1. Because your post implied that there was a preferred frame when you argued that "you don't age slower (sic), others age faster". This is a meaningless distinction.
2. Hmm? Humans have evolved for 1g environments-- it's not like you have to add in the Earth's 1g. The passengers won't know the difference between whether they're at home on the sofa or on their space ship accelerating towards relativistic speed. No difference in stress.
3. I said it wasn't exciting, not that it wasn't noticable. Exciting is 0.99c:)
2. You can reach relativistic speeds quite quickly even at low accelerations. 1g acceleration for a year will get you 1.0 c. Somehow, I think your body might be able to tolerate 1g.
3. If 'units' hasn't let me down, half a billion kph is 0.46c. Relativistic effects are pretty boring at that speed. The Lorentz factor is only about 1.155.
If you make yourself escape the "I gotta have the Series 37 because it comes with an espresso maker" mentality, lifetime subscription makes sense.
I have a series 1 tivo, purchased in 2000 with lifetime. I'm an early-adopter type, but that doesn't mean I have to buy each successive iteration! Lifetime was cheaper back then, but still, do you really expect your PVR to become obsolete in 3 years (provided you make your purchase with a >3-year lifespan in mind)? Technology moves fast, but not THAT fast.
I do a lot of EDA, mostly under Mentor Graphics PADS. It's quirky, it's buggy, but I find I'm used to it and I actually almost like it.
But for whatever reason, our department seems to be unable to purchase a seat (license) for it that doesn't expire. Which means that my designs become uneditable unless our department mails Mentor a check every 12 months. It makes me sick.
I really want there to be a good FOSS solution, and like the initial poster, I've found the existing products to be really inadequate. Some that I've tried (and I've forgotten names now) do not even properly maintain connectivity of parts as you move parts around in schematic capture. Others rapidly turn your schematic into a terrible mess because they don't redraw your lines right. These are not terribly hard things to implement, but it takes time, and evidently, the FOSS solutions haven't had that much time put in to them yet.
To those who believe that autorouting is "a high end feature for high end users", I disagree. If you're making a simple board that will never need to be revised, then sure, route it by hand. But if you're doing lots of revisions, an autorouter is a critical tool. When I design boards, I spend extra time making sure that the autorouter can always route it. That way, when I need to pop down a couple extra (whatevers), I just hit F9 and it reroutes again.
I think one of the most awful aspects is that gEDA keeps promising and promising. They have very nice screenshots and feature lists. But they seem to be vapor. I think they're inhibiting others from starting a good FOSS EDA package. Maybe.
If a handful of sensors and limited expandability is okay, then by all means, go for mindstorms! You get a plastic-molded case that keeps kid's fingers off it, and a more or less foolproof system.
At some point, you'll might be interested in doing something more serious or ambitious, which is where the OrcBoard comes in. If you're familiar with HandyBoards, the OrcBoard is designed as a modern replacement for them. Mindstorms is for a different audience.
The most common type of robot that people use OrcBoards for (and there are a number of researchers at MIT who use them in their robots), use a laptop for lots of CPU power, a camera and/or Sick laser scanner, and an orcboard (which handles the motors, quadrature phase decoders, control system, and low-level sensors like rangefinders.)
For a system that costs about the same as mindstorms, you get a lot more capability. But like I said, that comes at the cost: you CAN fry an OrcBoard, you're on your own for a case, and you might need to get your hands dirty by adding features or fixing bugs in the firmware or userland tools. Some of us like that, though!
Or you might be interested in the OrcBoard robotic controller, which is open source (schematics, layout, firmware, userland tools all GPL). It's being used by a number of robotics classes (6.188, 2.12), and a robotics competition (MASLab) at MIT.
It's a bit different than mindstorms in that it's designed to be used as a slave to a laptop or other more CPU-rich device. But you can use it in stand-alone mode too, if your robots are simple.
-Ed (disclaimer: creator of orcboard)
programming is like architecture
on
Is Programming Art?
·
· Score: 2, Insightful
I believe that computer programming is like brick-and-mortar architecture.
The vast majority of buildings are just buildings. But every once in a while, a building is a work of art.
One of the things I like about architecture (and computer programming) is that the buildings always serve a purpose. They don't arise out of the ether to express a purely abstract thought, but arise from the need to create something useful.
But don't delude yourself by thinking that you're an artist just because you're a computer programmer. The vast majority of buildings are cinder-block, minimum-cost affairs, and the same is true for code.
