OK, a virus is subcellular so it is biological but not really life, or perhaps it is a peculiar kind of parasite of symbiont.
I am wondering if a virus is really a separate entity or is really intrinsically a property of the host. Is a cold virus really just that, or is it a piece of human genetic machinery that has the capability of being shared between humans when one human picks their nose?
The reason I got to wondering is that it seems diseases that stick around have some kind of evolutionary purpose. Sickle cell anemia gene confers resistance to malignant malaria, cystic fibrosis gene confers resistance to cholera, and so on. Sexual reproduction was a big evolutionary breakthrough in allowing diversity in the offspring to adapt to environmental pressure.
Do you suppose a virus is a remnant of some other mechanism, such as conjugation where bacteria can swap genetic material? If a virus served no evolutionary purpose, it seems that individuals who were virus resistant would have a slight edge and over many generations viruses would die out. The fact that they are still here suggests that the cell mechanisms that support virus reproduction serve some other purpose than making a person sick.
The main issue I have with these hybrid drive systems is the cost of electricity extracted from a battery. Pulling some electric car numbers out of the air, a $3000 lead-acid battery pack stores 10 kWHr per charge times 300-500 charges or 60 cents to a dollar per kilowatt hour. Lets say a gasoline engine averages 1 lb fuel/hp-hour (it might peak at.45 lb fuel/hp-hour at peak load and optimum RPM). Lets say $2/gallon gas (yeah, yeah, I know people in the world pay more, but that is mainly tax policy -- the reason I ascribe dollars and cents costs is to keep track of all the hidden energy costs so we don't end up with the ethanol boondoggle that can use more fossil energy as input than it puts out). Then you are talking perhaps 40 cents per kWHr.
This means that even if increased the gas engine efficiency to 0 cents per kWHr (through smaller engine run at peak efficiency to only charge batteries) but ran all the power flows through the lead acid batteries, you costs would increase from 40 cents to 60 cents per kWHr delivered to the wheels.
Anyway, this argument is for series power flows. The traditional automatic transmission is a series power flow, at best 80 percent efficient. The automotive industry has latched on to parallel power flows as a way to boost efficiency. Modern automatic transmissions use split-torque direct drive and torque converter together with direct-drive clutches to boost this efficiency. The commercial hybrids today are parallel electric and gas engine drive. The Toyota Prius is gear shift free, but instead of the Diesel locomotive series drive of engine-generator-traction motor, the Prius has the gas engine, generator, and traction motor tied to a planetary gear set so there are parallel mechanical and electric torque paths, again to boost the efficiency.
If you have a parallel hybrid, obviously you need to use your 60 cents/kWHr battery electricity under circumstances where the gas engine electricity costs more than 60 cents/kWHr, or perhaps use it to resize the gas engine for greater efficiency while consuming the minimum of expensive 60 cents/kWHr juice. But given that batteries are 1) expensive, and 2) wear out (think laptop or cellphone or iPod battery), the hybrid vehicle is not a slam dunk.
Back in the days of PDP-11's used as lab data acquisition computers, there were these array processor boards use for FFT's and related calculations. When PC's came along, there was a generation of "array processor" board products based on DSP chips. Some were floating point, some fixed point, some were only accessible through a numeric subroutine library, others were programmable in hex.
My favorite was the IBM/Tecmar M-ACPA sound card. It sold for $495, at a 10 MHz clock (the TI TMS320C25 DSP), it was pipelined, and if you ordered instructions to take advantage of the pipeline, it could spit out a multiply-accumulate once for every clock cycle. It also came with no software to speak of, even to record and play sound, and the A/D was clocked at a fixed 44.1 kHz and the D/A at a fixed 88.2 kHz sampling rate. Unlike the Motorola DSP in competing products, the C25 could be halted, and the PC could read and write DSP memory through I/O ports whether the thing was halted or running, so you could just dump hex programs into the thing without having to bring up a bootstrap loader on the DSP, and you could start, halt, and inspect for debugging purposes, and it could generate an "A/D buffer full" interrupt for continuous A/D operation.
IBM supplied sample code where they used some simple macros in the PC macro assembler to generate op codes for the C25 DSP and wrote a simple PC-side loader in 8086 assembler. This was fairly easy to do because the C25 instructions are all fixed 16-bit words, and while the hacked-up assembler was all fixed address with no relocation, it all worked pretty reliably because you had complete control over all aspects of the software and hardware -- TI has some software tools for there TMS320CXX eval boards that were a POS because they were buggy as heck, but this setup was the ultimate in Keep it Simple and Stupid.
Digging into Crochiere and Rabiner on something called polyphase multirate FIR filter design, I had a bank of filters that upsampled and downsampled to support rates of 10, 11, 16, 20, and 22 kHz to match existing digitized speech files (the TI-MIT-NIST-DARPA speech database was sampled at 16, I had stuff sampled at 20). I also has subroutines for the card for real-valued forwards and inverse FFT, FIR filter, and second-order-section cascade IIR filter, and the whole wad of software fit in DSP memory at absolute addresses, and I had a little Turbo Pascal interface library to the whole thing, and I was king of the world.
Guess what. The M-ACPA card kind of went by the wayside by the mid 1990's when Windows 95 came along: there were much cheaper Windows sound cards and IBM never could get a Windows driver for the darned thing that didn't leave gaps (sound clicks) with every interrupt cycle. And the Pentium came along which just blew the thing away speed wise, and the DSP library could be written in C or other compiler language, portable to any other computer, and using floating point so one could stop worrying about scaling and overflow and fixed-point roundoff.
For all my work on the DSP library, I think I got at most about 3 years useful life out of it, and besides, I had the versioning problem of deciding who had one of these M-ACPA's installed and who didn't and reverting to a non-ACPA library for those who didn't (remember when the 8087 was optional and the pain that caused?). After that experience, I don't want to touch another array processor/DSP/GPU whatever: I am going to program whatever CPU is available in whatever compiler is available, and I am not going to mess in assembler with any strange instruction set enhancement promoted by disco dancers in clean room suits. I am just going to sit back and wait for Moore's law and wait for the CPU and compiler to catch up.
It seems that in the early days, NASA practiced a kind of bottom-up rather than top-down design. The 1.5 million pound thrust F-1 engine was in the test stand long before they decided on Nova/Saturn, direct-ascent/EOR/LOR modes of lunar travel, or how many engines went on the Saturn V. I guess the F-1 engine was some kind of crash program to catch up with the Russians, who had bigger boosters than Atlas and Titan and were besting us in space in the late 1950's. Once you had the F-1 engine, you could design a booster around it, and once you had the booster, it constrained your spacecraft. A lot of the Shuttle is constrained by the SSME, and perhaps work should go into the next generation engine and let the spacecraft grow up around it.
In the 1960's, there was a "mainline" nuclear program in the form of NERVA (Nuclear Engine for Rocket Vehicle Applications) and RIFT (Reactor In Flight Test). The concept used hydrogen as the fuel and a graphite-moderated fission reactor as the heat source, and they tested these engines either in Nevada or Idaho. One of the early tests burned up the reactor and sent flaming chunks of reactor up into the air (the tests had the rocket exhaust aimed skyward), but they refined the design so more of the reactor hung together, and they got impressive results in terms of thrust, and running time, and restarts.
The nuclear rocket program was just developing an engine, and the spacecraft and also the mission would grow up around that rocket, so they didn't have any clear plan as to what to do with it. But if the interest in Apollo hadn't evaporated and all the resources got diverted to Shuttle, the nuclear rocket was on track for a possible Mars mission or perhaps putting a large base on the Moon.
