Lauren Weinstein <lauren@vortex.UUCP> (07/31/83)
Greetings all. I just received some direct mail from
watmath!bstempleton on the continuing subject of "Usenet, Inc."
From the header, it is a little unclear if the article also went
to netnews, though it appears that was the intention in any case.
I am becoming a bit bored with this exchange, but let's give it
another try. I've edited my comments into Brad's message so if
you didn't receive the original via netnews, you'll get it
now (lucky you!)
I'm afraid that my comments may be becoming a bit more curt,
but even I have my limits...
Here we go...
--------
From watmath!bstempleton Sat Jul 30 17:39:41 1983 remote from allegra
To: Lauren Weinstein <lauren@vortex.UUCP>
Subject: Usenet, Inc. -- The Saga Continues... : net.news
Cc: news
References: <77@vortex.UUCP>
Well, I must not be expressing my points well, because there is more
opposition to this sort of company than I thought.
Indeed.
What makes you think that the usenet isn't paying a profit making
organization? What is the Bell System?
Yep. A common carrier. At least their basic communications
operations are regulated, and available to "all" without
relying on some central site or sites to manage most
communications. I'll trust Ma Bell (or her descendants)
one damn sight more than "Usenet, Inc." as we've heard
it described.
My original suggestion for usenet inc. was to form a company that acted
as little more than a transport mechanism (like phone lines), software
supporter, database maintainer and mail router. The idea here is that
the usenet inc machines would be on telenet or tymnet so that almost all
sites could reach them by a local call, and the transmission charges
would be reduced by the use of this data network. This would mean that
as long as the central machines had the capacity, any site could join
usenet for less, and without having to seek a megacorporation for a mentor.
Have you ever really tried using Telenet or Tymnet for such
applications? I have. Those networks are really oriented toward
normal terminal-oriented communications, not file transfer or
other protocol-locked applications. Yep, you can run X.25 under
certain circumstances to interconnect sites, but performance
varies widely in such applications. However, if it were determined
that a few, say, Telenet links between major sites could make
a substantial difference in costs (with an unclear effect on
reliability) then such links could be set up WITHOUT ANY DAMN
USENET, INC. However, any use of Telenet/Tymnet would require
explicit allocation of resources to Usenet, which we've already
determined could be a major problem for many sites.
The site would maintain the central mail routing database, and suddenly
mail would become much more reliable and easy to handle. Sites who connected
to this computer could be only 1 site away from the most central node on
the net, if this is important to them.
Capacity limits on both Telenet/Tymnet and on the central machine(s)
would almost certainly prevent the attainment of this goal.
These commercial networks do not behave well when dialup communica-
tions do not proceed in a "typical" manner (that is, a user sitting
there typing at relatively slow speeds). The sorts of operations
required by Usenet would almost certainly clog the commercial network
processors and result in very poor performance. My own experiments
with uploading materials into Telenet/Tymnet have demonstrated
these problems to my satisfaction, regardless of what their
marketing people say. Another problem is that these networks have
only limited resources. Usenet alone is growing at a rate
high enough that the commercial networks would rapidly slip
behind, and they have lots of other customers as well.
To summarize, these networks might be useful for a few
backbone interconnections, but not for general use as Usenet
communications paths.
At no time have I suggested censorship or control of what flows through
the net by this transport company.
Well, I'm glad to hear it.
What I have just said could be all the company does. Since Bell, Berkeley
and Dec in a way have already dedicated machines to unix networking, this
could easily be enough.
An invalid comparison. These sites are providing services
for their users and, out of charity, for the outside world
(to a certain extent). They are not selling access to these
machines commercially in the manner that Usenet, Inc.
envisions.
But I have also had ideas that can exist on top
of this structure. I am getting comments (some - some +) on them now.
They are:
1) Using the central company as a base for moderators. Our net is currently
distributed, and Lauren thinks this is not barrier to moderated groups.
I differ, as I think speed of reply is important. If the moderator is several
hops away (and some net hops insert delays of days) then it could be a week
before my submitted article makes it back to my half of the net.
Not at all. Most moderators would be located at fairly central
locations, and initial responses would be distributed
immediately. Latency would be minimal. It's only all of those
"late" responses that would take awhile to filter through the
pipe, and in most cases such "duplicates" would not normally
be distributed in any case (as I've discussed previously).
Some newsgroups will not fall into the "moderated" catagories
easily, and they would continue to be direct. Once again,
all of this can be worked out without Usenet, Inc.
This is the situation we had before we merged the arpanet digests with net
groups at sri.
Something fairly central will reduce this a great deal. You could
even have alternate routes for messages marked time critical. (You could
do this now, but the massive number of sites involved would make this
a very nasty task)
See above. Usenet, Inc. is not needed for such technical
solutions to what I have termed "accidental" problems.
2) The sender paying for mail. This is the way just about everything else
works in the communications industry. As long as you allow people to accept
collect mail, this should not cause too many problems. It does solve the
junk mail problem. Done right, the cost for the average electronic mail
message should be less than what the post office charges.
Well, Whoopdeedoo. Wow, I guess that since that's the way
"everyone" else does it (not even true) we just gotta follow right
along at the end of our leashes, eh? This "collect" mail
stuff is just another headache waiting in the wings that we
can live without. A person sending out a message would have no
way to know a priori how many messages he or she would be paying
for and how many would be accepted collect. Even most of the com-
mercial information utilities which are currently under development
will not attempt to explicitly charge users for submitting such
informational messages.
