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

CHAN variants

 Although this is not directly related to Peter's message, it has prompted
 me to ask you all about an extension of the channel concept.
 I am thinking about a variant of occam for hardware description (ocHam?)
 - and a variant of TCSP  (HCSP) to capture the semantics precisely.
 I have convinced myself that I need to capture CSP "events" that require
 the simultaneous involvement of more than two processes.
 Given that I want compositional semantics, and that we often need to "connect"
 several subcircuits, this seems very natural.
 The obvious way to extend the occam for this purpose is have an extended
 channel. I would have called it a SHARED channel if Geoff hadn't used
 that with different semantics in occam3. So my best terminology at
 present is
   SPLIT CHAN OF blah:
 I like short keywords so SPLIT has won for now. Or should this not be called
 a CHANNEL at all? Maybe it is a BUS, or a WIRE, although both of those seem to
 have too much of a hardware flavour, given that I want to support software/
 hardware codesign. (There will be mapping into ordinary occam when we abstract
 from hardware timing...)
 So comments invited:
 1) Is this a CHAN at all?
 2) If so, any advance on SPLIT?
 BTW, please don't ask me (yet) whether this could model a tristate bus, or
 whether there can be only one source. If I relax the former, then that really
 might be a case for regarding this as something different from a CHANnel.
Adrian Lawrence.
adrian.lawrence@xxxxxxxxxxxxxx or adrian.lawrence@xxxxxxxxxxxxxxx
  MicroProcessor Unit        | Computing Laboratory,
  13, Banbury Road,          | Wolfson Building,
  Oxford. OX2 6NN.           | Parks Road, Oxford.
  UK.                        | OX1 3QD. UK.
Voice: (+44)-1865-273274,(+44)-1865-283526  Fax: (+44)-1865-273275