Project Orion was a bunch of "renegade" atomic physicists engaged in their own venture capital-funded research (General Atomics in La Jolla). They were on the outside looking in for funding rather than the NERVA/RIFT program that had solid government backing. The original concept was to launch from the Earth's surface with nuclear-bomb drive, but it was amended to space launch on top of a Saturn, and at that point it was competing with NERVA/RIFT. Apart from nuclear fallout, Earth launch has an advantage that you can make the pusher plate large in diameter and give good specific impulse. Space launch meant the pusher plate was much smaller, constrained by the Saturn V, and the specific impulse than went to heck.
The one thing I never understood about NERVA was this. In theory it had twice the specific impulse as a hydrogen-oxygen rocket. In practice, it got that high specific impulse by using just liquid hydrogen as the fuel -- a bulky, low-density fluid. Also, the atomic reactor had to be heavy. It seems that for a mere doubling of specific impulse, you had to have a very large, bulky tank, and a heavy reactor, so I wonder how much edge over chemical rockets it really had.
Lets start with the Saturn V rocket. The thing was designed by the Huntsville Germans. When you think of German engineers, think meticulously designed and crafted, expensive as heck, and reliable. Did they ever lose a Saturn (Saturn V or Saturn Ib) in flight? Titan was much cheaper than Saturn but hasn't had quite the same record.
OK, now consider the Apollo CM with its ablative heatshield and low-lift blunt-body design. And with a Max Faget solid-fuel tractor escape rocket. Compare with Shuttle with wings, and tiles, and computers flying the thing and with the Shuttle parallel to the tanks where stuff can fall off or blow up. In the Challenger explosion, the crew capsule remained intact and killed the crew when it hit the water. If something happened to the Saturn rocket, the Apollo crew had an escape rocket, they had space suits to survice a cabin puncture, and they had parachutes to make a safe water landing.
Sure Apollo was primitive by comparison, primitive in the sense of Keep It Simple, Stupid (and Safe). Oh, and Apollo had redundant space crafts so even when the Service Module was blown to shreds (as a result of ground handling to empty a balky oxygen tank by running tank heaters until the insulation burned off), they brought back to crew, although one guy had a 103 F plus fever from a urinary infection because he didn't think they had enough electric power for him to take a leak often enough.
You can write applications Windows applications that do non-blocking I/O, and that the GUI has one thread has nothing to do with it. The real problem is that the kernel schedules non-preemptible tasks that are coarse-grained, and faster CPUs help because those coarse-grained tasks become fine-grained (in time) when the CPU churns through them faster.
That Explorer freezes on a file-tree expand is the fault of Explorer, and it would be Explorer's fault under BeOS if it were written that same way. The kind of fined-grained I am talking about is on the submillisecond level, important if you are doing video frames where a 10 ms lock will cause a frame to glitch. Those 10 ms locks are deep in the OS, and faster CPUs have reduced them to 1-2 ms locks. And as for hyper threading, its existence may encourage kernel designers to tune for allowing more preemption without worrying that the system will stall doing context switches.
Oppenheimer famously quoted from the Hindu scriptures "I am become death, the destroyer of worlds" upon the first A-bomb test at Almagordo. A lot of people took it to mean that Oppenheimer was reflecting on letting the nuclear genie out of the bottle, but apparently it had a more personal meaning.
Vishnu was trying to impress upon a prince that he needed to do his duty fighting a war the prince wanted no part of, so Vishnu appeared in one of his more terrifying avatars to convince the prince. I heard it said that Oppenheimer saw himself as the reluctant prince, who was required to do his duty in helping with the war effort, and the bomb test was the fearsome visage of Vishnu, urging him to do what came next, i.e., prepare for using the A-bomb in the war.
Nowadays we are into multi-culturalism, and besides, there are now many immigrants from India in various walks of life in American society, and then there was George Harrison and all of that, so the comfort level with Indian culture and Hindu religious icons is much better these days. But back then, Oppenheimer was already suspect for being somewhat left-of-center in his politics and for being somewhat of an egghead (to use swing-era slang), and being Jewish in America of that time already made a person suspect of not worshipping the same God, perhaps in the way being Muslim in America does today, and gosh, quoting some obscure Hindu scripture really put a person way in left field.
But the nagging, unanswered question I have is this: isn't "I am become death" ungrammatical or am I missing some fine point. I can understand "I am death" (present tense) or "I have become death" (past perfect? -- I am not up on grammer), but I always thought "I am become death" was the result of some mistranslation on the order of "all your base."
My experience is with Windows, but I suppose Linux is not that much different. The GUI is typically single-threaded and event-driven: the cooperative multi-tasking model. You can create all the threads you want, and I have even seen sample apps where those threads poke at the GUI, but to be technically rigorous, you really shouldn't do that. If you want to synchronize a worker thread with the GUI thread, all you need to do is SendMessage() (blocking call) or PostMessage() (non-blocking call), and the OS automagically takes care of the queueing and synchronizing to get this done.
While Java tries not to be tied to any one OS, you kind of see the OS poking through. Java too has a single-threaded GUI, and you are not supposed to invoke methods on any GUI object from other threads apart from InvokeNow() (guess what -- SendMessage()) and InvokeLater() (also guess what -- PostMessage()). The advantage of the single-threaded GUI is that any GUI method is in effect synchronized -- each GUI method is essentially its own critical section that won't get stepped on or poked at for the duration of its execution, and variables won't get changed unless you SendMessage() or PostMessage() (i.e. cooperative reentrancy) to somewhere else.
How does this work when you have a multi-threaded GUI -- are you declaring entire methods "synchronized" or have to have locks up the wazoo, or are there some easily-understood protocols?
Now apart from the single-threaded GUI, Windows has a way of "going away" for 10's of milliseconds at the system level -- disk reading is very coarse grained, and they say it is for performance reasons. These hyperthreaded Pentium 4's are creating very cheap context switches while the processors are getting so much faster that what used to be 10's of ms is now in single digits, so Windows and Linux and whatever are perhaps getting to multimedia Nirvana by brute computing power. Moore's law, yes BeOS can do it all on a 60 MHz Pentium I, but no one is running a 60 MHz Pentium I these days.
Can anyone explain to me OS X IB controllers?
on
Introduction To XAML
·
· Score: 1
One of the more intriguing concepts in Interface Builder is the IDE support for generating controller objects. Is there a description of IB controllers written more in geek-speak than PHB buzzword-speak? I am trying to figure out what they do without a Mac in front of me because I develop for Windows (like, maybe I can be persuaded to develop for the Mac, heh).
What I think those controllers are is what GoF calls mediators. The controller can be hooked up to data-model widgets as an observer, and it can be hooked up to display widgets so it can write stuff. Furthermore, it can be plugged in as an observer or plugged in to get access not just to a model or a view widget wholesale but to particular fields of such widgets using some kind of proxy objects that resemble C# delegates.
Why I think this is important is that VB, Delphi, and their ilk implement Mediator pattern, but the Mediator is wrapped up into the main form -- it is like the controller setup only the main form is the controller and you get only one controller for the entire form. A lot of the complaints about spaghetti code in VB, I believe, have to do with the main form acting as traffic cop for all the communication between widgets on the form. The IB controller seems to separate the mediator function from the main form and allow you to have as many mediator objects as you have data paths.
