software engineering isn't like any other type of engineering process. Not entirely true. A lot of software shops pretty much write the same program over and over again, and these shops can do accurate estimates if the management is competent. This is like a civil engineering firm that specializes in steel truss bridges -- give them the length of the spans, maximum weight, and roadbed width, and they can crank out the design in a quite predictable time and cost to build.
However, much software development is a wild plunge into the unknown, like building a bridge out of new materials using a whole new concept for holding up the bridge. And civil engineers just don't do that; the first steel bridges were copies of wooden bridges, and grossly overbuilt just in case, then once they got comfortable with steel they gradually moved to more efficient uses of it. Nor do automotive engineers design a whole new car from scratch, without referring 90% of the design back to things that worked OK on the last model... Previous designs are re-used to reduce the risk and uncertainty.
Software producers are caught in a bind here. If you are re-using a previous design, why should the customer pay a lot for copy-and-paste? It's just bad management if you know enough about the job from experience to estimate it accurately and still have to do much coding... And if a large part of the job is new and genuinely does require new code, then it's a high-risk project, and likely to cost more than the customer is willing to pay.
The Extreme Programming methodology has a way of dealing with this: basically, you only make predictions a few weeks in advance. Correct, given reasonably competent management, it should be possible to estimate what you can do in the next 3 weeks pretty accurately. The problem is, usually the customer wants to know the cost of a 6-12 month project up front, before they decide whether or not to even start it. Be honest about this ("You don't know exactly what you want yet, and even if you did no one can tell you what it will cost with any accuracy") and most customers will go to someone willing to _lie_ to them...
To accuratly plan a software release you must have the project, and all it's complexities and nuances down COLD. otherwise you are not giving an estimation, you are giving a guess based upon incomplete knowledge.
The bulk of the work of programming consists of getting all the complexities and nuances down cold. Once you really and completely understand what is required, coding is trivial.
This leads to a thoroughly unrealistic method of estimating software costs:
1) Work for months on the specs.
2) Get the customer to sign on to those incredibly detailed specs, even though he doesn't understand them.
3) Go and code it, no spec changes allowed.
8-)
The article mainly talks about the mathematics of estimating complexity. This is a lot like the proof that you cannot determine when or whether a computer program will end -- it's true for pathological programs, but it has little relevance for the real world. You try to write the code so the conditions for the program to end are clear. If it gets into an endless loop, you probably got a conditional expression backwards and you'll recognize it immediately once you figure out which loop is running endlessly... Likewise, there may be well-defined specifications for which it is impossible to estimate the coding time, but the usual problem is poorly-defined specs, which obviously makes any estimate a guess.
If you have new systems every 2 or 3 years, you don't really need to be concerned if it's built well enough to run reliably for 7 or 8 years, now do you? For desktops and for servers you can afford to have down now and then, no. For mission-critical servers, yes. You want a MTBF of decades (if possible), so there is a very small chance of it going down in the 2 or 3 years before it's replaced.
That's the hardware requirement -- very good odds that nothing will break. Software requirement is, first, the only reason to _ever_ reboot is if you had to shut down to replace hardware, and second, a way to automatically switchover to the backup server if there is a hardware failure.If you see a Windows system in that sort of application, the person in charge is an idiot -- simply because rebooting to install a security patch is unacceptable.
it's going to look a bit strange in the final cut if there are clouds in the sky and puddles everywhere in one scene but then it's gloriously sunny and dry two seconds later. Not to mention when your hero steps out the door and instantly looks like he's been standing in the rain all day. Not to mention the runny make-up... 8-) But if you want authentic Scottish or Irish scenery, you are going to spend a lot of time in the rain.
Re:Ternary has been known to be efficient...
on
Ternary Computing
·
· Score: 2
(1) AFAIK, there aren't any scientists and engineers working on ternary logic, just mathematicians. Can you point me to any circuit designs, or anyone working on this that is capable of designing practical circuits?
(2) Tristate is NOT ternary logic. The input circuits on tristate chips are just binary; I think a ternary input circuit would need two comparators, where a binary input uses one. On the output end, if you want the same performance as binary you need four transistors -- two to bring the output to the middle state from either direction, plus the normal high and low transistors. If you don't mind wasting power and slowing the transition to "middle", you could use two output transistors (like tristate) plus resistor(s) to bias it to the middle voltage when both transistors are off. (Do you know what Resistor-Transistor-Logic was?) (Some busses require resistor termination anyhow for proper behavior at high-speed, but in modern designs a capacitor and resistor termination, with no steady-state current draw, is usually preferred to just a resistor.)
I will charitably assume that you found the articles/.ed... Warner Brothers is trying to sell DVD's under a two tier pricing scheme, like $24 (Australian???) for home use, but $55 for rental disks. So they are trying to sue rental operators that rent the "home" disks. Aussie law says that they can't do that for film, but can for movies. If they have to pay that much more for the disks, they will have to increase the rental rates...
