email edit

[Jay R. Ashworth]

On Wed, Nov 01, 2006 at 12:50:31PM -0500, Brian K. White wrote:
> >50K email addresses off Usenet over a month, and test their edit
> >against them.  That's a sufficiently large corpus, I think, to catch
> >all the corner cases.
> 
> Indeed, almost anything can be a valid email.
> You don't need even one @

That's actually no longer true.  Most mailers can't handle
bang-addresses anymore, and there are vanishingly few mailers with UUCP
links configured at this point.

> >50K email addresses off Usenet over a month, and test their edit

> I wouldn't be surprised if you can have more than one @

You can't.

> You don't need even one .

You do.

> You can have any number of .'s

This is true.

> You can have one @ and no dots

Only on internal networks.

> then theres: network!hosta!hostb!hostc!user

See above; find me a live example of an address like this, and I'll buy
you dinner at Berns.

> and: john%node.bitnet at cunyvm.cuny.edu

This one, I did mention, but they're almost dead now, too; there aren't
too many networks without smart gateways these days.

> Those are all odd cases but perfectly correct if thats the way your 
> buisiness happens to be set up.

True.  But we're not programming a mailer...

> I think there are only a few bad characters you could really test for and 
> safely say it's bad... except then there is utf8 and utf16 multibyte 
> encodings coming our way pretty soon and any "bad" character might just be 
> part of a multibyte japanese address or something.

At the moment, RHS's must be ASCII.  International domain names are
available, but still kinda dodgy.

> The only correct way to validate an email address is to actually validate 
> it.
> Which requires the machine running fp to have a working mta.

That's the best way, but you can't really *validate* it, either; "VRFY"
only works on mailers run by Really Stupid Sysadmins.

> Or you could use one of the many 3rd party apps out there that claim to do 
> just this.

And they do that how?

> In the end, you _still_ can't really know because you have no control over 
> the recipient mail servers.

No, but we're only trying to trap typos.

> really, "none" is a completely valid email address and by rights I have no 
> right to delete it

For sufficiently small values of "valid", I guess.

> Thus it's not going to be handled properly by anything as simple as an edit.
> And I happen to rather not touch the data rather than touch it improperly.

There's something to be said for that viewpoint, certainly.

Cheers,
-- jra
-- 
Jay R. Ashworth                                                jra at baylink.com
Designer                          Baylink                             RFC 2100
Ashworth & Associates        The Things I Think                        '87 e24
St Petersburg FL USA      http://baylink.pitas.com             +1 727 647 1274

	"That's women for you; you divorce them, and 10 years later,
	  they stop having sex with you."  -- Jennifer Crusie; _Fast_Women_