It is generally understood that charging the sender does *not*
encourage fair or useful information flow. It sure hasn't
stopped junk mail in the "real" world. Give it up.
3) The sender paying for news rejected by a moderator. Again this is
an analog of the real world. If the editors of a magazine refuse your
story (ie. you can't get the recipients to pay) you can publish it yourself.
A fair bit cheaper on the usenet. I estimate with current net size we are
looking at about $4 to post a 1K byte article.
Postings where people feel it is obvious the poster should pay would also
be charged in this way. I refer to commercial product announcements,
job ads, that sort of thing.
This is getting repetitious. The people who read product
announcements and such often benefit from such reading.
Why should the poster be the only one who pays? And, now you've
put the moderator in the position of deciding who gets charged
and who doesn't. This would seem to mean that "the rich"
could post all the crap they wanted, while everyone
else would have to sit around counting their Usenet pennies.
What rot.
4) Active links to other nets and machines. Usenet inc. would go after
connections to "The Source", "Compuserve", perhaps an official CSnet or
ARPA (unlikely) connection. Also to the millions of microcomputer users
out there.
Oh boy. What fun. Heaven and Hell protect us.
I can't wait for those lightbulb jokes! Let's find a way
to solve the minor newsgroup problems we have now with a network of
this size before we start opening up the floodgates to the
unwashed masses. Even my IBM-PC UUCP will be under
controlled distribution via people with connections to
existing Usenet sites until I'm sure that the networks
can handle the load without serious problems. I don't want
to see the whole ball of wax pack it in!
5) Links to the postal services. Currently the post office will accept
electronic mail, print it in a remote city and deliver it for you.
Usenet inc. could provide this link so that you could send postal mail
from your terminal. This would be a big boon in border crossing mail,
which is usally verrrrrry slllloooooowwwww, not to mention expensive.
Compared to the one-day services, there would be no beating electronic mail.
The minimum quantity restrictions on this service are about to
be lifted. This means that any individual site (or even user
with a modem) will shortly be able to use this USPS service
economically, if they care to. We won't need Usenet, Inc. for that.
6) Unix consulting. In a central position, this company could offer
unix consulting of all sorts, from expert consulting to answering user
questions via fast net mail. A site without a guru could consider a direct
usenet inc. connection as a substitute for one at a fraction of the cost.
If you want to provide such a service, then go ahead. But such
services have been tried in the past, with only fair to poor
results. When people are paying real money for consulting,
they usually want the consultant there or on the phone in
a more interactive environment. The level of "consulting"
that can be reasonably accomplished over networks is about what
we see now in terms of UNIX-WIZARDS and direct mail.
I hope this clears things up. If this company gets of the ground without
the backing of some giant like AT&T, it is clear that it will only implement
what the most customers want. That's business, folks.
Yeah, right. Surely you jest? Once people are dependent on
the company for communications, the company will proceed as
*it* thinks best -- and people who aren't happy will be
invited to leave.
And now to some of Lauren's points:
1) By "Forbid" in the bill of rights note, I meant that the COMPANY would
not be allowed to do things (like certain forms of censorship, for example),
not that the customers would be barred.
Well, by the time we all finished our list of "forbids" I'll
bet you that there isn't damn much left for the company to
do other than collect the money!
2) You suggest big sites would not connect to further sites because they
have no charity? How does the net run now, then?
Must we go over this again? The network runs now as a
mass of independent communications whose costs are generally
being absorbed into other communications-related costs
of running a computer site. Once you've split off the costs
of Usenet and subject them to careful scrutiny, it is unlikely
that many sites would be willing to continue much of the "charity"
we now see. They would be extremely sensitive to the "value"
of network services in relation to costs, and as such only
high-value materials would continue to be distributed or submitted.
--Lauren--
Brad.
-------
Lauren here. On further consideration, it is becoming pretty clear
what's going on. Brad wants to start an information utility.
There's nothing wrong with that; I myself am involved with a company
that is investigating the same sort of possibility for largescale
implementation. However, Brad wants to take a "short-cut" and implement
his network upon the existing backbone of Usenet sites. That's the
problem. Don't try to push the existing network of Usenet sites, most
of whom are pretty happy with the way things are working (and would
be even happier with some "surigical" modifications to newsgroup
organization) into the commercial framework you'd like to see.
If you want to start a utility, do it the way the rest of us do --
start from scratch. Stop trying to "cash-in" on the nice, cooperative
systems we currently have running.
I have nothing against information utilities, but I want to see Usenet
continue as a friendly non-commercial cooperative as well. Let's try not
to muddle the two in the rush toward the bucks. Trying to build the
former out of the latter is simply wrong.
Unless strongly prodded, this will probably be my last statement on this
subject for awhile. I'm really getting tired of going through this
stuff again and again. So, if you see some messages crossing the net
about Usenet, Inc. and wonder why I haven't responded... it isn't
from any change of heart -- it's simply an attempt to avoid arthritis
of the fingers. Cheers.
--Lauren--smb@ulysses.UUCP (08/01/83)
I agree with Lauren. --Steve
bch@unc.UUCP (08/02/83)
I (as usual) agree with both Steve and Lauren.