One advantage of the one-size-fits-all VB approach is that the mediator needs to be an Observer, and if the mediator is the main form, you know that this Observer is going to stick around for the lifetime of the child widgets being observed. If you allow widgets to directly talk to each other, you have to have some kind of registration/notification mechanism in case widgets are created and destroyed at runtime. How does IB handle this? Or are the widgets on a form fixed by the NIB file and not allowed to be created or destroyed during program execution?
I recently found out that Matlab of all things can script Java objects. My PHBs are engineering students -- their tuition pays my salary, their evaluations influence my pay raises -- and yes, I would like them to program rather than rely on canned CAD and simulation software all their lives. They love Matlab, but knowing only Matlab will leave them developmentally disabled, the are required to learn Java, but I would like them to do more with it. Just wait until I show them how easy it is to use their beloved Matlab to create Java object instances and read and write data from those objects and use Matlab to make all kinds of pretty engineering graphs with that data.
For grins, I did some Googling about what it would take to poke at Java objects from C++. There are a number of commercial and open-source projects to do just that, but golly, the C++ program has to launch a Java JVM, and then you have to create C++ proxy objects for the Java objects, and then you have to fret lifetime management and omigosh, what about thread? Now since C++ and Java are vaguely similar C-like OO languages, you probably are better off doing everything in C++ or everything in Java. But the point is to do something like that, if there was a requirement for it (how about a legacy Windows app using student-written Java plugins -- remember they are required to learn Java, not C++), it would take the better part of the semester to get my flock of PHBs up to speed. The Matlab thing can be handled in one class period.
Ok, scripts and RAD and such will allow PHBs to sully the reputations of programmers by generating garbage. I will take that risk so I don't have to labor in the salt mines of coding -- I have paid my programming dues writing thousands of lines of Windows API code, and I want to write scripts and use RAD in my lazy old age.
UNIX is/was "user friendly" by a long shot compared to OS/360, and yes, I have OS/360 JCL experience. Essentially a card deck of JCL was like holy writ -- once you had a working JCL deck, the idea was to neither add nor subtract one jot nor tittle (i.e. diacritical marks in Aramaic script).
As far as RSX-11 and UNIX, there may be a tie, but the grand daddy of user friendly had to be TOPS-10 (PDP-10 time-sharing OS).
On the developer side, the would-have-been could-have-been that got pushed aside by UNIX would have to be some kind of Lisp or perhaps Smalltalk environment -- you know, a high-priced mega-pixel dedicated workstation running one of these deals where the language is the OS is the environment, where the debugger can trace into OS calls because the OS is the language is the environment, and where the system commands are Lisp expressions or Smalltalk commands -- Perl, Python, Matlab environments are perhaps the modern equivalent.
Would you be mad at me for saying Paul Wellstone (he also died in a pilot-error accident involving instrument flight conditions, although he hired a pilot instead of flew himself)?
Private aviation does not have a safety record on the level of cars, and I believe a lifetime of driving can give you 1 in a 100 odds of dying in a car crash, so what does that may private flying?
Let me tell you why I mentioned John F Kennedy Jr. The accident has had enough celebrity notariety that the details of it are widely known. And the details speak volumes of the safety problems of private aviation. I am sure every Harry and Moe hanging around the local FBO is saying, "Yeah, like he shouldn't have tried a water-crossing flight with his level of experience" or "he should have called Flight Planning and they would have told him VFR flight not recommended." Tom Wolfe called it "the Right Stuff" -- the notion that Harry bought the farm but I Larry have nothing to worry because I would never do anything that bone-headed as Harry in an airplane.
From what I had heard, if JFK Jr. had left according to his original plan, he would have been OK, but he got delayed waiting for his wife and/or sister-in-law to get their rears in gear, and he took off into the sunset where the ocean and sky melted into a pool of hazy red sunset. Guess what -- you do the most rigorous flight planning and you get thrown off your game by the requirements of passengers or the have-to-be-there demands of family.
The worst part of it is driving from New York to Cape Cod takes what, 4-5 hours? Yes he could have been in a car smashup, but private aviation is indeed a high-risk undertaking.
As far as the 747, I cannot remember riding on a plane, either pre or post 9-11, where just about every seat was taken. One of the advantages of computer databases and linear programs is that the airlines have figured out how to really pack us passengers in like sardines.
Fuel efficiency has a everything to do with the total weight, the lift/drag ratio, and the efficiency of the engines. A short range configured B-757 or Airbus A-320 comes in at around 1000 lb gross weight per seat (structure, fat U.S. passenger, baggage, fuel) while the longer range 747 comes in around 2000 lb/seat (more business class seats, cargo, more fuel load, the bigger plane takes more structure per seat than the smaller plane). The lift to drag work out to about 18:1 given current aspect ratios on wings -- boost the lift to drag and the weight goes up. A really good engine, like the PW 2037 on the 757 comes in at around.55 lb fuel per pound of thrust per hour. Assume jet fuel to be 6.5 lb/gallon (density.8). Assume cruise at 500 MPH. The 757 has 56 lb drag/seat, consumes 31 lb fuel/seat-hour, 4.7 gallons per seat-hour, or 106 seat-miles/gallon. The 747 comes in at half that -- around 53 seat-miles/gallon -- because it is heavier. Now the 757 is not going to achieve that on short range trips -- the taxi, climb, approach pattern fuel usage probably knocks it down to the 50-60 seat-miles/gallon range.
The Taurus gets 31 MPG average on the highway. It has about 100 lb of drag at the speeds I drive. If the engine was efficient as the PW 2037, the Taurus would be getting 60 MPG on the highway. The Taurus engine has to be oversized or I would have to be downshifting through several gears with every hill just like an 18-wheeler. Half of the Carnot-efficiency power is driving the wheels and overcoming rolling and aero resistance -- the other third is oil-bearing friction in the oversized engine with the remaining sixth the "pumping loss" for throttled engine operation. The 757 throttles its engines without pumping loss by flying at high altitude in thin air, and turbine engines don't have the friction losses of piston engines so they can operate "part load" using atmospheric "throttling" very efficiently.
Now for the train. The train is hyper efficient at slow speeds where aero drag is minimal and on flat level ground and with a minimum of starts and stops. When you make the train heavy (Federal safety standards), draggy (the only Amtrak train that is highly streamlined is the Acella), and have it climb the Western mountains, it starts using fuel.
L.A. Marre (1995) The Contemporary Diesel Spotter's Guide 2nd Edition, Kalmbach, Waukesha, Wisconsin, p 294 states that Amtrak experimented with a 9300 gallon fuel tender car to allow the Empire Builder to go Chicago-Seattle nonstop. Allow another 300 gallons for the tanks in two EMD F-40's, assume a 10 car train carrying 400 passengers (maybe 80 seats in a Hi-level coach, fewer in sleeping cars, no revenue seats in the lounge and dining cars), 2500 mile trip and you are at 81 seat-MPG at full load, and that train is generally sold out at least during the summer.
As I said, a single Diesel engine pulling a string of double-decker commuter cars probably checks in at 400 seat-MPG, but that is high-density seating, no dining car, slower speeds, and level ground. But if you pack a family of 4 into a minivan (25 highway MPG), you don't have to feel guilty about not taking the Empire Builder.
12 MWH/tonne of fuel is assuming 100 percent thermal efficiency. The TGV is electric powered (from a nuclear power plant in France?), but assume you generate that power in a Diesel-electric locomotive, perhaps 30 percent thermal efficiency is reasonable, and then you are in the range of 130 passenger-MPG, which is not too far off the capabilities of perhaps a propfan powered medium range airliner. If you are going to go that fast that close to the ground, it is going to take some power.