You probably recorded it onto mini-dv tape in the first place, so it's easy. After I do my edits and produce a final output, I lock up my tapes as my backup of last resort. (I could always reconstruct the final output using the tapes.)
I only keep final output on my primary storage (hard disk), but it's generally much smaller anyway.
Why are so many of the posters hostile to the suggestion that there could be a correlation between brain size and intelligence? A hostile response is not an objective response.
I don't know if they're correlated or not, but it seems plausible that there could be one. It may not be large. It may be a minor factor in comparison to the dozens of other factors that affect intelligence. Or, it may be entirely uncorrelated.
Just because a bunch of sexist supremecists long ago asserted that it was correlated in order to promote their own viewpoint doesn't mean they were completely wrong. There are a lot of assholes. Some of them are going to turn out to be right (if only in part) after all.
The reason that the latest Star Trek franchises have been unappealing to me is that the episodes have become too much like the other crappy social-dramas on TV, e.g., 90210, Dawson's Creek, OC.
Star Trek was good because it was different. Saying it was "intellectual" might go too far. But it scratched a different itch.
Even if US TV watchers *like* shows like 90210, I'm tempted to think that they must like some variety as well. Even those people who fully enjoy reality TV and American Idol must eventually want to watch something different.
By making Star Trek the same as all the other shows, they eliminate the appeal that would have brought a "cross-over" audience while simultaneously alienating the fans who liked it for what it was in the first place!
If Star Trek comes back, ditch the gratuitous action scenes. Ditch the scantily clad women (7, T'pol). Ditch the cliche of ugly and screeching bad guys who spit venom and have acid for blood (ok, that was Aliens, but the Borg queen was close). Ditch the sexed-up alternative universes. If I want these things, I'll watch Die Hard again or buy the Girls Gone Wild movies. Be different! Or just be sci-fi:)
If you're interested in a completely open-source robotics platform (where everything is open, including schematics, the firmware, the user libraries, *everything*), visit www.orcboard.org. There's no commercial manufacturer, but there's a community group-ordering effort.
The OrcBoard is used in the MASLab robotics competititon at MIT, and in MIT's cornerstone robotics class. Most folks use the OrcBoard with a linux laptop or embedded PC.
The OrcBoard is just the controller for a robot, not a robot itself, and can control a wide range of robot types and sizes, with lots of different sensor configurations. It's sort of a swiss-army knife for robot construction.
The two basic energy-consuming activities are pushing charge around to charge and discharge capacitance, and leakage current. The latter is clearly waste; it performs no computation. The former can be conceptually be reduced to zero by reducing the parasitic capacitance of wires and gates.
If we consider more exotic ("abstract") computing systems, then it really becomes clear that we've got a long way to go before we start bumping up against any theoretical limits.
The electoral college does NOT preserve the rights of the minority as you claim.
What it does is provide a major distorting factor that amplifies some folks' votes, while utterly negating others... without any particular preference for how they were voting.
For example, a handful of Florida voters became absolutely critical. You can fairly say their vote "counted". But my vote, being a non-democrat from Massachusetts, didn't "count" worth squat!
IMO, svn's use of berkeley DB as its backend, an opaque, non-human-readable, non-human-recoverable, non-machine-portable* database, is its biggest shortcoming...
I still use svn, though. I'm just glad to be able to rename directories.
I'd pee myself if someone forked svn and gave it a more friendly backend.
-Ed
*By this, I mean that you can't take the berkeley DB, copy it to another machine, and expect it to work... the internal byte order is machine specific.
A simple argument supporting the pro-Capitalism nature of Open Source:
No artificial measures have been used to "prop up" Open Source. Yet it exists in a Capitalist society. Free markets do not reach equilibriums instantaneously, so it is possible that the existence of Open Source is merely a bizarre transient. But every passing day is an indication that it is not.
On the other hand, artificial measures DO exist to prop up closed-source software. This directly hurts Open Source, yet Open Source is alive in spite of it. That's a pretty strong indication.
I can't give you a balance sheet showing how Open Source is "in the black", but if you believe in natural selection in the context of a free market, there's not really another explanation for the existence of Open Source.
Some folks think it's a typo, that it's supposed to be 65 miles, not 65K miles. No, 65K miles is more like it. You really want your elevator's center of mass to be in geosynchronous orbit... Space elevators to LEO tend to, uh, get wound around the earth right fast.
