Fidonet History

Historical Text by Ben Baker

Date: 02 May 87 08:04:00
From: Ben Baker
To:   David Garrett
 

You and Hartman asked for it, so here (as best I can remember) is a capsule
history of Fidonet.

TJ once told me he began work on Fido around Christmas, '83. He wasn't sure
exactly when.

In early '84, I was preparing to put a CBBS on line for the CP/M Sig of our
computer club at MAC when I was approached by the club Pres to make it a
club-wide BBS.  The club had a commitment from DEC for an indefinite load
of a Rainbow.  The theory was that the Rainbow had a Z-80 and could run
CP/M, so should be able to run CBBS.  When I finally got the manuals for
the Rainbow, I discovered that the Rainbow's Z-80 did not have access to
the I/O ports, so could NOT run CBBS!  I immediately began a frantic search
for BBS S/W which would run on the Rainbow, which led me to John Madill's
board in Baltimore.  There I engaged in a message exchange with Tom
Jennings, who was frantically searching for someone to write a comm driver
for FIDO_DEC.

No, I didn't do the comm driver (don't remember who did), but I did get a
copy of the first Rainbow version (which was, I think, originally intended
for John Madill's home machine -- don't know if he ever put it up).  By mid
March I had it running on the club's Rainbow 100.

I don't know when TJ began numbering Fido installations, but at that time
there were at least 6, but no more than 8.  He would not assign me a number
until he could "list" me in his informal Fido list, and I did not get a
phone line assigned to the system until sometime in April.  When I could
finally give him a publishable phone number, I was listed as "Fido 10," the
second St. Louis Fido (Tony Clark was Fido 4).  I began with a late Version
3, but by the time I was listed, I think I was running Version 5.

In May, Fido began to blossom, and by Memorial Day there were around 15
Fidos on line.  St. Louis had 5 of them -- 4, 10, 16, 17 and 22.  (TJ had
begun assignning Fido numbers when he mailed out diskettes -- many of them
never did come on line). Curiously, all but Tony Clark were running Fido on
Rainbows!

Sometime in late May or early June I was talking on the phone with TJ and
the subject of networking the BBSs together came up.  "Wouldn't it be neat
if one Fido could automatically call another and send it messages and files
-- automatic software updates!"  That night TJ logged into Fido 10 and
uploaded FIDO_DEC V6, a brand new program called FIDONET, and a new system file called "NODELIST.BBS."  With that, FidoNet was born.

Version 6 implemented a very primative amorphous network with just one
hard-wired schedule. Traffic level grew rapidly with everyone experimenting
with this new toy, and it soon became apparent that most of the time we
were butting heads, and many messages never went through.  We needed more
elaborate scheduling and some means of defining message routing.  But how
do you develop and do controlled testing on something like that without
spending a fortune on phone bills?  St. Louis which by this time had added
a 6th Fido (51), could model a real network with local phone calls!  No
other city could boast more than 2 Fidos.  That is how I became involved in
difining Fido's routing language with TJ.  TJ wrote -- we tested -- we fed
back results and needs -- TJ wrote, sometimes two releases in a day!  By
August we had version 7, with its routing language, ready for distribution,
and FidoNet began to change shape toward what we see today.

TJ was maintaining the nodelist.  When he received a change request, he
would write it down on a small slip of paper and stick it to the wall.
Frequently the slip would fall from the wall and disappear behind his
computer, never to be seen again.  By September the nodelist was a
shambles!

I'm not sure if we volunteered or WERE volunteered, but Ken and I agreed to
take over nodelist maintenance, and on September 21, 1984, we (well --
mostly Ken) published the first "St. Louis Nodelist."  It took us a couple
of weeks to weed out all the bad numbers and drop-outs, but by the middle
of October we had a pretty solid nodelist.  TJ had been bit once or twice
with fake node number requests.  (I'm sure many of you have heard a version
of the famous "little old lady." It actually happened.  He accepted a phone
request for a node number.  After several complaints from the net about
no-answer, he called the number during the day -- and got an earfull!)  So
we established our first FidoNet policy -- ya gotta request a node number
via net mail.  Of course, TJ was still passing out node numbers with
diskettes, and we still had a few bad ones.  It took another month to
persuade him to stop, and to publish "our policy" in the docs.

It was October or November that TJ published the first issue of our
irregular weekly newsletter, FidoNews.  I don't think he had ever intended
to continue with the newsletter very long, and in January he passed that
baton to Thom.    I remember at the time I had never heard of Thom Henderson!
Who the hell is he?  Ken didn't know either.  Hey Ken did you ever figure
out who this Thom Henderson is?  What kinda name it "Thom" anyway?

