Re: Renaming L2V

Hi Alexander

This is a personal response from one of the editors of the document, not in the name of the working group (which is no longer active). 

First, making changes to the actual document is no longer feasible, since such changes can be made only by an active working group and this one  is no longer active. So we will all have to live with the terminology and notation that is present until the next W3C WG is convened to revise RDF again.

The notation to which you refer was re-used from the older (2004) RDF specification, since there seemed to be no need to change it. You are the first person to comment critically on this in eleven years. 

Perhaps I can suggest a way to read it which will be more intuitive. As the text says, a datatype is thought of by the RDF sematnics as a complex kind of an object with a number of 'components'. It has a lexical space, a URI which identifies it, and a lexical-to-value mapping. Now, it is normal practice to describe such many-facetted things by 'selection functions' applied to a single argument. So for example if we were making a mathematical model of families, we might identify the family with a name, and then have functions which select the patriarch, the matriarch, the current heir, etc., so that we could write patriarch(Hayes), matriarch(Hayes), etc.. Note that the actual matriarch is the VALUE of the selection function called 'matriarch', not that function itself. This is what we did with datatypes: the URI identifes the datatype (conventionally called D in  the text) and the lexical-to-value mapping itself is indicated by L2V(D), which you could read as "the L-to-V mapping of the datatype D". 

You say: "... it is hard to express L2V(D) as defined in the RDF semantics to the "L2V_D(L)" that is returned from L2V(D) and performs the actual lexical-to-value mapping". But there is no problem here: the mapping that you describe as "L2V_D" is precisely L2V(D). That expression, "L2V(D)", is the (mathematical) name of the actual lexical-to-value map of the datatype D, the thing that is returned from L2V (when applied to D) and which in turn performs the actual lexical to value mapping. So why not use that name to refer to it? That would be correct, it is what is done in the specification, and would seem to also solve your problem, unless I am misunderstanding you.  

I hope this helps.

Best wishes

Pat Hayes

On Jan 10, 2015, at 7:27 PM, Alexander Darino <alexander.darino@ryoga.org> wrote:

> Hello Everyone!
> 
>     In looking over http://www.w3.org/TR/2014/REC-rdf11-mt-20140225/, I couldn't help but note that the name 'L2V' is somewhat misleading since it isn't a mapping L->V but rather D->(L->V). For clarity, perhaps it should be renamed to (something like) D2L2V, or perhaps L2V could be reworked to accept D as a parameter such as in L2V(D, L) -> V?
> 
>     I make this suggestion having found, in discussing with others and making illustrations of the various Interpretation structures and the relations between them, it is very awkward to express the intermediate steps made between the partial mapping of literals onto resources (IR) due to a naming conflict: L2V maps onto the actual lexical-to-value mapping for that data type, so it is hard to express L2V(D) as defined in the RDF semantics to the "L2V_D(L)" that is returned from L2V(D) and performs the actual lexical-to-value mapping (since, again, L2V as defined returns the mapping but is not itself the mapping). In other words, it is awkward to break down IL("sss"^^aaa) = L2V(I(aaa))(sss) into the partial relationships "sss"^^^aaa (from Literals) -> L2V -> Lexical-to-Value-Mapping -> IR since L2V and Lexical-to-Value mapping are not the same, and L2V itself is not the actual Lexical-to-Value mapping (ie. it is a misnomer; the name is misleading).
> 
>     (I acknowledge that my implicit definition of L2V_D(L) being returned from L2V(D) sufficiently disambiguated the two, but it doesn't address the fact that the name L2V is misleading in the original context, and it poses the additional problem of using terminology not specified in the original document.)
> 
> -- 
> Alexander Darino, Freelancer
> Ryoga Independent LLC
> 
> 
> 
> 

------------------------------------------------------------
IHMC                                     (850)434 8903 home
40 South Alcaniz St.            (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile (preferred)
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Monday, 12 January 2015 17:37:00 UTC