Re: defn variable vacuuous

Dan Connolly wrote:
> I'm transcribing the definitions into larch
> (cf http://www.w3.org/XML/9711theory/) and I got to...
> 
> "Definition: Query Variable
> 
> A variable in a query is name, used to define a graph pattern.
> Variables are associated to terms in a graph by a binding."
>  -- http://www.w3.org/2001/sw/DataAccess/rq23/#basicpatterns
> 
> That definition doesn't actually constrain the term
> Query Variable at all. There's no impact, formally.
> 
> I suggest getting rid of the Definition markup and just stating
> it as commentary.
> 

Ideally, I'd like a definition there, even though it is text.
How about something based on

"""
A symbol whose referent varies or is unknown. A place-holder, as opposed
to an abbreviation or name (a constant). ... A variable ranging over
individual objects from the domain of a system.
"""

http://www.earlham.edu/~peters/courses/logsys/glossary.htm#v

like:

"""
A variable is a placeholder in a query pattern for possible RDF terms.
A variables is asscoiated with at most one RDF term in a query solution.
"""
which disallows multivalue variables in a single query solution.

Just saying it as text misses the formal nature of variables.  Its
always difficult to handle the base definitions as they are going to
refer to concepts that are outside the doc.

	Andy


	Andy

Received on Monday, 6 September 2004 12:27:33 UTC