Blackwulf, if I understand your post, the movie company first sell the videos for (say) $100. Of course, hardly anyone except a video store could afford that. Then after a few months when sales fell off, they drop the price to (say) $20. This is a bit different from selling DVDS for two different prices _at the same time_.
It's not that you can't sell to different customers at different prices at the same time, but whether you should be able to get the courts to help you make sure your intended high-price customers don't go shopping in the low price section...
Yes, Australia is pretty isolated and would benefit from mach 5+ airliners. But the tickets are going to be very expensive. Do you get Concorde service down there? I think the Concorde would have to land for fuel twice to reach Europe or North America. And even with those undersized fuel tanks, and going just barely supersonic, there isn't enough lift capacity left for very many passengers, so a Concorde ticket costs twice as much as a 747 ticket. Scramjets will cost even more per passenger mile.
One thing I'm not sure about is the real requirements for lighting a scramjet. One article said Mach 5, another Mach 1. The present Aussie design obviously won't light at Mach 1, or they wouldn't have used such an extreme flight profile. (For a test run at Mach 1 or 2, you attach the scramjet to a supersonic fighter jet. Instead, they shot a rocket into space, but not angled to go into orbit. When it fell back, the scramjet was to light for a few seconds, then cut off before getting into too thick air. I'd think even Mach 5 was reachable more easily...) Maybe once they manage to light the thing, the test data will help them design for lower-speed ignition.
This is needed because a Concorde with an added set of scramjet engines wouldn't be able to carry passengers, and the Concorde maxes out at Mach 1.something. If you have to get higher and faster than a Concorde before you can light the scramjet, it will be competitive only with rockets. If you can light it at Mach 1 and near sea-level, you could launch your scramjet from a catapult. A catapulted passenger scramjet would still need a set of low-speed engines so it could break off from a landing and go around, but they wouldn't have to be capable of extended operation or of getting the plane much above landing speed.
I guess if I ran a pr0n site I could label the index page, advertise only unlabeled subdomains and be protected from lawsuits. There are restrictions on advertising too; I think SafeSurf intends that their filter can also ensure the kiddies never see the advertisements.
So what happens if the kiddie does somehow get hold of the URL to an unlabeled subdomain (copy bookmarks on daddy's computer, e.g.)? I think the filter would block unlabeled documents unless you worked your way down to them from an acceptably labeled domain.
Of course, this will play hob with following URL's to MOST subdomains...
Re:Ternary has been known to be efficient...
on
Ternary Computing
·
· Score: 1
armchair electronic hobbyists
6 years as an electronics repair technician, USAF.
BSEE. 13 years as a test engineer.
I don't design logic IC's, but I might know more about circuit design than you... And I don't see ternary logic in gate schematics. Each transistor is either on or off (in-between states suck lots of power, so you avoid them). The output circuits in both TTL and CMOS are usually two-transistor "totem pole", but they are quite incapable of outputting an in-between state. They can be tri-state, but the third state is to turn both transistors off, allowing other chips to put binary signals on the same wires. Changing that to output a middle state would require more complexity; designing the input circuit to recognize the middle state would be even harder, I think.
Yes, the Germans noted this with the first V2 launches. The rocket appeared to zigzag all over the sky, but (usually) it was just the winds pushing the smoke trail in different directions.
Re:Ternary has been known to be efficient...
on
Ternary Computing
·
· Score: 2
A sensible scheme would have been +ve rail, -ve rail, and ground.
First, in a ternary scheme, you have high, middle, and low voltages. You can call low, middle, or high "ground" but it makes no difference to the circuit design. Since transistors only work one way, what you seem to be suggesting would come out as a dual binary circuit -- one half processing plus or ground signals, the other half processing ground or minus signals. 100% more circuitry for 50% more information. It does save on wiring between circuits -- but in most cases for the cost of a ternary system you could instead make a binary circuit run twice as fast and get more data down the wire.
The other alternative is to simply ignore ground and reference everything to the supply rails. Analog op-amps do this. It really doesn't make any difference to low-power analog circuitry whether "ground" is considered to be one of the supply rails or a point half-way between. You could certainly use such circuitry to perform trinary logic, but in terms of price vs performance it's probably an anti-optimal design.
The fundamental advantage of binary digital circuits is that they are quite easy to stabilize. You want each transistor to go all the way on or all the way off, so as to give outputs that are as close as possible to the supply rails. For example, with a 3.0V supply, typical output voltages would be almost ground and almost 3V, or about 0.2 and 2.8V. In the most basic circuit (an inverter) inputs that aren't that sharply distinguished (say 1.0 and 2.0V) get amplified into 2.8 and 0.2V outputs, and this works even with widely varying transistor characteristics, and without requiring additional components for negative feedback. So binary circuits operate reliably even in a noisy world with only a few components for each bit.
