SAGE jumpstarted today’s technology and built IBM into a powerhouse | Hacker News Reader

SAGE jumpstarted today’s technology and built IBM into a powerhouse

arstechnica.com

76 pointsfurcyd31 comments

The manuals for the computer SAGE was built around are absolutely fascinating to read. They are basically an introduction to the digital computer written - genuinely - from scratch. As with many books and monographs from this period, they stand as an example of how pedagogy can actually get worse as a field progresses (authors start making assumptions).

http://www.bitsavers.org/pdf/ibm/sage/3-112-0_Theory_Of_Prog...

It's dated 1959, not that much has actually changed since then in terms of the rough outline of assembly programming.

We have domain specific languages, the AN/FSQ-7 was a domain specific computer (all 60K tubes of it).

I first learned about SAGE through an unlikely mechanism: sitting in the cockpit of a CF-100. There is a switch labeled "attack" that really confused me. It turns out the SAGE could be used to direct aircraft directly to a target.

After learning about this, I did some more research and it seems a signal was broadcast that could be received to direct planes to a target SAGE had identified. I don't have a link to the documents, but originally it was a relatively small transmitter that was upgraded to a much powerful system, I think around 30000 watts of output. It was interesting to me because today I assume an attacker would either jam such a system or actually interfere with it in such a way to send the interceptor off course. I never could find any info on the actual signal transmitted and if it had any form of security.

https://aviation.stackexchange.com/questions/33205/why-does-...

Another cool fact: at the opposite end of the SAGE project we have the "Safeguard" project which was completed in 1975. It spent less than 24 hours in operational capacity before being cancelled.

https://en.wikipedia.org/wiki/Stanley_R._Mickelsen_Safeguard...

One of the really interesting paragraphs from the article:

  The UK’s Royal Air Force, flying their new Vulcan bombers, got orders for the exercise, but the plucky Brits ignored some of the details. Flying their own attack profiles (essentially cheating) and using highly effective radar jammers, the Brits exposed wide gaps in SAGE capabilities. Despite a generally high success rate claimed for the fighters in “destroying” their targets, the best estimates were that only a quarter of the bombers were intercepted.

So not only was this a truly fascinating early computing system, but the signals seemed quite vulnerable to jamming.

At the time the USSR didn't really have any long range electronic warfare aircraft that could be used to effectively jam the SAGE transmitters before the bombers arrived.

Does anyone else find it terrifying how much could get done with so little in the old days? I know a lot of computing has changed and gotten a lot more demanding (like high res displayed, fast wireless networks) however things like this remind me why I refuse to install electron software. Or run web-apps for things that used to be a static html page (I’m thinking movie times).

I feel like we’re standing on the shoulders of giants twiddling our thumbs when I read stories like this.

So little? That machines you see in the pictures did cost millions of dollars each.

Each wire strand had to be carefully put in place and soldered or specially wire wrapped manually: https://en.wikipedia.org/wiki/Wire_wrap

If you make a mistake, it was so painful. Today we use PCBs and if we make a mistake, we just reuse 99% of the PCB design. Once it is done, repeating the PCB design cost you almost nothing.

In the past,when you needed to make something again, you needed to make all the connections again manually. That increased the cost so much and made stupid errors abundant as humans make stupid errors all the time.

It will also took a long time for a computer to be done.

People working there having access to computers were privileged people. Now I can control a huge CNC machine, the watering of the plants and a 3D printer with a computer that fits on my hand and cost peanuts.

That is magic for me.

> Each wire strand had to be carefully put in place and soldered or specially wire wrapped manually: https://en.wikipedia.org/wiki/Wire_wrap

It seems to be that wire wrapping isn't very popular nowadays, and that hobbyists tend to use breadboards [1]. Is there a particular reason for that?

[1] https://en.wikipedia.org/wiki/Breadboard

You’d still need to solder components to the posts, and only then you could do the wirewrap connections between those posts. It just doesn’t make any sense.

I believe even “back then” the wire wrap was only used on backplanes; modules plugged into those were through-hole PCBs plugged directly into the posts that were wrapped around in the other side.

Components are too small to solder or weld a wire to by hand.

Well, you could hot-wire some of these designs. Much harder to fixup a PCB in general, without throwing it away with all the components.

Quite the contrary, wirewrapped boards were easily patchable. And computers (not sure about SAGE, but this was common with Digital) used to be patched at customers premises; patches usually used differently colored insulation.

Well, maybe terrifying how much more we could do with modern hardware, but yes.

