Alternate historyPublication date: 15 September 2009
Originally published 2009 in Atomic: Maximum Power Computing
Last modified 03-Dec-2011.
When 3dfx started showing off the original Voodoo Graphics 3D accelerator card, many people thought the trade-show demonstrations had to be cheating, with an expensive hidden 3D workstation actually producing the graphics. (Ten years earlier, people had thought the same thing about the Amiga.)
But 3dfx's Voodoo Graphics card was real, and it was a few long steps ahead of anything 3dfx's competitors could manage.
If 3dfx had never existed, though, we'd still have 3D graphics cards today. They wouldn't necessarily be much slower than they are now, either. Almost-true-3D games like Doom and Duke Nukem 3D had already sold zillions of copies by the time the 3dfx cards arrived, and the original Quake came out the same year as the first Voodoo. And there were a hatful of other graphics-card companies - PowerVR, Rendition, S3, Matrox, and the obviously-never-going-to-be-major-players Nvidia and ATI - all beavering away at 3D accelerators for arcade games and home PCs.
This is how technology often develops. If someone sees a demand, they're usually not the only one to see it.
The precise sequence in which people come up with inventions, though, can sometimes make a very big difference.
Take trains, for instance.
Trains have some great strengths, but they're not good at climbing hills. Robert Stephenson didn't think a locomotive would ever be able to pull a load up a grade steeper than 1 in 100. Later locos proved him wrong, but unless they're fitted with cog-drive contraptions, even modern trains still need pretty flat tracks. A locomotive's pulling power is limited not by its engine-power or gearing, but by the small contact patches between the steel driven wheels and the steel rails. Steel-on-steel gives you a lot less traction than rubber-on-tarmac - or possibly even rubber-on-dirt.
Once the rails are built, a railway is a very economical way of moving people and goods. But the building stage is a nightmare. Laying flat tracks across Europe and the USA was a very, very serious undertaking.
When a railway crosses a river you need a bridge. Where it meets a mountain range you need cuttings, tunnels and probably a lot more bridges. The US railways had to cover huge distances, and cut costs by building lots and lots of wooden trestle bridges instead of sturdier, less flammable masonry or iron bridges, but every process for laying rails across a country was ruinously expensive.
(Around this time, there was a speculative frenzy in Britain called "Railway Mania". Shares in railways-under-construction traded far above any sane estimate of their value, and lots of shares were sold in completely fraudulent ventures. The most obviously fake proposed railways were the ones that promised long, as-the-crow-flies runs of track through roller-coaster countryside. They would have needed a dozen or two Millau Viaducts, at the very least.)
A lot of cuttings and tunnels and bridges could have been avoided, or at least been a great deal simpler and less expensive, if someone had come up with motorised road transport - trucks - before they invented trains.
Exactly one invention made trucks, and automobiles, possible: The pneumatic tyre.
There were, you see, roadgoing "steam carriages" long before the internal-combustion engine became practical. People were making steam carriages very early in the history of the railway - only shortly after someone figured out that locomotives probably shouldn't have legs. There's no reason why a steam-powered truck can't work well; steam cars were quite popular in the early days of the automobile, and a big heavy-duty steam truck with a trailer or three could run on widely-available fuels like coal or wood, just like a steam train, instead of refined petroleum.
(Steam cars generally did run on refined petroleum, because liquid fuel is easy to deliver to a burner. It's not practical to have to shovel coal into a firebox to keep your car going, but a huge steam truck could work OK that way.)
Early steam carriages, though, all ran on ordinary wooden cartwheels, with solid iron or steel hoop "tyres" shrunk onto them for durability, not traction. Cartwheels with or without metal tyres are only moderately dangerous when your carriage is pulled by sure-footed animals. If you start trying to get a vehicle to go, stop and turn entirely through its wheels, though, the miserably low strength and adhesion of those sorts of wheels puts very serious limits on your steam carriage's performance, and safety.
(Steam carriages also had a distressing tendency to explode, not least because boiler construction and operation was still a very young art. As everybody who's read J. E. Gordon's classic works on strong materials and structures knows, it actually took engineers a while to figure out that box-shaped boilers were a bad idea. Stationary, shipboard and locomotive boilers also didn't have to put up with rough treatment; jiggling a questionably-constructed boiler around on top of cartwheels running on rough roads is a good way to make it angry. But the same technological advancements that made locomotive and ship explosions a rarity would have worked on steam-carriages too, if only they had decent tyres.)
