- From: Drew McDermott <drew.mcdermott@yale.edu>
- Date: Fri, 18 Apr 2003 15:52:26 -0400 (EDT)
- To: arjohn.kampman@aidministrator.nl
- CC: www-rdf-rules@w3.org
Date: Fri, 11 Apr 2003 11:27:37 +0200
X-PH: V4.4@mr1
From: Arjohn Kampman <arjohn.kampman@aidministrator.nl>
X-Accept-Language: en-us, en
X-Archived-At: http://www.w3.org/mid/3E968A89.1080000@aidministrator.nl
Resent-From: www-rdf-rules@w3.org
X-Mailing-List: <www-rdf-rules@w3.org> archive/latest/373
X-Loop: www-rdf-rules@w3.org
Sender: www-rdf-rules-request@w3.org
Resent-Sender: www-rdf-rules-request@w3.org
X-YaleITSMailFilter: Version 1.0c (attachment(s) not renamed)
...
While implementing Andy's ResultSet vocabulary ... some question
concerning the encoding of unbound variables came up. According to the
document, unbound variables are assigned the value rs:undef. What's the
reason for doing this? IMHO, it would be more natural to exclude the
binding from the result set altogether. I.e.: no binding specified -->
variable is unbound.
If you're building a unification algorithm, I agree with you.
Otherwise, it's marginally easier and less bug-prone to require that
every variable be included in every solution.
--
-- Drew McDermott
Received on Friday, 18 April 2003 15:52:31 UTC