Analog circuits require in-between inputs to produce proportionally in-between outputs. Transistor gain and other characteristics vary quite widely from transistor to transistor, and vary in the same transistor with operating temperature. To reliably represent a middle voltage, you have to use resistors (resistor values are much more consistent) to create negative feedback loops. This feedback discards most of the transistor's gain, in order to ensure that the gain is set by the resistors and not by the transistor. The low effective gain then requires adding more transistor stages to get the required performance. So analog circuits require more transistors and many more resistors. Using one as a tri-level logic device does not make much sense; it is going to cost at least twice as much, with only 50% more information carried. If you really want to carry logic signals through analog circuitry, use more than three voltage levels to carry more information.
Ternary is quite nice mathematically, but the only practical application seems to be ternary decision trees and menus.
Right. You can do pretty good symmetric (secret) key crypto even with 8-bitters. So if you can figure out how to download and run a program in your 3G phone, you can send the symmetric key by other channels (hand carry, or RSA on a desktop), then use it encrypt and decode messages. What you can't do is what PGP does: create a random symmetric key for each message, and send that using RSA encryption...
Messages passed on a public system have to have un-encrypted destination headers, and in many cases the system will also require an accurate unencrypted source ID. They can't tell what you are saying, but they can tell who you are saying it to, and how often you exchanged messages, etc. Last week a man was convicted of murder, partly because his cell phone made a call from the murder scene at about the time of the murder...
For counter-terror, the issue is deciding whose phone needs to be watched -- and then identifying that phone. Disposable phones will certainly complicate that. Encryption isn't that much of a complication. If they are sending messages in the clear, they'll probably disguise the meaning anyway ("Meet you at the WTC on the 11th"), so the most important thing you'll find out is who your suspects talk to.
Toys which are so inflexible as to only do one trick (and insist on doing regardless of what the child wants) will end up at the bottom of the closet shortly after having demonstrated their one trick.
Toys like that aren't sold to children, they are sold to parents. I think to parents that spend too much time making money and not enough with the kids... You'd think they'd notice when the kids put the toy through its single trick three times, then play with the box.
Have you been in tourist class? You'd have to cut the number of seats almost in half to get room to swivel them. That will double the cost of tickets -- which would already cost much more than Concorde tickets, because this bugger needs bigger conventional engines than anything ever built to get going fast enough to light the scramjet...
Really, this isn't a supersonic airliner engine. It's a space shuttle engine.
There's a difference between a few seconds of sideways or forward force in a car and hanging from the seatbelt for several minutes while a plane decelerates. I don't have measurements, but I'd estimate the deceleration I experience in a jetliner on landing approach as.1g most of the way (with flaps and spoilers), rising to perhaps.2 to.3g when the wheels are on the ground and they use wheel brakes plus thrust reversers. It's not real uncomfortable, but the time duration is a lot shorter.
The prop planes that serve the smaller airports descend faster and with more deceleration. This gives me an upset stomach, much more than bumpy takeoffs or level flight, but I don't think this is due to deceleration pressure against the belt buckle. Maybe it's the cabin pressure increasing? Whatever, if the same conditions were maintained for much more than 10 minutes I would be using the little bags every time...
If you made the chairs swivel 180 degrees for descent and landing, then you could decelerate at 1 G -- only 57 miles. But the tickets would be very expensive...
This, of course, is the real problem with supersonic airliners on medium length runs. You go to the airport 2 hours before flight, then the airplane sits on the ground an hour or so after you board, you wait around for your baggage 1/2 hour after the flight, maybe you spend hours more in customs and immigration. So why are you going to spend a whole lot more to shave an hour or two off the flight time? The Concorde costs twice as much per passenger as a 747 and cuts transatlantic flights to 4 hours from 8, but it only gets a small percentage of the passengers.
With a scramjet airliner you'd have something even more expensive. If it can go trans-Pacific nonstop, that would be viable -- 20 hours Minneapolis to Singapore is way too long a time in an airplane... But flying NYC to Paris in under two hours is not worth that much when it takes 3 hours to get in the air, and hours more to get off the airplane and out of the airport.
This is definitely the best tall tale of the 20th century. (If you're browsing at too high a level to see the parent post, it just gives this URL:
http://www.geocities.com/Baja/Canyon/7665/)
How do I know it's a tall tale? They welded the railroad cars to the Impala's frame. Then they used air shocks to rase the Impala, so that when they let the air out, the car body would come down and press the crossbar down on the tracks to brake it. But just exactly where did they attach the shocks? The only way that would work is if the railroad cars were welded to the springs, but then I think it would come apart before it got very far. Also, the pipe-mount for the rocket was welded "to both the rear railroad car and the frame."
Other things. Assuming a JATO bottle sold as surplus in the 70's probably originated pre-1950, it probably wouldn't have been that powerful (to drive a 1-1/2 ton car 0-200 in 2.2 seconds). Few WWII planes were all that big, and I don't see where you could attach something that powerful without damaging the plane...
At multiple g accelerations, the issue isn't just people passing out or throwing up -- if you let the general public on the plane without physicals, some of them will die.
