Re: header-munging

["D. J. Bernstein" <[e-mail]>]

> Given the
> growing popularity of disconnected clients,

It's worth noting that many disconnected clients (1) let the user decide
when to connect to the ISP and (2) give the user an easy way to delete
messages before #1.

In this situation, the moment of sending is #1. Before the user tells
the client to connect, he hasn't authorized mailing.

I don't know whether these clients are more popular than the other type.

> But at present we have frequently occuring operational problems with
> Internet mail that are caused by a misapplication of this very
> technique. 

Then it should be easy for you to show us some real examples.

When you say ``this type of problem: joe at foo.com ---> [e-mail],''
I don't see what the problem is. The client specifically told Joe to
fill in his username at the ISP, so [e-mail] will work just fine.

Similarly, I don't see why ``joe@foo ---> [e-mail]'' is a
problem. That looks like a sensible default. What are you complaining
about?

``[e-mail] ---> [e-mail]'' is an example of why qmail
uses syntactic rules rather than DNS for qualification.

> And if there's a significant chance that the SMTP server is providing
> the *wrong* domain, it's botching the message.  

No. ``Wrong'' doesn't even make sense; do you expect the computer to
read the user's mind? The computer is providing a _default_. If the user
doesn't want that default, he can override it.

> If we really want users to be happy, we can fix up their envelope
> return addresses to be unreplyable.  That way, they will never be
> annoyed by nondelivery reports!

Perhaps you should stop using such an annoying report format.

---Dan