Thursday, May 04, 2006

The Cost of Big Science

A panel at the National Academy of Sciences suggests that the US should support Fermilab's bid to land the International Linear Collider.

Linear Collider. Ballpark Cost: $4 Billion to $10 Billion
via Chicago Sun-Times

As a scientist I should support such a project, especially one located in the suburbs of Chicago. But at what cost? As a perspective, next year's proposed budget for the entire National Science Foundation is just over $6 billion.

This ILC reminds me of the Superconducting Super Collider, a project that spent $2 billion dollars digging a hole in Texas that was killed by Congress in 1993 once the projected costs topped $12 billion.

Putting large dollars into a single basket will take away the incentive to increase basic research funding in other scientific endeavors. The main argument for the ILC at Fermilab is not that the research won't get done, it just won't get done in the US. So let the Europeans or the Japanese have the flashy expensive collider and let the US do what it does best—basic research advancing science over a large range of disciplines.


  1. Here's a different proposal: get rid of the Space Shuttle; then a divert a small portion of the ~$8 billion per year pissed away on that to the ILC, and the rest to space probes, small-scale basic research, and student travel to complexity conferences.

  2. Yet another one: acknowledge Irak was a mistake, withdraw all troops, save $100000 per minute, and let all complexity conferences be free for a year.

  3. Complexity theorists will be reassured to learn that the success of the International Linear Collider is not orthogonal to important algorithmic complexity issues.

    For the collider to work, that 28-mile-long particle beam needs to collide with itself within a spatial tolerance of a few nanometers.

    The algorithmic challenge is this: on a scale of nanometers, the earth is made of vibrating jello.

    So making the collider work is a major challenge in control theory ... not just optimal control theory, but robust optimal control theory. Specifically, the near-optimal yet robust control of systems whose parameters are known only within ranges.

    Many (even most?) such control problems are known to be in class NP-hard.

    So the ILC is a project within which applied complexity theorists (if there are such beasts) have a great deal to offer the physics community.

  4. It has been said that beauty lies in simplicity. One then is forced to wonder about the name computational complexity theory. If the techniques used were really beautiful, it would probably have been called computational simplicity theory....

  5. Ahh, but is it not said that within the heart of complexity lies a deeper simplicity?

    (Actually I don't know if it's said or not. But it sounds good.)

  6. Maybe if it were called "simplicity theory" then we'd have to call the those who study the subject something related...simpletons?

  7. Hi, I read your blog practically 100 times a day, if possible. I was just going through some of your old posts and I feel the way you wrote has evolved. Initially posts used to be less assertive, and about 80% of them technical/professional. Ofcourse less people commented, because blogging culture was not so well developed. This doesn't fit as a comment to this blog, but anyhow. I really like your blog and I hope your writing style keeps evolving for ever.

  8. ya iraq would be the big problem here. not a supercollider.

  9. These large scale colliders come at increased costs and reduced results. In other words the bang for the buck is going down.

    I also believe that if these scientists were starved from multibillion dollar funds, after a while they would find easier ways to produce the same results.

    Is discovering some arcane particle worth the combined math and computer science budget several times over? most certainly not.

    and yes, we should be out of Iraq too, and ground the ISS but that's a whole different bag of worms.

  10. Though I don't know enough about the projects that might be funded as alternatives to the ILC, I, too, wish that other worthy projects would be funded as alternatives to the Space Shuttle program.

    There are at least a couple of kickers I can think of, though. Would multiple smaller alternative projects actually be funded if these bigger items were killed, or do projects like the Space Shuttle function as Great Attractors, keeping the agencies where they're housed in the minds of those in the White House and Congress who draw up budgets? And what is the role of the Department of Defense in keeping the Shuttle alive? (Can the scientific community convince the DoD that there's something in the ILC for them? "Yah, should allow us to develop a really cool intercontinental death ray in about 5 years.")

  11. The DARPA Grand Challenge attracted a fair deal of media attention. Likely much more than a collider ever would.

  12. This week The Onion has a very clever article titled NASA Announces Plan To Launch $700 Million Into Space. As Homer Simpson said, it's funny because it's true. I liked the part about performing experiments on the effects of microgravity on $100 bills.

  13. Any physicist reading this?

    The NYT article talks about great possible insights. How great? How likely? How much of string theory alternatives is it likely to settle?

    The impact of large projects is hard to judge. It syphons money from worthwhile small ones, but it also focuses attention on spending money on science, which is a good thing. It might inspire young people to become scientists, and might even loosen purse strings (after all we would ask for small fractions of what the big project costs...)

    As a thought experiment: if the collider were to yield an experiment like Michelson-Morley, it would (in
    my opinion) be worth a couple of billion. The objections to the Space Shuttle are exactly that it is unlikely to produce anything.

    Here's another thought experiment: suppose Complexity Theory became VERY rich. All reasonable research projects funded, all deserving graduate students lavishly supported, conferences subsdized, etc. What promises of major breakthroughs could we offer for this? Feel free to speculate wildly!