And if the ribbon breaks, things generally aren't so bad. The portion of the elevator (including the counter weight) that's further from the earth will tend to move away from the earth. (If you spin in a circle with a rock in your hand, then let go of the rock, the rock goes away from you, not crashing in towards your head.) The nearer part will tend to fall, but it will tend to fall slowly and is relatively unlikely to cause damage. (At least, according to High lift systems, who came and gave a talk last year.) The elevator, since it's so huge, tends to not be terribly heavy. The system proposed by high lift systems
I believe Brad Edwards was involved in High Lift Systems, so I imagine the basic idea is the same.
If geo is ~20K miles, why does the elevator need to be so long? Does this mean that they're now thinking about a lighter counter weight? They used to talk about capturing an asteroid.
Re:not gonna happen, the lobbies are too powerful
on
Do-It-Yourself VOIP Telco
·
· Score: 2, Insightful
Yes, that's fine, if you're willing to have enough equipment for 5 times more capacity than demand.
~20megabits/second would fill a 1TB disk in under a week (4.6 days).
If computers become media centers, then 1 TB media center would be fairly stifling (compared to my 300 hour Tivo).
Gigabit ethernet is similarly explained. If you want a couple video feeds to coexist, 100mbit won't cut it.
Multiple cpus is a no brainer. CPUs are running out of steam; the road to better performance is multiple cpus. It's inevitable, and 5GHz is really a very modest increase in clockspeed over today.
Save your post and reread it in 4 years and feel a bit embarassed!
Why do people keep inventing new organizations that they want to give money to? Why should I pay some third party so that I can send email from myself to someone else?
No, no. The only thing that makes sense--if you want to consider a pay-for-email scheme-- is to pay the recipient. THEY are the one whose resources are being consumed. They are the one who can determine what price is a suitable deterrent for the spam that they receive. Nobody else can do it-- it's as simple as that.
It's true that the infrastructure to implement this system would probably require a third party financial clearing house, and they'd probably have to get a cut, but fundamentally, the payment must be receiver driven.
So imagine that to send an email, you contact an escrow service which gives you a token for your email: a promise of payment equal to the amount required by the recipient for delivery. The mail gets sent, the receiver can choose to collect payment or not to (friends don't pay friends to read their emails with this system). Obviously there are some technical challenges, but there are a lot of bright people out there.
Just, for god's sake, don't make me read another article where someone invents a company that they want to give money to when the recipient is the party being injured by spam!
I run a robotics competition at MIT, and we currently use 3.5" (roughly 4"x6"). form factor PCs. (http://maslab.lcs.mit.edu). The robots do fully autonomous exploration using vision, so CPU power is a real issue.
Our current boards are Geode-GX1 @ 300MHz. Read "Slow". But, they only consume about 0.7A @ 12V. Read "No 40 lb battery".
I've previously evaluated some Eden boards, including C-800 and 667 boards. BUT! Their power is about 2.3A @ 12V. Combined with our motors, that brings our run time under 40 minutes. Ouch.
Power consumption: what is the power consumption of these boards? It really needs to be 15 W or so to be practical for our application. But I bet it's more like 25 W. Can it at least be extremely underclocked?
No serial port? Are they out of their minds? If they're targetting embedded systems (at all), omitting a serial port is completely insane. So many peripherals are controlled via serial: GPS, actuators, sensors, industrial equipment, probably cash registers... Hopefully there's at least header for it, if not a proper DB-9. Anyone know?
It's not absolute voltage that's the problem, it's the voltage change required.
Each gate output has a capacitance. Every time the output of a gate changes, that capacitance has to be charged up or down, and this takes energy. If the change in voltage is big, it takes more energy. (.5*C*V^2, even).
Transitioning the output of a trinary circuit element from -3.3V to 3.3 (a change of 6.6V) requires a lot of energy.
Of course, this is obvious at a glance. a -3.3V to +3.3V is exactly the same thing as a 0 to 6.6V system because voltage is always relative.
Slashdot Mirror
1. Because your post implied that there was a preferred frame when you argued that "you don't age slower (sic), others age faster". This is a meaningless distinction.
:)
2. Hmm? Humans have evolved for 1g environments-- it's not like you have to add in the Earth's 1g. The passengers won't know the difference between whether they're at home on the sofa or on their space ship accelerating towards relativistic speed. No difference in stress.
