Request: Books/articles about history of early computing (pre 1950)

[pier4r]

Only recently I realized that instead of reading just little parts about the work of charles babbage, IBM and its tabulating machines, arithmometers and others, I could read actual books that goes in detail over those things.

In particular I would be interested to know what people were able to squeeze out from early computing machines.

Another point of interest is, even if the computing machine was used for the four basic operations, what were the heavy computations done with them? (I'd think logarithmic tables, firing tables for artillery units, and projects like manhattan project or design of airplanes. Still would be interesting to have a look at the formula involved and the procedure organization)

Hence the question: does someone has any suggestion were I could look for books or articles? Even better: titles? In the meanwhile I try again with google and google.scholar.

[Paul Dale]

Start early: Difference engines. This is a good read, although it isn't strong on mechanical detail.

Charles Babbage wrote a book. I've not read it, so cannot comment. Tim Robinson has build reproductions of both difference engines and is in the process of building an analytic engine. His web site has details and photographs.

Konrad Zuse did some amazing things pre 1950 but I'm not aware of books about his efforts. The Pascaline also predates Babbage by quite a lot but, again, I'm not aware of books detailing this. There are plenty of details of both on the internet.


Pauli

[EdS2]

Babbage's book is a great read but tells you nothing about his computers. I think at that point in his life he was dealing with a bereavement.

Zuse wrote The Computer - My Life which isn't too expensive second hand. But for a history of computing, I'd recommend
Stan Augartner's Bit by Bit (lots of pictures, some mistakes)
Paul Cerruzi's History of Modern Computing (lots of words)

I much enjoyed Hollingdale and Tootil's Electronic Computers (1965) when I read it back in the mid-70s.

Also interesting for a wealth of technical detail
Bowden's Faster than Thought (1963) which is a collection of papers by pioneers including Turing.

[pier4r]

Thanks ! I am taking notes

Anything about squeezing math from tabulating machines? I somewhere read that people were quite happy to have those.

[Don Shepherd]

Here is a terrific book you might like:

https://amazon.com/Computing-Before-Comp...+computers

[Accutron]

Die Rechenmaschinen by Ernst Martin. The original German language version was published in 1925, and the Charles Babbage Institute produced an English translation in 1992.

http://decadecounter.com/vta/pdf/Ernst%2...CR%204.pdf

[SlideRule]

Before the Computer
[attachment=5198]
James W. Cordata

Copyright @ 1993 by Princeton University Press
All Rights Reserved
Library of Congress Cataloging-in-Publication Data
Cortada, James W.
Before the computer: IBM, NCR, Burroughs, and
Remington Rand and the industry they created,
1865-1956/ James W. Cortada.
p. cm. - (Princeton studies in business and technology)
Includes bibliographical references and index.
ISBN 0-691-04807-X
I. Office equipment and supplies industry-United States-
History 2. Electronic office machine industry-
United States-History. I. Title. II. Senes.
HD980l.U542C67 1993
338.4'768-dc20 92-25399 CIP

BEST!
SlideRule

[Paul Berger (Canada)]

(09-20-2017 09:54 AM)pier4r Wrote:  Thanks ! I am taking notes

Anything about squeezing math from tabulating machines? I somewhere read that people were quite happy to have those.

Richard Feynman wrote about using tabulating machines to do math for the Manhattan Project in one of his books, but not in a lot of detail. These machines where programmed by plug boards that needed to be replugged for nearly every operation, so he first worked out a process of having several machine in a line each set up for one step in the calculation sequence, and the output cards from one step could be then just carried to the next machine to process the next step. After a while they even figured out how to wire the machines together to eliminate having to handle cards.

Paul.

[Alberto Candel]

Here are some I know about:

The Computer from Pascal to von Neumann, by H. Goldstine

A History of Algorithm: From the Pebble to the Microchip, by J-L Chabert et al.

The theory of mathematical machines, by F. Murray (who was a collaborator of von Neumann, but on operator algebras)

and the you also have the work of Leonardo Torres y Quevedo and of Vannevar Bush, for example Origins of Digital Computers: Selected Papers, by Randell.

