Since Jonathan Marder requested that this discussion be held in
bionet.plants, I have set the followup for this article to that group.
I wrote:
>> I wish that listserver groups were welcome to establish gateways in the
>> Bionet domain[...]
kristoff at NET.BIO.NET (Dave Kristofferson) replied:
>This is a subject which raises the hair on the back of my neck to some
>extent so pardon my passion in advance!
>>LISTSERV in its current form is **NOT** going to be part of the
>BIOSCI/bionet domain again. We have already gone through great pains
>to excise it in the past (note that there have not been any mailing
>loops for a long time!!!!!). If we have had any part in minimizing
>its use, then it will have been time well spent (more below).
Well, the number of LISTSERV groups (~3000) seems to be growing faster
than the number of Usenet groups. That's actually not true, because
there are many, many more Usenet groups that are not "propagated" beyond
the bounds of universities or geographical regions, while all LISTSERV
groups are counted in the LIST GLOBAL kept at BITNIC.educom.edu.
I agree completely that for the Bionet groups, BIOSCI's old system of
Usenet newsgroup and by-hand mailing list and listserver did not work.
However, the SAS-L/bit.listserv.sas-l and STAT-L/bit.listserv.stat-l
have generally performed excellently, with no routine maintenance by
anyone. Sure, there is occasional bounced mail, but nothing like what
the Bionet groups used to see on a regular basis. Electronic mail has
improved over all, all over the world. Recently, I got a submission
rejected by the LISTSERV that runs WISENET, with the explanation at the
top that my article duplicated a recent posting (I had forwarded an
announcement seen elsewhere). That's a nifty feature!
>> and not everyone will
>> have access to Usenet, no matter how useful Usenet is to those of us who
>> have it. And the more biologists who subscribe by e-mail, the more
>> drudgework Dave et al. have to do rather than providing more interesting
>> services, like better FAQ sheets.
>>Not so. We are shortly going to announce a new automated subscription
>service based on a program called "majordomo" which will handle
>subscription, unsubscription, and archive retrieval requests
>***without mangling mailing headers of postings*** the way that
>LISTSERV does (this affects the non-LISTSERV distribution schemes.
>LISTSERV may work great in isolation).
Not everyone will have access to Usenet. That is unquestionably true.
There are now many LISTSERV mailing lists gated into Usenet, most
at American University (write to net-admin at auvm.american.edu for info).
Experience with them over 5 years now has been quite satisfactory: just
ask the SAS-L and STAT-L groups how they like it. LISTSERVs do work,
however much some people don't like them.
I have never heard of "majordomo". Is this something written at BIOSCI,
or is it what other Internet mailing list owners are using? There are
quite a few systems for running automatic list maintenance, most of which
don't seem to work as well as LISTSERV, in my own experience.
>[...] we have no desire to have to tie our
>USENET newsgroup and mailing list administrative efforts in with a
>zillion mailing lists running under LISTSERV at sites all around the
>globe. This would make it necessary to contact too many different
>groups if something goes wrong (which *will* happen). We have just
>succeeded in reducing the number of BIOSCI administrative centers down
>to a a manageable number of *two* and have no intention of
>complicating the system all over again. That would show that we truly
>did not learn from the mistakes of the past.
I don't see how having LISTSERV groups gated into Usenet as part of the
Bionet domain has anything to do with administrative efforts at BIOSCI
for the non-LISTSERVed groups. BIOSCI and Bionet are not the same.
If anything, I am suggesting that BIOSCI get out of the business of
maintaining mailing lists as soon as possible. Most Usenet groups
manage to sustain themselves solely by the interest of those who
participate, and those that don't generate sufficient interest are
automatically removed.
[stuff deleted]
>There already are a number of listservers gatewayed into USENET.
My .newsrc file says there are 274 gated LISTSERV groups.
>We don't have the resources (0.5 FTE of me and 0.2 FTE of Kenton) to
>handle the kind of problems that this can create. We have a certain
>type of service (USENET and non-LISTSERV mailing lists) that we
>believe that we can competently provide. Trying to be all things to
>all people would result in a significant deterioration of our service.
I really must object here. Usenet is not a service that BIOSCI
provides to anyone, except perhaps those who happen to read Usenet
on computers owned or administered by BIOSCI. Bionet is a domain of
Usenet which BIOSCI was responsible for creating, simply by asking
for it. Someone else could as easily create new Usenet newsgroups
in some other domain, or even create some other domain, and someone
would have done it if BIOSCI hadn't. But now that Bionet is well
established as a Usenet domain, it would be nice to expand the domain
to include other groups. In no other Usenet domain does one person
have the final say-so as to what groups may be created and what their
names are to be, simply because they already take care of (all) other
groups in the domain.
It is very nice that in the Bionet newsgroups there is someone who
reliably explains how to subscribe to people who send subscription
requests to mailing list, not biosci at net.bio.net. But this service
is already duplicated by other people within the Bionet groups (I've
asked) and handling such requests is not a problem, no matter who
does it. BIOSCI does provide a true service by taking care of all
subscription requests (that's a real chore, I'm sure!), and I am glad
to hear that this service will be automated. But this service comes
at a cost: the Bionet newsgroups can only expand as fast and as much
as BIOSCI is able to find funding to employ individual people to do
what in the rest of Usenet is done by newsgroup participants for free.
They do it for free because no one has to do it too long: there
always seems to be someone who is willing to contribute some effort
to these administrative chores. By spreading the work around in this
way, no one ever has to do much. The biggest single administrative
chore that exists within Usenet is the writing of FAQs. The biggest
chore of all is that of the dedicated and truly noble computer system
administrators around the world who have written the free software
that we all use, and who keep it running smoothly for the rest of us.
>It is not my intent to try and "take over" the biological
>communication world and bring everything under the BIOSCI/bionet
>umbrella. We want simply to provide a useful service *that works
>consistently*. If people are wedded to LISTSERV, then I suggest that
>they try one of the other means of gatewaying in to USENET. If they
>want us to take over complete administration of their mailing lists
>and newsgroups then we will be happy to assist. We are not, however,
>going to create an adminstrative and service nightmare for *everyone*
>just because some groups don't want to give up the old software that
>has been in use since god-knows-when on BITNET.
Hey, be fair! It's entirely arguable that the LISTSERV software is
still in use after at least a decade *because* it *works consistently*.
And various Usenet reading programs have been around a long time to:
should we throw them out because they're "old"?
"nightmare for *everyone*" is rather misleading. As I said above,
many of the gated LISTSERV groups are entirely satisfied with the
arrangement. I am only arguing for the greater use of the Bionet
domain name, not Dave Kristofferson's valuable time and energy.
The whole point of domain naming conventions in Usenet is to collect
groups by topical subject or scientific discipline. The LISTSERV
group GIS-L recently became comp.infosystems.gis rather than
bit.listserv.gis-l, which would follow the standard convention for
gated LISTSERV groups. Names are important: it would be a shame
to gate PHOTOSYN into Usenet as bit.listserv.photosyn when a better
domain name exists.
Perhaps it's time to create a new domain: I propose "bio", and thus
bio.photosynthesis! I will be happy to propose this formally in
bit.admin if Jonathan Marder, owner of PHOTOSYN, likes the idea. Or
he could propose it himself. There are both procedures and precedents
for establishing LISTSERV gateways in domains other than bit.listserv,
and for creating new top domains. I believe that the ECOLOG-L list
would be interested in establishing a gate to Usenet as bio.ecology.
--
Una Smith Biology Department smith-una at yale.edu
Yale University
New Haven, CT 06511
