May 2015 - mathfun@mailman.xmission.com

[math-fun] EM propulsion drive
by Warren D Smith 21 May '15

21 May '15

Turns out Blacklight power has an enjoyable wikipedia page now: http://en.wikipedia.org/wiki/BlackLight_Power And among other details, this news story says they got $1.1 million from the State of New Jersey as a grant in 2013, since they were one of the most promising tech companies in the state: http://www.nj.com/middlesex/index.ssf/2013/12/20_middlesex_companies_receiv… which reminds me of the State of Utah funding a Cold Fusion Institute. Wikipedia claims Blacklight has received over $60M in funding so far. Nothing new here? Old story? Well, maybe. But I think they deserve some sort of hall of fame entry. I could be wrong, but I have a hard time believing many people could pull off this level of scam for this long. I couldn't, anyway.

1 0

Re: [math-fun] EM Propulsion
by Henry Baker 21 May '15

21 May '15

At 03:44 PM 5/20/2015, Eugene Salamin via math-fun wrote: >I believe there is an SEC regulation that allows wealthy investors to buy unregistered stock, on the grounds that they presumably understand their risk. > >If Randy Mills tried selling stock to the general public, indeed he would end up playing golf with Bernie Madoff. I'm not a lawyer, but my understanding of the SEC is that _so long as you disclose everything_ to potential investors, you can do almost anything you want with the money. If you disclose the fact that you're going to take the money, spend it all on a big party, and give the rest to the CEO for a bonus, that's completely OK with the SEC. Indeed, this is precisely what happened with many of the "dotcom" companies in 2000. If you go to the SEC.gov site an actually _read_ the IPO documents (which I often have to do for my living), you will be shocked by how much is really disclosed -- e.g., the most recent Fitbit IPO document. Outside the U.S., these types of disclosure are unheard of. Where Madoff went wrong was in lying to his investors about what he was "investing" the money in; in reality he was running an old-fashioned Ponzi scheme.

2 1

[math-fun] how can FullSimplify be so dumb
by Bill Gosper 20 May '15

20 May '15

In[356]:= FullSimplify[Sqrt[1 - Sin[\[Pi]/7]^2]] Out[356]= Sqrt[1 - Sin[\[Pi]/7]^2] and so smart? Out[353]= -1 - 2 Sin[\[Pi]/14] + 6 Sin[(3 \[Pi])/14] == 2 Sqrt[7] Sin[\[Pi]/7] In[354]:= FullSimplify[%] Out[354]= True --rwg

1 0

[math-fun] EM Propulsion
by Warren D Smith 20 May '15

20 May '15

The "Blacklight power" place was located down the road from where I used to work, and they got millions of dollars in grants from gullible power companies to develop their 110% bullshit, quack, awesome power source, which any competent sophomore could have told you was a complete fabrication. They still are on the internet even 20+ years later http://www.blacklightpower.com/ apparently still claiming to be a wonderfully successful innovative corporation. And why are they not in jail? My guess is that making utterly false claims of scientific breakthroughs, and asking for huge money for "future rights" to it, and getting it, all is not a crime? -- Warren D. Smith http://RangeVoting.org <-- add your endorsement (by clicking "endorse" as 1st step)

2 1

[math-fun] EM propulsion drive
by Stuart Anderson 20 May '15

20 May '15

http://www.nasaspaceflight.com/2015/04/evaluating-nasas-futuristic-em-drive/ A group at NASA’s Johnson Space Center has successfully tested an electromagnetic (EM) propulsion drive in a vacuum – a major breakthrough for a multi-year international effort comprising several competing research teams. Thrust measurements of the EM Drive defy classical physics’ expectations that such a closed (microwave) cavity should be unusable for space propulsion because of the law of conservation of momentum.

3 2

Re: [math-fun] "differential" in sports writing
by Dan Asimov 19 May '15

19 May '15

Isn't the usual jargon "point spread" ? --Dan Thane wrote: ----- I'm no fan, but how to reformulate this one without getting all wordy? "Golden State had a higher point differential vs. Memphis than it did vs. the Pelicans, despite losing two games to the Grizzlies." -----

8 10

Re: [math-fun] "differential" in sports writing
by Henry Baker 19 May '15

19 May '15

"Never accelerate when making a turn." (Or decelerate, either.) Due to my brother's participation in amateur auto racing, I studied some of this math out of curiosity. What the driver's ed teacher meant, but almost certainly couldn't put into proper mathematical language, was: Keep the _absolute value_ of the acceleration vector constant -- or at least < some bound. This is a very crude first-order estimate of the stickiness of the tires; they can't be expected to handle arbitrary accelerations in multiple directions at the same time. To aid a newbie driver, there are now accelerometers that have a little circle and a dot within the circle showing the current acceleration vector. The idea, of course, is to keep the dot within the circle. The best racing drivers can keep the dot _on_ the circle at nearly all times (this, of course, is course-dependent and auto-dependent). This is much harder than it seems, because the racing driver must also keep the suspension "loaded", otherwise there are unfortunate interactions between the suspension and the tires. Also, the moment of inertia of the race car itself about its vertical axis is non-negligible. A second-order approximation would utilize a non-circle, but the uncertainties are large enough that a simple circle is good enough. In 3 dimensions, the "constant absolute value acceleration" model is good for a solid fuel rocket, which can only be steered, but not throttled. I wanted to utilize the position curves as "splines" to try to piece together fastest-possible trajectories around a race course. I wasn't able to find closed-form solutions except for the most trivial cases. (I wasn't even trying to handle the 3D case, where the race car must climb or descend -- e.g., Laguna Seca, outside of Carmel, CA.) I can't say that I ever really conquered the math; there are a lot of sinh's and cosh's involved. Perhaps someone here has better links & references than I was able to find. At 09:28 AM 5/19/2015, James Propp wrote: >My driver's ed teacher taught us "Never accelerate when making a turn." > >I objected that, in the physics sense, it is impossible make a turn without accelerating. > >The teacher, who taught gym when he wasn't teaching driver's ed, wasn't amused.

