You underestimate the importance of HTML in creating the web.

Particle physics had progressed so fast since the 1940's that the particle physics community had developed a system of "preprints" in which people circulated drafts of their papers to colleagues at their institutions and others months before they were published in journals. The story goes that Tim Berners Lee got tired of e-mailing documents back and forth to colleagues at CERN and decided to invent HTML and code a bare bones browser to allow him to (we would today way) webcast his research. There is something about the pace of information exchange within CERN and in the particle physics community that supports the idea that HTML might have taken 5 more years to be developed elsewhere (and it would have been some university or other: USENET and the text-based tools to go with it, and GOPHER, developed in that environment).

The large particle physics laboratories do employ thousands of physicists, engineers and programmers specifically for particle physics experiments purposes, and that is a nonnegligeable fraction of the respective academic communities. If the large labs didn't exist these people would be competing for academic jobs elsewhere and it would result in more people going to industry, as well as fewer people getting doctorates.

If LHC funding hadn't gone through, CERN would have stagnated and maybe shrunk. You need far fewer people to run the existing facilities than you do to develop a new facility, and the LHC research programme is much more intense that what can be carried out at the existinc facilities (not that that isn't useful, too, but it's on a smaller scale in terms of people and resources).

Consider CERN and the LHC a Keynesian stimulus package for physics and engineering.

We have met the enemy, and he is us — Pogo

Three related reasons: Sputnik angst, there was a race with the Soviet Union in every field, and areospace technology development for military purposes.

But the fact is, the Apollo program was a one-shot thing. It was wound down and the US lost its ability to fly to the moon. It also discontinued its high-payload rockets in favour of the Space Shuttle, so now the rocket market is occupied by the European Ariane and the Russian Proton.

The Soviet manned space program made more scientific and technical sense than the American one, and the ISS owes more to the Russian Soyuz and Mir than to the American Skylab, which was also discontinued and folded into the Shuttle.

We have met the enemy, and he is us — Pogo

(1) Spend money on chip research for what? What would the chips be used for? The great proliferation of electronics came on the back of very advanced requirements for components for space programs, etc. One could argue that only once they had been developed for such purposes were it possible to consider their use for more mundane matters. The personal computer only became possible with a maturation of integrated circuit technology, computation infrastructure, and computational techniques that allowed for cheap mass manufacture. The driver for this technology were expensive research programmes in fields requiring processing of large data sets, such as say high energy physics research. Forget about the direct spinoffs, I would argue that the influences of these expensive programmes are far more subtile.  Technological diffusion requires that the basic building blocks are already lying about looking for a different application. You don't start developing processor technology because you think that in 20 years you'll be able to make a machine to play games on.

(2) Missile programmes you say? Because military programmes with large destructive potential are soooo useful while high energy physics, space exploration and the like are vanity projects! And you know how the military loves to share the technology it develops and would not like to keep it secret. One of the great advantages of the large lab high energy physics environment is exactly that it is not a military programme. We don't build things to kill people, I think this is a plus. Further, there is not a tendency to keep progress secret. Quite the opposite in fact, thus a greater chance that progress made here can defuse faster and wider.

(Disclosure, I work at CERN.)

It would be interesting to compare commercial to military / government revenues over time. I should study the topic further because it occurs at the intesection of several topics I'm interested in.

The early commercial viability of mainframes is a good point that I managed to forget. I'll still make my vague 20 year claim, though.

I agree that it all happened shockingly fast.

Most of what's happened has been faster and cheaper, but not so truly innovative.

I disagree. Reading IEEE mags since I became an EE major in college, there has been some stunning work over the years in the semiconductor physics realm that has been required to get the commercially viable transistor sizes we have today. From the computing point of view, though, I agree what you're saying.

you are the media you consume.

An interesting episode in the killing of the SSC was how, during the Congressional hearings, a prominent physicists (someone big, like John A. Wheeler or Murray Gell-Mann or Steven Weinberg) was asked by a congressman whether the LHC would provide evidence of the existence of God. The negative answer did not help.

We have met the enemy, and he is us — Pogo