For me, as a software developer, it's about detailed design rather than implementation. 1837-s design is much more impressive than 1941-s implementation (100+ years later - you'd better build something with all that time).
And the same museum at Bletchley Park (there are two) is building a functioning replica of the EDSAC, one of the first programmable, fully electronic computers:
(Whether EDSAC was "the first" depends on how you define "computer", and what you think it means for a machine to be in operation, among other things, but it's one of several plausible claimants. The Z3, cited elsewhere here, was a programmable computer, but not fully electronic -- there was a lot of relay circuitry, which kept the clock rate down to about 10 Hz.)
From the title, I was honestly expecting the article to be about this project instead.
Slightly OT but reminded me of an article from HN from a couple of years ago about old(er) computers that are still in operation so I've reposted it [0] for those who missed it last time around.
> Gleason realized early on that he couldn’t make his portion of ENIAC run actual calculations—such an endeavor would require all 40 panels, not to mention thousands of new components and technical know-how that had long been forgotten. But he resolved to make the computer at least appear like it was hard at work figuring out the best flight paths for howitzer shells.
That's a bit disappointing. I guess it's nice you can at least see it though.
Time to correct the record: the World's first electronic, programmable computer was build by a German by the name of Konrad Zuse in May 1941 [1] called Z3 [2]. He designed a language for it called Plankalkül [3].
Eniac was announced in 1946. Incidentally Turing wanted to be build one in 1945, called Electronic Brain. His paper "on computable numbers" was published in 1936.
Name operating since Floating Point Binary Eletric Programmable Turingcomplete
Zuse Z3 May 1941 yes yes no yes yes
Atanasoff-Berry Summer 1941 no yes yes no no
Colossus 1943 no yes yes with rewiring no
Mark I 1944 no no no punch cards yes
Zuse Z4 March 1944 yes yes no punch cards yes
ENIAC 1946 no no yes with rewiring yes
ENIAC 1948 no no yes with a resistor matrix yes
Interestingly, the Z3 was not meant to be Turing complete (by Zuse) (that is, he did not intend the functionality that Turing completeness describes. Living in Germany back then he did not know about Turing nor about the "on computable numbers" paper).
Only in 1998 did people find out that the Z3 was indeed Turing complete.
What's also impressive is that Zuse didn't know about Charles Babbage either.
In college, I had a very old professor from... I think it was Romania, or possibly Bulgaria. Anyway, one day he was saying he had worked for Konrad Zuse, and I blurted out "Wait, you worked for the N..." and stopped there, thinking better of it. He laughed and claimed it had been after the war. I did later learn not to spit out every thought that entered my head.
I just saw some ENIAC panels at the Computer History Museum in Mountain View, CA and I'm hoping to visit the Moore School of Engineering in Philadelphia soon to see the parts they have. Now I'll have to visit Oklahoma and see the restored portion they have. (I've also got cousins there, so that's another reason to go.) I'm a big fan of ENIAC and the other early computers. As for "first", I don't think any computer should be called out as the first, each of them contributed something to the machines we have today.
It is no small source of acrimony in the University of Cambridge that Manchester beat them to the first stored-program computer by one month with the Baby. However, once the Baby had been demonstrated, it was promptly dismantled, leaving Cambridge's EDSAC with the title of the world's first practical stored-program computer. The EDSAC was then used for nine years at the Mathematical Laboratory before being shut down in favour of the EDSAC 2.
Back when I was doing my degree, the 50th anniversary of the EDSAC came around, and we were given a project of creating a simulator for the machine. This we did, even including the rather interesting sound that the original machine made from a loudspeaker connected across the main mercury delay line. Yes, RAM was implemented as a long tube of mercury with sound pulses travelling down it and fed back into the beginning again.
The Whirlwind Project seriously considered using a microwave "delay line", a round trip microwave connection a long distance across the state of Massachusetts. Ended up using Williams tubes (https://en.wikipedia.org/wiki/Williams_tube) as did the Mark 1 and a number of other early computers (their random access had an advantage over delay lines), and eventually invented 3D magnetic core memory.
It has also been proposed to use a laser and one of the reflectors on the moon. That would give you a 300,000 km storage line. That proposal wasn't serious, though, if only because it is a challenge to keep that reflector in your field of view.
Another impractical delay line, but with variable delay, is the internet. You can hide some bits in a DNS request or in a Google search.
Sorry for not joining the positivity parade, but taking the control panels, repainting them and then wiring up the lights to blink in pretty ways seems substantially wrong in a number of ways...
I guess "resurrected" is being used in the "brain dead and slightly rotten body" zombie sense of the word :(
The ENIAC panels at the Smithsonian were at one time powered up, with a functional counter. There was a little control box on a stand where you could push some buttons and watch it count. But they had to power it down decades ago. For the fiftieth anniversary, they powered it up briefly.
34 comments
[ 3.2 ms ] story [ 78.9 ms ] threadFirst computer not first vaporware
Also of note is that by the end of the war the GC&CS had 10 working Colossi.
http://www.tnmoc.org/special-projects/edsac
(Whether EDSAC was "the first" depends on how you define "computer", and what you think it means for a machine to be in operation, among other things, but it's one of several plausible claimants. The Z3, cited elsewhere here, was a programmable computer, but not fully electronic -- there was a lot of relay circuitry, which kept the clock rate down to about 10 Hz.)
From the title, I was honestly expecting the article to be about this project instead.
Slightly OT but reminded me of an article from HN from a couple of years ago about old(er) computers that are still in operation so I've reposted it [0] for those who missed it last time around.
https://news.ycombinator.com/item?id=8657823
That's a bit disappointing. I guess it's nice you can at least see it though.
Eniac was announced in 1946. Incidentally Turing wanted to be build one in 1945, called Electronic Brain. His paper "on computable numbers" was published in 1936.
[1] http://en.wikipedia.org/wiki/Konrad_Zuse [2] http://en.wikipedia.org/wiki/Z3_%28computer%29 [3] http://en.wikipedia.org/wiki/Plankalk%C3%BCl
http://en.wikipedia.org/wiki/Calculating_Space
http://people.idsia.ch/~juergen/digitalphysics.html
Only in 1998 did people find out that the Z3 was indeed Turing complete.
What's also impressive is that Zuse didn't know about Charles Babbage either.
http://www.oldmouse.com/mouse/misc/telefunken.shtml
I'm kind of curious what other parts of computer history are overlooked because they happened outside of the anglosphere.
http://www.oldmouse.com/mouse/firstmouse.shtml
Englebart's work wasn't secret; various other people may have constructed their own mice after 1963, and yes, that sort of thing is interesting.
http://www.computer50.org/mark1/new.baby.html
http://en.wikipedia.org/wiki/Manchester_Mark_1
Disclaimer: University of Manchester Alumni
Back when I was doing my degree, the 50th anniversary of the EDSAC came around, and we were given a project of creating a simulator for the machine. This we did, even including the rather interesting sound that the original machine made from a loudspeaker connected across the main mercury delay line. Yes, RAM was implemented as a long tube of mercury with sound pulses travelling down it and fed back into the beginning again.
Another impractical delay line, but with variable delay, is the internet. You can hide some bits in a DNS request or in a Google search.
http://en.wikipedia.org/wiki/Manchester_Small-Scale_Experime...
I guess "resurrected" is being used in the "brain dead and slightly rotten body" zombie sense of the word :(
[1] http://xkcd.com/356/