From: Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu>
I think that having all bounces (for whatever reason) archived is
fine; I think having it as "web pages somewhere" is overkill.
both the volume of spam, and the ratio of spam to legitimate content are
so high, that I'm not sure how much longer it will be practical to
archive it. if we were to archive rejected messages, it should probably
only be for a few weeks.
My rolling 40 day log of all spam sent to my traps or real addresses
contains about 34,484 samples in a total of about 242 Mbytes or an
average of about 7 KBytes/spam. (Each sample is truncated to ~32 KBytes.)
Judging from DCC numbers from a bunch of medium sized ISPs, the typical
consumer mailbox receives about 10 messages/day (more than 5, less
than 20), of which about half are spam. (Never mind that judicious
"unsubscribing" can reduce that by about 50%.)
So in round numbers, assume 10 spam/list/day. If the IETF has 100
mailing lists, archiving all IETF list spam would involve 1000 spam/day
or about 7 MBytes/day or 2.5 GBytes/year.
I remember quite well when a large disk drive was a ~2 meter cube,
weighed over a ton, used 220 volts and compressed air, and held only
48 MByte, but those days are long past. 2.5 Gbytes would fit on a
single DVD, not to mention modern magnetic media.
I keep unique (as determined by DCC checksums) copies of all of the
spam sent to my trap addresses permanently on CDROM. I have ~1000
trap addresses, although only a several dozen are hit more than half
dozen times/week. (These records help me answer complaints about my
blacklist.)
Vernon Schryver vjs(_at_)rhyolite(_dot_)com