Zusie – My Relay Computer | Hacker News Reader

Zusie – My Relay Computer

nablaman.com

160 pointsnew29921 comments

Hi everyone, I am the author of Zusie. I was alerted by a friend about this thread. Thanks for all the interest!

The machine is actually more or less finished, even though the web page doesn't say so. The only thing I was going to add, but haven't gotten around to yet, is a mechanical master oscillator instead of the current solid-state one. I might also add a proper relay based memory someday (I still have thousands of relays in boxes.. :)

A very fascinating project, one I would never have the perseverance to see through. Thanks for posting it.

It does, however, end with the machine being "not quite finished" in 2011, can anyone find an update?

The last update on the News page[0] is dated "June 13, 2011", so it seems not. A shame, it's such an interesting project.

[0] http://www.nablaman.com/relay/progress.php

If you go to the downloads page[1], all the files have last modified date of 21-nov-2013. I don't know if that's due to actual changes, or something else. I tried sending him a message through the contact form.

[1]: http://www.nablaman.com/relay/software/

no, unfortunately I don't think he ever finished it. Though as he notes Zuse built complete relay computers in the early 1940s (using discarded telephone switching relays!). It's interesting that this development in computing has kind of been overlooked (I'd guess because it took place in 1940s Germany).

I came across his work again as I was putting together a relay based oscillator this weekend [1]. I wonder if I have enough relays to put together a half adder. :)

[1] http://41j.com/blog/2014/12/electromechanical-oscillator/

To qualify the "overlooked" part: In Germany, his work is part of basic computing history every Computer Scientist should have at University. So his general work is far from forgotten.

Yep. In Germany. Pretty much nowhere else.

The tragedy is that Zuse tried to patent a lot of his inventions. His submissions were lost during the war, and when he resubmitted them after the war, his applications were struck down due to American and British prior art.

At least in the last 20 years or so, I think the Z3 is part of standard history-of-computing introductions in American CS curricula. We definitely learned about it when I was doing a CS degree in the early 2000s.

That's really great to hear. Aside from this web page I'd unfortunately never heard of his work before. If you have any references I'd be interested in reading them.

It's interesting that the wikipedia pages on the Manchester Baby, ENIAC, and Z3. All refer to them as "first computers" by one definition or another. I guess the ones you end up hearing about are to a degree culturally determined.

I assumed something similar might have been developed in Japan. And it seems they also developed a relay based computer in 1952 which I found interesting:

http://museum.ipsj.or.jp/en/computer/dawn/0005.html

There is at least the Konrad Zuse Internet Archive, even partly in english: http://zuse.zib.de/. I have no own references apart from lecture slides I have no access to anymore.

The thing with the first computer is similar with cars. France has a very different definition of what was the first car than germany - distinguishing between the first car and the first "modern" car - the same will probably be true for the USA and UK.

Well (according to Wiki, at least, https://en.wikipedia.org/wiki/Z3_(computer)#The_Z3_as_a_univ... ) the Z3 was only accidentally Turing-equivalent, and the hack for writing arbitrary programs on it was only discovered in 1998 (and wouldn't have been practically useful anyhow). The Manchester Baby was operational and ran its first program before the completion of the redesign and rebuild that made ENIAC Turing-equivalent. However, it had only been built as a technology demonstrator/testbed, and it was soon disassembled. (I say Turing-equivalent, but Turing's work wasn't the inspiration for any of these efforts to build a fully-programmable computer.)

This one is cool, almost no semiconductors: http://www.northdownfarm.co.uk/rory/tim/tim-8.htm

That capacitor memory array is very neat!

It blows my mind to think that, with the invention of the relay in 1835[0] and the invention of boolean logic in 1847[1], a binary computer like this could have been built almost a hundred years before it actually happened. There was just never the right person looking at both parts of the puzzle.

This is my go-to answer for the old "if you could go back in time, what present technology would you invent" dinner party question.

[0] https://en.wikipedia.org/wiki/Relay#History [1] https://en.wikipedia.org/wiki/George_Boole#Symbolic_logic

I wonder what it would and could have been used for. Seems like it would have been too large and expensive to be cost effective for most things, wouldn't it?

Well, the earliest IBM systems were used for processing census results, so that's one possibility. Ballistic trajectories for artillery were first done with analog computers, so probably not a good match. If you look up the historical usage of log table books, there should be some very good possibilities.

I made one of these, to sell: http://relaysbc.sourceforge.net/

I need to make some more of them..

What's the price per unit?

These look gorgeous, I searched ebay for the boards, ;)

Drum memory might be a nice touch.

"The movie Tron: Legacy [1], which revolves around a world inside a computer system, features a character named Zuse [2], presumably in honour of Konrad Zuse."

[1] https://en.wikipedia.org/wiki/Konrad_Zuse#Zuse_Year_2010

[2] http://www.imdb.com/character/ch0209988/

His 'Zusie Simulator' is also interesting to play with - watch the bits change as the program runs.

http://www.nablaman.com/relay/sim/zusie.html

Reply on news.ycombinator.com