Content filtering at wire-speed demands big iron. Latency will always be introduced, and to keep it to a minimum, the cost of equipment rises sharply. The most effective way is at the application level (notwithstanding the above bypass methods), but that will need even more grunt, and will generate massive log files coupled with the age-old problem of what to do with such a huge store of data.

On top of all this, the ISPs should be quaking for fear of potential litigation:

“So and so died because he got a web page that told him how to top himself. The ISP is supposed to stop this.”

“I should have had that information yesterday, Now I’ve lost a 10 million dollar contract”

- hypothetical I know – but you get the picture. To avoid this, then the ISPs will need to adopt the Wilson’s parking stance: “All care – no responsibility”. Which of course adds to the inherent technical problems of meeting accuracy levels. It’s not fair or correct that a carrier be responsible for the content. As has been pointed out many times – the post office will never be subject to this kind of scrutiny. Why the internet?

Much better would be legislation for web sites to encode their own classification byte and let the end-point filter it. Then a spider could keep some kind of eye on things, and public-reporting of breaches could target offending operators at the source where it belongs.