Didn't they, like, try to dig a "whole to China", or at least drill a hole down to the crust/mantle boundary named after that Croatian dude whose name I can't pronounce let alone spell? And wasn't there an NSF grant to dig the whole, and they subcontracted the drilling to Halliburton or Brown and Root or somebody, and they burned through all the money before they got very far and the grant didn't get refunded?
I looked into private aviation as transportation, and for anyone thinking about it, buy yourself the plushest luxury car (Lexus, Mercedes), and you will end up getting to your destinations much safer, not that much slower, and will on a life-cycle basis end up spending much less money. Interesting you mention Martha's Vineyard -- the safety issue and JFK Jr. comes to mind.
There was a Golden Age of private aviation -- perhaps the mid 1960's, when a Piper or a Cessna was competitive with a luxury car. What has happened since then is that liability insurance has driven the light plane manufacturers into the ground and priced light planes out of the market.
We can all get mad at lawyers and call for tort reform and exemptions for plane manufacturers. But flying a light plane is a much riskier activity than driving a car, and the high liability insurance making planes really expensive is society's way of saying that we place a high value on human life, or at least on human life lost in transportation accidents, and the legal system coupled to the market system has perhaps made the correct decision in trying to get people to drive rather than fly themselves.
You mention the "steam gauges" and the need for a glass cockpit in a light plane. The "steam gauges" are there because they are simple and reliable -- and perhaps safer unless there is an enormous breakthrough in light plane avionics.
The engine controls are very primitive and manual: throttle, mixture control, and in some cases, propeller speed: not much more sophisticated than a lawn mower. If you have a turbo engine, you have manual control over boost pressure and have to follow rules for both advancing and retarding the throttle so as to not ruin the engine. In the 1980's there as some attempt at modernization: Porsche came out with an engine with electronic controls and "single-lever power control." But I don't know if this changed the general market trend that light plane manufacturers went out of business or went high-end (half-million dollar plus airplanes), and the only affordable planes are the ones stamped "Experimental" (i.e. I built it myself so I can't sue anybody).
Planes, trains, and automobiles.
on
The Future of Flight
·
· Score: 2, Interesting
The auto/airplane fuel economy comparisons have some variables. I can pretty much get 31 MPG highway cruising in my Ford Taurus while a packed 747 can get 60 seat-MPG. But if it is my wife and I, we are getting better than 60 person-MPG in the Taurus.
The airplane has some interesting scaling laws. The 747, say, gets 60 seat-MPG (you probably have to look at some actual data rather than max range data because max range has reserve for headwinds, diversion to alternate airport, but the 60 seat-MPG is not far from the mark). I am guessing, but I would say a 757 may get close to 100 seat-MPG New York-LA (the 757 is a lighter plane which is lighter per passenger, carries less fuel than a trans Pacific flight, the PW 2037 is a really efficient engine). But if you operate the 757, say Minneapolis to Detroit, I would bet the fuel mileage may drop to 40-50 seat MPG because of the energy cost of climbing to cruising altitude, low-and-slow flight in the landing pattern.
From reading a trucking magazine at a truck stop, I learned that an 18-wheeler with good aerodynamic fairings and a late-model Diesel engine gets about 6 miles per gallon. If you figure 8 miles per gallon for a bus, an intercity bus can give over 300 seat-MPG. As for the train, I am willing to guess that for a transcontinental train like the Chicago-Seattle Empire Builder, what with sleeping and dining cars, U.S. standards that passenger cars be built like tanks to withstand derailments (European passenger cars are much lighter), and all of the mountain passes on the western routes, the train is probably little different than the airplane and at the 50-60 seat-MPG range.
These Diesel commuter trains where they have one locomotive pulling 6 or 7 100-passenger double-decker cars may get something in the 500-600 seat-MPG. But if you go for acceleration and frequent stops such as a subway train, or if you go for high speed, I bet your seat-MPG start working their way back to the car and airplane range.
Interesting that you should mention high-speed rail. Part of what makes my Taurus get 31 highway MPG is that I pretty much stick to 65 MPH. All those dudes passing me doing 75 or 80 are probably getting more like 25 MPG or less. I read that the highly-streamlined but quite fast TGV's are maybe a factor of 2 better than air - I am guessing maybe about 100 seat-MPG. And that 300 MPH Japanese maglev may be comparable to air, perhaps in the 40-50 seat-MPG range.
If you start going fast, think of going at airplane speed but doing it at ground level where the air is thicker. Part of how jets get their efficiency is that they fly high -- the thin air reduces the power on their turbine engines without the losses of throttling, and the thin air allows them to go fast without too much drag. They have to pay for that efficiency with a long climb to cruising altitude. Coupled with large-long range planes being heavier, there is probably a sweet spot in efficiency for perhaps a 1500-3000 mile trip.
Perhaps I wasn't clear. Microsoft is not the only commercial developer out there collecting tribute money. For example, I will discuss with a colleague using something like LabView to develop a custom instrumentation/measurement system vs rolling one's own widgets to display wave plot, meters, and the like. I will say, "If we went with LabView, we would have to pay them tribute money" meaning that we would have to pay a royalty for redistributing their runtime. This is by no means denigrating or criticizing National Instruments for charging for their software or for charging developers for use of their runtime. But if you are going to use their stuff, you have to figure the royalties as a cost of doing business just like if you engaged in certain activities on some Mafia bosses' turf, legal or otherwise, you need to plan on paying them the required tribute money.
OK, I am the spoiled 3 year old who didn't get the toy I wanted, but what does that make, say Ford Motor Company? Ford puts a lot of effort into their own R&D and patents because they don't want to pay tribute money (i.e. patent royalties) on anything going into a car because that really kills them on their bottom line. I guess the tribute money they pay for office-worker desktops doesn't bother them, although they are big enough they probably could have their own guys customize a Ford Linux Desktop, although Bill is probably smart enough to cut them a deal on a Windows site license below that cost.
Just because I call it tribute doesn't mean it doesn't make sense under circumstances to pay it. The reason I call it tribute is that I want to draw a distinction between, say my heating contractor charging me $700 to fix my central air (my heating contractor is a pirate, but the money I am paying is not tribute because parts are installed in my system and a guy has to come over to the house and do stuff) and what economists call an economic rent.
Tribute money would be if say, my central air ran this control program called Carrier 98SE, and every couple of days the AC would just quit and I would have to go in the basement, throw a switch, wait 10 minutes, and watch a message flash on the screen telling me that I shouldn't have turned the AC off by turning off the power. I could pay $100 to install a control program called Carrier XP, only I had to connect the air conditioner to the phone to be allowed to turn it on, and by connecting the AC to the phone, not only would the AC run constantly until the coils froze up, ten of my friends would have their AC do the same thing in the next week.
Visual Basic (and Delphi) famously let you connect events in child widgets to functions of the main form. Visual Basic doesn't have a way to use the Designer to connect child widgets to each other and while Delphi allows such connections, it doesn't get used that much. The result is that a Visual Basic main form is one big honkin' GoF Mediator Pattern -- every last thing that happens in a widget signals the main form, and the main form code is this rats' nest of figuring out what to do with all of those signals.
As far as I can tell, Mac OS-X Cocoa Interface Builder has these controller objects so the rats' nest of event processing code can go into one or more controller objects. You may say this is no big deal, Qt has had its "signals and slots" for some time now. True, but I think the innovation is that Apple allows you to do the Visual Basic-y think with signals and slots: automatically generate code for one or more controller classes rather than funnelling everything through the main form.
