On Fri, Jun 16, 2000 at 01:43:17PM -0400, WebWiz wrote:
> Christopher Masto wrote:
> >> My recommendation is "accept them all, and let the
> >> MTA system sort 'em out".
>
> That's a really bad idea when you consider the fact that the administrative
> email address is the ONLY way we have to send a password to a user who has
> forgotten it. Certainly you'll agree that this happens on a semi-regular
> basis.
That is a separate problem, which I do think needs to be addressed.
It is too easy to generate a catch-22 situation with the current scheme.
> Do you really want to have to tell a customer that you can't get his
> password for him because he entered "<no email>" for his email address and
> you accepted it into the system without complaint or warning?
Does it make much difference whether they enter "<no email>", spell
the address wrong, use a hotmail account that gets closed, or any
other of a thousand possible mistakes? It may be useful to require an
"@", but that's about it. Unless you're going to mail a password to
the address and wait for a response before you'll register the domain,
there's not much point in applying the rather extreme parsing it takes
to verify syntactical compliance with RFC822. I know there's a
regular expression. I also know how long it is. I also know the
owners of the (valid) e-mail addresses, "fred&barney@stonehenge.com",
and "*@qz.to".
FWIW, we register most domains with "hostmaster@domain" as the contact
and recommend the users leave it there, so it can be redirected at will
by whoever controls the name (hence ultimately mail) servers. As always,
a clever fool (the most common type) can easily defeat this. I've
probably made it clear by now, though, that I don't believe in chasing
the impossible goal of 100% foolproofing.
-- Christopher Masto Senior Network Monkey NetMonger Communications chris@netmonger.net info@netmonger.net http://www.netmonger.netFree yourself, free your machine, free the daemon -- http://www.freebsd.org/
This archive was generated by hypermail 2.1.3 : Tue Oct 19 2004 - 23:35:38 EDT