bts@evergreen.cs.unc.edu (Bruce Smith) (08/07/89)
When people talk about the history of computing or the history of computer science, the discussion is usually about hardware. Some AI folks talk about philosophical traditions and, some NA folks talk about mathematics before computers. What about the history (more generally) of software before computers? For instance, how did people produce tables of functions? I'm not asking whether they used Taylor series, but rather how did they manage the computations. Did someone shut a mathematician in a closet and not let him out 'til it was finished? Or, did they hire an army of clerks and give each instructions on what numbers to add, what numbers to multiply and to whom to pass on their portion of the answer? Other examples might include "ancient" business data processing, etc. Algorithms for simple arithmetic don't count (I think), as once learned most people forget they're executing a program and develop personal variations on the general methods. I want to focus on people "mechanically" executing non-trivial programs, before computers. I'd appreciate any (fairly simple) references on this subject. Maybe about the level of Scientific American's history articles. This is posted to a couple of education newsgroups, because one reason for asking is to find material suitable for introductory CS courses. It's posted to sci.math, because this is probably just as much a question in the history of (applied) mathematics.
hubey@pilot.njin.net (Hubey) (08/07/89)
In article <9086@thorin.cs.unc.edu> bts@evergreen.cs.unc.edu (Bruce Smith) writes: > I'd appreciate any (fairly simple) references on this subject. > Maybe about the level of Scientific American's history articles. > This is posted to a couple of education newsgroups, because one > reason for asking is to find material suitable for introductory > CS courses. It's posted to sci.math, because this is probably > just as much a question in the history of (applied) mathematics. I can't supply references--not exactly anyway--but I remember reading years ago that the Romans had worked out different algorithms for multiplying different numbers. They had special matrix-like algorithms. i.e. recall they did not use Hindu-Arabic numerals. The best I can do for reference is that I think the book had something to do with the 'Great Books of the Western World' series. regards mark -- hubey@OSultrix.montclair.edu hubey@pilot.njin.net hubey@apollo.montclair.edu VOICE: 201-893-5269 ...!rutgers!njin!hubey
dmcallis@bambam.UUCP (David McAllister) (08/08/89)
From article <9086@thorin.cs.unc.edu>, by bts@evergreen.cs.unc.edu (Bruce Smith): > When people talk about the history of computing or the history > of computer science, the discussion is usually about hardware. > Some AI folks talk about philosophical traditions and, some NA > folks talk about mathematics before computers. What about the > history (more generally) of software before computers? I read a Scientific American article recently from a few years back (July 1987?) about the history of the U.S. Census Bureau. I think it's exactly the stuff you're looking for. It had information on their hardware, things like mechanical card counters, etc., and it also described algorithms that people used to get whatever results they needed. A perfect example of the kind of thing you're interested in is the radix sort, a software algorithm modeled after a physical process. To sort a stack of cards, often computer punched cards, there were slots and holes punched in the top of the cards that rods could be slid through, pulling up cards with holes and leaving those with slots. Then the cards on the rod were placed in a group behind the other ones and the rod was slid into the next hole. This process continued through all the holes, sorting all the cards according to the binary number in the slots and holes in the card. This same process is used for the nlog(n) computer radix sort. -- David McAllister, SLC, Utah | "Slow neutrons at play" Evans & Sutherland Computer Corp. | bambam!esunix!dmcallis@cs.utah.edu |
cik@l.cc.purdue.edu (Herman Rubin) (08/08/89)
In article <9086@thorin.cs.unc.edu>, bts@evergreen.cs.unc.edu (Bruce Smith) writes: > When people talk about the history of computing or the history > of computer science, the discussion is usually about hardware. > Some AI folks talk about philosophical traditions and, some NA > folks talk about mathematics before computers. What about the > history (more generally) of software before computers? > > For instance, how did people produce tables of functions? I'm > not asking whether they used Taylor series, but rather how did > they manage the computations. Did someone shut a mathematician > in a closet and not let him out 'til it was finished? Or, did > they hire an army of clerks and give each instructions on what > numbers to add, what numbers to multiply and to whom to pass on > their portion of the answer? Essentially the algorithm was designed and then either done serially or in parallel. Little was passed on from one person to another, other than tables of coefficients, etc. We know that at least one of the early tables of common logarithms was done by computing 10^(p/q), q a power of 2, to sufficient accuracy, and then using interpolation. The Ptolemaic school, around 2000 years ago, computed tables of the trigonometric functions by successive bisection and the addition formulas. If the quadrant is divided into 96 parts, no interpolation is needed. Other tables were computed by power series, etc. Do you really think that computational procedures started with computers? There are very few computational methods which arose AC, and most of them would have been appreciated BC, where C stands for computer. -- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907 Phone: (317)494-6054 hrubin@l.cc.purdue.edu (Internet, bitnet, UUCP)
