The risks of retro computing



#2

This article appears in the current Risks Digest and which might be of interest to some of you here.
http://catless.ncl.ac.uk/Risks/24.35.html#subj9


#3

Interesting read.

#4

There are lots of old (and not-so-old) codgers spending lots of time (and LOTS of money) restoring WWII vintage aircraft. They certainly are focusing on the thing, rather than the thing-makers, but they also are experiencing fresh, new thing-maker excitement in the process.

And for some of us (perhaps especially for those of us detached from the realities of a thing's creation and maintenance!) a particular aircraft (for example) is itself is a thing of inherent beauty and inestimable worth. (I have an inexplicable weakness for the Douglas A-20 Havoc/Boston.) To experience the thing first hand is to touch something that transcends its physical actuality.

In my ignorance, an IBM 1401 doesn't qualify as an inherently beautiful object. But I'm sure there are those of us out there who will be moved in some strange way, should the device be fully recreated.

Edited: 1 Aug 2006, 4:52 p.m.


#5

Ditto for the Lockheed P-38.

tm

#6

Perhaps the difference is in the experience which results.

Going for a ride in a P38 or P41 (I think my Dad worked on the engines for those in WWII!) or other restored "warbird" is something you can not experience in any other way.

You could (in theory, at least) experience the slowness or other artifacts (lack of divide/multiply, say) of the IBM 1401 by emulating it in software.

Us Cessna 172 drivers (or Learjet drivers, for that matter) could never get the feel of the Mustang by any conceivable alteration of the airplanes now available.


#7

The experience of watching the top-of-stack registers flashing on the display panel of a Burroughs B6700 cannot be replicated in software.

Each bit in the registers was represented by an incandescent lamp (lit when 1) and the display was arranged in the conventional (for Burroughs) pattern for the 48-bit registers. The default idle pattern was the Burroughs big "B", although customers sometimes changed it to suit their requirements: e.g., "$" for a financial institution or a light bulb for Ford.

In addition to showing whether the system was idle, you could also tell whether the system was thrashing and some programs had recognizable display patterns. Watching the flashing display could be mesmerizing. The entertainment value of modern computer control panels is nil.


#8

I'm reminded that, during one project for development of a new timeshare O.S. on the academic computing system (an HP-2000, IIRC) at The Evergreen State College (at Olympia, WA), the first (and perhaps only completed) task of the project was the new idle loop -- making the front panel lights flash in an interesting (yes, even mesmerizing) fashion.

User Interface, first & foremost!

Edited: 2 Aug 2006, 4:23 p.m.

#9

Speaking of vintage planes here's an interesting development: T-51 Mustang Airplane

T-51 Mustang[/link]

T-51 Mustang

#10

The guy hasn't a clue...

#11

He's not even familiar with the machine he discusses. The IBM 1401 *did* have arithmetic circuitry; it was the IBM 1620 that did not. They were very dissimilar machines.


#12

from the article "The risks..."

..."truly global encyclopedia in the form of software simulators "...

Nonpareil anyone?

...but many of us are interested in (and willing to pay some money for) current-hardware incarnation of such simulators!

Eric, did you note? Of course.

#13

Ha ha ha ha ha ha!

Someone ought to introduce the writer to us and this entire website!!

:D

#14

This guy has to be a software engineer. No offense to anyone on this list that would place themselves in that pigeon hole, and recognizing that generalizations about the discipline are likely to fail with respect to any given individual, particularly one that hangs out here, and apologizing inline for long, run-on sentences, (and for parenthetical, but non-grammatical emoticons, which AOL did not invent 8), taking all that into account, someone who lives for software is far more likely to miss the value of a hardware restoration or reconstruction. (Whoa, what was that? A period.)

Think about, for example, the Analytical Engine. (I thought someone had attempted a partial construction based on Babbage's design, but I can't find a reference in Wikipedia.) Would a physically embodied Analytical Engine have a different worth than its designs, or the plans for a Turing-Complete programming language demonstrated by Lady Ada, or simulations of the device on modern computers? Of course it would have different worth. Apart from the working out of practical difficulties that Babbage faced in the construction, the artifact itself would be an invaluable addition to any museum. (I would love to hear the thing in action. Yes, I've read The Difference Engine.)

