Does anyone use ISP email addresses anymore? (It's just not very smart to rely on.)
I concur. I tell everybody I end up helping to NEVER use an email account from your ISP.
A person should expect to fire their ISP every year and adopt a practice that makes that trivial to do. Hotmail and Gmail are pretty much the Global Email Providers that most likely won't be going away.
Even though my ISP provides me with an email automatically, I don't check it and I told them to use the email address I give them to send me mail. Any ISP that does not understand their role in the relationship in the 21st century is an ISP that does not understand that they are simply packet passers.
As to the general concept of blocking by IP, there are unintended consequences and complexities to that idea.
One example of good idea gone bad is where a manager told his IT staff to block all traffic from China. Worked great until the evening, when Google shifts its load over its chinese servers. So attempts to visit google were failing because responses were coming back from China.
Furthermore, the magic and mystery of the IP address has changed over the years. Once upon a time, every PC got an IP address that was directly reachable from any other PC on the internet. Nowadays with firewalls and Natural Address Translation (NAT), the IP address your PC actually has is likely to be 192.168.1.X where X is a number 1-255. Your firewall or router may have an internet-reachable IP address, but nothing in your house or company does.
This means that multiple users inside your firewall all show up as the same IP address when they both browse this site.
It's also not as simple as every address that starts with 101 to 125 is a chinese address.
It's also not hard for a Chinese spammer to lease a linux server in the US and then setup email accounts on that and to send out from there. Or to even remote-control browse from the server to american servers (so it will have an American IP address). There's no magic backtrace that could show that the server is being driven remotely from the perspective of Enworld.
There's better ways to block bots. Using the goofy graphic text you have to re-enter is effective, and you're actually helping a document scanning project (those snippets come from an OCR project that has stumbled on those characters).
Requiring an initial post in an Introduction area is also probably reasonable. As is doing an email verification to prove your really you. Both mostly require a human to take the time to go through the process.