Up until about 4 years ago, I used to work for a major/small satellite company. One of my last tasks was to update the processor board used in many satellites... you guessed it, it was based on the 8086. What was my upgrade? I added an 8087 to it! These math coprocessors are even rarer - we bought the last 50 bare die in existence to eventually custom-package in a special high-density radiation shielding ceramic package.
It was about that time that I decided that the company was going in the hole. It's not that the 8086's were particularily good processors... True, they are made with a bigger geometry and suck more power -- things that make them generally more radiation resistant than anything produced in the 1990's. But, they were never designed to tolerate radiation. (NASA isn't stupid - they have high-performance radiation tolerant parts like the RAD6000).
Since my company wasn't making even minimal internal investments (they had a '386 based system that they built but never applied power), I decided that, for my career, I should leave. I notice now that they are hiring people with 5 years of PowerPC experience -- eventually they must have decided to get with the times, but since they didn't keep their employees current, they shot themselves in the foot and now have to hire outsiders.
p.s. I'm back on the job market - anyone need a kick-ass PowerPC engineer?:)
Summary: 2 guys in spacesuits taking a giant helium balloon to 40km. In the vacum of space, they'll descend at 1600-1800 kph (994-1118 mph), becoming the first man to break the sound barrier unaided (that's cool!). Plenty of video taken for marketing opportunities.
The funny thing is that both of these ballons claim to be going to the "edge of space" at 40km, but the traditional definition (and the one used for the X-prize) is 100km. By that standard, I guess I have beach-front property!
The 34010 kicked butt! It was used by Atari's Hard-driving game. It had a lot of neat features, including hardware X/Y addressing (i.e. move x,y,pixel), bit-level addressing (you could twiddle any bit in memory, or write a word/byte on any boundry), and built-in simple graphics operations (copy a block of memory, xor source & destination, use larger of the two, subtract, union, difference, add but don't overflow, etc)
But what was *REALLY* cool was the math coprocessor, the 34020. It was blazingly fast (almost, but not quite as fast as the industry-crushing i860 IIRC), but it featured a programmable microcode so you could create your own instructions and get every ounce of performance out of the machine. I'm still looking for a processor that will allow that... we're getting those with modern NPUs (cradle, intel IXP1200), but these generally lack floating point functionality.
It's always been amusing when I see an article where icopyright offers to allow linking to an article for $5. Example. Their other services, such as reprints ($250, hosted at icopyright for a guaranteed amount of time, not at the original source) seem reasonable, but this linking-to comes with no guarantees. Other than they won't try to sue you, I guess. Their description:
HTML Link permission allows you to link to a specified Web page. Clicking on the licensed HTML link, whether embedded in a logo, in text, or in some other object, results in the immediate display of the Web page. Note: linked-to content is not guaranteed to be hosted by the Web site owner for any specific length of time; refer to the publisher's License Agreement for specific terms of use.
I wonder how they would treat thier example. By clicking above, you get a page that has a link to the original article. So, linking to icopyright is just one step removed, and (so far) free. I wonder if icopyright takes this linking-to seriously. By allowing linking to their site, they can generate revenue for themselves, but at the same time, they diminish the "protection" they offer to their customers,
It took them a day to find a million sites? They have to be going for a sensationalist angle -- anyone can find 4,290,000 sites in 0.04 seconds, but extending their search time and lowering the quantity of their results implies that they've gotten higher quality results -- though I doubt it. These results -- warez+zip (43,600 hits), warez+rar (24,700 hits), and warez+iso (32,000 hits) -- show that even simple qualifiers result in far fewer possible sites.
You'd be far worse seeing the result on a 5" black and white TV, or listening to it in mono. With DVD's, the studios can't control what you use to view it, so until they way to force you to see it only on an approved viewer, they have no leg to stand on. However, that may not be too far away when we get to a HDTV/digital-only world when analog tv is outlawed in 2006.
The first was a Wittnauer (I don't think they make it anymore). It was unfinished titanium, and got a lot of scratches during the year and a half that I wore it (it has a 10-year pacemaker battery!)
