[philip.j.fry]

Originally Posted by stuartbe

The trouble is that they will get complaints from people if they start blocking mails. I have sugested before that they run some form of AI software like spam asasin and simply flag the mail. The end user could then set up a rule to zap the crap - so to speak.

Problem with that is that AI routines are traditionally very processor intensive, especially heuristics like those used in SA. It's fine for a home machine but lets face it, NTL's mail servers can't cope as it is without the extra burden of spam filtering.