[cpp-threads] RE: "Agenda" for august 23-25 concurrency meeting

[Nick Maclaren]

I humbly apologise to Herb for maligning his terminology!

I had been following only the NON-lock postings, as I thought that we
had settled the lock matter, and so thought that we were talking about
the other issue!  To summarise my views, here they are:

I agree with Herb and others that transitivity is essential, and that
anything that has release/acquire semantics on objects of kind X should
imply full transitivity on ALL objects of kind X.  As I posted, I have
written code that had only pairwise semantics (message passing) and had
major difficulties in handling globally shared/updatable objects.

But we need to be very careful that we don't confuse this with the
associated concept of when two transitive relations are consistent.
I know that I have significant problems handling two transitivity
graphs in my head at one time, but unfortunately that is the way that
we have to work when handling two different kinds of object.

This won't be the last time that I get confused :-(


Regards,
Nick Maclaren,
University of Cambridge Computing Service,
New Museums Site, Pembroke Street, Cambridge CB2 3QH, England.
Email:  nmm1 at cam.ac.uk
Tel.:  +44 1223 334761    Fax:  +44 1223 334679