The second one is a venturav-matic watch, and it's had the honor of being on my wrist for 3 years. Usually I get bored of watches, or they get scratched, so that's quite a feat! This watch has a special nitrogen coating that seems unique to ventura (I'd love to see it on non-watch products). The surface has been hardened to the hardness of saphire. Saphire is just below diamond on the hardness scale, and, yes, it scratchs glass. The watch is absolutely scratchless. It has a small ding (.5mm dia) that occured in a hangglider emergency landing (although "survived a plane crash" sounds much sexier!). But, the ding is exactly that - not a scratch. Since only the surface is hardend, the material is still soft underneath and can be dented.
Although the watch is big (pure mechanical, automatic winder), it's still light. As an engineer, I love the see-through back!! Check out the 3d viewer of it.
It's a bit spendy, but most of the price is the mechanicals inside. Ok, a good chunk (1/5-2/3) is to the retailer, but most of the manufacturing cost is probably labor.
Re:How about information, not matter
on
Time Travel
·
· Score: 2
A practical application:
A car company offers Crash Advance Warning (CAW) on their new line of cars. A black-box type instrument on the car detects a crash and records your final utterances before the crash. Presuming the box survives the impact, it would automatically place a phone call to the car manufacturer, describing the nature of the collision (rear end, roll over, etc.) and that snippit of sound ("oh sht!" or "big funky bus, passengers should've seen this coming"). This would be sent back in time at service center, and then broadcast to your CAW receiver, where you will be given advance warning of your crash.
Of course, depending on god's sense of humor, there may be nothing that you can do about it. You pull over, and a giant meteorite still smashes your car.
Re:How about information, not matter
on
Time Travel
·
· Score: 1
Moving matter would be one way to fulfill my request to send information back through time, but I didn't want to limit it to that. Anyway someone figures out how to do it, and I'll be happy!
How about information, not matter
on
Time Travel
·
· Score: 2
I'd be content just to have information travel through time - not matter. It sounds like that would be a far simpler proposition. If you can send back individual particles to exact locations and/or times, you'll have information travelling... it sounds like a quick way to make a "simple" experiment useful. Of course, it would have to be a significant amount of time to be worthy... at least a few hours.
One of us has got to dress up like Ronald Mallett-- all out, with a mask and everything, plus a scorched labcoat and frizzy hair-- and show up at his doorstep.
Actually, it can predict the future! The sensing microphones are closer to the noise source than the cancelation speaker, so they receive the noise before the speaker has to generate it. At 340 m/s, 1 foot is about 0.1 milliseconds "into the future"; most speakers have a decent frequency response in the range of 1 kHz (1/0.1 msec) and could easily respond in time.
Like the poster above said, it looks like it uses 8 noise-sensing microphones, 8 speakers, and then 8 microphones in the quiet zone to measure the sucess of the speakers and automatically make adjustments.
In a 3d world, these sense microphones are essential because they are used to compensate when the environment changes. Stand up, and it needs to compensate. Move a chair, it does, too.
It looks like the system would try to minimize the noise at these 8 places, and not worry about everything else. I'm guessing that these 8 microphones would have to be in relatively open air (kindof like your head is when you walk around - not in corners), and not near any local noise sources (such as the fan on yuor computer). Or, they may be built into the device, on a framework attached to the speakers and the front microphones.
The TI 320c32 is a decent floating point processor - 20-30 MIPS / 40-60 GFLOPS. But, (IHALEWD - I have a little experience with dsps), I don't think it's enough horsepower to entirely model a room and do super-sophisticated stuff, but it looks like it would generally get good results.
Automatically adjusting filters are neat - they can be tuned relatively easily, and since a large number of coefficients, they can describe some pretty complicated behaviour without a lot of programming. Of course, AFAIK, they assume linearity; if the air conditioner vibrations rattle a coffee cup on your desk, that's non-linear and probably won't get fixed.
And, of course, this system may require some sort of calibration -- like a firecracker that you light near the noise source. (ok, maybe just something that makes test tones or a sharp snap). The microphones can get a picture of the room by listening to all the reflections (and by knowing what the initial sound was).
is what happens at the finges of this "shadow of silence". Does it start to break down such that the anti-noise and the noise become in phase again, and you get an area of double-the-noise?
This is a lot more complicated than headphones. Headphones are relatively one-dimensional (one microphone, one speaker, one eardrum per circuit) - the only thing you have to worry about is not generating feedback.
This seems to be a more complicated 3-dimensional solution, and it'll have much more complicated problems. Does this cancel noise effectively in corners? Will a computer monitor cast a non-noise-canceled shadow? Is there a limit to the noise source (can it be all around you, or must it be generated in one specific place?)