Then, you've also got to decelerate. You can't do that at even 1g, because the passengers and crew would basically be hanging from their seatbelts. I think the maximum tolerable for the "fit" general public would be.1 to.2g. Deceleration would take 11 to 23 minutes and cover 286 to 573 miles. On a NYC to Chicago flight, most of the flight would be deceleration... NYC to LA would work, but the tickets would be extremely expensive. And that's the reason that the Concorde is not more used on trans-atlantic flights -- few people find it worthwhile to spend maybe $1,000 extra just to save four hours. The Concorde was a gov't subsidized prestige project; I don't know if they've ever made back the R&D costs.
Accelerating at 1/10g (3.2 ft/sec/sec) to keep the passengers comfortable:
Mach 5 = 3000mph at high altitude = 4400 ft/sec
Time to reach this speed: 4400/3.2 = 1375 sec = 23 minutes
Distance covered while accelerating: 1/2(4400)(1375) = 3,025,000 feet = 573 miles
And your deceleration is going to be at least as long, so it's not worthwhile on trips much shorter than LA-NYC.
If the customers can be less comfortable, accelerate at 1g and your acceleration covers 57 miles in 137.5 seconds. But the ticket is going to be very expensive -- the plane has to go faster than most fighter jets BEFORE it can turn on the scramjets. So it needs ungodly big conventional engines, or else rockets. And the seats have to swivel around because you really don't want to hang from your seatbelt in 1g deceleration for 2 minutes.
Someone mentioned accelerating at 8g. This only takes 7 miles -- you could launch from an electromagnetic catapult to scramjet speed, IF the scramjets will run in thick air at low altitude. Of course, your passengers need a thorough physical, and there is still that deceleration issue.
Scramjets are not for airliners. The military applications should be obvious -- and you can use something like the space shuttle boosters in that case. Other than that, they might be useful for space launches -- get to Mach 5 with solid-fuel rockets, then use the scramjets to get to the outer fringes of the atmosphere at Mach 15 or so before you have to switch back to rockets. (The trouble with rockets is that they carry their own oxygen, which outweighs the fuel...)
I'd call it excellent intelligence-spoofing on both sides. The MIG-25 was apparently first built to intercept a rumored bomber version of SR-71. I don't think we actually flew SR-71's into Soviet airspace very often because they would have eventually managed to knock one down, but they definitely couldn't stop them the first time, and with nukes once is enough. The MIG-25 flew high and fast, but not very far -- so it could nail an SR-71 if they put the MIG-25 base in the right place... Then we didn't make an SR-71 bomber (and it probably wouldn't have worked anyhow, from what I know now. It had no payload, no maneuverability, and it leaked fuel all over the runway on takeoff, so you just hoped nothing went wrong and it outran the fire. If other reliability issues were handled like that...) Instead we tried for improved penetration at low-level (F-111, cruise missiles), and eventually for stealth technology.
So the MIG-25 got the secondary job of making the Americans nervous, and it certainly did since it's defects were no more obvious to us than the SR-71's were to the Russians. But the steel construction isn't that unreasonable -- it made it heavier, cutting down on fuel or weapons capacity, but this didn't affect the design mission. The vacuum-tube electronics wouldn't be burned out by EMP when the first nuke dropped. The most reassuring thing we learned from Belenko may have been that this complex piece of precision equipment was serviced by drunks...
The Navy likes manual controls because they _work_. Sailors are more likely to survive hits and still do their job than electronic controls -- and sailors don't corrode away in salt spray and tropical heat, but electronics sure will. So you definitely should have manual controls for everything possible and drill the crew on them now and then.
That said, electronics could be used much more than it is to bring information to the few men actually controlling the ship, and to give them more direct control. E.g., one reason that sub surfaced underneath a Japanese trawler was that the sonarman couldn't get through the crowd of civilians to mark contacts on the whiteboard. Whiteboard??? Why not a big flat-panel display so the sonar & radar operators can mark all contacts without leaving their stations?
Finally, I was under the impression that even in WWII, subs had mechanical interlocks on the torpedo tube doors and firing mechanism. That is, you're blocked from opening the inner door if the outer door is open and vice versa, and whatever it is that switches the torpedo on shouldn't work until the inner door is closed and the outer door open. Plus the crewman should check the indicator for the outer door before firing. But obviously even mechanical interlocks break sometimes, and the crew must have ways of defeating them when needed.
All this is second-hand at best. I was an electronics tech in the Air Force myself (relatively cushy job, right -- that's why I picked that branch to enlist;-). My son is on the USS Ogden (LPD -- that's a troop transport with a well-deck for landing craft). It's definitely heavy on the manual control, but then it's 50 years old!
any such aircraft must... also be capable of taxing. I'd prefer it was capable of taxiing. I already pay too many taxes to the gov't without airliners getting into it too... 8-)
Slashdot Mirror
software engineering isn't like any other type of engineering process. Not entirely true. A lot of software shops pretty much write the same program over and over again, and these shops can do accurate estimates if the management is competent. This is like a civil engineering firm that specializes in steel truss bridges -- give them the length of the spans, maximum weight, and roadbed width, and they can crank out the design in a quite predictable time and cost to build.
