Re: PAR SKIP or PAR PAR SKIP?

[johan.sunter@xxxxxxxxxxx]

Hi Oyvind,

>PROC RESCHEDULE()
>  PAR
>    SKIP
>    -- SKIP -- My add
>:
>
>Taken from Intrinsics.occ.
>Why don't I need to have the second SKIP?

Do you mean why is this enough to have the transputer reschedule or
do you mean why doesn't Occam force me to have two processes? I'll
presume you mean the last alternative.

Warning: Intuitive, informal (and non-formal) statements coming up...

To me the only difference between the PAR and the SEQ is that the SEQ
construct implies a fixed order of execution, from top to bottom, and
the PAR doesn't.

>This seems to be the same:
>
>PAR
>  PAR
>    SKIP
>
>SEQ SKIP makes sense to me. But PAR SKIP doesn't, mentally.


So to me SEQ SKIP makes as much sense as PAR SKIP. The SKIP is just
another process (that happens to do very little indeed). Also, any construct
in Occam (as far as I can remember) is also a process, so 'PAR SKIP' is
a process. That makes the above code fragment correct as far as syntax
is concerned. Good to see that the LTSA analyzer agrees with my intuition
that both have the same behaviour.

Cheers,

Johan

P.S.: The answer to the first alternative would be that this is caused by
      the implementation of the compiler. This has nothing to do with
      Occam of course.

P.P.S.: Ideas about the integration of CSP and OO are still very welcome...

Dr. Ir. J.P.E. Sunter
Philips TASS
Building HCZ-1
PO. Box 218
5600 MD Eindhoven
The Netherlands

Tel. +31-40-2755288
Fax. +31-40-2755419
E-mail: johan.sunter@xxxxxxxxxxx