1. Those digital outputs were on players, not the discs themselves. (Saved by my own poor choice of words -- that's a talent I should further develop!!)
2. I should have been more specific. What these guys put out is digital uncompressed music - not the original compressed data that was on the disc. To make a copy, this music has to be recompressed again. With lossless compression, this would be no problem -- all the compression/decompressions cycles would yield perfect copies. But, minidisc uses lossy [i.e. f'(f(x)) != x] compression that isn't totally symetric [i.e. f'(f(f'(f(x)))) != x ], so a little noise is introduced into each generation, even if its transfered digitally. more info here. I'd love to hear that you can read/write the compressed digital data!
I didn't know that there were digital outs [even if they're not the ones I was thinking of]! At $460, the rack unit isn't even outrageously expensive.
Ignore all the other users claiming it's SACD and DVD - they are too new to make any meaningful comparisons yet. Minidisc, however, has been promoted as a smaller-size compression-based consumer format. Copy protection is built-in; there are no digital outs on any minidiscs.
Stereo isn't good enough? What do you want, a surround-sound 5.1 channel setup? It takes quite a geek to step into a shower and think "this sounds good, but could use an over-spigot center speaker, plus two more astride the shampoo rack, plus a subwoofer built into the cabinent space under sink."
If I were you, I'd concentrate on reducing the background noise before I consider a second speaker.
Now, if you can just find a source for quality quadraphonic or 5.1 MP3's...
Not a troll, but why would you want to do 3d interlacing? AIUI (as I understand it), normal 2d interlacing was done to work around the limitations of phospher, TV scan rates, and bandwidth. The phospher needed to be scanned at 60 Hz (either there was no long-persistance phosper available so it would flicker, or long-persistance phospher was available and had a ghosting problem), but horizontal scan rates were limited to too few lines at 60 Hz.
With a new technology, and especially with all the cool stuff we can put into receivers today (full-frame buffers, image enhancement, MPEG decompression), it wouldn't make sense to be bound to a 1941 standard. (see also)
When I'm microwaving something late at night, the beeper is waay to loud. It's got a clock, it knows it's 2 in the morning, why doesn't it silence the beeper?
I thought the microwave that my parents got could solve the problem. It actually has a microphone. But, no, it doesn't use it to make sure the beep is an appropriate level - it uses it for a stupid voice recorder function. I mostly use it record sounds of small animals being microwaved (ribbit.... ribbit.... ribitribitribit!!) No! not actual animals!
Ok, so my parent's microwave has an appointment reminder. So, one night we were sleeping in the kitchen (don't ask) and needed to wake up early to catch a plane. We thought we could finally use this. We entered the time (keypad entry '3' '3' '0' 'AM'), and the day, and then it asked us what kind of appointment it was. Why'd it need to know? Who cares. We pushed the button for "doctor's appointment". It then responded "alarm set for 2:45am". Somehow, it determined that 45 minutes was the time needed to drive across town, find a parking space, and walk to rest of the way to the doctors office. I suspect that it had different times for different events, but we were too tired to try to figure it out. We canceled it and set it again for 4:15.
After my microwave is done, it'll beep every 2 minutes to remind me that my food is ready (it even says "food is done" - is that an undercooked chicken lawsuit waiting to happen?). But, as far as I can tell, it'll do this forever (my roommates haven't let me test this hypothesis yet). Even when my food is cold and dried out and totally unedible, it'll still beep. Maybe just beep a couple of times at 2 minute intervals, and then after it's cold, change to 20 minute intervals and leave the light on so I'll notice it when I'm good and ready.
Makes me long for the old-style microwaves with a "time" dial and a "power" dial, a start button, and a single bell when it's done.
One reason: Ford doesn't have to deal with the orders-of-magnitude in product variations that occur in computers, so counting vehicles means more than counting computers.
The limits of vehicles seem to be about 10:1 at most (2 passenger car : 50 passenger bus), 50:500 hp, maybe 1 ton car: 5 ton bus.
In computers you've got a much wider dynamic range... 1 user:1000 user, 32kHz (Sega VMU):2.2 GHz (P4), 1K:16GB memory, 56kbps modem : 2 Gbps fibre channel, etc.
