Re: ISSUE-101 (plainliterals): equating plain literals with no language tag to xsd:string

On 14 Mar 2008, at 10:23, Michael Schneider wrote:

> Oops, this time it was me who overlooked a newly raised issue... by  
> Alan. :)
>
>> ISSUE-101 (plainliterals): equating plain literals with no
>> language tag to xsd:string
>>
>> http://www.w3.org/2007/OWL/tracker/issues/
>>
>> Raised by: Alan Ruttenberg
>> On product:
>>
>> RDF specifies that plain literals with no language tags are to
>> be equated with xsd:strings.
>
> Hm, can someone tell me where RDF does specify this?
>
> Perhaps I misunderstand this issue, but I thought it was a well  
> known fact that xsd:string builds a separate datatype from the  
> plain literals (with or without a language tags), and that there is  
> no defined way to compare elements from both sets.

I thought it was well known the other way :) But it's not so easy for  
find controlling text in the spec. The closest I found with a quick  
perusal is:
	http://www.w3.org/TR/rdf-mt/#D_entailment

"""In particular, the value space and lexical-to-value mapping of the  
XSD datatype xsd:string sanctions the identification of typed  
literals with plain literals without language tags for all character  
strings which are in the lexical space of the datatype, since both of  
them denote the Unicode character string which is displayed in the  
literal; so the following inference rule is valid in all XSD- 
interpretations."""

This text asserts that it follows from the semantics that they are  
equivalent (and gives an argument for that assertion). Whether you  
believe that assertion and argument is a different story :) Seems  
solid to me.

> What I can find is that the RDF Concepts spec says in [1]:
>
>     Two literals are equal if and only if all of the following hold:
>
>         * The strings of the two lexical forms compare equal,  
> character by character.
>         * Either both or neither have language tags.
>         * The language tags, if any, compare equal.
>     [!] * Either both or neither have datatype URIs.
>         * The two datatype URIs, if any, compare equal, character  
> by character.

That's not semantic equality (i.e., sameAs). Consider that for this  
equivalent relation "01"^^xsd:integer is not equal to (sameAs)  
"1"^^xsd:integer. (or consider various ways of spelling boolean true)

Consider as well that the nearby notion of graph equivalence is *not*  
defined in terms of entailment (even for simple entailment as this  
notion requires a bijection but simply equivalent graphs do not.)

>> RIF makes the same choice. I think we should too.
>
> Provided that I have understood the issue, I disagree. In this case  
> the main premise of this issue is wrong, so I would rather opt to  
> reject this issue.

Hope this helps. The concept document would have been more helpful to  
entailment minded folks if it had used "isomorphic" instead of "graph  
equivalent" there and some qualifier on "equal" for literals (or at  
least pointing out that "lexically unequal" literals may be "value  
equal"). (And then there's value equal under coercion :)).

(BTW, it having struggled with things like this that is one of my  
strong reasons for wanting the WG to be *very* cautious about  
generating potentially redundant and alternative presentations and  
explanatory text. It's not good to muddle the specification  
picture...we should strive to make the *specification* as  
approachable as possible while it being consistent with other  
necessary spec virtues (like unambiguity).)

Cheers,
Bijan.

Received on Friday, 14 March 2008 10:58:21 UTC