However, much software development is a wild plunge into the unknown, like building a bridge out of new materials using a whole new concept for holding up the bridge. And civil engineers just don't do that; the first steel bridges were copies of wooden bridges, and grossly overbuilt just in case, then once they got comfortable with steel they gradually moved to more efficient uses of it. Nor do automotive engineers design a whole new car from scratch, without referring 90% of the design back to things that worked OK on the last model... Previous designs are re-used to reduce the risk and uncertainty.
Software producers are caught in a bind here. If you are re-using a previous design, why should the customer pay a lot for copy-and-paste? It's just bad management if you know enough about the job from experience to estimate it accurately and still have to do much coding... And if a large part of the job is new and genuinely does require new code, then it's a high-risk project, and likely to cost more than the customer is willing to pay.
The Extreme Programming methodology has a way of dealing with this: basically, you only make predictions a few weeks in advance. Correct, given reasonably competent management, it should be possible to estimate what you can do in the next 3 weeks pretty accurately. The problem is, usually the customer wants to know the cost of a 6-12 month project up front, before they decide whether or not to even start it. Be honest about this ("You don't know exactly what you want yet, and even if you did no one can tell you what it will cost with any accuracy") and most customers will go to someone willing to _lie_ to them...
The problem is, you don't get paid for coding up a small working prototype in order to do an estimate. So my estimating technique is:
Figure the time to do the parts I understand.
Count the parts I don't understand. Allow a very long time for each of them.
Add it all up, then multiply by 3
To accuratly plan a software release you must have the project, and all it's complexities and nuances down COLD. otherwise you are not giving an estimation, you are giving a guess based upon incomplete knowledge.
The bulk of the work of programming consists of getting all the complexities and nuances down cold. Once you really and completely understand what is required, coding is trivial.
This leads to a thoroughly unrealistic method of estimating software costs:
1) Work for months on the specs.
2) Get the customer to sign on to those incredibly detailed specs, even though he doesn't understand them.
3) Go and code it, no spec changes allowed.
8-)
The article mainly talks about the mathematics of estimating complexity. This is a lot like the proof that you cannot determine when or whether a computer program will end -- it's true for pathological programs, but it has little relevance for the real world. You try to write the code so the conditions for the program to end are clear. If it gets into an endless loop, you probably got a conditional expression backwards and you'll recognize it immediately once you figure out which loop is running endlessly... Likewise, there may be well-defined specifications for which it is impossible to estimate the coding time, but the usual problem is poorly-defined specs, which obviously makes any estimate a guess.
If you have new systems every 2 or 3 years, you don't really need to be concerned if it's built well enough to run reliably for 7 or 8 years, now do you? For desktops and for servers you can afford to have down now and then, no. For mission-critical servers, yes. You want a MTBF of decades (if possible), so there is a very small chance of it going down in the 2 or 3 years before it's replaced.
That's the hardware requirement -- very good odds that nothing will break. Software requirement is, first, the only reason to _ever_ reboot is if you had to shut down to replace hardware, and second, a way to automatically switchover to the backup server if there is a hardware failure.If you see a Windows system in that sort of application, the person in charge is an idiot -- simply because rebooting to install a security patch is unacceptable.
it's going to look a bit strange in the final cut if there are clouds in the sky and puddles everywhere in one scene but then it's gloriously sunny and dry two seconds later. Not to mention when your hero steps out the door and instantly looks like he's been standing in the rain all day. Not to mention the runny make-up... 8-) But if you want authentic Scottish or Irish scenery, you are going to spend a lot of time in the rain.
(1) AFAIK, there aren't any scientists and engineers working on ternary logic, just mathematicians. Can you point me to any circuit designs, or anyone working on this that is capable of designing practical circuits?
(2) Tristate is NOT ternary logic. The input circuits on tristate chips are just binary; I think a ternary input circuit would need two comparators, where a binary input uses one. On the output end, if you want the same performance as binary you need four transistors -- two to bring the output to the middle state from either direction, plus the normal high and low transistors. If you don't mind wasting power and slowing the transition to "middle", you could use two output transistors (like tristate) plus resistor(s) to bias it to the middle voltage when both transistors are off. (Do you know what Resistor-Transistor-Logic was?) (Some busses require resistor termination anyhow for proper behavior at high-speed, but in modern designs a capacitor and resistor termination, with no steady-state current draw, is usually preferred to just a resistor.)
Oops. Aussie law says that they can't do that for film, but can for software.
