[Nottingham] Just how complicated can email addresses be?!

[Michael Erskine]

The pragmatic approach is to specify for your individual usage case
what "valid" needs to achieve: e.g. if you're just looking for typos
in a web form then it makes sense to use a quick javascript regex to
filter out obvious invalid email addresses without requiring a http
POST -- ever notice how often there are two email address fields in
web forms? To catch exactly this sort of human data entry fault! Once
the form is submitted and you can dedicate some server-side scrutiny
to it, it makes sense to use a rigorous, modern, widely-used,
opensource library function rather than attempting to catch all cases
yourself. I trust a hundred minds working on common task more than I
trust my own mind working on that same task. That said, an address
that is entirely "valid" may not be a real address at all and that is
why we have confirmation emails: "reply to this email to confirm your
address" -- it works, that's why it is so common!

Regards,
Michael Erskine.