dietz@SLB-DOLL.CSNET (Paul Dietz) (02/12/86)
> 1. SPSs. Not even Gold will argue that SPSs won't be important within 20 > years. Now most feasible SPS designs call for solar panels on the order of > kilometers across; you aren't going to ship that up in one shot. I'm sure he would argue that. Gold has a radical hypothesis that huge amounts of abiogenic methane are trapped deep in the earth, and believes there may be huge amounts of geopressured methane in the deep crust. If true, this would mean fossil fuels could supply us with energy for generations. Also, the NAS study of an earth-launched SPS concluded that it wouldn't be competitive with more conventional power sources, even if launch costs from earth are reduced by orders of magnitude and cheap, lightweight, efficient GaAs solar cells can be developed. Face it, no one is going to build SPS unless it promises *much* lower cost energy: it's just too risky. A cheap SPS needs cheap extraterrestrial material, which means an automated or remote controlled lunar mine. > 2. The Space Station. All right, Gold and Van Allen probably think it's a > boondoggle, but as a platform for launching planetary probes it can't be > beaten. I suggest that empty space would make just as good a launching platform. The only sorts of manipulations you would reasonably perform on a probe in orbit are things like mating together prefab modules and fuel tanks or loading fuel (chemical or nuclear); that could be done just as well without a space station. > Further, I don't know of any geologist who would be prepared to > argue that he could learn more from a series of unmanned probes in LEO than > a fully-equipped lab manned by geologists in LEO. But could we learn more from 8 billion dollars worth of unmanned scientific instruments in LEO, or from a few hundred million $ in instruments attached to an $8 billion space station? It's not obvious that the much greater expense of the space station buys you much. > Hence we should > expect that the Space Station and SPSs will be constructed primarily by > humans; and that in turn means we need to send more shuttles up now. Rick, I think you're being selectively myopic in your projections of future technology. By 25 years from now computers should have as many bits of RAM as the human brain has synapses (assuming costs continue to go down by a factor of 4 every 3 years or even somewhat slower). Logic chips should be housing 10**8's of transistors. If ESA thinks teleoperators can be made now then what will the situation be when computing power is 100 to 1000 times (or more) cheaper? Building an SPS in space from earth launched material needs launchers to LEO at around $10/lb, and even then it's just competitive with earth-based power supplies. I don't know anyone predicting we'll have launchers this cheap in 25 years. The only reasonable scheme for exploiting lunar resources I've seen needs teleoperators on the moon, otherwise the system requires far too much mass to be imported from earth. Your argument seems to be: there's a lot that should be done in space RIGHT NOW, and there's no time to develop teleoperators. The counterargument is that there's little that can be reasonably (read: ecomonically) done in space right now, or even soon (until somewhat cheaper launchers are developed), and any big project in space will be very expensive, so why not use some of the funds to develop teleoperators?
dietz@SLB-DOLL.CSNET (Paul Dietz) (02/13/86)
>(1) It wasn't NSA, it was OTA, that did the SPS study, and it showed that >SPS would breakeven in 25 years. I believe it was NAS (not NSA) that did the study. Even under very favorable assumptions (low interest rates, $10/lb to orbit launchers) SPS was only just competitive. NASA has no plan at all to make a $10/lb to orbit launcher anytime soon (certainly not in 25 years). Was that breakeven point the PAYBACK time on the construction costs? >(2) I had heard that story about Gold, but I'd hesitated to mention it: >it seemed to much like ad hominem criticism "Gold is the nut that thinks >that the earth's a large methane ball..." The theory is certainly controversial, but that doesn't mean it's wrong (do not take this as an endorsement of the theory). See the Feb. 86 issue of The Atlantic magazine for a (biased) review of Gold's ideas. Wrong or not, because of this (and because of the generally poor reputation of SPS in the non-rabid-space-fan community) Gold would, I think, disagree that SPS will be important in 20 years. >(3) I would be very surprised if we could get 10^8 transistors on a (~) >1 cm x 1 cm chip. At 1 micron cmos, we can get about 5E5; to get to >1E8, we'd need to go to about 10 nanometer feature size, with appropriate >scaling on things like well size. At that point, quantum effects start >to really screw you. The Japanese think they can push DRAM technology to at least 64 megabits per chip, using smaller feature sizes, trenched capacitors and vertical cells, and other tricks. A review article on semiconductor trends in Science a few years ago predicted a quarter billion transistors on a chip by the end of the century. That's less than 25 years away, so I'm being conservative. >(4) AI hasn't done appreciably better in the decade 1975-85, which saw >a hundredfold increase in logic chip densities. Why do you think that's >going to change? Where did I mention AI? >(5) Manipulators aren't that light. I'm not at all sure that >human + life-support + fail-safe is all that much heavier than even a dumb >teleoperator. The arms are fairly light, and can be built of graphite composites (say) for space applications. In space the arms need not support large static loads, giving further weight savings. How much does the shuttle arm weigh, for example? I think you really must provide some justification on this one. >My argument is that there's a lot waiting for us in space, and there will be >more in the future. For that future, we SHOULD develop teleoperators of >the sort you've mentioned, and of the sort of I've mentioned in messages to >you. We should also work, hard, on reducing costs to orbit: I agree that >that's the primary task. We should also develop an manned infrastructure in >space so that we'll be able to do those things I've mentioned, whether the >teleoperators are here or not. In a world of unlimited resources, we could do this. Resources are restricted, we have to focus on the critical paths. Putting men in space now is not, in my opinion, on the critical path. > Hence the question >is, do we get more for $8G of manned spaceflight or a much smaller amount >of spending on other research? I think the answer is clear. What you mean "research"? Money spent on manned spaceflight is (now) operating costs, and doesn't produce much in the way of new scientific or technical information. >Your argument and Gold's is based on a false premise, anyway: namely, that >there's some choice between manned and unmanned programs. I'll remember that as I watch the TV pictures coming back from the NASA Halley probe (whether such probes have been well directed is also a good question; complex probes to the outer planets are probably not needed for space exploitation anytime soon). I don't believe in throwing money at space. I believe in spending money to achieve good solutions to well thought-out goals. I don't think a space station fits this description. Research and development of teleoperators would.