[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

the semantics of deckchairs

   I am not sure who said that and in what context. 

Probably me. 

   That is all about the semantics of hiding and offered events not
   actually accepted (in my denotational terminology). The point is
   that a CSP(P) process can undergo internal transitions as a result
   of an offer, even if the offer is not accepted.

Sounds like a move towards a refusal testing model.  Hmm.  It also
sounds like people spending a lot of time trying to (re-)invent useful
variations of the failures representation.

And, in the process, missing the point entirely.

I don't have time to give a one-week course on the subject in email
form.  (Although I might be able to arrange a substantial discount if
you wanted to attend one.)

   In almost all realistic circumstances, "overtures" should be

This is complete nonsense.  ;^).  Sorry, but it really is.  

In arriving at an adequate representation, you may need to consider
whether unsuccessful (in some sense) attempts at communication should
be recorded.  

 * If the answer is yes, then you will need to use one or more classes
   of CSP event to represent them.

 * If they can effect a relevant change (perhaps exceptional) in the
   receiving object, then you will need to share one or more of these
   events with the process representing that object.

And that's the end of the matter.  No need to invent a new semantic
model, particularly if the accompanying mathematical theory may prove

   Perhaps that comment was picking up on the difference between the
   two sets of laws?

The laws of CSP are there to help you learn about the operators, and
to build an intuition to support the decisions that you have to make
in constructing a model (and a semantic model, for that matter).  

To present two sets of laws would be confusing and counter-productive
[this kind of behaviour is one of the reasons for the relative lack of
adoption of formal techniques in the 90s - academics presenting the
theory at the expense of the practice] and we don't do it around here.

So no, that comment wasn't picking up on the difference between two
sets of laws.  It was explaining why failures is an appropriate and
powerful abstraction.  And how to use CSP events (atomic,
instantaneous, synchronisations) to represent patterns of interaction
at different levels of abstraction.

Yours (refusal) testily