3. I said it wasn't exciting, not that it wasn't noticable. Exciting is 0.99c
Wow. IANAP, but somehow I picked this up:
1. There are no preferred frames.
2. You can reach relativistic speeds quite quickly even at low accelerations. 1g acceleration for a year will get you 1.0 c. Somehow, I think your body might be able to tolerate 1g.
3. If 'units' hasn't let me down, half a billion kph is 0.46c. Relativistic effects are pretty boring at that speed. The Lorentz factor is only about 1.155.
If you make yourself escape the "I gotta have the Series 37 because it comes with an espresso maker" mentality, lifetime subscription makes sense.
I have a series 1 tivo, purchased in 2000 with lifetime. I'm an early-adopter type, but that doesn't mean I have to buy each successive iteration! Lifetime was cheaper back then, but still, do you really expect your PVR to become obsolete in 3 years (provided you make your purchase with a >3-year lifespan in mind)? Technology moves fast, but not THAT fast.
Plus I hate recurring charges.
doesn't it nearly incite you to violence that you can call any widget methods from arbitrary threads? It drives me crazy. (Or has this been fixed?)
I do a lot of EDA, mostly under Mentor Graphics PADS. It's quirky, it's buggy, but I find I'm used to it and I actually almost like it.
But for whatever reason, our department seems to be unable to purchase a seat (license) for it that doesn't expire. Which means that my designs become uneditable unless our department mails Mentor a check every 12 months. It makes me sick.
I really want there to be a good FOSS solution, and like the initial poster, I've found the existing products to be really inadequate. Some that I've tried (and I've forgotten names now) do not even properly maintain connectivity of parts as you move parts around in schematic capture. Others rapidly turn your schematic into a terrible mess because they don't redraw your lines right. These are not terribly hard things to implement, but it takes time, and evidently, the FOSS solutions haven't had that much time put in to them yet.
To those who believe that autorouting is "a high end feature for high end users", I disagree. If you're making a simple board that will never need to be revised, then sure, route it by hand. But if you're doing lots of revisions, an autorouter is a critical tool. When I design boards, I spend extra time making sure that the autorouter can always route it. That way, when I need to pop down a couple extra (whatevers), I just hit F9 and it reroutes again.
I think one of the most awful aspects is that gEDA keeps promising and promising. They have very nice screenshots and feature lists. But they seem to be vapor. I think they're inhibiting others from starting a good FOSS EDA package. Maybe.
If a handful of sensors and limited expandability is okay, then by all means, go for mindstorms! You get a plastic-molded case that keeps kid's fingers off it, and a more or less foolproof system.
At some point, you'll might be interested in doing something more serious or ambitious, which is where the OrcBoard comes in. If you're familiar with HandyBoards, the OrcBoard is designed as a modern replacement for them. Mindstorms is for a different audience.
The most common type of robot that people use OrcBoards for (and there are a number of researchers at MIT who use them in their robots), use a laptop for lots of CPU power, a camera and/or Sick laser scanner, and an orcboard (which handles the motors, quadrature phase decoders, control system, and low-level sensors like rangefinders.)
For a system that costs about the same as mindstorms, you get a lot more capability. But like I said, that comes at the cost: you CAN fry an OrcBoard, you're on your own for a case, and you might need to get your hands dirty by adding features or fixing bugs in the firmware or userland tools. Some of us like that, though!
Or you might be interested in the OrcBoard robotic controller, which is open source (schematics, layout, firmware, userland tools all GPL). It's being used by a number of robotics classes (6.188, 2.12), and a robotics competition (MASLab) at MIT.
It's a bit different than mindstorms in that it's designed to be used as a slave to a laptop or other more CPU-rich device. But you can use it in stand-alone mode too, if your robots are simple.
-Ed
(disclaimer: creator of orcboard)
I believe that computer programming is like brick-and-mortar architecture.
The vast majority of buildings are just buildings. But every once in a while, a building is a work of art.
One of the things I like about architecture (and computer programming) is that the buildings always serve a purpose. They don't arise out of the ether to express a purely abstract thought, but arise from the need to create something useful.
But don't delude yourself by thinking that you're an artist just because you're a computer programmer. The vast majority of buildings are cinder-block, minimum-cost affairs, and the same is true for code.
Um, how about on mini-dv tape?
You probably recorded it onto mini-dv tape in the first place, so it's easy. After I do my edits and produce a final output, I lock up my tapes as my backup of last resort. (I could always reconstruct the final output using the tapes.)