2 1

[math-fun] Wine Klein Bottle
by Henry Baker 19 May '15

19 May '15

FYI -- $95 (about par for wine decanters), NFI -- http://www.kleinbottle.com/wine_bottle_klein_bottle.html

2 1

[math-fun] "differential" in sports writing
by Warren D Smith 19 May '15

19 May '15

> I'm no fan, but how to reformulate this one without getting all wordy? > > "Golden State had a higher point differential vs. Memphis than it did vs. > the Pelicans, despite losing two games to the Grizzlies." --Even though Golden State lost 2 basketball games to Memphis, that in no way stopped it from parking its bus in such a way that its rear axle was located 95 inches above Memphis's, and 53 inches above the Pelican's. This height ordering remained valid even after each bus collapsed into 3 separate Schwarzschild singularities.

1 0

[math-fun] Replicoids
by Warren D Smith 19 May '15

19 May '15

> So P only proves that the machine it runs on is a UTM and it also verifies > that it is > correct. But what about the AI program that you want to run on the UTM to > operate your > robot? If it can do mathematical induction it will run into Goedel's > incompleteness. Right? > > Brent --I don't really know what "run into Godel incompleteness" means. The most obvious problem if trying to write something like "an AI" as a proven-correct-program, is what does "correct" even mean? What statement is one trying to prove? But for many subprograms that would be no obstacle. And obviously these programs would be being written by other programs, all smarter than me... But to return to the world I understand... What did Godel prove? Really, his result was basically just Turing's theorem "there exist Turing machines for which it is undecidable whether they halt" all dressed up in math. I.e. you encode the machines as statements in number theory, etc and it transmutes to "there exist statements in number theory which are true, but for which there is no proof." (Historically it did not happen that way, but if it had it would have been simpler.) So what? There exist plenty of programs, which cannot be proven correct, even though they are. So what? If you wanted to fulfil my vision of writing an AI (and it writing better versions of itself, which in turn do so again...) with the programs coming with correctness proofs, then you would not use programs in that nasty class. You would only write programs in the nice subclass of programs that ARE provably correct. And for which you actually have a proof. And this is quite a wide class. Some examples to help illustrate how wide a class it is: By hooking up a counter to a (verified) proof-verifier, you would get a provably correct program that would output every mathematical theorem, with proof. Or would output every provably correct computer program. Or would find a proof or disproof of the Riemann hypothesis (if a proof or disproof exists) otherwise run forever. Another provably correct program exists, which emulates every neuron inside Brent Meeker. So, I mean, the subclass of provably-correct programs is quite a large one. It certainly is large enough to include programs that (if run on a fast enough computer) would pass a "Turing test." Penrose in ENM had nutty notions that since humans solve instances of undecidable problems, that "proves" their neurons work using quantum gravity, and "proves" human brains have super-Turing power and cannot be emulated by Turing machines and "proves" quantum gravity (whatever it is) is beyond understanding by a Turing machine. That all is complete garbage. Penrose simply did not understand the basics of computer science dating from the 1920s or 1930s and made a fool of himself. Because he was such a famous genius, nobody in the publishing house tried to stop him. If I'd tried the same thing (which I wouldn't have since I am not that incompetent) then my manuscript would hopefully have been savagely reviewed and rejected. If you want a little story about the dysfunctionality of science, and/or me: I actually did once write a paper demonstrating that the Navier Stokes equations would (if taken seriously as "laws of physics") enable a physical system to solve the general halting problem in 1 hour. Or certain other equally remarkable consequences. My paper was correct. It was, however, rejected, by a referee who, like Penrose, did not understand what undecidability even was. But like Penrose, he thought he understood it. I then got a computer scientist expert in the decidability area, who'd written a book on that topic and whose office was right down the hall from the editor of the journal, to write a letter saying "referee=idiot, Smith=correct" to that editor. But this failed; my paper was never accepted. The editor wrote to me saying my anonymous referee was a Truly Great Expert and could not be contradicted by any mere computer scientist down the hall from him. The Navier Stokes equations are, however, not the true laws of physics obeyed by real fluids, and that very much matters. Real fluids are made of atoms, for example. Indeed, one can write down truer laws than the NS equations, and prove they are simulable by a Turing machine, even though the NS equations are not.

1 0