Slashdot Mirror
Building a 5-Ton Calculator From 19th-Century Plans
alphadogg writes "Starting in May, many will have the opportunity to see computing done the old-fashioned way: with lots of gears, a big crank, and some muscle. The Computer History Museum in Mountain View, CA, will unveil a new construction, the first in the US, of the 19th-century British mathematician Charles Babbage's Difference Engine No. 2, an improved version of his earlier design for a mechanical digital calculator. It weighs in at two tons more than the Difference Engine built in 1991 at London's Science Museum. Microsoft millionaire Nathan Myhrvold commissioned and paid for the US model."
When you can make an difference engine out of LEGO, it really doesn't seem that impressive to build a five ton one. Babbage's analytical engine, however... that would be an interesting piece.
File under 'M' for 'Manic ranting'
Here are some links :
http://www.fourmilab.ch/babbage/
The obligatory 99-bottles-of-beer-on-the-wall in punched card Analytical Engine assembly language :
http://99-bottles-of-beer.net/language-babbage's-analytical-machine-79.html
Hmmm, I dare say that's shorter than the C# version, if you remove the comments. Oh and it will run Linux, if you have enough coal and are willing to wait a few years for X to load. ;) (it does have a graphical output device) As for a beowulf cluster, that might help performance, although your interconnect mechanism would probably be pneumatic ie. tubes (that's what the Internet is made of anyway right?) and the cluster size would require a few tens of millions of units. ;)
jdb2
Well, it sounds like it's turning complete (the first machine to be), so in theory it can run any program runnable today.
http://en.wikipedia.org/wiki/Babbage#Analytical_engine
As a practical matter you may want to invent a time machine and perpetual power source first.
I am not an expert. If I am misled in something, please correct me.
What happens when you divide by zero on a calculator using a physical engine?
I've done that.
The particular calculator in question would spin madly, with the result digit dials working like a cross between an odometer and a clock movement, until you hit the button that aborts the process. (The abort apparently consisted of changing the divisor to a large number. It took close to a minute as the machine would do a trial subtraction, undo it, shift the register bar one to the left, and repeat until it got to the last digit.)
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
"On two occasions I have been asked, "Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?" I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question."
What truly happens to an impossible sum?
Does it dry up
like a slashdotter in the sun?
or does it fester like sco
and then run?
does it stink like an overused meme?
or crust and sugar o'er--
like a deferred dream?
maybe it just sags like a 5-ton calculating machine under a heavy load
or does it explode?
Those who advocate genocide deserve every protection afforded by law, and none afforded by common human decency.
Here is the pic of the UK version.
See what one (difference engine #2) looks like running, impressive!
This one is in mechano parts (Erector Set for us Americans)
http://www.meccano.us/difference_engines/rde_2/index.html
"Enjoy what you're doing! If it becomes drudgery, you're doing it wrong!" - Jim Butterfield
I'm am one of the Docents for the Difference Engine #2, and although the team making it work is WAAAAY more competent to comment, lemme put out a few FACTS, at the risk of "flame wars of death"... The Engine is a single function calculator that can iterate the values of a 7th order polynomial approximation to an arbitrary mathematical function. After about an hour of VERY careful setup, any set of coefficients could be entered, allowing almost any function to realized. It uses a technique called "finite differences" that allows the calculation to be performed using only addition (and 10's compliment coefficients to represent negative numbers). Our working plan is to set it up to do a table of logarithms, much like Babbage's own table, produced well before he thought of Difference Engine #2. The polynomial approximation for logarithms is quite accurate over the space from 1.0 to 1.6, 6000 iterations of the Engine. (It takes four turns of the crank or about 6 sec. per iteration.)
The calculation section has about 4,000 parts, and a very elaborate printer mechanism has another 4,000, and was designed to produce sterotype molds of a complete page of a book of tables.
It is a WONDEROUS device to behold! There are 52 distinct stages in it's control graph (EXACTLY like a modern timing diagram, just vertical...) An elaborate nest of 14 cams control the complex sequence of events to do an iteration, which is !pipelined!. The sinuous ripple carry mechanisms on the back side are HYPNOTIC, as are the forward and backward movements of the intra-column sector gears.
