Re: Simultaneous destructors

Jason Duell wrote:
>> I can offer a hand-wavy argument about why I think that state divergence
>> makes protocols harder to reason about.
>>
>> is a protocol that allows state to permanently diverge.  When reasoning
>> about this protocol, I need to keep the "branch" of the divergence in
>> mind, and consider two paths (states synced, states not).  Now, I could
>> take this same pattern, recursively graft it into S1 and S2, and now I
>> have two more branch points.  So there are 6 stable states in the
>> protocol, and my C++ code has to keep track of two levels of possibly
>> different paths taken through parent/child.  Then recursively graft ...
> 

BTW, I think a better way of saying this is "with divergent state, 
reasoning about the protocol becomes an O(n^2) operation in the number 
of states, rather than O(n)."

> Perhaps we could set a limit on how many such divergences a protocol
> is allowed
> to have.  We certainly don't need to support the sort of example
> you're giving.
> 

Seems like a slippery slope to me, since that limit would have to 
monotonically increase with time.

>> It seems desirable to me that we do not require acknowledging every message:
>> when the channel has data available, it sends that data along and doesn't
>> cache it.
> 
> I agree with bsmedberg here.  But I can start by seeing how much of a
> pain it is
> to cache data if we can't modify IPDL (or do it in time: it hopefully
> won't be
> long).
> 

I think it would be an interesting experiment to try.  Probably only 
need to spec out the code, short of writing it, to see how much of a 
pain queueing dataavailable would be.

Cheers,
Chris