Filter list members,
I have heard the current thinking on vacation is to track the
response interval for each vacation action. Thus, a person
may receive more than 1 vacation response in a given interval
if their messages were to cause different vacations actions
to be executed in a sieve script.
My understanding is that vacations actions would be considered
the same if they contained the same response string, even though
they may be different statements within the script. It was suggested
that a hash code of the response string could be computed to reduce
the storage requirements. The hash code was equal, the vacation
string and action are assumed to be the same.
This requirement increases the burden on the implementation, which
has to save the address and time of each vacation response, now
has to save also the hash code of the vacation text.
This is another trade off between MUST and SHOULD. I would prefer
a statement like...
Sieve implementations SHOULD track vacation response times for
each vacation action based on the text string. Vacation actions
with the same text are treated as the same. Some implementations,
due to constraints, MAY treat all vacation actions, regardless of
the text, as the same.
Do we really want to make this MUST?
Greg Sereda