While I'm not firmly in the camp that "everything was better in the old days", it clearly wasn't, I do wonder if we've taken a detour. There's nothing wrong with taking some of the added speed we gain from modern hardware, and use it to great language and frameworks that get us through implementation faster, but are we developing things faster? Best I can tell, SAGE took four years to build, that's pretty fast for what it is and given that much of the hardware also had to be designed and build. I don't think the modern day equivalent could be build that fast, regardless of much better tooling.

Part of the problem, as I see it, is misdirected effort. Most problems we're trying to solve is attacked with way more effort and ambition that required. Electron is a great example, it really does solve a problem and it truly does speedup development. That great, but why use it to build chat applications? Chat was basically solved by the year 2000. I use Google Chat of work, it's only better than IRC in two key feature. 1) I can instantly setup a video call. 2) I can bombard my coworkers with stupid reaction gifs. That is it. Add TLS to IRC, and it is better in most other aspects.

Or how about webshops? We've been able to buy stuff only since 1995, has the experience really improved? No, not really, not in any meaningful sense. Yet somehow I now need a Kubernetes cluster, even tough my servers are a 1000 times faster. Can't a modern server handle at least the same amount of traffic? Why do I need 5MB of Javascript to buy a book?

We're spending too much time solving problems that we successfully handled decades ago, and we're often claim to have a better solution, when it's really just prettier.

Maybe you'd spend less time twiddling your thumbs if you installed some of the excellent Electron apps our there?

I would be I am bored of killing processes in task manager.

Introducing - Electron killer! Easily kill electron processes. Just install our 350mb Electron app...

> Electron killer! Easily kill electron processes.

You could even call it 'Positron' or something like that.

Weekend puzzle: Find and terminate all Electron/CEF/Node bloatware on a system without killing any dedicated web browsers.

Easy enough when using w3m

> Just install our 350mb Electron app...

Added bonus: repeat this process monthly to ensure you stay up to date!

Want to uninstall? Upgrade to PRO

Not only that, its follow up system SACCS, gave birth to JOVIAL, one of the first system programming languages, in 1958.

https://en.wikipedia.org/wiki/JOVIAL

Maybe the title should be changed to uppercase "SAGE". The first thing that came to my mind was the enterprise software company: https://www.sage.com/

My first thought was the UK's "Scientific Advisory Group for Emergencies", who have become a household name in Britain during the pandemic, and are most famous for having been repeatedly wrong about everything yet for some reason still taken seriously by many.

'Independent SAGE' were worse in my opinion, they were shamelessly trying to piggyback on SAGE's credibility with the public with their intentionally misleading name. SAGE are the government's official scientific advisory group, Independent SAGE are little more than a lobby group for excessive use of the precautionary principle typo-squatting on an official organisation.

But yeah, I'm not a fan of a lot of decisions SAGE made either. I think the decision by SAGE to use behavioural psychology to 'increase the sense of personal threat' and other manipulative tactics like that was unethical for example, and of course the modelling turned out to be quite far out from what actually transpired. I definitely think the press deserves a good chunk of the blame though, they latched onto the worse case scenarios SAGE modelled regardless of their likelihood and caused preventable panic in the process which isn't necessarily SAGE's fault.

I guess it's all a moot point now, the recent revelations of the PM and other high-ranking politicians hosting parties in Downing Street with impunity while we were in the first 'nobody's allowed to leave the house but to work or shop' lockdown has massively reduced the any chance of further restrictions I think regardless of what SAGE or other scientists say. Regardless of anyone's opinion on their merit as a policy, lockdowns always depended on public consent which just isn't there any more.

Or SAGE Computer Technology waaay before that, who made some neat early 68000 based systems.

See https://en.wikipedia.org/wiki/SAGE_Computer_Technology

My first thought was

https://en.m.wikipedia.org/wiki/SageMath

but it made no sense.

My dad was programmer on the SAGE project in the early 60s. He there were dozens of techs with shopping carts full of vacuum tubes going up and down the aisles 24 hours a day to keep it running. He also the circuit for storing one bit took up a 6 inch sided cube of space.

I went to college in Corvallis OR, which is near Camp Adair. One of these concrete bunkers still stands there. One of my roommates was a local and said it used to house a supercomputer. I guess he was right..

> A breathtaking amount of storage: 64K of 32 bit words!

So here total memory is 64K and each Kilo byte is a 32 bit word?

It's 64,000 'positions' of 32-bit words. That'd be 256,000 modern 8-bit bytes.

For comparison, the original Macintosh (25 years later) had 128KB. SAGE really did have a crap-ton of memory for its time.

Reply on news.ycombinator.com