Thanks for replying to my questions, ForumAdmin.
3) though I don't see this small-scale test as conclusive proof that all will scale up just as flawlessly, there's certainly now a dent in the technical arguments against the filtering. Not that the technical argument would be the most important one, but still...
3) It isn't "small scale" for either us or any other ISP of a similar or smaller size which is every ISP in Australia except for 10 or so. Having said that we are told that the largest production use of the solution we tested has almost 1,000,000 internet users so that would make the test results apply to ALL ISPs in Australia with the exception of BigPond (unless my understanding of the Australian ISP market is way out of date).
Sorry for not being more explicit - I was in a more technical question over in the other "secondary" thread - anyway, with "small scale" I meant to refer to the size of the filtering target/list, rather than that of the customer base. I am not convinced that the impact of growing the list by one or two orders of magnitude can be easily extrapolated from this small one in a linear fashion.
The second issue, which I can't imagine possibly being caused by the trial, but seemed to coincidentally fall in it, was that grabbing new mail seemed to take considerably longer, and the bulk of my emails are all text-only.
Funny, I observed the same thing. Wouldn't know how it could be caused by the filtering, and attributed it to "something" at my hosting provider, but fetching mail was noticeably slower than usual for me as well.
Not sure about the latency from here - it wasn't as fast as it gets but there's quite a bit of variation there anyway and I did not do any systematic testing or/and extensive surfing throughout the trial.