From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-0.8 required=5.0 tests=BAYES_00,INVALID_DATE
	autolearn=no autolearn_force=no version=3.4.4
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!inmet!stt
From: stt@inmet
Newsgroups: comp.lang.ada
Subject: Re: Ada timing
Message-ID: <20600011@inmet>
Date: 20 Oct 89 20:18:00 GMT
References: <679@sce.carleton.ca>
Nf-ID: #R:sce.carleton.ca:-67900:inmet:20600011:000:1116
Nf-From: inmet!stt    Oct 20 16:18:00 1989
List-Id: <comp.lang.ada>



It is not currently well defined in Ada exactly
when the delay starts for a timed entry call, though
I suspect a goal of Ada9X is to define this well.
The place where this becomes interesting is obviously
in a physically distributed system.  

Personally, I have always favored the interpretation where
the delay starts when the call "arrives" at the acceptor,
since this makes the conditional entry call equivalent
to a timed entry call with a delay of 0.  
The delay is therefore interpreted as
a limit on the queue waiting time, rather than on
any aspect of the communication overhead of the entry call.

Of course, there are situations where the other interpretation
(where the communication delay contributes to the expiration
of the delay) is more useful, but since it is so hard
to define well in the absence of "global time,"
and to implement efficiently and faithfully, and
since it means that a conditional entry is *not* the same
as a timed entry with delay of zero, I don't think it
is the right answer.

S. Tucker Taft   stt%inmet@uunet.uu.net   uunet!inmet!stt
Intermetrics, Inc.
Cambridge, MA  02138