Dave Crocker writes:
> A side discussion about history might be interesting, but I'll suggest
> that it should not really be a factor for the current discussion.
> The reasons are:
> 1. It has a substantial installed base of use now.
Reason enough for me.
> 3. It is not generally recognized as causing any problems now.
I concede that most people don't think lost DSNs are a significant problem.
Let's be careful with generalizations here. Depending on what you mean by
"lost", I think there are significant problems and moreover ones we should
be trying to address.
However, I see no-MX fallback policies as having little if any impact in this
area. In fact I'll go further and say that if the goal is to improve the
reliability of email error reporting (and there is no doubt it is in trouble)
various things being proposed here would be in fact be actively
counterproductive in this regard.