25 Jan
2014
25 Jan
'14
9:57 a.m.
On 25/01/2014 01:53, Steve Lafferty wrote:
I appreciate all the hearty discussion in response to
the article:
http://www.tronola.com/html/who_built_the_first_minicomput.html
There are some points that I would like to respond to:
> It perpetuates the myth that there was such a thing as a "minicomputer"
before the late 60's. There have been several threads now on the origin of the term...
In many ways I think the reverse is true. Apart from being being
physicality larger than a later "Mini Computer" many early computers
are, in terms of capability and architecture, essentially "Mini
Computers". All that happened was that some got smaller and cheaper and
at some point the term "minicomputer" appeared....
All we are really arguing about was when they became small and cheap
enough to be called "mini-". To me its "mainframes" that are the
divergent branch in computing and to me what is interesting, I guess in
the same way in which other look for the "missing link" between apes and
humans, is how and when they diverged far enough to be a "Mainframe"
But to go back to early computers. Take the Ferranti Pegasus from 1953
for example. Taking up just three large cabinets for the computer the
Pegasus is physically similar in size to a large PDP-11, but being
Valve, serial and with 333Khz clock is probably slower than a
Straight-8. I don't have my Pegasus history book to hand, but it was
certainly also marketed almost as a Mini being touted as low cost and
easy to use. It was built on a production line, with I think around 150
being built. The example at MOSi was used in a design department at a
ship builders to do engineering calculations, something I think early
DEC's were aimed at.
In construction its has much in common with an early 8, except the 500
or so boards in the Pegasus are larger and only contain two gates, or a
single "49-bit" register, but in essence and exchangeable standard card.
Reasonable image of a two gate board here:-
https://www.cs.auckland.ac.nz/historydisplays/FourthFloor/Packaging/Packagi…
https://www.cs.auckland.ac.nz/historydisplays/FourthFloor/Packaging/Packagi…
And this lovely page from a Ferranti Brochure
http://www.cbi.umn.edu/hostedpublications/Tomash/Images%20web%20site/Image%…
shows how it all fitted together. Sadly the Teleprinter isn't directly
connected but only prints the 5-hole tape that comes from the high speed
punch. So print spooling, but with real spools of paper tape, so I guess
I have made it into a mainframe....
Any way I wouldn't call it a minicomputer as its valved but if it were
TTL I think it would qualify. It certainly shows most of the other
important attributes of a Mini as early as 1953.
I guess there are a couple of points I really want to state...
1. What is a "minicomputer" is a rigidly defined area of doubt and
uncertainty...
2. Considering the taxonomy of one computer type in isolation isn't
really a good idea. The picture is complex....
Dave
G4UGM
--- In no way did I mean to suggest that the TERM
minicomputer was used contemporaneously with the machines which I found to be classifiable
as minis. Rather, I am trying to clarify what that term has come to mean. From there, I
proceeded to look for the first machines that fit that definition. Now, I certainly
don't claim that the definition I propose is the very best one and would appreciate
helpful comments on improving it.
There have been numerous attempts to redefine the
boundaries based on [various criteria] ALL such redefinitions, that ignore the fundamental
nature of being MARKETING terminology...
---Marketing people might have originated
the term minicomputer but like all words, it is defined by the people who use it. We do in
fact use it and others seem to know what we mean. My task was to clarify what people do
mean by it.
Likewise, the use of the word "first"
is fraught with danger without completely ARBITRARY further stipulations...But, the
boundaries are undefined, or rather, everybody has a unique definition of the boundaries
of their own.
--- I totally agree that "first" would be meaningless
without first defining what you are saying is first. Rather than using arbitrary criteria
with which to define mini, I tried to find ones which could be justified and that many
could agree upon. For each factor, I looked for good common-sense reasons that it needed
to be so and stated those. If I am off base on some, I would appreciate feedback on what
you think would be better.
Are we talking about working prototype?
announcement? orders taken? first delivery? full retail availability?
--- It was
stated that first production ship date would determine priority. Please see the article
for the justifications for that. I think you will find that this is what makes sense.
Should limited production machines be excluded?
--- The requirement was simply that it must have been manufactured in quantity.
Fortunately, the exact quantities shipped did not become an issue for minis. The lowest
quantity machine was probably the LINC, of which some 50 or so were made.
Thanks,
Steve L.
http://www.tronola.com/