'Modern' Computers Turn 60 Years Old

Christian Smith writes "Stored program computers are 60 years old on Saturday. The Small Scale Experimental Machine, or 'Baby,' first ran on the 21st of June, 1948, in Manchester. While not the first computer, nor even programmable computer, it was the first that stored its program in its own memory. Luckily, transistors shrank the one tonne required for this computing power to something more manageable."

A tonne?

What's that in Volkswagen Beetles?

Re:A tonne?

1.21 Old, 0.74 New.

Re:

That's proof it ain't a Japanese car. Else the new one would be at the very least at factor 2. And next gen it would be 5.

Evolution

In 60 years we've gone from computers the size of a room to a laptop computers thin enough to fit in an interoffice envelope. Where will we be in another 60 years, or even ten for that matter? It's somewhat scary that we've created a technology that advances much quicker than ourselves. It's just a matter of time before we are the number two species ( if you can call a computer a species ) walking the planet.

Re:Evolution

"In 60 years we've gone from computers the size of a room to a laptop computers thin enough to fit in an interoffice envelope."

Yeah...but, those old tubes used to make the data 'feel' warmer.

:-)

Re:Evolution

1. Transistors get far smaller

2. ???

3. We are slaves to robotic overlords

Maybe if you use Will Smith's humor, or a recursive time-travel paradox, to distract us from the "???" it could work as a plot.

Not so fast...

Luckily, transistors shrank the one tonne required for this computing power to something more manageable

The poster apparently hasn't checked the specs needed to run Vista.

More seriously :

In 60 years we've gone from computers the size of a room to a laptop computers thin enough to fit in an interoffice envelope. Where will we be in another 60 years, or even ten for that matter?

You can bet that the developer will definitely find use for additional power, as machine performance increases.

What has caused the computers to shrink to envelop-size isn't as much the increased performance/size ratio. It's the market.

If Moore's law stated (roughly paraphrasing) that computer performance doubles each 2 years, one should expect the computer to reduce their size by half in that time frame. But that didn't happen. Because most of the time people only one to use the additional performance to have the same box as before but faster.

Only from time to time the users' interest shifts.
Desktop replaced microcomputers and mainframes, not (only) because suddenly the circuits could have been made smaller, but mainly because there was an increased interest in having a computer in each house.

Today's UMPC appeared only because the public is starting to have interest into something that is small and cheap. With the increase of circuit density, building pocketable devices that have the same power as computers from a couple of years before has been possible for quite long time. PDA have been around for a few years and some have quite decent performance. But the demand only started arising now.

So what will happen in 10 years ?
It all depends on the market then.
The technology will be around that could fit the processing power of today's big cluster into a chip as small as a pen.
But then it all depends of buyers choice. If suddenly pen--sized computer are the latest trends, you'll see them around. Probably with geeks claiming that 2018 will finaly be year of the Linux PenComputer, because Windows 8.0 just can't run on them.

But if UMPC are still the trend, you'll only see the same form factor as before, only with 40x processing power than today - three quarter of which will be taken by a combination of the bloated operating system and the DRM lock mechanisms.

That Depends

US or European? Saftey equipment varies and so do curb weights.

Re:

I would imagine the weight he wanted was the weight of the beetle filled to the maximum amount possible with live clowns. Sometimes the clowns are considered 'airbags' and so may qualify as safety equipment, but your mileage may vary. I don't think he was asking about how much a curb weighs though. that would probably depend on the height, width, length, and composition of the curb piece in question.

Re:and year . . .

The schools these days . . .

let's try it this way:

Mommy hippo weighs 1200 pounds.

Daddy hippo weighs a time and a half as much as Mommy.

How much does Mommy hippo weigh? :)

Re:

Tonne= UK Ton= US

No it isn't. A tonne is 1,000kg everywhere. A ton is 2,240lb in the UK and 2,000lb in the US. I can't believe you'd post something as uninformed as that while logged in.

Logging

Because of the limitations of the display the team tested the machine using prime numbers.

"If you give it a prime number to try then the highest factor of that is one," said Mr Burton.

"If what they saw when they ran the program was a one - in other words a dash when everything else was dots - then bingo they knew it was working."

If only they'd had log4baby they could've tried factoring more interesting numbers.

No, it doesn't run Linux....

not a hope of backporting to this one...

Re:No, it doesn't run Linux....

Yeah, but imagine a beowulf cluster of those. The real estate alone would be staggering.

Re:

I didn't get to welcome any overlords yet!

Not entirely true

It used 24 bit address space, expandable to 32, so although it only physically posessed 32 words of memory, it could easily have supported a modern operating system if the memory had been built for it. And you didn't mind the response times.

Re:

Bill, is that you?

Way to go

"If you give it a prime number to try then the highest factor of that is one," said Mr Burton.

First program EVER and it already had a bug in it.

Re:

He meant the highest factor except for the number itself. Otherwise it doesn't matter what number you put in, the highest factor is always that number.

50th anniversary programming contest in '98

They had a programming contest 10 years ago. A pot-noodle timer won and was loaded on the rebuilt machine in a big celebration.

Read more:

Manchester Celebrates the 50th Anniversary of the First Stored-Program Computer [computer50.org]

The 1998 Programming Competition [computer50.org]

Simulators [computer50.org] so you can try your hand at programming a 60-year-old computer.

Re:Zuse did it first

Zuse's machine didn't have memory, which is part of how they're defining "modern computer."

Re:

Well, according to wilkipedia it had memory

"Improving on the basic Z2 machine, he built the Z3 in 1941. It was a binary 64-bit floating point calculator featuring programmability with loops but without conditional jumps, with memory and a calculation unit based on telephone relays. The telephone relays used in his machines were largely collected from discarded stock."

Re:Zuse did it first

More specifically, it didn't store software in memory.

Re:

The Z3 has a register (the "memory" cited in Wikipedia), but does not have memory for storing a program. (They're using the Von Neuman criteria for defining a computer.)

Re:Zuse did it first

The Baby is set apart from other early machines by two major features.

Memory was what we would call dynamic RAM. The storage element was special CRT's called Williams tubes which were the first all-electronic memory device (flip-flops we not economically viable for storing data). Williams tubes were randomly accessible and, used charges to store bits, and were therefore volatile. The volatile characteristic means that bits had to be refreshed by reading, or they would evaporate due to charge leakage. This is the same reason modern RAM chips have a periodic refresh cycle. This isn't a functional parallel, just a historically interesting one. FWIW, mercury delay lines are volatile, too, but not because of charge leakage. Programs were read into RAM from which they were executed.

The other feature of the Baby which was adopted into subsequent designs was conditional jumps - sort of like goto's. The relative jump is a jump to a calculated address. Without the ability to hop around the program space based on whether statements are evaluated as true or false precludes easy implementation of things like for loops and arrays. In 1998, the Z3 was mathematically proved to be capable of conditional jumps, but this was not an intent in its design and didn't lead anywhere.

The Baby had only seven instructions (take that, Microchip PIC!):

Jump (indirect), Jump Relative (indirect), Load Negative, Store Accumulator, Subtract, Skip if Accumulator < 0, Halt

A very good and hard to find page with info on the Mark I <URL:www4.wittenberg.edu/academics/mathcomp/bjsdir/madmmk1.shtml/>