lvc@cbnews.ATT.COM (Lawrence V. Cipriani) (08/08/89)
>In article <9086@thorin.cs.unc.edu>, bts@evergreen.cs.unc.edu (Bruce Smith) writes: > For instance, how did people produce tables of functions? I'm > not asking whether they used Taylor series, but rather how did > they manage the computations. Did someone shut a mathematician > in a closet and not let him out 'til it was finished? ... Well, actually yes! At least one mathematician, it was either Euler or Gauss, paid an autistic savant to compute tables of logarithms. It benefited both the mathematican and the savant. The mathematician had better things to do than compute tables of logarithms, and the savant, well, he couldn't do much of anything else. I read the tables were used for years, and had several decimal places of accuracy, and only minor and rare errors were ever found in them. Also there were programmable weaving looms years before Countess Ada Lovelace and the Analytic Engine. The programs were encoded by holes on cards. I think the looms were called Jaquard looms. This is an example of software before computers (feminists in the audience should note for future reference, the first programmer was not a woman! :-) -- Larry Cipriani, att!cbnews!lvc or lvc@cbnews.att.com
aed@j.cc.purdue.edu (Dan Hartley) (08/08/89)
In article <9086@thorin.cs.unc.edu> bts@evergreen.cs.unc.edu (Bruce Smith) writes: > When people talk about the history of computing or the history > of computer science, the discussion is usually about hardware.... > > For instance, how did people produce tables of functions? I'm > not asking whether they used Taylor series, but rather how did > they manage the computations. Did someone shut a mathematician > in a closet and not let him out 'til it was finished? Or, did > they hire an army of clerks and give each instructions on what > numbers to add, what numbers to multiply and to whom to pass on > their portion of the answer? ... Following the adoption of the metric system in France (late 1700s), a fellow by the name of de Prony tackled the immense job of recalculating the trig tables to reflect the 100-degree quadrant. Prevailing on the mathematicians of the time, he came up with a (labor-intensive) procedure for constructing the new tables. How did he do it? Well, it seems that the powdered wig had fallen into disfavor, leaving legions of unemployed Paris hairdressers... Anyway, his method relied partly on some number of independent computations to assure accuracy. He managed this by having the calculations repeated by isolated teams around the country, since the art of copying from your neighbor's paper was already well developed by that time :-). In keeping with the tradition of fickle government funding that continues to this day, no monies were appropriated for the publishing of whatever tables were produced. -- Daniel E. Hartley Manager of Operations and Facilities ARPANET: aed@j.cc.purdue.edu Purdue University Computing Center BITNET: HARTLEY@PURCCVM Mathematical Sciences Building, room 228 Phone: (317) 494-1787 ext. 208 West Lafayette, IN 47907
lsheldon@cup.portal.com (Laurence Larry Sheldon) (08/09/89)
Somewhere between the Job and Jobs there was some fascinating computing going on that hardly ever gets talked about (or written about) except when I talk or write--analog 'puting. The first 'puters I worked on were mechanical analogs, where all the math functions were in gears, cams, levers, and all sorts of wonderful stuff. Then there were the electro-mechanical and electronic analogs--when machines really started to be quick. That was when machines were _MACINES_ . Anybody remeber what an amplidyne is?
bjornl@tds.kth.se (Bj|rn Lisper) (08/09/89)
In article <9086@thorin.cs.unc.edu> bts@evergreen.cs.unc.edu (Bruce Smith) writes: %For instance, how did people produce tables of functions? I'm %not asking whether they used Taylor series, but rather how did %they manage the computations. Did someone shut a mathematician %in a closet and not let him out 'til it was finished? Or, did %they hire an army of clerks and give each instructions on what %numbers to add, what numbers to multiply and to whom to pass on %their portion of the answer? I've heard that Napier spent 40 years of his life calculating logarithm tables. (I think he was a 17th century Scotch matematician.) Bjorn Lisper
lvc@cbnews.ATT.COM (Lawrence V. Cipriani) (08/10/89)
Here are corrections to the following article I received via email. In article <8856@cbnews.ATT.COM>, lvc@cbnews.ATT.COM I wrote: > Well, actually yes! At least one mathematician, it was either Euler or Gauss, It was Gauss. > paid an autistic savant to compute tables of logarithms. The savant was a computing savant named Dase, not an autistic savant, and in fact was an engineer. Gauss was a computing savant as well. > ... > I think the looms were called Jaquard looms. Correct spelling is "Jacquard". >Larry Cipriani, att!cbnews!lvc or lvc@cbnews.att.com At least I got my signature right! -- Larry Cipriani, att!cbnews!lvc or lvc@cbnews.att.com
steve@hubcap.clemson.edu ("Steve" Stevenson) (08/10/89)
From article <BJORNL.89Aug9170721@tarpon.tds.kth.se>, by bjornl@tds.kth.se (Bj|rn Lisper):
> .... (I think he was a 17th century Scotch matematician.)
Scotch is a drink. :-)
Napier's Bones play a role in computing.
--
Steve (really "D. E.") Stevenson steve@hubcap.clemson.edu
Department of Computer Science, (803)656-5880.mabell
Clemson University, Clemson, SC 29634-1906