I only keep final output on my primary storage (hard disk), but it's generally much smaller anyway.
Why are so many of the posters hostile to the suggestion that there could be a correlation between brain size and intelligence? A hostile response is not an objective response.
I don't know if they're correlated or not, but it seems plausible that there could be one. It may not be large. It may be a minor factor in comparison to the dozens of other factors that affect intelligence. Or, it may be entirely uncorrelated.
Just because a bunch of sexist supremecists long ago asserted that it was correlated in order to promote their own viewpoint doesn't mean they were completely wrong. There are a lot of assholes. Some of them are going to turn out to be right (if only in part) after all.
The reason that the latest Star Trek franchises have been unappealing to me is that the episodes have become too much like the other crappy social-dramas on TV, e.g., 90210, Dawson's Creek, OC.
:)
Star Trek was good because it was different. Saying it was "intellectual" might go too far. But it scratched a different itch.
Even if US TV watchers *like* shows like 90210, I'm tempted to think that they must like some variety as well. Even those people who fully enjoy reality TV and American Idol must eventually want to watch something different.
By making Star Trek the same as all the other shows, they eliminate the appeal that would have brought a "cross-over" audience while simultaneously alienating the fans who liked it for what it was in the first place!
If Star Trek comes back, ditch the gratuitous action scenes. Ditch the scantily clad women (7, T'pol). Ditch the cliche of ugly and screeching bad guys who spit venom and have acid for blood (ok, that was Aliens, but the Borg queen was close). Ditch the sexed-up alternative universes. If I want these things, I'll watch Die Hard again or buy the Girls Gone Wild movies. Be different! Or just be sci-fi
If you're interested in a completely open-source robotics platform (where everything is open, including schematics, the firmware, the user libraries, *everything*), visit www.orcboard.org. There's no commercial manufacturer, but there's a community group-ordering effort.
The OrcBoard is used in the MASLab robotics competititon at MIT, and in MIT's cornerstone robotics class. Most folks use the OrcBoard with a linux laptop or embedded PC.
The OrcBoard is just the controller for a robot, not a robot itself, and can control a wide range of robot types and sizes, with lots of different sensor configurations. It's sort of a swiss-army knife for robot construction.
It's virtually all waste heat.
% 2C 1%2C0.25%2CDownload/http://citeseer.ist.psu.edu/ca che/papers/cs/3548/http:zSzzSzwww.neci.nj.nec.comz SzhomepageszSzwdszSzfundphys.pdf/smith95fundamenta l.pdf
= /n ature/journal/v406/n6799/full/4061047a0_fs.html
The two basic energy-consuming activities are pushing charge around to charge and discharge capacitance, and leakage current. The latter is clearly waste; it performs no computation. The former can be conceptually be reduced to zero by reducing the parasitic capacitance of wires and gates.
If we consider more exotic ("abstract") computing systems, then it really becomes clear that we've got a long way to go before we start bumping up against any theoretical limits.
These papers might be helpful.
http://citeseer.ist.psu.edu/rd/27839242%2C24929
http://www.nature.com/cgi-taf/DynaPage.taf?file
The electoral college does NOT preserve the rights of the minority as you claim.
What it does is provide a major distorting factor that amplifies some folks' votes, while utterly negating others... without any particular preference for how they were voting.
For example, a handful of Florida voters became absolutely critical. You can fairly say their vote "counted". But my vote, being a non-democrat from Massachusetts, didn't "count" worth squat!
IMO, svn's use of berkeley DB as its backend, an opaque, non-human-readable, non-human-recoverable, non-machine-portable* database, is its biggest shortcoming...
I still use svn, though. I'm just glad to be able to rename directories.
I'd pee myself if someone forked svn and gave it a more friendly backend.
-Ed
*By this, I mean that you can't take the berkeley DB, copy it to another machine, and expect it to work... the internal byte order is machine specific.
A simple argument supporting the pro-Capitalism nature of Open Source:
No artificial measures have been used to "prop up" Open Source. Yet it exists in a Capitalist society. Free markets do not reach equilibriums instantaneously, so it is possible that the existence of Open Source is merely a bizarre transient. But every passing day is an indication that it is not.
On the other hand, artificial measures DO exist to prop up closed-source software. This directly hurts Open Source, yet Open Source is alive in spite of it. That's a pretty strong indication.
I can't give you a balance sheet showing how Open Source is "in the black", but if you believe in natural selection in the context of a free market, there's not really another explanation for the existence of Open Source.
