
>like to see appropriate health warnings, may be in reference or guide, to
>the effect that, currently, the numbers in the cardinality constraints
>should be kept small. The test document extra credit section could better
>document our combined implementation experience.
>
>Jeremy

Jeremy-
   I misread how the problems were encoded (thanks to Bijan for 
explaining this to me) - let's wait to see what Racer and some others 
do
  -JH


-- 
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	  *** 240-277-3388 (Cell)
http://www.cs.umd.edu/users/hendler      *** NOTE CHANGED CELL NUMBER ***
Received on Wednesday, 10 September 2003 22:24:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:02 GMT