[SlideRule]

One more...
The Universal Computer
[attachment=5199]
the Road from Leibniz to Turing
Copyright © 2000 by Martin Davis
All rights reserved.
Printed in the United States of America.
First Edition
The universal computer : the road from Leibniz to Turing I by Martin Davis.
p. em.
Includes bibliographical references and index.
ISBN 0-393-04785-7

BEST!
SlideRule

[Karl-Ludwig Butte]

(09-20-2017 09:07 AM)EdS2 Wrote:  Babbage's book is a great read but tells you nothing about his computers. I think at that point in his life he was dealing with a bereavement.

Zuse wrote The Computer - My Life which isn't too expensive second hand. But for a history of computing, I'd recommend
Stan Augartner's Bit by Bit (lots of pictures, some mistakes)
Paul Cerruzi's History of Modern Computing (lots of words)

I much enjoyed Hollingdale and Tootil's Electronic Computers (1965) when I read it back in the mid-70s.

Also interesting for a wealth of technical detail
Bowden's Faster than Thought (1963) which is a collection of papers by pioneers including Turing.

There is a fantastic detailed technical description of Konrad Zuse's machines in this book. Unfortunately it is available in German only.

Best regards

Karl

[KeithB]

(09-20-2017 05:39 PM)Paul Berger (Canada) Wrote:  

(09-20-2017 09:54 AM)pier4r Wrote:  Thanks ! I am taking notes

Anything about squeezing math from tabulating machines? I somewhere read that people were quite happy to have those.

Richard Feynman wrote about using tabulating machines to do math for the Manhattan Project in one of his books, but not in a lot of detail. These machines where programmed by plug boards that needed to be replugged for nearly every operation, so he first worked out a process of having several machine in a line each set up for one step in the calculation sequence, and the output cards from one step could be then just carried to the next machine to process the next step. After a while they even figured out how to wire the machines together to eliminate having to handle cards.

Paul.

They also figured out a way to correct errors without stopping the whole process. They would just "inject" the correct cards in the middle of the process.

I only saw the movie, but I guess the "Hidden Figures" book might have a lot about early efforts with computing at NASA. I was hoping that the book she used to teach herself FORTRAN had been written by McCracken. No such luck, though. 8^)

[Don Shepherd]

Quote:I was hoping that the book she used to teach herself FORTRAN had been written by McCracken. No such luck, though. 8^)

I looked high and low for that book she used in the movie, without success. I think it must have been a movie prop. The earliest book I could find titled "Understanding FORTRAN" was by Mary McCammon in 1968, and its cover is nothing like the cover of the book in the movie.

[toml_12953]

(09-21-2017 07:20 PM)Don Shepherd Wrote:  

Quote:I was hoping that the book she used to teach herself FORTRAN had been written by McCracken. No such luck, though. 8^)

I looked high and low for that book she used in the movie, without success. I think it must have been a movie prop. The earliest book I could find titled "Understanding FORTRAN" was by Mary McCammon in 1968, and its cover is nothing like the cover of the book in the movie.

That might not mean much. Different editions and print runs of books will frequently have different cover styles. To coin an old adage: "You can't judge a book by its cover"

[pier4r]

Is "Nigel (UK)" by chance the author of this nice site?
http://www.vintagecalculators.com/html/m...ators.html

PS: once again, I do not know why but sites that are very simple but with a lot of interesting content are way more appealing than facebook-like sites with almost nothing inside. Long live to those 1995 looking sites that have a lot of info. (this is also a reason I like wikis)

[pier4r]

Recently I found this (mentioned by wikipedia).

Punched card methods in scientific computation

[pier4r]

Given my explorations so far, I learned the following (hopefully I will add link later)

- Computing did not "exploded" mostly during and after WW2 as I thought before (how much knowledge I missed and I still miss. Online museums are great); a great deal of tools where operational to make computations faster and less error prone already before.

A branch where computation was needed a lot, although not necessarily for complicated formulas but rather for the sheer amount of data to process - naive me, I never considered it - was accounting for banks, governments and bureocracy of various types.