Slashdot Mirror
Up until about 4 years ago, I used to work for a major/small satellite company. One of my last tasks was to update the processor board used in many satellites... you guessed it, it was based on the 8086. What was my upgrade? I added an 8087 to it! These math coprocessors are even rarer - we bought the last 50 bare die in existence to eventually custom-package in a special high-density radiation shielding ceramic package.
:)
It was about that time that I decided that the company was going in the hole. It's not that the 8086's were particularily good processors... True, they are made with a bigger geometry and suck more power -- things that make them generally more radiation resistant than anything produced in the 1990's. But, they were never designed to tolerate radiation. (NASA isn't stupid - they have high-performance radiation tolerant parts like the RAD6000).
Since my company wasn't making even minimal internal investments (they had a '386 based system that they built but never applied power), I decided that, for my career, I should leave. I notice now that they are hiring people with 5 years of PowerPC experience -- eventually they must have decided to get with the times, but since they didn't keep their employees current, they shot themselves in the foot and now have to hire outsiders.
p.s. I'm back on the job market - anyone need a kick-ass PowerPC engineer?
This sounds similar to this australian attempt: check this article and this (better) article.
Summary: 2 guys in spacesuits taking a giant helium balloon to 40km. In the vacum of space, they'll descend at 1600-1800 kph (994-1118 mph), becoming the first man to break the sound barrier unaided (that's cool!). Plenty of video taken for marketing opportunities.
The funny thing is that both of these ballons claim to be going to the "edge of space" at 40km, but the traditional definition (and the one used for the X-prize) is 100km. By that standard, I guess I have beach-front property!
This is Sandia. One of those governement labs with supercomputers and stuff. Like ASCI Red, the world's 3rd fastest supercomputer, for example.
They've got an OC-48 2.5 Gbps link to San Francisco. That was in 2000, they may have upgraded since then...
yeah, I know, they may have outsourced the web server to a 56k modem line, but somehow I doubt it...
That brings back memories!
The 34010 kicked butt! It was used by Atari's Hard-driving game. It had a lot of neat features, including hardware X/Y addressing (i.e. move x,y,pixel), bit-level addressing (you could twiddle any bit in memory, or write a word/byte on any boundry), and built-in simple graphics operations (copy a block of memory, xor source & destination, use larger of the two, subtract, union, difference, add but don't overflow, etc)
But what was *REALLY* cool was the math coprocessor, the 34020. It was blazingly fast (almost, but not quite as fast as the industry-crushing i860 IIRC), but it featured a programmable microcode so you could create your own instructions and get every ounce of performance out of the machine. I'm still looking for a processor that will allow that... we're getting those with modern NPUs (cradle, intel IXP1200), but these generally lack floating point functionality.
Liposuction fee = $2049
Stem cell processing = $1200
Stem cell storage = $100/yr
The self-storage option begins to look attractive!
I wonder how they would treat thier example. By clicking above, you get a page that has a link to the original article. So, linking to icopyright is just one step removed, and (so far) free. I wonder if icopyright takes this linking-to seriously. By allowing linking to their site, they can generate revenue for themselves, but at the same time, they diminish the "protection" they offer to their customers,
It took them a day to find a million sites? They have to be going for a sensationalist angle -- anyone can find 4,290,000 sites in 0.04 seconds, but extending their search time and lowering the quantity of their results implies that they've gotten higher quality results -- though I doubt it. These results -- warez+zip (43,600 hits), warez+rar (24,700 hits), and warez+iso (32,000 hits) -- show that even simple qualifiers result in far fewer possible sites.
Sorry, but a folding screen is not new... They ALL fold; this is one of the first to unfold.
How about this novel or this one?
You'd be far worse seeing the result on a 5" black and white TV, or listening to it in mono. With DVD's, the studios can't control what you use to view it, so until they way to force you to see it only on an approved viewer, they have no leg to stand on. However, that may not be too far away when we get to a HDTV/digital-only world when analog tv is outlawed in 2006.
I've had two titanium watches....
The first was a Wittnauer (I don't think they make it anymore). It was unfinished titanium, and got a lot of scratches during the year and a half that I wore it (it has a 10-year pacemaker battery!)
The second one is a ventura v-matic watch, and it's had the honor of being on my wrist for 3 years. Usually I get bored of watches, or they get scratched, so that's quite a feat! This watch has a special nitrogen coating that seems unique to ventura (I'd love to see it on non-watch products). The surface has been hardened to the hardness of saphire. Saphire is just below diamond on the hardness scale, and, yes, it scratchs glass. The watch is absolutely scratchless. It has a small ding (.5mm dia) that occured in a hangglider emergency landing (although "survived a plane crash" sounds much sexier!). But, the ding is exactly that - not a scratch. Since only the surface is hardend, the material is still soft underneath and can be dented.