I think we were in Fido Version 8 when, in the Spring of 85, we were
rapidly approaching Fido's 250 node limit.  And a nodelist that size was
becomming difficult for one man to manage and still find time to kiss his
wife occasionally.  Our computer club and the local DECUS chapter brought
TJ to St. Louis to speak at a joint meeting on April 10th amd the next day
we had an all-day meeting at Ken's house.
After an 11 hour session we codified what was already taking place.  With
the advent of a routing language,  FidoNet was collecting itself into local
groups or "nets," usually around a node willing to foot the bill for
long-distance calls.  So we created the net/node addressing scheme.  Node
numbers within a net would no longer have to be unique on FidoNet, only
within the local net.  Thus the "network host" could maintain his own net
list.  But that still left about 100 or so nodes scattered throughout the
hinterlands and not alligned with anybody.  The net implied routing.  How
about a different kind of "net" that did NOT imply routing -- a Region.  TJ
reached into his knapsack (hey, that's the way he travels, knapsack and
scateboard) and pulled out two or three hugh U.S. maps.  We spread one out
on the floor and with a felt pen, began carving.  We divided up the country
into ten pieces we hoped represented nore-or-less equal populations (at
10pm on a Thursday night we were not in a scientific mood) and dreamed up
names for the ten new "regions."

TJ went home and got back in the "a version a day" mode.  Ken and I put a
freeze on the nodelist and began creating net and region files, and
assigning new net addresses.  By early May, the software was beginning to
stablize and we "went public." As I recall, we set June 15 as the cut-over
date to the new addressing scheme (with a silent prayer that we could get
everything in place by then).  We found ten people willing to be regional
coordinators.  We unfroze the nodelist and gave hosts a formula for
assigning node numbers (until the cut-over, they still had to be unique).

Finally the the fateful day came for us to all use the '3' command and set
our new net addresses.  I was expecting total chaos.  I was not at all
prepared for just how smooth the transition happened!  Oh sure, there were
a few stragglers and even a few drop-outs, but still, one day we were an
amorphous network and the next FidoNet was partitioned into local nets and
regions -- and the mail kept flowing as if nothing had happened!  It took a
good deal of coordinated effort by a great many people -- and it proved we
COULD function as a body!

It was about that time that TJ first suggested a membership association.
After all, we had proved we were an organization, so why weren't we an
officially sanctioned organization.  I was originally cool to the idea.
Providing tee shirts and bumper stickers was not the kind of service
foundation I thought appropriate, so I dragged my heels.  With the
Tsimpidis affair still fresh in my mind, I saw a need for a strong
collective voice, but I didn't have any idea how to get there.

I'm sure there were events of momment, but I don't recall much more as 85
slid quietly into 86.  A 500 node limit came and went, almost without
notice.  TJ said "This new version (11) can handle 1200 nodes.  That ought
to hold us for quite a while."  We coined the name "International FidoNet
Association" and used it a first line in the mailing address. FidoNet began
appearing more frequently in national publications -- like it or not, we
were a growing force and we were being noticed.  Ken began receiving
donations in the name of IFNA, and they helped defray the costs of our
new-found recognition.

Two things happened in 86 to crystalize the IFNA concept and one to
definitely polarize it.
First, an April conversation between Ken and his accountant went something
like this:  "You've got to pay income tax on these 'donations.'"  "But
that's not my money!"  "I know, but what IS IFNA?  Can you prove to the IRS
that it exists?" "Well. . .er. . .uh. . ."  Total receipts for 85 were only
a few hundred dollars, but still, that's a non-trivial tax burden and 86
revenues had already exceeded 85's.
Then in May we were asked by COSUG, "How would you like to help us put on a
Sysops' Conference?"  Sounded like a good idea to us and we immediately
went to work on it.  Then in July they said "Looks like we might have a
small surplus.  We will gladly share it with IFNA, but we can only do that
if IFNA is a bona fide not-for-profit corporation.

So, with some trepidation, Ken filed IFNA's incorporation papers in late
July or early August.  On reflection, we should have said "Keep the money
-- let's see what happens at the conference first."  Marvelous thing --
hindsight.

Then came the conference in August.  From that momment to this our history
becomes a blur to me.  I recall that a self-appointed IFNA spokesman put us
in deeper, hotter water every time he opened his mouth.  I recall that,
with no authorization save the aforementioned spokesman's, IFNA memberships
went on sale.  I recall a disasterous "business meeting."  I recall Ezra
putting out the fire under the tar pot.  I recall a by-laws committee, a
New Hampshire meeting, a Chicago meeting, flames, counterflames.  I recall
twice throwing in the towel and twice being persuaded to reconsider my
action.

But can I put order to all of that?  'Fraid not -- it's all a blur.
Another historian will have to pick up from the conference; one with
clearer recollections (or perhaps records).

Has it all been worth it?  For me, the first two years were an unqualified
success.  As to the last year, only time will tell.  I think we now have
the skeleton of a potentially successful -- and useful -- organization.
Now, lets get some meat on the bones.

                            Ben

I have told this abreviated history from my own perspective. I have left
out many people and events really important to the development of FidoNet.
The list is long and I will not attempt to enumerate them for fear of
ommitions.  You know who you are.  Most of you know who 'they' are.  I
would simply say to all of you -- THANK YOU.
 

Back Back to FidoNet History