Re^4: About validating mail id

[tirwhan]

OK, I acknowledge I probably deserved getting the nits picked on my nitpicking node...

Just because an MX record isn't propagated to you doesn't mean it doesn't exists.

Which, to me, is irrelevant. If I can't resolve the MX and can't send mail to the A record of the domain, the email address is invalid (to me). If your point was that there is no such thing as a globally valid email address then we are in violent agreement.

a valid address doesn't have to contain a '@'

This is incorrect for email addresses. You can have a syntactically valid mailbox address (which may be delivered locally or rewritten/forwarded to a valid email address by your MTA) without the @, but a syntactically valid email address requires the @ to separate the local and domain part. See RFC5322.

But it doesn't mean that an address that is syntactically (according to RFC 822 or any of its successors) invalid isn't deliverable.

Weeell, most MTAs I know do check for syntactic validity(at least in the SMTP transaction, if not the message headers) and will not deliver to a syntactically invalid address. So regarding a syntactically invalid email address as generally undeliverable is good practice IMO, special circumstances for weird mail setups notwithstanding.

[JavaFan]

The section of the RFC you point at defines the grammar for addr-spec. I guess it depends on whether you call an addr-spec, or an address an "email address". It seem you do the former, I do the latter. I consider "Foo bar" <[e-mail]> to be an email address. The RFC itself doesn't mention the word "email" earlier than Appendix C. Acknowledgements