I will charitably assume that you found the articles /.ed... Warner Brothers is trying to sell DVD's under a two tier pricing scheme, like $24 (Australian???) for home use, but $55 for rental disks. So they are trying to sue rental operators that rent the "home" disks. Aussie law says that they can't do that for film, but can for movies. If they have to pay that much more for the disks, they will have to increase the rental rates...
Blackwulf, if I understand your post, the movie company first sell the videos for (say) $100. Of course, hardly anyone except a video store could afford that. Then after a few months when sales fell off, they drop the price to (say) $20. This is a bit different from selling DVDS for two different prices _at the same time_.
It's not that you can't sell to different customers at different prices at the same time, but whether you should be able to get the courts to help you make sure your intended high-price customers don't go shopping in the low price section...
Yes, Australia is pretty isolated and would benefit from mach 5+ airliners. But the tickets are going to be very expensive. Do you get Concorde service down there? I think the Concorde would have to land for fuel twice to reach Europe or North America. And even with those undersized fuel tanks, and going just barely supersonic, there isn't enough lift capacity left for very many passengers, so a Concorde ticket costs twice as much as a 747 ticket. Scramjets will cost even more per passenger mile.
One thing I'm not sure about is the real requirements for lighting a scramjet. One article said Mach 5, another Mach 1. The present Aussie design obviously won't light at Mach 1, or they wouldn't have used such an extreme flight profile. (For a test run at Mach 1 or 2, you attach the scramjet to a supersonic fighter jet. Instead, they shot a rocket into space, but not angled to go into orbit. When it fell back, the scramjet was to light for a few seconds, then cut off before getting into too thick air. I'd think even Mach 5 was reachable more easily...) Maybe once they manage to light the thing, the test data will help them design for lower-speed ignition.
This is needed because a Concorde with an added set of scramjet engines wouldn't be able to carry passengers, and the Concorde maxes out at Mach 1.something. If you have to get higher and faster than a Concorde before you can light the scramjet, it will be competitive only with rockets. If you can light it at Mach 1 and near sea-level, you could launch your scramjet from a catapult. A catapulted passenger scramjet would still need a set of low-speed engines so it could break off from a landing and go around, but they wouldn't have to be capable of extended operation or of getting the plane much above landing speed.
I guess if I ran a pr0n site I could label the index page, advertise only unlabeled subdomains and be protected from lawsuits. There are restrictions on advertising too; I think SafeSurf intends that their filter can also ensure the kiddies never see the advertisements.
So what happens if the kiddie does somehow get hold of the URL to an unlabeled subdomain (copy bookmarks on daddy's computer, e.g.)? I think the filter would block unlabeled documents unless you worked your way down to them from an acceptably labeled domain.
Of course, this will play hob with following URL's to MOST subdomains...
armchair electronic hobbyists
6 years as an electronics repair technician, USAF.
BSEE. 13 years as a test engineer.
I don't design logic IC's, but I might know more about circuit design than you... And I don't see ternary logic in gate schematics. Each transistor is either on or off (in-between states suck lots of power, so you avoid them). The output circuits in both TTL and CMOS are usually two-transistor "totem pole", but they are quite incapable of outputting an in-between state. They can be tri-state, but the third state is to turn both transistors off, allowing other chips to put binary signals on the same wires. Changing that to output a middle state would require more complexity; designing the input circuit to recognize the middle state would be even harder, I think.
Yes, the Germans noted this with the first V2 launches. The rocket appeared to zigzag all over the sky, but (usually) it was just the winds pushing the smoke trail in different directions.
A sensible scheme would have been +ve rail, -ve rail, and ground.
First, in a ternary scheme, you have high, middle, and low voltages. You can call low, middle, or high "ground" but it makes no difference to the circuit design. Since transistors only work one way, what you seem to be suggesting would come out as a dual binary circuit -- one half processing plus or ground signals, the other half processing ground or minus signals. 100% more circuitry for 50% more information. It does save on wiring between circuits -- but in most cases for the cost of a ternary system you could instead make a binary circuit run twice as fast and get more data down the wire.
The other alternative is to simply ignore ground and reference everything to the supply rails. Analog op-amps do this. It really doesn't make any difference to low-power analog circuitry whether "ground" is considered to be one of the supply rails or a point half-way between. You could certainly use such circuitry to perform trinary logic, but in terms of price vs performance it's probably an anti-optimal design.
The fundamental advantage of binary digital circuits is that they are quite easy to stabilize. You want each transistor to go all the way on or all the way off, so as to give outputs that are as close as possible to the supply rails. For example, with a 3.0V supply, typical output voltages would be almost ground and almost 3V, or about 0.2 and 2.8V. In the most basic circuit (an inverter) inputs that aren't that sharply distinguished (say 1.0 and 2.0V) get amplified into 2.8 and 0.2V outputs, and this works even with widely varying transistor characteristics, and without requiring additional components for negative feedback. So binary circuits operate reliably even in a noisy world with only a few components for each bit.
