On 28/08/2018 00:04, Gene Heskett wrote:|
My knowledge is based on a conversation I had with my then isp in about 1993 or so, so its entirely possible that the protocol has been changed since then. What they had then struck me as very very wastefull of resources. Because I was such a PITA, they actually built another machine for NNTP and had at&t bring in another oc3 circuit to feed it. I had what was a full house Amiga 2000 with 64 megs of ram on a PP&S 040 board, had a pair of 1GB scsi seagates, their machine had a 47GB drive, which was filled in just an hour or so, so the expire was set at 8 hours. So the last thing I did at night was dial them up and grab what I wanted that was new, and the first thing in the morning, the same. Sheer economics has likely driven some major changes in how NNTP works today. And I expect thats a hell of a lot better for the average ma & pa isp. By the time I built a new machine and put red hat 5 on it, in 1998 I think, NNTP had degenerated to 90% spam, so I never rejoined that pool party, it was too polluted for me. Email was easier to filter, and here I am still, almost 20 years later, and older too, I'll be 84 in a couple more months if I don't miss morning roll call first.
As in my reply to Michael Stone, posted just now, you are conflating NNTP with Usenet. The problems you describe above are all to do with Usenet, not with the NNTP protocol per se.
NNTP is not a bandwidth hog. It is not now and never has been.
Usenet was (and still is) a bandwidth hog, but it would have been so no matter protocol was used to transmit it.
-- Mark Rousell