On Tue, 26 Oct 1999, Philip Hazel wrote:
> about what happened earlier. There is not much point in doing half this
> job. I have made a note to think about it, but no promises.
OK, thank you. It is the intermittent failures that could not
be reproduced by testing that were bugging me here. I see that
the implementation is non-trivial, so "no promises" is fine.
>
> --
> Philip Hazel University of Cambridge Computing Service,
Thank you again,
Hugh