Solid rubber tyres had existed for years by the time the steam carriage was invented, but solid rubber tyres are scarcely better than steel-rimmed cartwheels for a vehicle with a built-in engine, as opposed to one that's pulled by animals. They protect the road from damage, not the vehicle and passengers.
The basic cartwheel structure actually hung on well into the early automobile days. The original Ford Model T ran on wooden "artillery wheels", which are essentially just cartwheels upgraded to tolerate lateral loads. But the Model T had pneumatic tyres, which make all the difference.
A fellow called Robert William Thomson invented a rather impractical pneumatic tyre in 1846. His "Aerial Wheels" were a step in the right direction, and apparently even quite durable, with the critical advance of a rubber inner tube filled with compressed air, but with an outer casing made of leather. The Aerial Wheel wasn't meant for wheel-driven vehicles, though, because those barely even existed in 1846. The first pedal-driven "boneshaker" bicycles, which took their name from their steel-shod suspension-less design, were still 20 years away, and the famous penny-farthing arrived about another decade later. Thomson intended the Aerial Wheels only to make horse-drawn vehicles ride more smoothly and make a lot less noise, and they apparently achieved this very well.
All of the pieces were in place for proper pneumatic tyres in the 1840s, but nobody put them together. If someone like Thomson had constructed an inner-tube tyre with an outer layer made from tough rubber instead of leather, the job would have been done. You need to vulcanise rubber to make it tough enough to work as tyre for powered vehicles, but Charles Goodyear figured out how to do that around 1840. News didn't move as fast then as it does now, but even if it took a few years for someone to learn about the Aerial Wheel and vulcanised rubber, and a few more years to come up with a working product, you'd think there would have been a rubber pneumatic tyre ready to be displayed at the Great Exhibition of 1851.
But it was actually not until 1887, some 41 years after Thomson's Aerial Wheel, before John Boyd Dunlop invented the thin-walled inflatable bicycle tyre that led to proper car tyres.
(The final part of the recipe for the "modern" tyre is internal reinforcement to stiffen the tyre and stop it from, for instance, folding over on hard cornering. It took until 1904 for the first cord-reinforced tyres to hit the market. The modern "steel-belted radial" tyre has steel-wire reinforcements that're wound at right angles to the tyre's direction of travel, which has a number of benefits, including better road traction and lower rolling resistance. Older "bias-ply" tyres have only diagonal windings, which makes them less rigid; this is actually a good thing, for low-speed off-road motoring.)
So far as I can see, there really wasn't much stopping someone from coming up with a quite effective pneumatic-tyred steam carriage in, say, 1850, when the USA only had a twentieth of the railways it actually sprouted by 1890. If this had happened, a lot of passengers and goods would have ended up being moved via steam-truck on cheap roads, rather than via iron-horse on expensive rails.
And if a tyre-inventor had put their thinking caps on a few decades earlier still, the train may never have taken off at all. We'd doubtless still have some sort of trains today, but we probably wouldn't have nearly as many railways, and the 19th century would have been very different. It would have given an easier outlet to the desperate pressure to have some way, any way, to move great quantities of goods and people overland. Nobody would have bothered spending fortunes to cut tunnels through mountains when they could just run a road over them.
As it was, the railways stole a march on the roadways. By the time someone made a car that was steered with a wheel instead of a tiller - but which still ran on bloody cartwheels - the world already had enough railways to reach to the moon.
In computing, there was a critical trains-or-trucks moment in 1981. That was the year when IBM made their original "Personal Computer", and went completely against their nature by giving the new PC an "open architecture". Mere months later, Compaq made the first "clone" PC, and the rest was history.
If IBM hadn't made this strangely charitable move, the PC platform simply wouldn't have existed. Perhaps that would have made some kind of Apple the global standard; perhaps Acorn or Commodore or, heck, Sinclair or someone would have ended up dominating. Perhaps by now there'd still be five different equally-popular platforms, so economies of scale didn't apply as much and a decent desktop computer would still cost $5000.
Could be worse, though. You could be trying to get to a LAN party in a steam carriage.