Some folks think it's a typo, that it's supposed to be 65 miles, not 65K miles. No, 65K miles is more like it. You really want your elevator's center of mass to be in geosynchronous orbit... Space elevators to LEO tend to, uh, get wound around the earth right fast.
And if the ribbon breaks, things generally aren't so bad. The portion of the elevator (including the counter weight) that's further from the earth will tend to move away from the earth. (If you spin in a circle with a rock in your hand, then let go of the rock, the rock goes away from you, not crashing in towards your head.) The nearer part will tend to fall, but it will tend to fall slowly and is relatively unlikely to cause damage. (At least, according to High lift systems, who came and gave a talk last year.) The elevator, since it's so huge, tends to not be terribly heavy. The system proposed by high lift systems
I believe Brad Edwards was involved in High Lift Systems, so I imagine the basic idea is the same.
If geo is ~20K miles, why does the elevator need to be so long? Does this mean that they're now thinking about a lighter counter weight? They used to talk about capturing an asteroid.
Yes, that's fine, if you're willing to have enough equipment for 5 times more capacity than demand.
Consider an HDTV feed:
~20megabits/second would fill a 1TB disk in under a week (4.6 days).
If computers become media centers, then 1 TB media center would be fairly stifling (compared to my 300 hour Tivo).
Gigabit ethernet is similarly explained. If you want a couple video feeds to coexist, 100mbit won't cut it.
Multiple cpus is a no brainer. CPUs are running out of steam; the road to better performance is multiple cpus. It's inevitable, and 5GHz is really a very modest increase in clockspeed over today.
Save your post and reread it in 4 years and feel a bit embarassed!
Wouldn't spelling as a-t it out be shorter/faster?
at = dit dah, dah
@ = dit dah dah dit dah dit
Or does an inter-symbol pause really take as along as 3 symbols?
Why do people keep inventing new organizations that they want to give money to? Why should I pay some third party so that I can send email from myself to someone else?
No, no. The only thing that makes sense--if you want to consider a pay-for-email scheme-- is to pay the recipient. THEY are the one whose resources are being consumed. They are the one who can determine what price is a suitable deterrent for the spam that they receive. Nobody else can do it-- it's as simple as that.
It's true that the infrastructure to implement this system would probably require a third party financial clearing house, and they'd probably have to get a cut, but fundamentally, the payment must be receiver driven.
So imagine that to send an email, you contact an escrow service which gives you a token for your email: a promise of payment equal to the amount required by the recipient for delivery. The mail gets sent, the receiver can choose to collect payment or not to (friends don't pay friends to read their emails with this system). Obviously there are some technical challenges, but there are a lot of bright people out there.
Just, for god's sake, don't make me read another article where someone invents a company that they want to give money to when the recipient is the party being injured by spam!
Diamandis not only predicts it, he diamands it.
ugh. why did i post this?
detecting and decoding are two very different things.
I run a robotics competition at MIT, and we currently use 3.5" (roughly 4"x6"). form factor PCs. (http://maslab.lcs.mit.edu). The robots do fully autonomous exploration using vision, so CPU power is a real issue.
Our current boards are Geode-GX1 @ 300MHz. Read "Slow". But, they only consume about 0.7A @ 12V. Read "No 40 lb battery".
I've previously evaluated some Eden boards, including C-800 and 667 boards. BUT! Their power is about 2.3A @ 12V. Combined with our motors, that brings our run time under 40 minutes. Ouch.
Power consumption: what is the power consumption of these boards? It really needs to be 15 W or so to be practical for our application. But I bet it's more like 25 W. Can it at least be extremely underclocked?
No serial port? Are they out of their minds? If they're targetting embedded systems (at all), omitting a serial port is completely insane. So many peripherals are controlled via serial: GPS, actuators, sensors, industrial equipment, probably cash registers... Hopefully there's at least header for it, if not a proper DB-9. Anyone know?
It's not absolute voltage that's the problem, it's the voltage change required.
Each gate output has a capacitance. Every time the output of a gate changes, that capacitance has to be charged up or down, and this takes energy. If the change in voltage is big, it takes more energy. (.5*C*V^2, even).
Transitioning the output of a trinary circuit element from -3.3V to 3.3 (a change of 6.6V) requires a lot of energy.
Of course, this is obvious at a glance. a -3.3V to +3.3V is exactly the same thing as a 0 to 6.6V system because voltage is always relative.