Personally, however you choose to implement denials of multiple calls, my
thought is to just do it. User calls and connects with PM, user's friend
calls and connects with PM, your method detects *possible* multiple login
and double checks (best make sure), the first call is dropped, and the
process continues.
No warning messages, as a matter of psychology. User thinks "Hey let's see
if I can let my friend use my account?" System gives them a kick with no
warning and the user feels like a deer in the headlights. "Scratch that
idea" and askes friend to stop using account.
Haven't lost anyone to this yet.
We are unlimited, so one account, one call and violaters will be persecuted.
All callers connect via PPP/SLIP and then have to telnet in (or call a term
server if that is all they want). I find this a better solution than
prefixes or suffixes. Bogus shells limit services, so PPP/SLIP only cannot
telnet or ftp to the system.
The fact that we do not run accounting eliminates another solution. Those
that charge for time, either per hour or blocks, can include checks when
they run billing on the accounting logs. A message could then be generated
that would advise the user that allowing a friend(s) to use thier account
will cost THEM.
It seems that most want someone else to do the work for them, and expect
that they will not have to do very much thinking.
Everyone has their only style and each system has thier own way of doing
things, so there will *never* be and easy solution for everyone.
My .02 to the thread.
------------
Jeff Mountin
sysop@mixcom.com
MIX Communications
Serving the Internet since 1990
Sure my business card says "Senior Network Administrator"
They still make me do just about anything.