The reason for the link is that Apple is this amalgam of the corporate and Marin County buzzword cultures, and they go far beyond calling a program a solution. On the other hand, buried under the buzzwords seems like some neat stuff. Should I be wary of Apple?
I prefer to call payment for such things "tribute money" - site licenses, royalties, upgrade charges. I count as tribute money any and all fees paid for things on account of "intellectual property" (OK already, patents, copyright and trademark laws then).
Spare me the explanations of the poor starving software developers; I am fully aware that a software developer seeks renumerations for one's labors, and charging license fees and upgrade fees is a way to amortize the effort required to develop a complex piece of software. That doesn't change the fact that license fees are a kind of economic rent (i.e. money you can rake in because the law grants you a limited monopoly -- you can say that software won't get developed in the absence of such a monopoly, but that doesn't change the material facts that "intellectual property" law has the intent of granting limited monopolies to facilitate collecting economic rents).
I prefer the term tribute money to "Microsoft tax" because "tax" suggests governmental power and some sense of the consent of the governed. Microsoft is not to be dignified by considering it a government -- it is more like such extragovernmental entities such as high-seas pirates, Mafia bosses, feudal lords, and Delaware corporations in that money payed to them to avoid punishment (i.e. lawsuits, getting wacked) is to be called tribute and not a tax.
I also differ with the common usage of "pirate" to denote someone who avoids paying tribute money. I use the term "pirate" to describe contruction contractors that you bring into your house for remodeling and repair work. The reason contractors are pirates has less to do with the amount of money you pay them than the part about when you let them into your house they control every aspect of your life. Yes, it is about the money because whatever contract you sign, there is some uncontrolled eventuality that you have to agree to spending more money once work commences, but even if you are rich enough that the money spent is a minor concern, you become their pirate-hostage regarding letting them in and out of the house at their whim and work schedule.
So construction contractors are pirates simply on the basis that their clients are pirate hostages, and money spent for the XP upgrade when 98 was working just fine for you, thank you, is tribute money.
My FBO dude is a retired Army MP, he has a good safety record by controlling when the rental plane keys get handed over the counter, regardless of what the CFIs have to say about conditions, and I am pretty sure he wouldn't sell me the gas if I was in such a spot.
Everyone is focused on the notion that McMurdo is trying to teach this guy a lesson. What about the idea of keeping the guy from taking another shot at getting himself killed? The guy has already crossed oceans and circled the globe. Well, some 70 years ago there was this hot babe with this rich hubby who was crossing oceans and circling the globe who got herself killed doing something dumb, reserve fuel and navigation-wise. Someone should have refused to sell her fuel.
The real crisis in the recording industry is not illegal copying but Britney Spears. Because you have paid for music up front you will start downloading Britney's tracks, which you wouldn't have dreamed of doing in the absence of the tax. That is the only way the industry will stay solvent given all the drek they are peddling.
There is a simple solutions: you fellows from Canada will drive south to stock up on hard drives while we Yanks will drive north to get Grandma's blood pressure medicine.
Another question: who is getting all this money from the blank-CD tax? Is it all Canadian artists, or is all the money going to the record companies in Los Angeles? If you are levying these taxes to fund copyright holders who are mainly south of the border, why are you getting out of it?
Slashdot Mirror
I am wondering if a virus is really a separate entity or is really intrinsically a property of the host. Is a cold virus really just that, or is it a piece of human genetic machinery that has the capability of being shared between humans when one human picks their nose?
The reason I got to wondering is that it seems diseases that stick around have some kind of evolutionary purpose. Sickle cell anemia gene confers resistance to malignant malaria, cystic fibrosis gene confers resistance to cholera, and so on. Sexual reproduction was a big evolutionary breakthrough in allowing diversity in the offspring to adapt to environmental pressure.
Do you suppose a virus is a remnant of some other mechanism, such as conjugation where bacteria can swap genetic material? If a virus served no evolutionary purpose, it seems that individuals who were virus resistant would have a slight edge and over many generations viruses would die out. The fact that they are still here suggests that the cell mechanisms that support virus reproduction serve some other purpose than making a person sick.
This means that even if increased the gas engine efficiency to 0 cents per kWHr (through smaller engine run at peak efficiency to only charge batteries) but ran all the power flows through the lead acid batteries, you costs would increase from 40 cents to 60 cents per kWHr delivered to the wheels.
Anyway, this argument is for series power flows. The traditional automatic transmission is a series power flow, at best 80 percent efficient. The automotive industry has latched on to parallel power flows as a way to boost efficiency. Modern automatic transmissions use split-torque direct drive and torque converter together with direct-drive clutches to boost this efficiency. The commercial hybrids today are parallel electric and gas engine drive. The Toyota Prius is gear shift free, but instead of the Diesel locomotive series drive of engine-generator-traction motor, the Prius has the gas engine, generator, and traction motor tied to a planetary gear set so there are parallel mechanical and electric torque paths, again to boost the efficiency.
If you have a parallel hybrid, obviously you need to use your 60 cents/kWHr battery electricity under circumstances where the gas engine electricity costs more than 60 cents/kWHr, or perhaps use it to resize the gas engine for greater efficiency while consuming the minimum of expensive 60 cents/kWHr juice. But given that batteries are 1) expensive, and 2) wear out (think laptop or cellphone or iPod battery), the hybrid vehicle is not a slam dunk.
My favorite was the IBM/Tecmar M-ACPA sound card. It sold for $495, at a 10 MHz clock (the TI TMS320C25 DSP), it was pipelined, and if you ordered instructions to take advantage of the pipeline, it could spit out a multiply-accumulate once for every clock cycle. It also came with no software to speak of, even to record and play sound, and the A/D was clocked at a fixed 44.1 kHz and the D/A at a fixed 88.2 kHz sampling rate. Unlike the Motorola DSP in competing products, the C25 could be halted, and the PC could read and write DSP memory through I/O ports whether the thing was halted or running, so you could just dump hex programs into the thing without having to bring up a bootstrap loader on the DSP, and you could start, halt, and inspect for debugging purposes, and it could generate an "A/D buffer full" interrupt for continuous A/D operation.
IBM supplied sample code where they used some simple macros in the PC macro assembler to generate op codes for the C25 DSP and wrote a simple PC-side loader in 8086 assembler. This was fairly easy to do because the C25 instructions are all fixed 16-bit words, and while the hacked-up assembler was all fixed address with no relocation, it all worked pretty reliably because you had complete control over all aspects of the software and hardware -- TI has some software tools for there TMS320CXX eval boards that were a POS because they were buggy as heck, but this setup was the ultimate in Keep it Simple and Stupid.
Digging into Crochiere and Rabiner on something called polyphase multirate FIR filter design, I had a bank of filters that upsampled and downsampled to support rates of 10, 11, 16, 20, and 22 kHz to match existing digitized speech files (the TI-MIT-NIST-DARPA speech database was sampled at 16, I had stuff sampled at 20). I also has subroutines for the card for real-valued forwards and inverse FFT, FIR filter, and second-order-section cascade IIR filter, and the whole wad of software fit in DSP memory at absolute addresses, and I had a little Turbo Pascal interface library to the whole thing, and I was king of the world.
Guess what. The M-ACPA card kind of went by the wayside by the mid 1990's when Windows 95 came along: there were much cheaper Windows sound cards and IBM never could get a Windows driver for the darned thing that didn't leave gaps (sound clicks) with every interrupt cycle. And the Pentium came along which just blew the thing away speed wise, and the DSP library could be written in C or other compiler language, portable to any other computer, and using floating point so one could stop worrying about scaling and overflow and fixed-point roundoff.
