Re: T2 Clarify TimeToLive

[Dan Weinreb <dlw@exceloncorp.com>]

I don't think we should conflate TimeToLive with business-level
concepts such as "This special offer is only extended for 30 days".
First, you might want to receive a message about a special offer even
if the offer period is over, just so that you'll know that there was
such an offer.  Second, you might want to send mail about a special
offer that lasts for year, but have a TimeToLive much less than a
year, because TimeToLive is really driven by the needs of the message
protocol rather than considerations up at the business level.

The spec should make clear that if a message's TimeToLive has passed,
then the message is discarded and forgotten.  For example, if a To
Application asks the To MSH for an incoming message, the To MSH must
check the TimeToLive, and if the time has passed, the To MSH must
discard this message, and behave as if this message had never been
present at all.

TimeToLive is particularly important because it saves you from
having to persistently remember the MessageID of every message
you have ever received (as I have discussed in earlier mail).

-- Dan