Paul Russell wrote:

> The web interface should be modified to:
>
> (a) Provide more robust syntax checking that recognizes either
>     <[log in to unmask]> or [log in to unmask] as the subscriber's
>     address, regardless of its placement in relation to the subscriber's
>     name; or
>
> (b) Accept only the format used on the email 'add' command, i.e., the
>     email address followed by the name.

I'd settle for (c) Parse as it does now, then if this results in an @
sign in the person's "name", reject the whole thing as an error.  In
my experience, people who use computers are used to dealing with errors
and figure "if it's wrong it'll tell me".  And I'm not convinced it's
only people who use the web interface who make the error you describe.
I know that we see that mistake all the time around here.

I suppose perhaps L-Soft is trying to be flexible by allowing for the
possibility that a person - or more likely an organization - might
actually -have- an @ sign in their name.  Personally I think that would
be pretty unusual compared to the common mistake you describe, but what
do I know?  :)

Another popular mistake around here is including the person's name on
the DELETE command when deleting via email, because they want the
DELETE command to be just like the ADD command.  Thankfully, this
-does- result in an error (at least on 1.8d which is where we are) but
the error message (something about Global, Local and Test) could be a
wee bit more enlightening.  :)

Margaret King
Michigan State University