[Collins] Art Collins and Collins Radio

[Dr. Gerald N. Johnson]

While working on my MSEE at Iowa State after I took a leave of absence 
from Collins, I created a power network analysis program called PWRMAT 
that handled matrices with complex coefficients for IBM 360. I also 
created a version that ran on an IBM 1600 but it took lots of overlays 
and compiling the box of cards to an executable took all day they told 
me later. It was especially adept and transmission line analysis 
including coupling between adjacent lines and allowed for transpositions 
and their affect on coupling. It basically handled 10 x 10 matrices with 
complex coefficients along with some of the fancy matrix handling for 
the tranmssion line cases.

In the army, the NVL bought a small 360 as a vision analyzer but I got 
to use it until it was proven to work. PWRMAT took more memory than it 
had (128K for the program, 96K in the computer) and I had to teach the 
IBM field reps how to write Job Control Language to use it. After I got 
tossed off it, I got to send my programs to Johns Hopkins from Ft. 
Belvoir so I might have gotten two or three day turnaround.

I think Art's goal was not really to compete in general purpose 
computing but to do primarily airline reservations, a niche which he 
never personally used because he had use of the company airplanes for 
any traveling he wanted to do. The Collins Airline with scheduled 
flights between plants daily. And also served to display the latest in 
Collins avionics to potential customers.

On 7/5/2014 10:07 PM, David Thompson wrote:
> I remember seeing one of the Collins Computer Booklets while I was at
> LSU in about 1967 or 8. It was used as a recruiting booklet by Collins.
> At about this same time IBM replaced the old 7094's in the LSU computer
> lab with a IBM 360/65 which was Gene Amdahl's first in the 360 series of
> Enterprise Computers which ran from the 360 series through the later
> 370, and now 390 series.
>
> I went to work for Burroughs Corp which I found in a conflict between
> the old accounting machine type computers and the newer B series which
> competed head on with the IBM 360. We installed a B2500 at a big Wire
> and Cable Company south of Atlanta. The top of the B line was the B5000
> which was installed at Georgia Tech as a B5500. Then in 1971 Control
> Data Came along and installed the first of the Cyber Super Computer at
> Georgia Tech. I met my wife who worked for Control Data in Wind Tunnel
> testing at Lockheed Georgia .

After I got out of the army one of my grad research assistantship 
projects was trying to move network analysis (predecessors to SPICE) 
programs to the IBM 360. One I recall from a Burroughs computer was 
impossible to move easily even with Fortran source code. The programmers 
used a single matrix with connections in binary and integer and values 
in floating point because apparently Burroughs math processing read 
integer or floating point just based on content, and the 360 wanted them 
kept in separate matrices.
>
> The advances of these companies wiped out Collins attempt in the
> Computer Market. Only IBM is still a player in the market although
> Seymour Cray left Control Data to form his own Company
> Cray Research and is still selling his big vector machines to DOD, NASA
> and the like and his competition is several Japanese Companies.

I think Collin's lack of knowledge of computation wiped out that attempt 
more than the competition.

Today a room full of fast PCs compete pretty good with Cray's when the 
processing can be split into a few thousand parallel tasks. My programs 
don't split at all, I've been thinking sequential logic too long. But 
many of my programs are still running at weather.net supply animated 
radars on demand and the data for the html displays.
>
> Dave K4JRB
>
73, Jerry, K0CQ