Analog circuits require in-between inputs to produce proportionally in-between outputs. Transistor gain and other characteristics vary quite widely from transistor to transistor, and vary in the same transistor with operating temperature. To reliably represent a middle voltage, you have to use resistors (resistor values are much more consistent) to create negative feedback loops. This feedback discards most of the transistor's gain, in order to ensure that the gain is set by the resistors and not by the transistor. The low effective gain then requires adding more transistor stages to get the required performance. So analog circuits require more transistors and many more resistors. Using one as a tri-level logic device does not make much sense; it is going to cost at least twice as much, with only 50% more information carried. If you really want to carry logic signals through analog circuitry, use more than three voltage levels to carry more information.
Ternary is quite nice mathematically, but the only practical application seems to be ternary decision trees and menus.
Right. You can do pretty good symmetric (secret) key crypto even with 8-bitters. So if you can figure out how to download and run a program in your 3G phone, you can send the symmetric key by other channels (hand carry, or RSA on a desktop), then use it encrypt and decode messages. What you can't do is what PGP does: create a random symmetric key for each message, and send that using RSA encryption...
Messages passed on a public system have to have un-encrypted destination headers, and in many cases the system will also require an accurate unencrypted source ID. They can't tell what you are saying, but they can tell who you are saying it to, and how often you exchanged messages, etc. Last week a man was convicted of murder, partly because his cell phone made a call from the murder scene at about the time of the murder...
For counter-terror, the issue is deciding whose phone needs to be watched -- and then identifying that phone. Disposable phones will certainly complicate that. Encryption isn't that much of a complication. If they are sending messages in the clear, they'll probably disguise the meaning anyway ("Meet you at the WTC on the 11th"), so the most important thing you'll find out is who your suspects talk to.
Toys which are so inflexible as to only do one trick (and insist on doing regardless of what the child wants) will end up at the bottom of the closet shortly after having demonstrated their one trick.
Toys like that aren't sold to children, they are sold to parents. I think to parents that spend too much time making money and not enough with the kids... You'd think they'd notice when the kids put the toy through its single trick three times, then play with the box.
Have you been in tourist class? You'd have to cut the number of seats almost in half to get room to swivel them. That will double the cost of tickets -- which would already cost much more than Concorde tickets, because this bugger needs bigger conventional engines than anything ever built to get going fast enough to light the scramjet...
Really, this isn't a supersonic airliner engine. It's a space shuttle engine.
There's a difference between a few seconds of sideways or forward force in a car and hanging from the seatbelt for several minutes while a plane decelerates. I don't have measurements, but I'd estimate the deceleration I experience in a jetliner on landing approach as .1g most of the way (with flaps and spoilers), rising to perhaps .2 to .3g when the wheels are on the ground and they use wheel brakes plus thrust reversers. It's not real uncomfortable, but the time duration is a lot shorter.
The prop planes that serve the smaller airports descend faster and with more deceleration. This gives me an upset stomach, much more than bumpy takeoffs or level flight, but I don't think this is due to deceleration pressure against the belt buckle. Maybe it's the cabin pressure increasing? Whatever, if the same conditions were maintained for much more than 10 minutes I would be using the little bags every time...
If you made the chairs swivel 180 degrees for descent and landing, then you could decelerate at 1 G -- only 57 miles. But the tickets would be very expensive...
This, of course, is the real problem with supersonic airliners on medium length runs. You go to the airport 2 hours before flight, then the airplane sits on the ground an hour or so after you board, you wait around for your baggage 1/2 hour after the flight, maybe you spend hours more in customs and immigration. So why are you going to spend a whole lot more to shave an hour or two off the flight time? The Concorde costs twice as much per passenger as a 747 and cuts transatlantic flights to 4 hours from 8, but it only gets a small percentage of the passengers.
With a scramjet airliner you'd have something even more expensive. If it can go trans-Pacific nonstop, that would be viable -- 20 hours Minneapolis to Singapore is way too long a time in an airplane... But flying NYC to Paris in under two hours is not worth that much when it takes 3 hours to get in the air, and hours more to get off the airplane and out of the airport.
This is definitely the best tall tale of the 20th century. (If you're browsing at too high a level to see the parent post, it just gives this URL:
http://www.geocities.com/Baja/Canyon/7665/)
How do I know it's a tall tale? They welded the railroad cars to the Impala's frame. Then they used air shocks to rase the Impala, so that when they let the air out, the car body would come down and press the crossbar down on the tracks to brake it. But just exactly where did they attach the shocks? The only way that would work is if the railroad cars were welded to the springs, but then I think it would come apart before it got very far. Also, the pipe-mount for the rocket was welded "to both the rear railroad car and the frame."
Other things. Assuming a JATO bottle sold as surplus in the 70's probably originated pre-1950, it probably wouldn't have been that powerful (to drive a 1-1/2 ton car 0-200 in 2.2 seconds). Few WWII planes were all that big, and I don't see where you could attach something that powerful without damaging the plane...
But it's a hell of a story.