So it seems that devices for mass computations were:
- mathematical tables (I still remember them shipped with books in 2000)
- slide rules (from ~1650 to: until pocket calculators were available en masse)
- tabulating machines (with punch cards) for massive basic operations.
- desktop mechanical calculators, mostly models with addition, subtraction, multiplication and division . I am not (yet) aware of mass produced models with more functions (aside from a square root function and a memory register in few models).
A couple of links:
http://www.vintagecalculators.com/html/m...hotos.html
http://www.oldcalculatormuseum.com/links.html
- actually electronic computers were ultra expensive, so very few institutions could buy them.
- electronical desktop calculators, available in the first half of 1960, were not better than the mechanical calculators. They had the four operations, they were expensive but the big plus - I think - is that they were silent (check the sound of a mechanical calculator multiplying on youtube). The speed was not that much faster in human terms.
- Thanks to a video of Gene (whoever cares for the channel "hpcalc.org" on youtube: thank you! Really neat idea) on youtube I got to know that the most of electronical calculator were simple (that is, not much more powerful than a mechanical version). https://youtu.be/g1FnEmeUHjs
- As I wrote, my current knowledge is surely partial but so far it seems that the electronic desktop calculators that were not expensive, were not better than the hp 35 (the first scientific handheld calculator). That still shocks me.
I was thinking that after some years (from 1960 to 1970, 10 years) the scientific functions were really common.
Once again a video of gene shows that the scientific functions were far from being a commodity, at least for commodore calculators: https://www.youtube.com/watch?v=t55vJDNbHDA

Furthermore I found a video on youtube (when you know more, please share!) in which the content delights me because it is not only advertisement, it actually contains a mini lesson in physics and the plotter is amazing!
https://youtu.be/JmTyrS-jfi0

[pier4r]

adding http://www.hpmuseum.org/forum/thread-9930.html

"Calculation and Computation in the Pre-electronic Era"

[pier4r]

And finally one of my questions gets answered, if not completely, something close enough.

The problems solved on the Eniac. Sure it is not representative of all the work done on pre 1950 computing machines, but it is something. Also some descriptions are interesting.

A Survey of Eniac Operations and Problems 1946-1952
https://apps.dtic.mil/dtic/tr/fulltext/u2/1003735.pdf
Impressive how the US military gives a lot of documents in the public domain. Every nation - in particular universities and schools - should do it, at least for the work of their students if not up to Phd theses.

In short it computed mostly firing tables, that when they try to be precise are quite complicated and time intensive for humans with mechanical desktop calculators (see http://ds-wordpress.haverford.edu/bitbyb...ng-tables/ ) .

It had an uptime around 75% , and they even compute problems to identify outliers in a distribution. Or runs on random numbers!
I believe on earlier machines they tried to solve similar problems although of smaller scale. Indeed in the bit by bit site it is mentioned that to compute one trajectory of a projectile from a gun an analog computer had to run for 30 minutes and each setup needed around two hours.

Today we get impatient after 2 minutes. I wonder the level of debug needed before trying to waste days of computations.

[EdS2]

(12-30-2018 08:02 PM)pier4r Wrote:  A Survey of Eniac Operations and Problems 1946-1952
https://apps.dtic.mil/dtic/tr/fulltext/u2/1003735.pdf

Nice find!

ENIAC took 70 hours over a holiday weekend in 1949 to calculate just over 2000 digits of pi - it only had 200 digits of memory, so the calculation repeatedly streamed inputs from and output to punched cards. Half the time was spent in checking each operation by calculating the inverse.

(ENIAC was originally a hard-wired machine, but by sacrificing a factor of six in the speed it had been modified to fetch instructions from ROMs intended for use as function lookup tables. Each digit of the ROM was a ten-position rotary knob. I don't know how big the pi program was. I think the information can be found, behind a paywall.)

Part of the motivation was to see how evenly spread the digits of e and pi are: e is unusually even whereas pi looks more like random digits. (As it turns out, the two calculations were very thorough proofs of reliability too.) Original paper can be seen, in part, at http://goo.gl/LuOUnb and some accessible information at http://www5.wittenberg.edu/news/2013/03_14_pi.html

By 1983 you could buy a PC/XT, and if you'd had the 125-byte program found here you could compute over 9000 digits at a rate of about 32 digits per minute.

If you have IEEE access you may be able to read a paper about the ENIAC calculation at
http://doi.ieeecomputersociety.org/10.1109/MAHC.2011.61

There's a popular writeup here too:
https://www.pocket-lint.com/laptops/news...niac-first