Re: additional issue-57 use case: polysemy

On Jun 5, 2012, at 3:22 PM, Larry Masinter wrote:

> I added a new use case
> to the use case list and to the matrix
>> Thanks to Larry Masinter for this one. I have no doubt distorted the
>> use case description in trying to make it fit in with the comparison
>> framework, so he is not to blame for any deficiencies in what is
>> written.  Also the way I have filled in the matrix is somewhat
>> arbitrary, and I welcome arguments for changing its matrix entries.
> I wasn't arguing for a new use case, but for a coherent perspective for analysis. So the use case is fine, but its lack wasn't the problem.
> There are two branches of analysis: using semantics and knowledge representation, and using communication theory.
> Both of these are internally coherent. However, mixing them is, in general, not coherent .
> If you start out talking about identification or denotation or semantics or meaning, then fine, but there's really no place in a description of semantics for a "protocol" (UDDP), since a protocol is an operation,  computational behavior whose behavior depends on the operation of the network,

OK, but...

> and not a mathematical function.

Wrong. It is a mathematical function, just a more complicated one than one might have hoped. It takes the state of the network (or something like it) as an argument. 

There is really no limit to what can be treated as a mathematical function. Check out continuation semantics for stateful program descriptions, for example. 

>     If you start out talking about protocols and messages and transmission of representations of state, then switching to a semantic theory leaves all of the context out.

That might be true for some semantic theories, but I see our task (my task, maybe) as being to make the semantics more able to handle the complications that real life forces upon us. This can be done, and its a serious mistake to give up before we have even started. 

> (I think doesn't help and can't, it makes several unwarranted assumptions

Can you briefly say what they are?

> and doesn't even touch the temporal elements.)

It allows for that possibility. (There is a temporal example in the document.) The RDF world has to be able to handle both temporal aspects *and* the idea of a permanent record of transient events or temporally sensitive data. 

However, what triggered my comment was the idea of two sources using the same IRI with different intended meanings., which is exactly the situation this 'context' proposal is designed to accommodate. This is a rather different topic from that of handling the temporal details of network protocols.



IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile

Received on Wednesday, 6 June 2012 01:55:30 UTC