(PART 1)
Hello,

This is just a creative idea of mine, feel free to add more to it. I think it would be helpful to those server communities out there that are experiencing significant RMT spam via /tell.

Here are the things I think a good filter should do (based on my past experience coding programs to identify hacking attempts and identifying original DNS request based on a log):

-determine the message sent is a RMT spam message if it fits ALL of the following criteria:
1) sender and receiver are not on each other's friendlist
2) sender and receiver do not belong to the same FC
3) sender and receiver do not share membership in any LS
4) length of message exceeds X characters
5) message contains known RMT URL address
6) message contains rare-in-conversation symbols known to be used for the sake of bypassing URL filter*
7) words "sale", "delivery", and any other frequently used words used to advertise RMT is present in the message*