--On January 20, 2014 2:28:27 PM -0500 Paul Koning
<paulkoning at comcast.net> wrote:
On Jan 20, 2014, at 1:18 PM, William Donzelli
<wdonzelli at gmail.com>
wrote:
I suspect
that most people won't see much value in preserving bisync
knowledge and hardware, but in my biased experience, there was a lot
of it in the pre-TCP/IP world.
There is a tendency with many people (some professionals) to think
that computer networking is, was, and will always be the ARPAnet, the
Internet, and all things TCP/IP.
Of course, the ARPAnet wasn?t always TCP/IP. ARPAnet bits
predating that switch would be seriously interesting. Do IMPs still
exist? I think so. Software to run them? Software for your PDP-11
to talk to the IMP-11 Unibus to IMP interface?
Another predecessor of the Internet as we know it was the Merit
Computer Network in Michigan. There were probably other interesting
regional networks, but Merit built its own hardware and had its own
packet switching software and protocols before the Internet existed.
It also ran the NSFNet backbone for a time, of course, but that was
later. I expect that the hardware is all gone, but the software is
probably still around.
A predecessor of Merit at Michigan was the ConComp project which built
a machine known as the Data Concentrator. [1] This is also
interesting, partly because it was the first non-IBM product ever
connected to an IBM 360 channel. It didn't do networking in the
Internet sense, but provided terminal access to mainframe computers for
years. The software for this might still be around somewhere too,
papers about it certainly are.
Dave's history at the link below is slightly misleading. The Data
Concentrator actually predates the Model 67 at Michigan. It was first
attached to a Model 50 in 1966. I was present for the event (since I
wrote the supervisor it was talking to) and remember that it soon put
the Model 50's microcode into a loop which was an interesting
experience.
Mike
[1] <http://www.eecis.udel.edu/~mills/gallery/gallery7.html>