RE: More fulltext advocacy (was Re: Lee's feature proposal)



> -----Original Message-----
> From: public-rdf-dawg-request@w3.org [mailto:public-rdf-dawg-
> request@w3.org] On Behalf Of Kjetil Kjernsmo
> Sent: 5 May 2009 10:49
> To: public-rdf-dawg@w3.org
> Subject: Re: More fulltext advocacy (was Re: Lee's feature proposal)
> 
> On Tuesday 05 May 2009 11:17:30 Seaborne, Andy wrote:
> > A special case is use in a filter (that does just test for a pattern
> being
> > higher than some threshold so it is a constraint and is more like a
> regex).
> > The filter form is unnecessary unless non-conjunctive combinations of
> text
> > checking and other filter expressions are really imporant.
> 
> Could you please explain this a little further? I have thought of the
> freetext
> matching I've done as a filter, and that it would have been
> syntactically
> more elegant to write it as such, so I do not quite understand the
> context
> they are unnecessary.
> 
> Someone has to be asking the stupid questions here, sorry. :-)

No problem - Orri covered this as well.

A filter

  ... rdfs:label ?x .
  FILTER (match(?x, '+word', 0.6)) 

is the same as 

  ... rdfs:label ?x .
  ?x pf:textMatch ( '+word' 0.6 ) .

FILTER (A&&B) is the same as FILTER(A) . FILTER(B) so if the match() is in filter it can be replaced by a property function form unless you want a complex filter expression where the match is deep in the expression.

 Suppose you want a complex expression ((?x > 15 && "match 'foo') or ?x < 5") (yes - you can rewrite to union and expand but it's ugly and error prone).  It's not an easy rewrite so if the syntax can handle it, the user can concentrate on the real work.

 Andy



> 
> Kind regards
> 
> Kjetil Kjernsmo
> --
> Senior Knowledge Engineer
> Mobile: +47 986 48 234
> Email: kjetil.kjernsmo@computas.com
> Web: http://www.computas.com/

> 
> |  SHARE YOUR KNOWLEDGE  |
> 
> Computas AS  PO Box 482, N-1327 Lysaker | Phone:+47 6783 1000 | Fax:+47
> 6783
> 1001
>

Received on Tuesday, 5 May 2009 09:59:24 UTC