"I work for a large business ISP in the UK which also hosts some large home-user ISPs, and the team I work for deals with our support mailbox." the reader wrote. "We get approximately twenty to thirty e-mail a day, and maybe five calls a night, that take a course as follows:

End user: Hi, Someone's trying to hack me, and they're on your network, their IP is 123.123.123.123. TELL THEM TO STOP IT OR I'LL TAKE LEGAL ACTION!

Support: That IP address belongs to a web proxy that you use as part of your contract with (an ISP we host).

End user: Then tell them to stop it!

Support: 'They' are not doing anything. 'They' is a server cluster. What told you about this?

End user: My firewall.

Support: Which one is it?

End user: Norton Personal Firewall.

Support: OK. Can you try a different one, and see if the same thing happens, please?

End user: Ha! I already have! But nothing showed up, so I guess it was inferior. NPF is an industry standard firewall, in use in enterprise solutions!

Support: Not ours. Our standards are Cisco and Nokia's.

End user: Anyway, I hope they stop it.

Support: I'm sure they will."

The reader finally decided to start tracking what firewalls were involved in such complaints. "After a particularly busy night on the phones, we counted these calls up and of the 40 or so we received in one night, EVERY SINGLE USER had NPF," he wrote. "I then asked our Phone Robots -- ticket loggers/call answerers -- to keep a record of numbers of calls, originating IP, and firewall being used to report this. That's right, every single one was Norton Personal Firewall."

What's the reader to conclude? "Either; a) NPF is badly coded, and sees normal broadcast traffic as intrusion attempts," he says. "Or, b) ISP's the world over clearly have no control whatsoever over their autonomous servers. We could say b), but Symantec would probably try to sell us a firewall."