bstempleton@watmath.UUCP (Brad Templeton) (06/08/83)
Somebody named Gene Spafford recently rather foolishly used net.test (When are people going to learn that net.test should never, ever be used except by the authors of the news software, and that's about two people right now) to test signature files, which are undocumented for some reason. He has put in a 5 or 6 line signature with a dozen ways to reply to him. You will note my file is one line long. This is because these signatures are getting rediculous, and rather a waste of other people's money since they are sent over hundreds of links. Please folks, let the news software and the mail aliasing or internet programs worry about how to reply. We already have a huge header full of such information. Please don't include it again. -- Brad Templeton - Waterloo, Ont. (519) 886-7304
spaf@gatech.UUCP (06/09/83)
I guess I should apologize to Brad Templeton if my ".signature"
file upset him so much as to label me as "foolish." Most
people need more information than that to make such a judgement.
However, his comments do raise a few points:
1) I didn't know that "net.test" was only for "a select few";
is that why he was reading it? I wanted to test something,
and the newgroup was there.
2) The article cancellation feature in "readnews" does
not seem to work. I tried to cancel my submission after
I had seen what it did.
3) Why is the ".signature" file undocumented in "postnews"?
That same file is used by the MMDF "send" program.
4) I use the same signature file for mail over 3 different
networks and that requires 3 different types of return
address. Since uucp mail is not always reliable, and since
there are two distinct paths to here, I have 2 uucp addresses.
The first line is my "signature" and I've been using it over
CSNet, etc. for many, many months.
5) If you're concerned about costs you might complain
about some of the other stuff that goes out over news.
By the way, if you have another bad day and want to grump about
something like this, may I suggest we meet at sunup in net.flame
with metaphors drawn?
Not afraid to sign myself:
"The soapbox of Gene Spafford"
Spaf @ GATech (CS Net)
Spaf.GATech @ UDel-Relay (ARPA) School of ICS
...!{sb1, allegra}!gatech!spaf (uucp) Georgia Tech
...!duke!mcnc!msdc!gatech!spaf Atlanta, GA 30332
--
"The soapbox of Gene Spafford"
Spaf @ GATech (CS Net)
Spaf.GATech @ UDel-Relay (ARPA) School of ICS
...!{sb1, allegra}!gatech!spaf (uucp) Georgia Tech
...!duke!mcnc!msdc!gatech!spaf Atlanta, GA 30332dje@5941ux.UUCP (06/09/83)
I would very much like to avoid the necessity of writing out primary and
alternate paths to my home system every time I sign a message. Unfortunately,
many of my MAIL submissions never get through to their destination, and
many MAIL messages sent to me never reach me. I have no way of finding out
where the messages are getting lost, since no failure notifications ever
reach me.
Maybe this is a problem with MAIL/uucp rather than with NEWS. We users
should not have to know about the topology of the network in order to send
or receive mail. Especially since there is no way to ascertain the set
of uucp links to and from remote systems (perhaps there should be a special
uuname option e.g. "uuname -s harpo" to tell which systems harpo can
send/forward mail to -- but this would entail either a massive data base
coordination among systems or else built-in, hidden remote dialup access).
What is a good, workable solution to this? Until I can be assured of
RELIABLE mail delivery in both directions (and why shouldn't the network be
treated as a utility I can count on, just like my telephone?), I see no
better choice than posting multiple paths to my system address.
Any help, practical suggestions and ideas are welcome. Reply to the net
if of general interest, or to me:
Dave Ellis / Bell Labs, Piscataway NJ
...!harpo!npoiv!npois!houxm!5941ux!dje
...!{ariel,lime}!houti!hogpc!houxm!5941ux!dje