Although the watch is big (pure mechanical, automatic winder), it's still light. As an engineer, I love the see-through back!! Check out the 3d viewer of it.
It's a bit spendy, but most of the price is the mechanicals inside. Ok, a good chunk (1/5-2/3) is to the retailer, but most of the manufacturing cost is probably labor.
A practical application:
A car company offers Crash Advance Warning (CAW) on their new line of cars. A black-box type instrument on the car detects a crash and records your final utterances before the crash. Presuming the box survives the impact, it would automatically place a phone call to the car manufacturer, describing the nature of the collision (rear end, roll over, etc.) and that snippit of sound ("oh sht!" or "big funky bus, passengers should've seen this coming"). This would be sent back in time at service center, and then broadcast to your CAW receiver, where you will be given advance warning of your crash.
Of course, depending on god's sense of humor, there may be nothing that you can do about it. You pull over, and a giant meteorite still smashes your car.
Moving matter would be one way to fulfill my request to send information back through time, but I didn't want to limit it to that. Anyway someone figures out how to do it, and I'll be happy!
I'd be content just to have information travel through time - not matter. It sounds like that would be a far simpler proposition. If you can send back individual particles to exact locations and/or times, you'll have information travelling ... it sounds like a quick way to make a "simple" experiment useful. Of course, it would have to be a significant amount of time to be worthy... at least a few hours.
One of us has got to dress up like Ronald Mallett-- all out, with a mask and everything, plus a scorched labcoat and frizzy hair-- and show up at his doorstep.
His probable reaction: They will discover a cure for my baldness!!
"since it can't predict the future"
Actually, it can predict the future! The sensing microphones are closer to the noise source than the cancelation speaker, so they receive the noise before the speaker has to generate it. At 340 m/s, 1 foot is about 0.1 milliseconds "into the future"; most speakers have a decent frequency response in the range of 1 kHz (1/0.1 msec) and could easily respond in time.
Like the poster above said, it looks like it uses 8 noise-sensing microphones, 8 speakers, and then 8 microphones in the quiet zone to measure the sucess of the speakers and automatically make adjustments.
In a 3d world, these sense microphones are essential because they are used to compensate when the environment changes. Stand up, and it needs to compensate. Move a chair, it does, too.
It looks like the system would try to minimize the noise at these 8 places, and not worry about everything else. I'm guessing that these 8 microphones would have to be in relatively open air (kindof like your head is when you walk around - not in corners), and not near any local noise sources (such as the fan on yuor computer). Or, they may be built into the device, on a framework attached to the speakers and the front microphones.
The TI 320c32 is a decent floating point processor - 20-30 MIPS / 40-60 GFLOPS. But, (IHALEWD - I have a little experience with dsps), I don't think it's enough horsepower to entirely model a room and do super-sophisticated stuff, but it looks like it would generally get good results.
Automatically adjusting filters are neat - they can be tuned relatively easily, and since a large number of coefficients, they can describe some pretty complicated behaviour without a lot of programming. Of course, AFAIK, they assume linearity; if the air conditioner vibrations rattle a coffee cup on your desk, that's non-linear and probably won't get fixed.
And, of course, this system may require some sort of calibration -- like a firecracker that you light near the noise source. (ok, maybe just something that makes test tones or a sharp snap). The microphones can get a picture of the room by listening to all the reflections (and by knowing what the initial sound was).
cool. Thanks for the info. Good luck archiving your MDs!
is what happens at the finges of this "shadow of silence". Does it start to break down such that the anti-noise and the noise become in phase again, and you get an area of double-the-noise?
This is a lot more complicated than headphones. Headphones are relatively one-dimensional (one microphone, one speaker, one eardrum per circuit) - the only thing you have to worry about is not generating feedback.
This seems to be a more complicated 3-dimensional solution, and it'll have much more complicated problems. Does this cancel noise effectively in corners? Will a computer monitor cast a non-noise-canceled shadow? Is there a limit to the noise source (can it be all around you, or must it be generated in one specific place?)
questions... looking forward to the answers!
