2012 — Write something funny...
... for the Christmas issue, will you? A simple enough request I received back in 1995 from the editor of the IBM Hursley Lab's in-house staff magazine. I rediscovered this text today while looking — obviously — for something completely different. Which I still haven't found. But for the record, as it were, here's what I wrote:
Well, I intended to, but I'm not sure it came out funny. Being a writer, I tend to read a lot. I also take an interest in the history and development of this curious industry of ours. Discounting the efforts of Hero of Alexandria (which I won't dwell on, mainly because I couldn't persuade the lad to show me how to use his Encarta CDROM) it seems one Wilhelm Schickard of Tübingen (just down the road from Böblingen) produced the first digital calculator. Sadly, it was made of all-too-combustible wood, and was burnt. By the 1660s, calculators (while not yet given away with 20 gallons of ozone-depleter) existed in some abundance, though they invariably had trouble with the carry system. (Some things never change.)
Fast forward (not that that was a term that existed in some abundance) to the 1820s, to spend a little time with a chap called Babbage and an Engine that made a Difference. Babbage, who was reading widely and keeping up with all the latest work in the physical sciences (some things do change) managed to get his hands on some government finance (some things... stop it! Ed.) for the first Difference Engine. In 1834, he was to recall...
"The first idea which I remember of the possibility of calculating tables by machinery occurred either in the year 1820 or 1821: it arose out of the following circumstances. The Astronomical Society has appointed a committee consisting of Sir J Herschel and myself to prepare certain tables; we had decided on the proper formulae and had put them in the hands of two computers1 for the purpose of calculation. We met one evening for the purpose of comparing the calculated results (you had to make your own entertainment in those days...) and finding many discordancies, I expressed to my friend the wish, that we could calculate by steam, to which he assented as to a thing within the bounds of possibility." (At which point, I presume, they put the kettle on.)
So this first Difference Engine was for making tables (though it could also have been used for solving equations). Existing navigational tables were full of errors, to the point where saving just a few ships from wrecking would easily cover the cost of the project.
Babbage had in mind two main ways of storing numbers. The one involving digits on toothed wheels needed a lathe of more facilities than his possessed — though quite what an educated (nay, technically vital) gentleman of the day was doing with a lathe of any sort struck at least one of his biographers as quite remarkable. Anyway, he outsourced several parts (again, ahead of his time, though actually driven more by paranoia about theft of his intellectual property), assembling the model himself and, by 1822, was able to show that his basic principle was correct. He announced the project in an open letter to Sir Humphrey Davy, President of the Royal Society.
A copy of this letter (presumably not a Xerox) found its way to the Lords of the Treasury who, on 1st April 1823, asked for "The opinion of the Royal Society on the merits and utility of this invention". By 1st May, the Royal Society delivered the opinion that Mr Babbage was "highly deserving of public encouragement in the prosecution of his arduous undertaking" and, on 27th June 1823, Babbage was happily telling Herschel of the £1500 he had been promised (and the £1000 advanced) by the Chancellor of the Exchequer, a Mr Robinson (later Lord Goderich, and finally Earl of Ripon, for those who want the minutiae).
By 1828 the money had run out, and the Duke of Wellington (then Prime Minister) had to be lobbied — hard — for further funding. By February 1830 another £7500 had been liberated, and work on the Engine resumed. In May 1832, the costs passed the £12,000 mark and by 1838 they had passed £17,000 but, after 1834, Babbage actually devoted the rest of his computing career (as opposed to his many other interests) to the Analytical Engines that (in theory, at least) foreshadowed many key concepts of the stored-program computer. In passing, note that Babbage's treatment by successive administrations was to inspire (if that's le mot juste) another writer chappie by the name of Dickens to invent the "How not to do it Office" in "Little Dorrit" (but that's literally another story).
So much for the hardware side of things. Contrast it, now, with the situation in the closing years of the 20th Century and a 1995 take on software design and development from a book called "The trouble with Computers" (Wasn't that Star Trek? No, that was "Tribbles" - Ed.)
The author argues the case for the existence of something of a gap2 between the money spent on computers and the increased productivity delivered by them.
Anyway, talking of design... "Somebody usually talks to customer representatives, typically managers of data processing departments. But if any of these people ever did the job the program is supposed to aid, it wasn't very recently, and they probably had different talents from the majority of users — who were not promoted... Typically in software design, a system is developed that performs, in a strictly technical sense, many functions imagined useful. Then an interface, a set of controls by which people can make it work, is attached. If people can't make the system behave, courses are offered and books published. The users are redesigned. Meanwhile, managers are taken to task for failing to rearrange work flow and organisational structure properly."
This (says the author) is "blaming the user".
He goes on to describe the "waterfall model" of software development, too: "It starts upstream with requirements that feed a tumbling chaos of planning and programming that — usually — plunges into chaos just before it comes to rest... The 'get it right the first time' philosophy in software engineering is astonishingly and radically at odds with other forms of engineering."
What a different world from that of the elegantly-machined Difference Engine!
There you have it. Two vastly different writers. One documenting the hassles over many years of getting hardware development funded, not helped by the fact that his ideas were so far ahead of their time. The other, 130 years or so later, bewailing the lack of usability and poor return on the systems so many of us spend so much time and money on as we earn our daily bread. Both well worth reading3 with the mince pies during the holiday break.
Funny old world, isn't it? Happy Christmas!