For all my work on the DSP library, I think I got at most about 3 years useful life out of it, and besides, I had the versioning problem of deciding who had one of these M-ACPA's installed and who didn't and reverting to a non-ACPA library for those who didn't (remember when the 8087 was optional and the pain that caused?). After that experience, I don't want to touch another array processor/DSP/GPU whatever: I am going to program whatever CPU is available in whatever compiler is available, and I am not going to mess in assembler with any strange instruction set enhancement promoted by disco dancers in clean room suits. I am just going to sit back and wait for Moore's law and wait for the CPU and compiler to catch up.
In the 1960's, there was a "mainline" nuclear program in the form of NERVA (Nuclear Engine for Rocket Vehicle Applications) and RIFT (Reactor In Flight Test). The concept used hydrogen as the fuel and a graphite-moderated fission reactor as the heat source, and they tested these engines either in Nevada or Idaho. One of the early tests burned up the reactor and sent flaming chunks of reactor up into the air (the tests had the rocket exhaust aimed skyward), but they refined the design so more of the reactor hung together, and they got impressive results in terms of thrust, and running time, and restarts.
The nuclear rocket program was just developing an engine, and the spacecraft and also the mission would grow up around that rocket, so they didn't have any clear plan as to what to do with it. But if the interest in Apollo hadn't evaporated and all the resources got diverted to Shuttle, the nuclear rocket was on track for a possible Mars mission or perhaps putting a large base on the Moon.
Project Orion was a bunch of "renegade" atomic physicists engaged in their own venture capital-funded research (General Atomics in La Jolla). They were on the outside looking in for funding rather than the NERVA/RIFT program that had solid government backing. The original concept was to launch from the Earth's surface with nuclear-bomb drive, but it was amended to space launch on top of a Saturn, and at that point it was competing with NERVA/RIFT. Apart from nuclear fallout, Earth launch has an advantage that you can make the pusher plate large in diameter and give good specific impulse. Space launch meant the pusher plate was much smaller, constrained by the Saturn V, and the specific impulse than went to heck.
The one thing I never understood about NERVA was this. In theory it had twice the specific impulse as a hydrogen-oxygen rocket. In practice, it got that high specific impulse by using just liquid hydrogen as the fuel -- a bulky, low-density fluid. Also, the atomic reactor had to be heavy. It seems that for a mere doubling of specific impulse, you had to have a very large, bulky tank, and a heavy reactor, so I wonder how much edge over chemical rockets it really had.
Lets start with the Saturn V rocket. The thing was designed by the Huntsville Germans. When you think of German engineers, think meticulously designed and crafted, expensive as heck, and reliable. Did they ever lose a Saturn (Saturn V or Saturn Ib) in flight? Titan was much cheaper than Saturn but hasn't had quite the same record.
OK, now consider the Apollo CM with its ablative heatshield and low-lift blunt-body design. And with a Max Faget solid-fuel tractor escape rocket. Compare with Shuttle with wings, and tiles, and computers flying the thing and with the Shuttle parallel to the tanks where stuff can fall off or blow up. In the Challenger explosion, the crew capsule remained intact and killed the crew when it hit the water. If something happened to the Saturn rocket, the Apollo crew had an escape rocket, they had space suits to survice a cabin puncture, and they had parachutes to make a safe water landing.
Sure Apollo was primitive by comparison, primitive in the sense of Keep It Simple, Stupid (and Safe). Oh, and Apollo had redundant space crafts so even when the Service Module was blown to shreds (as a result of ground handling to empty a balky oxygen tank by running tank heaters until the insulation burned off), they brought back to crew, although one guy had a 103 F plus fever from a urinary infection because he didn't think they had enough electric power for him to take a leak often enough.
Give me Apollo primitive over Shuttle any day.
Don't know how the boldface got in there -- can I say it was a type? A bug?
That Explorer freezes on a file-tree expand is the fault of Explorer, and it would be Explorer's fault under BeOS if it were written that same way. The kind of fined-grained I am talking about is on the submillisecond level, important if you are doing video frames where a 10 ms lock will cause a frame to glitch. Those 10 ms locks are deep in the OS, and faster CPUs have reduced them to 1-2 ms locks. And as for hyper threading, its existence may encourage kernel designers to tune for allowing more preemption without worrying that the system will stall doing context switches.
Nowadays we are into multi-culturalism, and besides, there are now many immigrants from India in various walks of life in American society, and then there was George Harrison and all of that, so the comfort level with Indian culture and Hindu religious icons is much better these days. But back then, Oppenheimer was already suspect for being somewhat left-of-center in his politics and for being somewhat of an egghead (to use swing-era slang), and being Jewish in America of that time already made a person suspect of not worshipping the same God, perhaps in the way being Muslim in America does today, and gosh, quoting some obscure Hindu scripture really put a person way in left field.
But the nagging, unanswered question I have is this: isn't "I am become death" ungrammatical or am I missing some fine point. I can understand "I am death" (present tense) or "I have become death" (past perfect? -- I am not up on grammer), but I always thought "I am become death" was the result of some mistranslation on the order of "all your base."
While Java tries not to be tied to any one OS, you kind of see the OS poking through. Java too has a single-threaded GUI, and you are not supposed to invoke methods on any GUI object from other threads apart from InvokeNow() (guess what -- SendMessage()) and InvokeLater() (also guess what -- PostMessage()). The advantage of the single-threaded GUI is that any GUI method is in effect synchronized -- each GUI method is essentially its own critical section that won't get stepped on or poked at for the duration of its execution, and variables won't get changed unless you SendMessage() or PostMessage() (i.e. cooperative reentrancy) to somewhere else.
How does this work when you have a multi-threaded GUI -- are you declaring entire methods "synchronized" or have to have locks up the wazoo, or are there some easily-understood protocols?
Now apart from the single-threaded GUI, Windows has a way of "going away" for 10's of milliseconds at the system level -- disk reading is very coarse grained, and they say it is for performance reasons. These hyperthreaded Pentium 4's are creating very cheap context switches while the processors are getting so much faster that what used to be 10's of ms is now in single digits, so Windows and Linux and whatever are perhaps getting to multimedia Nirvana by brute computing power. Moore's law, yes BeOS can do it all on a 60 MHz Pentium I, but no one is running a 60 MHz Pentium I these days.
What I think those controllers are is what GoF calls mediators. The controller can be hooked up to data-model widgets as an observer, and it can be hooked up to display widgets so it can write stuff. Furthermore, it can be plugged in as an observer or plugged in to get access not just to a model or a view widget wholesale but to particular fields of such widgets using some kind of proxy objects that resemble C# delegates.
Why I think this is important is that VB, Delphi, and their ilk implement Mediator pattern, but the Mediator is wrapped up into the main form -- it is like the controller setup only the main form is the controller and you get only one controller for the entire form. A lot of the complaints about spaghetti code in VB, I believe, have to do with the main form acting as traffic cop for all the communication between widgets on the form. The IB controller seems to separate the mediator function from the main form and allow you to have as many mediator objects as you have data paths.
One advantage of the one-size-fits-all VB approach is that the mediator needs to be an Observer, and if the mediator is the main form, you know that this Observer is going to stick around for the lifetime of the child widgets being observed. If you allow widgets to directly talk to each other, you have to have some kind of registration/notification mechanism in case widgets are created and destroyed at runtime. How does IB handle this? Or are the widgets on a form fixed by the NIB file and not allowed to be created or destroyed during program execution?
