This is going much different than expected. It is no issue whatsoever to
limit my "online time" to something like 20 minutes a day. Apparently,
my usage patterns don't require a lot of sustained "online time" in the
first place. It's much more important that I'm able to connect several
times a day, so I can repeatedly fetch mails, read news, publish twtxt
replies, and so on.

If the Internet really is *unavailable* due to an ISP outage, I do feel
"cut off from the world". But as long as I can connect as I wish, it's
fine.

So ... maybe the cost per connection has to be accounted for after all.
I don't remember if this pricing model existed throughout the entire
1990'ies or if it only existed in the beginning. I just tried to find
out, but this information is pretty much lost these days.

Not sure how to proceed. Should I abort the experiment? I haven't really
learned that much thus far. Maybe I'll just keep going and see what
happens next.