Hacker News Reader

Comment by juiceandjuice | Hacker News Reader

I'd like to mention that this sort of research at SLAC and other Office of Science facilities and programs in the DOE, is being targeted for heavy cutbacks. A bill passed the house proposing a 20% cut ($1Bn) from the Office of Science budget, but the consensus is that it won't pass the senate or the president.

More extreme examples come from people like Rand Paul who have proposed eliminating the DOE completely.

View on HN

I agree that this kind of research is needed, but surely you aren't suggesting that the DOE is the only entity capable of investing in research of this kind?

Investments in research which have a time to profitability longer than 20 years, which is the length of the patent term in the US and most of the world, cannot be supplied by for-profit companies. That pretty much leaves government or philanthropic entities. Philanthropic entities do not invest at the level justified by the expected benefit to society of the research. If you want funding at the appropriate level, you need to either do it through the government or you need to modify patent law to enable corporations to capture profits with very large time horizons.

Is there a reason that the DOE is a particularly bad agency, among potential government agencies, for doing this investment?

Wouldn't it be better to fund this through NSF so that it could be part of a cohesive portfolio? Instead of the through the DOE which adds massive beurocratic overhead/duplication of effort. Wouldn't one science funding research body be sufficient?

Also, as for nuclear weapons research wouldn't that be best carried out by the dod, perhaps throug DARPA?

There's a lot of things wrong with what you are assuming, and assuming that the DOE adds massive bureaucratic overhead is a bit naive, frankly.

The NSF doesn't have a physical presence, and given that particle physics is a very natural extension of nuclear physics, that's why the DOE does particle physics (which includes lightsources). DARPA is purposefully limited in scope, they also have very few facilities (really, they are sort of like the NSF for weapons). Nuclear Weapons research is a small (but very important) part of the DOE's mission, but weapons and reactors are so closely tied together that it's really impossible to separate the too. There's certain levels of civilian and military separation that have been put into place by law. Many of the labs do weapons (usually conventional, only two do nuclear) research, many do only science, some do both, some do reactors and some do accelerators. The DOE services everyone though, defense, civilian, academic, everything, and that's largely why they are there and set up the way they are. It sort of ends up being this buffer zone between pure science and defense, and that's generally a good thing I think.

Just to add to what juiceandjuice said: I get the impression you are arguing based off of the names of the various government agencies. This is silly; the names are historical. What's important is whether the agencies are organized properly to bring the people with the correct skills to bear on the particular investment problems. I have zero expertise in how large agencies should be organized, but I can assure you that the DOE has decades of experience funding all sorts of physics research. Moving funding jurisdiction to another agency would be an expensive shift with no obvious benefit.

Let's say IBM wants to build a huge free electron laser, and synchrotron light source, maybe mimic SLAC.

NSLS-II, currently being built at BNL, has a total project cost of $912 million. LCLS, at SLAC, has a total project cost close to $350 million or so, but it's been estimated that they saved several hundred million reusing previous facilities from the linac (plus land) So, $1bn to build, plus operational costs which would be pricy. And you want one of each because a FEL is different than a synchrotron.

Now, for your $2bn+ops costs, you can service maybe 500 experiments a year, maybe a bit more, but less than the demand. At the labs, proposals for experiments that are generally picked are picked based off of merit, usefulness, promise, etc... not necessarily based on who can pay to use the facilities and what sort of profits could be tied to the research. You know... science, often done by tons of professors who get their money from the government as well.

Admittedly, the budget hits probably wouldn't affect the operations of these facilities too much, but they would definitely affect the people who propose to use these facilities. For example, the guys who did this research probably don't get their funding out of the operating costs of the SSRL.

So, yes, the government is the only entity that has invested this kind of money (largely through the DOE) to be able to do this on the scale that they do. I don't see any company picking up the slack anytime soon.

Lists of light sources: http://en.wikipedia.org/wiki/List_of_synchrotron_radiation_f...

http://sbfel3.ucsb.edu/www/fel_table.html

What's the right way to account for opportunity costs? How do we know that $2 billion spent here is better than investing $100,000 in seed money for 20,000 startups?

Room temperature superconductors are cooler (hah!) than 20,000 social networking photo apps.

So you're saying that social networking is really hot right now? Thanks for the tip! ;-)

If you question the government's ability to decide if it should go to hard science or 20K startups, then how can you trust their ability to choose which 20K startups?

In general, if you don't trust the government's ability to make a decision, then how can you trust their ability to delegate those decisions? This is a decision in and of itself.

>the DOE is the only entity capable of investing in research of this kind?

who else other than government has this kind of money for such fundamental research? (i.e. a lot of money without any feasible productization in the next at least 10 years) May be IBM Almaden comes closest, and yet it is still very far.

Reply on news.ycombinator.com