Re: Should rdf:value have a semantics??

[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


----- Original Message ----- 
From: "ext pat hayes" <phayes@ai.uwf.edu>
To: "Patrick Stickler" <patrick.stickler@nokia.com>
Cc: <w3c-rdfcore-wg@w3.org>
Sent: 04 December, 2002 22:14
Subject: Re: Should rdf:value have a semantics??


> Heres a rewrite of the section which hopefully will overcome the 
> objections. Feel free to suggest modifications (including, delete it 
> altogether :-)

Delete it and deprecate rdf:value. See comments below...


> The intended use for rdf:value is to indicate that some entity is 
> associated with a particular value, 

Well, this seems to preclude the scoping usage where the semantics of
rdf:value is essentially equality. 

> usually indicated by a literal, 

Including, or rather, usually a typed literal. Right?

> from some predefined range of possible values. 

OK, that would seem to suggest a typed literal, since the 
set of simple string literals is infinite and any subset
constitutes a datatype (IMO) where the L2V mapping is simply 
equality. E.g. xsd:language

> ...and whose other 
> properties indicate the context in which the value is to be 
> understood. 

Including rdf:value? So rdf:value denotes a context?

As Graham points out, how does one reliably know which property
of such a resource is "primary".

I've always understood rdf:value as having "equality" to
some degree as part of its semantics, but even that is hard
to pin down with some usage.

> ...a plain literal indicating a textual 
> representation of the amount 

Whoops! Wrong. A plain literal is *not* a lexical form. If it is
a lexical form, then it is fair to presume that the datatype is
known, and that it denotes the value to which the lexical form maps 
to. Dangerous language there...


> ... there is no way to give a formal 
> specification of this intended usage.
> 
> Users are cautioned that any such usage will be context-dependent and 
> is liable to be misunderstood if removed from its context. A single 
> triple involving rdf:value has no particular meaning in isolation. 
> The use of rdf:value in this way can often be replaced by the use of 
> more complex RDF constructions or more explicit user-defined 
> vocabulary in order to avoid such ambiguities.


This pretty much sums it up for me.

I suggest that rdf:value be deprecated as having no true utility,
since it has no consistent semantics, and offer recommendations
as to how to better model present usage of rdf:value, such as
the use of URIs or explicit properties that more accurately
capture the intended semantics.

Folks can still continue to use it, since deprecation is not
deletion, but it's not considered "best practice" given the
potentially conflicting application-specific interpretations 
and ambiguity arising from its use.

Patrick

Received on Thursday, 5 December 2002 03:59:16 UTC