But I imagine that a lot of greenhouse gases would be emitted by the machining of parts, and the transporting, assembly and presentation of the results, not to mention the grave danger of crushed limbs the sure-to-be-impressive weight of the thing would pose.

OK, so there never was a complete Analytical Engine, so the construction would be new, and therefore perhaps more valuable in the estimation of the Risks author. What about the Difference Engine, then? I saw a reconstruction of that in IBM's Pallisade's NJ facility. Does that artifact add anything to the world over and above the ideas it embodies? I think the answers are the same as for the Analytical Engine.

The point is, old machines disappear into the relentless maw of ever-accelerating progress. Conserving working computers obviously has merit. Even if the fantastically compressed history of the Computer Age gets overshadowed by subsequent advances in other areas (like Biology) or by subsequent failures of Civilization (such as "take-your-pick-of-doomsday-scenarios") the conservation of these artifacts will have worth to someone, if only to the optimistic conservators themselves.

Regards
Howard


#15

Babbage never completed the design for the Analytical Engine. He kept changing it, and there is no coherent set of drawings that would produce a working machine.

On the other hand, his Difference Engine #2 design was complete, and had only a small number of minor design errors. His son built a subset of it, but the complete machine was not built until a team led by Doron Swade of the Science Museum in London built one from 1989 to 1991. AFAIK, this is the only working full-scale Difference Engine based on Babbage's design, although a scaled down machine based on the same principles has been constructed by Tim Robinson using Meccano.

I don't know what you saw at IBM's Palisades NJ facility, but I would guess that it is a scaled down model of some sort.

There has been some discussion of the possible construction of a second Difference Engine #2 based on the work of Doron Swade and his team, but AFAIK nothing has yet come of that.

I've thought a little bit about trying to build a binary mechanical computer (vs. Babbage's decimal designs), but I don't really know enough about mechanical engineering to do it.


#16

Quote:
AFAIK, this is the only working full-scale Difference Engine based on Babbage's design, although a scaled down machine based on the same principles has been constructed by Tim Robinson using Meccano.

Tim has parts of an analytic engine also constructed using Meccano.
He also isn't the only person to have build a Meccano difference engine, I know of (and have seen in operation) one other. I'd also like to have a go at building one myself at some stage :-)

Tim Robinson's web site of mechanical computing using Meccano:

    http://www.meccano.us/

Regards,

- Pauli

#17

Looking at the model of Difference Engine number 1 at Robinson's Meccano site, it's clear what I saw at Pallisades was a model of that version. I can't tell at what scale, however. The model was quite large, about four feet high if I recall correctly. It was in a glass case and looked gorgeous. It looked like it was made out of brass and steel. This was in 2003, and I was being brainw^H^H^H^H^H^Horiented into IBM at the time, and nobody seemed to know much about the model, or about any of the other, more accessible artifacts strewn around. There were models of several devices by Da Vinci in my wing, plus pieces of the Semi Automatic Ground Environment machinery, an old truck used to deliver batteries or something like that and lots of other great stuff. I was definitely impressed, and delighted.

Regards,
Howard
(Edited fur speling erurs)

Edited: 3 Aug 2006, 2:33 a.m.

#18

Quote:
I've thought a little bit about trying to build a binary mechanical computer (vs. Babbage's decimal designs), but I don't really know enough about mechanical engineering to do it.

I've been told that the UK government at least, and probably others, put its boffins onto solving the problem of producing mechanical logic gates during the cold war, in order to produce computers that would survive EMPs should things turn hot. The conclusion was that it couldn't be done.


#19

Back in the late '70s, I operated a die-casting machine that used
pneumatic binary logic for its controls. Nothing very complicated,
but I do remember that the logic board included AND, OR (or maybe
it was NAND and NOR), and NOT gates, and some type of flip-flop.
It had pneumatic timers too, but I believe that those were analog.

So if "pneumatic" can be considered an example of "mechanical",
then I'm sure that "mechanical" logic gates have been built.

Given that each of those pneumatic logic gates would have had a
volume measured in cubic inches, a calculator or general purpose
computer built of those would be quite bulky.