At multiple g accelerations, the issue isn't just people passing out or throwing up -- if you let the general public on the plane without physicals, some of them will die.
.1 to .2g. Deceleration would take 11 to 23 minutes and cover 286 to 573 miles. On a NYC to Chicago flight, most of the flight would be deceleration... NYC to LA would work, but the tickets would be extremely expensive. And that's the reason that the Concorde is not more used on trans-atlantic flights -- few people find it worthwhile to spend maybe $1,000 extra just to save four hours. The Concorde was a gov't subsidized prestige project; I don't know if they've ever made back the R&D costs.
Then, you've also got to decelerate. You can't do that at even 1g, because the passengers and crew would basically be hanging from their seatbelts. I think the maximum tolerable for the "fit" general public would be
Accelerating at 1/10g (3.2 ft/sec/sec) to keep the passengers comfortable:
Mach 5 = 3000mph at high altitude = 4400 ft/sec
Time to reach this speed: 4400/3.2 = 1375 sec = 23 minutes
Distance covered while accelerating: 1/2(4400)(1375) = 3,025,000 feet = 573 miles
And your deceleration is going to be at least as long, so it's not worthwhile on trips much shorter than LA-NYC.
If the customers can be less comfortable, accelerate at 1g and your acceleration covers 57 miles in 137.5 seconds. But the ticket is going to be very expensive -- the plane has to go faster than most fighter jets BEFORE it can turn on the scramjets. So it needs ungodly big conventional engines, or else rockets. And the seats have to swivel around because you really don't want to hang from your seatbelt in 1g deceleration for 2 minutes.
Someone mentioned accelerating at 8g. This only takes 7 miles -- you could launch from an electromagnetic catapult to scramjet speed, IF the scramjets will run in thick air at low altitude. Of course, your passengers need a thorough physical, and there is still that deceleration issue.
Scramjets are not for airliners. The military applications should be obvious -- and you can use something like the space shuttle boosters in that case. Other than that, they might be useful for space launches -- get to Mach 5 with solid-fuel rockets, then use the scramjets to get to the outer fringes of the atmosphere at Mach 15 or so before you have to switch back to rockets. (The trouble with rockets is that they carry their own oxygen, which outweighs the fuel...)
I'd call it excellent intelligence-spoofing on both sides. The MIG-25 was apparently first built to intercept a rumored bomber version of SR-71. I don't think we actually flew SR-71's into Soviet airspace very often because they would have eventually managed to knock one down, but they definitely couldn't stop them the first time, and with nukes once is enough. The MIG-25 flew high and fast, but not very far -- so it could nail an SR-71 if they put the MIG-25 base in the right place... Then we didn't make an SR-71 bomber (and it probably wouldn't have worked anyhow, from what I know now. It had no payload, no maneuverability, and it leaked fuel all over the runway on takeoff, so you just hoped nothing went wrong and it outran the fire. If other reliability issues were handled like that...) Instead we tried for improved penetration at low-level (F-111, cruise missiles), and eventually for stealth technology.
So the MIG-25 got the secondary job of making the Americans nervous, and it certainly did since it's defects were no more obvious to us than the SR-71's were to the Russians. But the steel construction isn't that unreasonable -- it made it heavier, cutting down on fuel or weapons capacity, but this didn't affect the design mission. The vacuum-tube electronics wouldn't be burned out by EMP when the first nuke dropped. The most reassuring thing we learned from Belenko may have been that this complex piece of precision equipment was serviced by drunks...
The Navy likes manual controls because they _work_. Sailors are more likely to survive hits and still do their job than electronic controls -- and sailors don't corrode away in salt spray and tropical heat, but electronics sure will. So you definitely should have manual controls for everything possible and drill the crew on them now and then.
;-). My son is on the USS Ogden (LPD -- that's a troop transport with a well-deck for landing craft). It's definitely heavy on the manual control, but then it's 50 years old!
That said, electronics could be used much more than it is to bring information to the few men actually controlling the ship, and to give them more direct control. E.g., one reason that sub surfaced underneath a Japanese trawler was that the sonarman couldn't get through the crowd of civilians to mark contacts on the whiteboard. Whiteboard??? Why not a big flat-panel display so the sonar & radar operators can mark all contacts without leaving their stations?
Finally, I was under the impression that even in WWII, subs had mechanical interlocks on the torpedo tube doors and firing mechanism. That is, you're blocked from opening the inner door if the outer door is open and vice versa, and whatever it is that switches the torpedo on shouldn't work until the inner door is closed and the outer door open. Plus the crewman should check the indicator for the outer door before firing. But obviously even mechanical interlocks break sometimes, and the crew must have ways of defeating them when needed.
All this is second-hand at best. I was an electronics tech in the Air Force myself (relatively cushy job, right -- that's why I picked that branch to enlist
any such aircraft must ... also be capable of taxing. I'd prefer it was capable of taxiing. I already pay too many taxes to the gov't without airliners getting into it too... 8-)