Avoid CHM on May 10, it's gonna be a madhouse! But this is pretty close to the top of the list of "1000 Geeky Things to See Before You Die", oh, and by the way, there's all the other ABSOLUTELY WAY COOL stuff at CHM, wanna see an Apple I signed by "the Woz"...
YOU GOTTA SEE THIS! chmguy
All those moments will be lost in time, like tears in rain. Time to die.
Basically, that term is meant to tell you that a particular system can perform any calculation that you could perform on a Turing Machine, which is a minimalist calculating machine devised by Alan Turing, to explore what "computability" means, in a mathematical sense.
The Turing Machine is very simple, but given unbounded time and storage, it is believed to be able to calculate anything that can be described by a discrete set of steps (i.e. an algorithm).
Where this gets interesting for evaluating computer systems is that, if you can prove that it's possible to simulate a Turing Machine in a particular hardware/software system, that means that you can use that system to implement any algorithm that you can implement on any other deterministic computer system.
This doesn't say anything about the efficiency of that implementation. For example, it's easy to write a program that emulates an old 8-bit processor on a modern 32-bit CPU. It's nearly as easy to write a program that runs on an 8-bit CPU and emulates a Core2 Duo processor.
One key difference between those two emulators would be performance - you can easily emulate an 8-bit processor running at several megahertz on a 32-bit CPU, but emulating a 32-bit CPU on an 8-bit processor will be orders of magnitude slower than real time.
Actually, it is at least possible that it was intended as a trick question. You know, one where if you say "yes" then you've just said "yep, I'm a con artist." Admittedly, it's a stupid one even as trick questions go, but still, there might be some purpose behind it.
To put things into perspective Babbage got funding for one machine, never finished it, decided he's rather begin designing the version 2 model, asked for more funding, repeat ad nauseam. Pretty much it was _the_ original computing vapourware. Pretty soon he got no more funding, but that never stopped him from asking for more and hyping his unproven creation to the parliament.
He also seems to have descended into a nerd-like bitterness, in which he took such questions out of context as proof that everyone else is a drooling idiot and that's why they don't see he's right. And in that he also included such questions as, basically, "well, what _can_ it do?" and "what's the business advantage for making one of these?" Stuff that you'd get asked by any business nowadays too. He took them as proof that his contemporary Englishmen were narrow minded and lacking in vision.
It may seem obvious in retrospect that his design was right, but at the time it was everything except obvious. It was a _monumental_ expense with the economy and technology at that time, even compared to paying armies of people to calculate those by hand. And it was anything but proven. Noone knew if it would even work at all. Again, the first round of funding he got, produced nothing tangible.
Also regarding the parliament at the time, they were not as obtuse as you (or Babbage) seem to think. They funded a lot of research, actually. The nautical clock, for example, was paid for by the parliament, and that was quite the iterative development. The first couple of versions not only were too inexact to be any use, but at least the first one didn't even compensate for the ship's rolling around. But nevertheless, that guy had _something_ working to show for his work, and kept getting more money to keep working. Babbage had nothing except his claims.
Now before I sound too damning to Babbage, it wasn't only his fault. He got into a conflict with the company actually building it, and that was the chief reason why the V1 was never completed. But, still, seen from outside, he never had anything working to show, and even more damning he just unilaterally scrapped the design in the middle of the project and began designing an even more overengineered V2 instead.
So, anyway, given that he was technically hyping vapourware, I can see a smart-arse member of the parliament trying to catch him with a trick question. Again, it _is_ a dumb one, but it's not the same class of dumb as actually thinking that a machine can magically guess the right answers when fed wrong data.
(But then again, I see a ton of PHBs and businesses nowadays believing just that about electronic computers, so maybe it was just a dumb question after all.)
A polar bear is a cartesian bear after a coordinate transform.
Even being modded funny, an emulator would be very interesting as an "educational tool" about how the machine works.
So say we all