For grins, I did some Googling about what it would take to poke at Java objects from C++. There are a number of commercial and open-source projects to do just that, but golly, the C++ program has to launch a Java JVM, and then you have to create C++ proxy objects for the Java objects, and then you have to fret lifetime management and omigosh, what about thread? Now since C++ and Java are vaguely similar C-like OO languages, you probably are better off doing everything in C++ or everything in Java. But the point is to do something like that, if there was a requirement for it (how about a legacy Windows app using student-written Java plugins -- remember they are required to learn Java, not C++), it would take the better part of the semester to get my flock of PHBs up to speed. The Matlab thing can be handled in one class period.
Ok, scripts and RAD and such will allow PHBs to sully the reputations of programmers by generating garbage. I will take that risk so I don't have to labor in the salt mines of coding -- I have paid my programming dues writing thousands of lines of Windows API code, and I want to write scripts and use RAD in my lazy old age.
As far as RSX-11 and UNIX, there may be a tie, but the grand daddy of user friendly had to be TOPS-10 (PDP-10 time-sharing OS).
On the developer side, the would-have-been could-have-been that got pushed aside by UNIX would have to be some kind of Lisp or perhaps Smalltalk environment -- you know, a high-priced mega-pixel dedicated workstation running one of these deals where the language is the OS is the environment, where the debugger can trace into OS calls because the OS is the language is the environment, and where the system commands are Lisp expressions or Smalltalk commands -- Perl, Python, Matlab environments are perhaps the modern equivalent.
Private aviation does not have a safety record on the level of cars, and I believe a lifetime of driving can give you 1 in a 100 odds of dying in a car crash, so what does that may private flying?
Let me tell you why I mentioned John F Kennedy Jr. The accident has had enough celebrity notariety that the details of it are widely known. And the details speak volumes of the safety problems of private aviation. I am sure every Harry and Moe hanging around the local FBO is saying, "Yeah, like he shouldn't have tried a water-crossing flight with his level of experience" or "he should have called Flight Planning and they would have told him VFR flight not recommended." Tom Wolfe called it "the Right Stuff" -- the notion that Harry bought the farm but I Larry have nothing to worry because I would never do anything that bone-headed as Harry in an airplane.
From what I had heard, if JFK Jr. had left according to his original plan, he would have been OK, but he got delayed waiting for his wife and/or sister-in-law to get their rears in gear, and he took off into the sunset where the ocean and sky melted into a pool of hazy red sunset. Guess what -- you do the most rigorous flight planning and you get thrown off your game by the requirements of passengers or the have-to-be-there demands of family.
The worst part of it is driving from New York to Cape Cod takes what, 4-5 hours? Yes he could have been in a car smashup, but private aviation is indeed a high-risk undertaking.
Fuel efficiency has a everything to do with the total weight, the lift/drag ratio, and the efficiency of the engines. A short range configured B-757 or Airbus A-320 comes in at around 1000 lb gross weight per seat (structure, fat U.S. passenger, baggage, fuel) while the longer range 747 comes in around 2000 lb/seat (more business class seats, cargo, more fuel load, the bigger plane takes more structure per seat than the smaller plane). The lift to drag work out to about 18:1 given current aspect ratios on wings -- boost the lift to drag and the weight goes up. A really good engine, like the PW 2037 on the 757 comes in at around .55 lb fuel per pound of thrust per hour. Assume jet fuel to be 6.5 lb/gallon (density .8). Assume cruise at 500 MPH. The 757 has 56 lb drag/seat, consumes 31 lb fuel/seat-hour, 4.7 gallons per seat-hour, or 106 seat-miles/gallon. The 747 comes in at half that -- around 53 seat-miles/gallon -- because it is heavier. Now the 757 is not going to achieve that on short range trips -- the taxi, climb, approach pattern fuel usage probably knocks it down to the 50-60 seat-miles/gallon range.
The Taurus gets 31 MPG average on the highway. It has about 100 lb of drag at the speeds I drive. If the engine was efficient as the PW 2037, the Taurus would be getting 60 MPG on the highway. The Taurus engine has to be oversized or I would have to be downshifting through several gears with every hill just like an 18-wheeler. Half of the Carnot-efficiency power is driving the wheels and overcoming rolling and aero resistance -- the other third is oil-bearing friction in the oversized engine with the remaining sixth the "pumping loss" for throttled engine operation. The 757 throttles its engines without pumping loss by flying at high altitude in thin air, and turbine engines don't have the friction losses of piston engines so they can operate "part load" using atmospheric "throttling" very efficiently.
Now for the train. The train is hyper efficient at slow speeds where aero drag is minimal and on flat level ground and with a minimum of starts and stops. When you make the train heavy (Federal safety standards), draggy (the only Amtrak train that is highly streamlined is the Acella), and have it climb the Western mountains, it starts using fuel.
L.A. Marre (1995) The Contemporary Diesel Spotter's Guide 2nd Edition, Kalmbach, Waukesha, Wisconsin, p 294 states that Amtrak experimented with a 9300 gallon fuel tender car to allow the Empire Builder to go Chicago-Seattle nonstop. Allow another 300 gallons for the tanks in two EMD F-40's, assume a 10 car train carrying 400 passengers (maybe 80 seats in a Hi-level coach, fewer in sleeping cars, no revenue seats in the lounge and dining cars), 2500 mile trip and you are at 81 seat-MPG at full load, and that train is generally sold out at least during the summer.
As I said, a single Diesel engine pulling a string of double-decker commuter cars probably checks in at 400 seat-MPG, but that is high-density seating, no dining car, slower speeds, and level ground. But if you pack a family of 4 into a minivan (25 highway MPG), you don't have to feel guilty about not taking the Empire Builder.
12 MWH/tonne of fuel is assuming 100 percent thermal efficiency. The TGV is electric powered (from a nuclear power plant in France?), but assume you generate that power in a Diesel-electric locomotive, perhaps 30 percent thermal efficiency is reasonable, and then you are in the range of 130 passenger-MPG, which is not too far off the capabilities of perhaps a propfan powered medium range airliner. If you are going to go that fast that close to the ground, it is going to take some power.
Didn't they, like, try to dig a "whole to China", or at least drill a hole down to the crust/mantle boundary named after that Croatian dude whose name I can't pronounce let alone spell? And wasn't there an NSF grant to dig the whole, and they subcontracted the drilling to Halliburton or Brown and Root or somebody, and they burned through all the money before they got very far and the grant didn't get refunded?
Well I guess I haven't kept up with what is happening in the light plane market these days -- perhaps there are brighter horizons ahead.
There was a Golden Age of private aviation -- perhaps the mid 1960's, when a Piper or a Cessna was competitive with a luxury car. What has happened since then is that liability insurance has driven the light plane manufacturers into the ground and priced light planes out of the market.
We can all get mad at lawyers and call for tort reform and exemptions for plane manufacturers. But flying a light plane is a much riskier activity than driving a car, and the high liability insurance making planes really expensive is society's way of saying that we place a high value on human life, or at least on human life lost in transportation accidents, and the legal system coupled to the market system has perhaps made the correct decision in trying to get people to drive rather than fly themselves.
You mention the "steam gauges" and the need for a glass cockpit in a light plane. The "steam gauges" are there because they are simple and reliable -- and perhaps safer unless there is an enormous breakthrough in light plane avionics.