I've seen hydraulically operated hydraulic valves, so it's easy
for me to imagine hydraulic binary logic gates.

And of course I've seen plenty of machines that used "relay
logic", but I'd call these "electro-mechanical" rather than
"mechanical".

But even if such devices did survive an EMP, I have doubts about
how useful they'd be after a nuclear war.

Regards,
James

#20

quote: "I would love to hear the thing in action."

Howard, you bring in a new dimension- audio. I an a retired telephone employee. Before electronic switching, one could hear incredible sounds in an electro-mechanical 5X-bar central office. You were all alone in the building at the peak busy period, mid-afternoon. The entire X-bar machine, which filled the building, with all its racks of switches, seemed to be alive. The sound was not too loud. It would rise and fall like the waves on a beach reaching a crescendo then almost silence. Awesome.

tm


#21

I've heard others talk about that sound. I doubt a recording would do it justice. It must have had quite a physical presence.

The metaphor of communication playing out in the chatter of thousands of relays is a satisfying one. Now we don't have anything so crass as the motion of half a gram or so of metal in an electomagnetic field to sustain our discussions. Our symbols are embodied in the motion of fewer and fewer electrons in thinner and thinner paths, or varying phase or intensity in fiber optic wave guides at higher and higher densities and speeds. Too bad what we have to say hasn't undergone a similar revolution.

The appeal of hearing analogues of electronic switching strikes me as similar to the delight I feel when viewing lots of "blinkenlights" on an old machine. There's also something similar about the affection I feel for front-panels with toggle switches. I'm not old enough to have actually used any system like that. Those I have touched have been museum pieces. But there is satisfaction in representing a binary or octal number in the positions of such switches. I guess I feel that if I can touch it, it a) must be real and b) must be - quite literally - within my grasp.

It's an illusion, of course, but a comforting one.

Regards,
Howard

#22

Quote:
The sound was not too loud. It would rise and fall like the waves on a beach reaching a crescendo then almost silence. Awesome.

Whereas the sound of a roomful of data entry clerks typing in credit card transaction details on IBM ultra-clickety keyboards was unbelievably horrible -- like a bunch of irregularly-driven ratchets crackling up and down your skin.

And don't ask what thousands of chicks sound like in an enclosed room in behavioural analysis classes -- all I remember is my hearing slowly returning over lunch.

Still, all better than someone playing the bagpipes in a small restaurant, something that some visiting American academics had arranged at a conference I attended in Edinburgh. I was smart enough to ask my wife to step outside with me for a moment when I saw the piper arrive, and the look of stunned amazement on the faces of those who stayed inside for the experience amuses me to this day.

What was the question again?


#23

Frank--at the risk of one-upmanship I must tell you this story. Back in the 60's, Shakey Johnson the founder of the Shakey Pizza Parlor chain, took a bunch of his friends to Hawaii for fun. He also invited along a contingent of the Lockheed Bagpipe Band. Imagine hearing those pipes at 30,000 feet in a DC-8!

tm


#24

It would be better in a DC3 prop plane. A band of pipers might be the only acoustic ensemble that could make itself heard over the engine noise and vibration in a DC3.

Come to think of it, the sounds would probably be harmonically compatible. 8)

Regards,
Howard

#25

Quote:
Imagine hearing those pipes at 30,000 feet in a DC-8!

I bet a few dental fillings were lost on that journey. Still, it sounds like a great way to get reluctant parachutists to leave the plane.

As a Scot, I'm thoroughly familiar with the skirl of the pipes, and also with the ideal performance set-up: a lone piper, standing on a mountain-top across the valley from you, playing some modern song that everyone thinks is traditional, such as 'Flower of Scotland'. Brings a tear to the eye. Any closer, and there's danger of a bringing a tear to the eardrum.

Here's a secret that non-pipers might not know: at home, pipers practise using just a chanter, rather than the whole bagpipe. That way, their neighbours only think they have some weird clarinet player next door who can't play more than an octave or so, and their spouse doesn't have to wrestle the dog to the floor while wearing ear defenders and a gum shield.