Actually, I was right, for 2 reasons...
1. Those digital outputs were on players, not the discs themselves. (Saved by my own poor choice of words -- that's a talent I should further develop!!)
2. I should have been more specific. What these guys put out is digital uncompressed music - not the original compressed data that was on the disc. To make a copy, this music has to be recompressed again. With lossless compression, this would be no problem -- all the compression/decompressions cycles would yield perfect copies. But, minidisc uses lossy [i.e. f'(f(x)) != x] compression that isn't totally symetric [i.e. f'(f(f'(f(x)))) != x ], so a little noise is introduced into each generation, even if its transfered digitally. more info here. I'd love to hear that you can read/write the compressed digital data!
I didn't know that there were digital outs [even if they're not the ones I was thinking of]! At $460, the rack unit isn't even outrageously expensive.
They do, and it's called minidisc.
Ignore all the other users claiming it's SACD and DVD - they are too new to make any meaningful comparisons yet. Minidisc, however, has been promoted as a smaller-size compression-based consumer format. Copy protection is built-in; there are no digital outs on any minidiscs.
You forgot...
5. Remote Control ($30-$40)
"stereo sound or better"
...
Stereo isn't good enough? What do you want, a surround-sound 5.1 channel setup? It takes quite a geek to step into a shower and think "this sounds good, but could use an over-spigot center speaker, plus two more astride the shampoo rack, plus a subwoofer built into the cabinent space under sink."
If I were you, I'd concentrate on reducing the background noise before I consider a second speaker.
Now, if you can just find a source for quality quadraphonic or 5.1 MP3's
Not a troll, but why would you want to do 3d interlacing? AIUI (as I understand it), normal 2d interlacing was done to work around the limitations of phospher, TV scan rates, and bandwidth. The phospher needed to be scanned at 60 Hz (either there was no long-persistance phosper available so it would flicker, or long-persistance phospher was available and had a ghosting problem), but horizontal scan rates were limited to too few lines at 60 Hz.
With a new technology, and especially with all the cool stuff we can put into receivers today (full-frame buffers, image enhancement, MPEG decompression), it wouldn't make sense to be bound to a 1941 standard. (see also)
Someone, please explain this to me!
When I'm microwaving something late at night, the beeper is waay to loud. It's got a clock, it knows it's 2 in the morning, why doesn't it silence the beeper?
I thought the microwave that my parents got could solve the problem. It actually has a microphone. But, no, it doesn't use it to make sure the beep is an appropriate level - it uses it for a stupid voice recorder function. I mostly use it record sounds of small animals being microwaved (ribbit.... ribbit.... ribitribitribit!!) No! not actual animals!
Ok, so my parent's microwave has an appointment reminder. So, one night we were sleeping in the kitchen (don't ask) and needed to wake up early to catch a plane. We thought we could finally use this. We entered the time (keypad entry '3' '3' '0' 'AM'), and the day, and then it asked us what kind of appointment it was. Why'd it need to know? Who cares. We pushed the button for "doctor's appointment". It then responded "alarm set for 2:45am". Somehow, it determined that 45 minutes was the time needed to drive across town, find a parking space, and walk to rest of the way to the doctors office. I suspect that it had different times for different events, but we were too tired to try to figure it out. We canceled it and set it again for 4:15.
After my microwave is done, it'll beep every 2 minutes to remind me that my food is ready (it even says "food is done" - is that an undercooked chicken lawsuit waiting to happen?). But, as far as I can tell, it'll do this forever (my roommates haven't let me test this hypothesis yet). Even when my food is cold and dried out and totally unedible, it'll still beep. Maybe just beep a couple of times at 2 minute intervals, and then after it's cold, change to 20 minute intervals and leave the light on so I'll notice it when I'm good and ready.
Makes me long for the old-style microwaves with a "time" dial and a "power" dial, a start button, and a single bell when it's done.
One reason: Ford doesn't have to deal with the orders-of-magnitude in product variations that occur in computers, so counting vehicles means more than counting computers.
The limits of vehicles seem to be about 10:1 at most (2 passenger car : 50 passenger bus), 50:500 hp, maybe 1 ton car: 5 ton bus.
In computers you've got a much wider dynamic range... 1 user:1000 user, 32kHz (Sega VMU):2.2 GHz (P4), 1K:16GB memory, 56kbps modem : 2 Gbps fibre channel, etc.