Thu, 5 Aug 1993 20:17:55 -0400
|
On Aug 5, 15:01, "Lawrence E. Rondot" wrote:
} Subject: Re SCIFAQ-L errors
} Are we all arguing about the same thing? Chris Lewis sent my list a FAQ
} one night by hand. The following morning (5 hrs later) his Usenet process
} also posted the FAQ. The second posting was rejected as a duplicate.
Important distinction: I've never sent your list anything, by hand or
automatically. I only send to USENET.
That incident was indeed a finger problem on my part, but it still begs
the question of whether a LISTSERV should ever bounce a duplicate back
at a USENET-origin posting. The SCIFAQ-L incident is much different.
Una must have his checksum period pretty long to catch Detweiler's
postings. I don't know what his interval is, but I presume that it's at
least a week. Few FAQs have that short an interval, but a couple do
(they have usually been set that low after experimentation with a
specific group. I had to make the wiring FAQ go out once a week because
the repeated Qs that were in the FAQ got ridiculous). Certainly not 13
parters like L. Detweiler's.
} As I understand the process, Listserv accepted the first FAQ and passed it
} to the list membership. As part of the posting process it added a CRC
} calculated from the body of the message to what I assume is a circular
} queue of 90 entries. Why 90? Because I set my list to accept 90
} messages per day with the "Daily-Threshold= 90" statement in the list header.
} This normally defaults to 50, but this list/Usenet group generates on an
} average day about 75 messages.
This sounds like a good setup. Actually, I'd personally suggest that
you set the checksum count up much higher. Come up with a minimum acceptable
repeat time (perhaps a week or more) that your readership sees reasonable.
But don't do this until Eric fixes the bounce-usenet postings problem.
} Are we arguing about rejecting the posting after a month because
} it is a duplicate realistic? Does this happen often? I would think that it
} would not be much of a problem unless the FAQs were the only thing posted
} to a list (is this the case for the SCIFAQ-L list?).
This is my impression.
--
Chris Lewis; [log in to unmask]; Phone: Canada 613 832-0541
Psroff 3.0 info: [log in to unmask]
Ferret list: [log in to unmask]
|
|
|