[RFCI-Discuss] about FQDN for our smtp servers

[Yiorgos Adamopoulos]

mouss wrote:
> but at the same time, it says the client SHOULD use a "canonical" helo. 
> why should a client developer do any effort to follow this SHOULD if the 
> same RFC says the server shouldn't refuse it because of that?

Because the developer must follow the robustness principle:

- Be liberal in what you accept (server)
- Be conservative in what you sent (client)

Keep in mind that both the initial RFC and the robustness principle were 
devised by the same person, the late Jon Postel.

The contradictions you have found exist mostly because of this reason 
and copying from:

http://en.wikipedia.org/wiki/Robustness_Principle

"Counter-intuitively, Postel's robustness principle ("be conservative in 
what you send, liberal in what you accept") often leads to deployment 
problems."

More on this on RFC3117

./Yiorgos