Re: URI as sort v. URI as spelling for constants

> 
> Since I didn't manage to express my comment very clearly on the telecon 
> here's another attempt ...
> 
> Recap:
>    We started out discussing issue 30: what is the definition of rif:uri?
> 
>    That turned into: do we really want a rif:uri sort at all?
> 
>    Which led to: but what is the alternative?

Dave,

One cannot solve conceptual issues just by switching syntax.

The thingie

  subclass Const
        property name: xsd:string

in the current WD1 is a clear mistake.
What you are proposing might be a fix for the current abstract syntax,
but first we need to decide what it is *supposed to be*.

I am waiting to hear from Jos how he thinks things might work otherwise.
I sent some clarifications in a previous message.


	--michael  


> The point I was trying to make is that the alternative is to hardwire 
> use of URIs to denote individuals, functions and predicates into the 
> syntax as lexical forms rather than using sorts.
> 
> Thus instead of the current asn:
> 
> [[[
> ...
> class TERM
> 
>      subclass Var
>          property name: xsd:string
> 
>      subclass Const
>          property name: xsd:string
> 
>      subclass Uniterm
>          property op: Const
>          property arg: list of TERM
> ]]]
> 
> One would have the asn:
> 
> [[[
> ...
> class TERM
> 
>      subclass Var
>          property name: xsd:string
> 
>      subclass Const
>          subclass Literal
>              property lexicalForm: xsd:string
>              property sort: xsd:anyURI [*]
> 
>          subclass Symbol [**]
>              property uri: xsd:anyURI
> 
>      subclass Uniterm
>          property op: Symbol
>          property arg: list of TERM
> ]]]
> 
> It seems to me that what we set out to do was to find a way of saying
> "you can use URIs to denote symbolic constants such as predicates, 
> functions and individuals".
> 
> The original argument for using sorts for this was to permit other 
> things to also be used as to denote these. Which is why I asked (well 
> tried to ask) "so what are these other things?". I'm not yet convinced 
> we do need anything else, especially for RIF Core.
> 
> Does this make sense?
> 
> Dave
> 
> [*] Note that this is not a sort this is a lexical form. Note also that 
> xsd:anyURI includes both relative an absolute URIs so that the URI as 
> written in a RIF document might be spelt "foo" or "#bar" and might not 
> look like "http://example.com/foo#bar".
> 
> [**] I use "Symbol" here rather than "Individual" (as used in Adrian's 
> version [1]) since it can be used to as a predicate and function symbol 
> as well as an individual.
> 
> [1] http://lists.w3.org/Archives/Public/public-rif-wg/2007Mar/0009.html
> 
> 
>

Received on Tuesday, 20 March 2007 21:02:22 UTC