#26

I like bagpipes in strictly measured doses. The distant sound rising and falling, preferably over a body of water, is one of the most evocative (in a good way) things I've ever heard. There's something down at the level of my DNA that responds (positively) to that type of sound. There was a mini-craze in the USA for pipe music after the RSDG PB performance of "Amazing Grace" made it to the silver screen in IotBS. I was a fan before then.

Regardless, I do agree that the GHB particularly are obnoxious when heard up close. Distance, and modulation by water, cement mixers, jack-hammers and the like tends to smooth out some of the more unpleasant clashing harmonics, not to mention allowing the listener to fill in melodic lines that are actually missing or misplaced in the original sound. That's why I figure that the bone-jarring vibrational modes put out by two enormous propeller engines on a DC3 might actually improve the bagpipe listener's experience. I might prefer that experience to repeatedly hitting myself on the head with a sack of ball bearings, for example.

Well, maybe ..

Regards,

Howard

#27

Another risk of retro computing:

you establish your own computer museum in your basement.

I'vd had at one time a 286, 486, and a PIII Celeron collecting dust at the same time. Fortunately, we needed a replacement 486 in the lab to run an old piece of equipment, as the one it came with broke down, so I gladly volunteered my old Gateway 486.

But here's where retro machines... up to a point, of course... are still VERY useful- in many research labs. PCs upgrade technologically every three hours and the software for much equipment is written to run on only one kind or one class of CPUs. When your (exorbitantly) expensive equipment ages, yet remains highly functional and useful, then it makes no sense to junk it just to get marginally faster or graphically cooler software... and the PC to run it on.

For data collection and treatment, most ordinary scientific calculations, you might be surprised that even a 8088 could still do the job... if you can get (or if your have the time, write) software or find some way to hook up current storage devices for it.


#28

Hi Ed.


Quote:
I'vd had at one time a 286, 486, and a PIII Celeron collecting dust at the same time. Fortunately, we needed a replacement 486 in the lab to run an old piece of equipment, as the one it came with broke down, so I gladly volunteered my old Gateway 486.

I ran into a similar situation last year at work. Many years ago, we used a tape drive for our backups that would only run in a true dos environment. Not a problem since DOS was what we were using at that time. Of course, one of the Engineers came to me requesting that I resore an old project. All our machines at work are on Windows.

I ended up taking the tape drive home and using one of my old abandoned 386 machines in the garage was able to restore the data. Forturnately, I had the foresigt to save an image of the tape backup program bootable Disc to a CDROM - I save image copies of every Floppy disc that I've ever used just in case I every need them to recreate them.

While I had the tape drive installed and working, I just went ahead and did a resore of all the old tapes we had, then burned them to DVD.

Bill


Possibly Related Threads…
Thread Author Replies Views Last Post
  HP Prime - incalculable risks for students bluesun08 12 3,607 11-27-2013, 07:43 AM
Last Post: Martin Paech
  Computing pi with the PC-1300S Kiyoshi Akima 0 1,131 11-17-2013, 12:24 AM
Last Post: Kiyoshi Akima
  Books on computing Elementary Functions Nick_S 4 1,659 09-14-2012, 07:33 AM
Last Post: David Hayden
  OT: Retro programmable makes comeback. Egan Ford 2 1,112 01-27-2010, 07:49 AM
Last Post: Bill (Smithville, NJ)
  WOW talk about your retro calculator HP 9100A Donald Williams 8 2,253 10-15-2009, 08:16 PM
Last Post: Mark Edmonds
  Computing Square Roots Thomas Klemm 11 2,941 03-30-2009, 03:26 PM
Last Post: Thomas Klemm
  Computing many digits of Pi Juan Pablo Martinez (Spain) 21 4,969 09-24-2007, 10:58 AM
Last Post: Egan Ford
  Computing PI -- a simple programming problem Katie Wasserman 41 9,079 10-20-2006, 10:34 PM
Last Post: Palmer O. Hanson, Jr.
  computing chances Bram 4 1,448 04-16-2006, 01:25 PM
Last Post: Bram
  1980s Vintage computing - new HP-9915 page and disassembler John 0 677 12-19-2005, 09:00 PM
Last Post: John

Forum Jump: