weemba@brahms.BERKELEY.EDU (Matthew P. Wiener) (04/13/86)
After watching for a while, I'd like to present various suggestions for improving(?) the network. (1) Have posting software force the poster to list his entire article, signature included. This may eliminate/reduce doubled signatures, overly large signatures, followups with the smallest of responses, the wrong editor buffer response, etc. (2) Have posting software apply some crude heuristics to recognize tests, and warn the poster that tests go to net.test with minimal distribution. While these are bytewise rather minor, they generate far too much obnoxious ill will between naive newcomers and the hardcore net.police. (3) Give backbone sites a super kill file. If it is felt that too much noisy traffic is generated by one person, then let sites choose so on an individual basis to cut their costs. If a few key sites killed run-on puns, for example, they would die. It is no longer a question of free speech but of expensive speech. (4) Don't name any groups talk.* or trash.* ! How about vol.*, as in volatile (or volume)? (5) Before posting, the software inquires newsgroup by newsgroup for a [yfn] answer: y == yes, post to this newsgroup, f == and also include this group in the followup line, n == not to this newsgroup. Too often inappropriate newsgroups drag along because posters don't even notice. As the default, only the first newsgroup listed gets a followup, not all of them. (6) Pnews offers me the choices of send, abort, edit, or list. I'd also like kill, which must be spelled out, meaning don't even bother saving a copy. (7) ftp-able archives longer than most sites' expiration dates would be nice. Say six months. Do they in fact exist? ucbvax!brahms!weemba Matthew P Wiener/UCB Math Dept/Berkeley CA 94720