comp.sys.3b1? (was Re: Monthly posting about Unix-PC network)

[John R Ruckstuhl Jr]

In article <1990Nov3.045636.12735 at cs.ucla.edu>, kirkaas at makaha.cs.ucla.edu (paul kirkaas) writes:
> Why not just move the whole unix-pc tree under comp.sys.unix-pc, which
> is clearly where it belonged in the first place?

Surely you'll get many replies, saying essentially: "unix-pc has a
separate hierarchy because it evolved separately, ie is separate for
historical reasons", and "it hasn't been integrated with the 'comp'
hierarchy because there is no true benefit to doing so".  Some will
argue that "no one can have real troubles getting unix-pc if they make
just a small effort".  Likely, several people will generously offer a
feed of the unix-pc newsgroups, or suggest ways to gain access - perhaps
via the digest that Scott Hazen Mueller kindly mails.

I now offer a mild rebuttal in the form of a counterexample from my
personal experiences.

At my previous residence in Colorado Springs, my 3b1, sphere, began
exchanging news with hp-lsd.  I was (am) a novice at news
administration, and it took one month for me to realize that although
hp-lsd claimed to be forwarding me unix-pc, I was seeing only parts of
some discussions in that newsgroup.  Soon I discovered an explanation -
I was seeing only those unix-pc articles which were cross-posted to
comp.sys.att (a newsgroup in the "standard" distribution).  I studied
the paths of articles.  The path was usually pretty consistant for about 
6 hops, from hp-lsd to hp-sdd (California), because most sites in HP ran
"notes" instead of some form of "news", and didn't have the ability to
exchange articles by NNTP (so HP sites exchanged with each other rather
than exchange with other well-connected sites).  hp-sdd was the
demarcation - they exchanged news with the "outside" world.  So any of 6
upstream sites could be the bottleneck.  I wrote system administrators
upstream and asked politely about unix-pc  (notes doesn't support the
"sendsys" control message, so I couldn't get this information 
automatically).  Some never responded.  Some responded that they didn't 
carry unix-pc, but would start (thanks hp-sdd, hpcc01, hp-col).  Mike 
Rodriquez at hplabs said, essentially, "I *won't* carry it because I 
don't see that it is beneficial to HP".  I wrote back asking him to 
reconsider, hoping that he wouldn't apply a criterion to unix-pc that he 
wouldn't apply to, say, rec.arts.tv.soaps.  This logic, and a "spirit of 
USENET" appeal failed to persuade him.

Concurrently with investigating the HP path, I looked at alternative
sources in Colorado Springs.  DEC had an entry in the maps, so I called
the system administrator listed.  He'd relocated, and his replacement
was out of town for 2 weeks.  I was unable to find his alternate.  I
reached him a few weeks later, and he did indeed receive unix-pc from
some DEC-hub somewhere, but he was uncomfortable UUCP'ing outside DEC
for "security" reasons, and I didn't feel I had the right to argue with
him, since supporting news exchange is a courtesy rather than an
obligation (in my opinion).
The local branch of University of Colorado was well-fed from UC Boulder,
but according to the persons I spoke with, they couldn't provide
telephone access because they didn't have the proper hardware.

All other sites in Colorado Springs that I knew about (from the
regularly posted UUCP maps) were fed solely by hp-lsd.

I then made contact with a system administrator at Cray (thanks, David
Keaton, Peter Hill) who had an !independent! feed via NNTP with
University of Denver who exchanged news with UC Boulder, who exchanged
news with ncar who I !knew! was well-connected.  They generously allowed
me to poll them for unix-pc.  I was elated!  I saw unix-pc articles
which hadn't been cross-posted to comp.sys.att, so I believed I finally
had a healthy feed.

But two weeks later, I became suspicious again when someone referred to
a posting in unix-pc.sources that I never saw.  I did a "sendsys".
Everyone in ncar!boulder!upikes!aspen!sphere looked fine!  So I wrote
these system administrators, and determined there was a bottleneck at
upikes.  Aha!  A week later, after a few unreturned letters and calls, 
I spoke with Randy Hagan at upikes who was quite pleasant, and explained
that awhile ago, he'd removed unix-pc from the active file (Ah, I wasn't
a sophisticated enough news administrator to think of that possibility).
But how did I get the occasional articles which had been posted only to
unix-pc, I wondered.  He'd accidently left unix-pc.bugs active, so I was
just seeing articles which had been (cross-)posted to that newsgroup.
He soon reactivated all of the unix-pc newsgroups, and !finally! I had a
healthy newsfeed.

The interval between original attempt to subscribe to unix-pc and 
acquiring a healthy newsfeed was 6 months (obviously I didn't work on
this full time, but I think I did make considerable effort).

I think this is perhaps the most long-winded article I've ever posted.
Readers, please accept my apologies.

1.  The bottlenecks such as at hplabs and upikes would be much less 
    likely if unix-pc newsgroups were part of the "standard" 
    distribution rather than an "alternative" hierarchy.  I believe
    moving into "comp" would increase distribution of the newsgroups.
    I think this alone justifies such a move.

2.  I will gladly concede that it is easy for anyone to gain access to
    unix-pc if one is willing to pay, but I assert that news exchange
    via local telephone call to a friendly neighbor is preferred to a 
    toll call or paid subscription to a service.  Some very, very,
    generous individual in ?RI? once offered to call 3b1's that couldn't
    get a local feed (did I really read that right?), but some of us
    prefer not to receive calls because we share one telephone line for
    data and voice.

My experiences have made me more knowledgeable, but I wish them on no one.
I hope that my letter convinces others that it is possible to have
difficulties acquiring a free feed for unix-pc, and that in the interest
of greater distribution of the newsgroup, they will vote for a move if
and when such a voting opportunity occurs (even if they themselves do
not anticipate access difficulties).

> This would also allow us to separate out discussions of 3b1's and 3b2's
> and 6386's, which have nothing really to do with each other.

The mechanism to separate those discussions do with other machines 
already exists, viz unix-pc, u3b, and comp.sys.ibm.pc.

Best Regards,
John.
-- 
John R Ruckstuhl, Jr	ruck%sphere at cis.ufl.edu, sphere!ruck
University of Florida 	ruck at cis.ufl.edu, uflorida!ruck