The engine controls are very primitive and manual: throttle, mixture control, and in some cases, propeller speed: not much more sophisticated than a lawn mower. If you have a turbo engine, you have manual control over boost pressure and have to follow rules for both advancing and retarding the throttle so as to not ruin the engine. In the 1980's there as some attempt at modernization: Porsche came out with an engine with electronic controls and "single-lever power control." But I don't know if this changed the general market trend that light plane manufacturers went out of business or went high-end (half-million dollar plus airplanes), and the only affordable planes are the ones stamped "Experimental" (i.e. I built it myself so I can't sue anybody).
The airplane has some interesting scaling laws. The 747, say, gets 60 seat-MPG (you probably have to look at some actual data rather than max range data because max range has reserve for headwinds, diversion to alternate airport, but the 60 seat-MPG is not far from the mark). I am guessing, but I would say a 757 may get close to 100 seat-MPG New York-LA (the 757 is a lighter plane which is lighter per passenger, carries less fuel than a trans Pacific flight, the PW 2037 is a really efficient engine). But if you operate the 757, say Minneapolis to Detroit, I would bet the fuel mileage may drop to 40-50 seat MPG because of the energy cost of climbing to cruising altitude, low-and-slow flight in the landing pattern.
From reading a trucking magazine at a truck stop, I learned that an 18-wheeler with good aerodynamic fairings and a late-model Diesel engine gets about 6 miles per gallon. If you figure 8 miles per gallon for a bus, an intercity bus can give over 300 seat-MPG. As for the train, I am willing to guess that for a transcontinental train like the Chicago-Seattle Empire Builder, what with sleeping and dining cars, U.S. standards that passenger cars be built like tanks to withstand derailments (European passenger cars are much lighter), and all of the mountain passes on the western routes, the train is probably little different than the airplane and at the 50-60 seat-MPG range.
These Diesel commuter trains where they have one locomotive pulling 6 or 7 100-passenger double-decker cars may get something in the 500-600 seat-MPG. But if you go for acceleration and frequent stops such as a subway train, or if you go for high speed, I bet your seat-MPG start working their way back to the car and airplane range.
Interesting that you should mention high-speed rail. Part of what makes my Taurus get 31 highway MPG is that I pretty much stick to 65 MPH. All those dudes passing me doing 75 or 80 are probably getting more like 25 MPG or less. I read that the highly-streamlined but quite fast TGV's are maybe a factor of 2 better than air - I am guessing maybe about 100 seat-MPG. And that 300 MPH Japanese maglev may be comparable to air, perhaps in the 40-50 seat-MPG range.
If you start going fast, think of going at airplane speed but doing it at ground level where the air is thicker. Part of how jets get their efficiency is that they fly high -- the thin air reduces the power on their turbine engines without the losses of throttling, and the thin air allows them to go fast without too much drag. They have to pay for that efficiency with a long climb to cruising altitude. Coupled with large-long range planes being heavier, there is probably a sweet spot in efficiency for perhaps a 1500-3000 mile trip.
OK, I am the spoiled 3 year old who didn't get the toy I wanted, but what does that make, say Ford Motor Company? Ford puts a lot of effort into their own R&D and patents because they don't want to pay tribute money (i.e. patent royalties) on anything going into a car because that really kills them on their bottom line. I guess the tribute money they pay for office-worker desktops doesn't bother them, although they are big enough they probably could have their own guys customize a Ford Linux Desktop, although Bill is probably smart enough to cut them a deal on a Windows site license below that cost.
Just because I call it tribute doesn't mean it doesn't make sense under circumstances to pay it. The reason I call it tribute is that I want to draw a distinction between, say my heating contractor charging me $700 to fix my central air (my heating contractor is a pirate, but the money I am paying is not tribute because parts are installed in my system and a guy has to come over to the house and do stuff) and what economists call an economic rent.
Tribute money would be if say, my central air ran this control program called Carrier 98SE, and every couple of days the AC would just quit and I would have to go in the basement, throw a switch, wait 10 minutes, and watch a message flash on the screen telling me that I shouldn't have turned the AC off by turning off the power. I could pay $100 to install a control program called Carrier XP, only I had to connect the air conditioner to the phone to be allowed to turn it on, and by connecting the AC to the phone, not only would the AC run constantly until the coils froze up, ten of my friends would have their AC do the same thing in the next week.
Visual Basic (and Delphi) famously let you connect events in child widgets to functions of the main form. Visual Basic doesn't have a way to use the Designer to connect child widgets to each other and while Delphi allows such connections, it doesn't get used that much. The result is that a Visual Basic main form is one big honkin' GoF Mediator Pattern -- every last thing that happens in a widget signals the main form, and the main form code is this rats' nest of figuring out what to do with all of those signals.
As far as I can tell, Mac OS-X Cocoa Interface Builder has these controller objects so the rats' nest of event processing code can go into one or more controller objects. You may say this is no big deal, Qt has had its "signals and slots" for some time now. True, but I think the innovation is that Apple allows you to do the Visual Basic-y think with signals and slots: automatically generate code for one or more controller classes rather than funnelling everything through the main form.
The reason for the link is that Apple is this amalgam of the corporate and Marin County buzzword cultures, and they go far beyond calling a program a solution. On the other hand, buried under the buzzwords seems like some neat stuff. Should I be wary of Apple?
Spare me the explanations of the poor starving software developers; I am fully aware that a software developer seeks renumerations for one's labors, and charging license fees and upgrade fees is a way to amortize the effort required to develop a complex piece of software. That doesn't change the fact that license fees are a kind of economic rent (i.e. money you can rake in because the law grants you a limited monopoly -- you can say that software won't get developed in the absence of such a monopoly, but that doesn't change the material facts that "intellectual property" law has the intent of granting limited monopolies to facilitate collecting economic rents).
I prefer the term tribute money to "Microsoft tax" because "tax" suggests governmental power and some sense of the consent of the governed. Microsoft is not to be dignified by considering it a government -- it is more like such extragovernmental entities such as high-seas pirates, Mafia bosses, feudal lords, and Delaware corporations in that money payed to them to avoid punishment (i.e. lawsuits, getting wacked) is to be called tribute and not a tax.
I also differ with the common usage of "pirate" to denote someone who avoids paying tribute money. I use the term "pirate" to describe contruction contractors that you bring into your house for remodeling and repair work. The reason contractors are pirates has less to do with the amount of money you pay them than the part about when you let them into your house they control every aspect of your life. Yes, it is about the money because whatever contract you sign, there is some uncontrolled eventuality that you have to agree to spending more money once work commences, but even if you are rich enough that the money spent is a minor concern, you become their pirate-hostage regarding letting them in and out of the house at their whim and work schedule.
So construction contractors are pirates simply on the basis that their clients are pirate hostages, and money spent for the XP upgrade when 98 was working just fine for you, thank you, is tribute money.
Everyone is focused on the notion that McMurdo is trying to teach this guy a lesson. What about the idea of keeping the guy from taking another shot at getting himself killed? The guy has already crossed oceans and circled the globe. Well, some 70 years ago there was this hot babe with this rich hubby who was crossing oceans and circling the globe who got herself killed doing something dumb, reserve fuel and navigation-wise. Someone should have refused to sell her fuel.
The real crisis in the recording industry is not illegal copying but Britney Spears. Because you have paid for music up front you will start downloading Britney's tracks, which you wouldn't have dreamed of doing in the absence of the tax. That is the only way the industry will stay solvent given all the drek they are peddling.
Another question: who is getting all this money from the blank-CD tax? Is it all Canadian artists, or is all the money going to the record companies in Los Angeles? If you are levying these taxes to fund copyright holders who are mainly south of the border, why are you getting out of it?