On another list, I made some statements regarding the efficiency or lack
thereof, regarding the way Listserv postings are distributed ON individual
hosts after they pass through Bitnet. Now, I am wondering if perhaps our
Listserv, or our VM system itself, is being run properly in this regard. If
this is a problem peculiar to this system, I would certainly like to know
about it and have our systems people fix it.
If, for example, 200 people on this system are signed up to a particular
listserv distribution list, our system will end up holding 200 copies of
every posting sent to that list; one for each subscriber. There are in fact,
several lists which are very popular here and very prolific so this example
is realistic. Are we doing something very wrong here on our system? It
seems wasteful to have many copies of the same posting in this system's spool
area when one copy for central access would do quite nicely. My
understanding, however, is that this is just the way things are done on VM
and that there's nothing anyone can do to change the situation. Perhaps I am
wrong. If so, I'de like to know so I can ask our systems people to fix it.
The reason I am posting this here because several people thought I was insane
when this topic came up elsewhere. You know how insanity is, the victim is
always the last to know. :) I heard that some sites have implemented a
bulletin board system to centralize incoming Listserv postings. Is this what
most sites do? Maybe there's an outside chance that its not necessary. I
could probably implement this type of facility here, but it would take a
while because I'me already too busy with other matters. Is there perhaps
some way of improving the internal mechanism on VM systems to change the mode
of distribution of incoming list postings that's not commonly known?
From my account, I can easily see these replicated postings on our system's
reader so this is definitely not a figment of my imagination as some people
on another distribution list implied. I even wrote a SAS program a couple of
years ago which summarizes our spool file situation. About a year or two
ago, I had to use this program often to help head off problems resulting from
the large volume of spooled files here. This was on our previous system. We
still have many users who frequently have several hundred spooled files. The
max is something like 30,000. If there's a way to improve this situation
without implementing any user restrictions, I would certainly like to know.
With about 3,000 users on this system, this situation can cause problems,
although its been rare lately thanks to our bigger system.
My disclaimer follows. Some of you probably saw my postings elsewhere
regarding Listserv. Let's not continue that thread here, or privately. For
the record, I happen to like and respect Listserv a lot. We intend to keep
Listserv and Bitnet at Temple. Its also time to toot my own horn since I am
feeling a little defensive right now. It is accurate to say that the reason
why we have continued to run a Listserv here at Temple is almost entirely a
result of my efforts at popularizing our Listserv in the past. The view
among some managers here a year or two ago was that Listserv was silly
because it had little to do with academic pursuits. It took a lot of
posturing on my part to encourage those who held this view to reconsider. My
goal now is simple. I want to get an answer to the above question. I will
not participate in any long-winded debates on anything at this time. Any
responses to my question will be appreciated.
Take care everyone,
Stan Horwitz
Listserv Postmaster
Temple University
Acknowledge to: OASIS@TEMPLEVM (or VM.TEMPLE.EDU)
|