There are not many new Xbox 1's and the x86 inside them does not compare well in terms or raw horsepower even to current ARM processors that draw a tiny fraction of the power they use to heat their surroundings.
They can't build a decent website. I sincerely doubt they will be able to build a decent console.
Pity, because the idea is good - a small inexpensive machine with reasonable graphics performance that could be used for just about anything. Not necessarily a console, a computer or a set-top-box but, depending on the user, a combination of the three.
I can imagine that during a crisis of some sort you may need to function with less than ideal number or representatives, but I see no such crisis at hand and said representatives should not be allowed to declare it by themselves. Also, there should be a penalty for invoking such a condition - like not being able to run for any election office or to work in any government position for a given period.
IIRC, in one of the Rama books, there is a species that considers participating in a war is a crime punished by death. It doesn't prevent one from happening, but it sure makes one unwilling to declare war unless for a very, very good reason.
A whole lot of wars would be prevented by laws much gentler than that.
There is one good reason to send humans to live outside Earth, but it is still very, very expensive. At these prices, we won't even consider it seriously until we are confronted with a very real threat of extinction.
You should not forget the Soyuz vehicles were improved since the early 60s while the shuttle had, IIRC, 20 years less time for design changes. The latest Soyuz is many generations beyond the two first-generation units that actually killed astronauts and while some mishaps do happen, nobody got killed in their latest design.
Also, the design is much simpler. When reliability is at stake, I go for the simpler design. Shuttles are impressive machines, but there is a job being done and a 737-sized fragile beast is not the best tool for sending people to orbit.
I won't count that. The Apollo 1 vehicle had a more than one atm pure Oxygen atmosphere that made everything inside it flammable and a hatch that was very hard to open (I think that was because the combustion lowered internal pressure below 1 atm exerting a force on the hatch that astronauts were not able to counter). After the accident and the design changes, the normal air atmosphere and explosives around the hatch made it a lot less unsafe to fly (nobody in his right mind would claim space travel is safe).
There is no reason not to attach an Apollo-like capsule to a interplanetary spacecraft that has a nuclear-thermal (or anything "N") engine strapped on it. Still, you are right - the eco-folks will make it nearly impossible to have a spacecraft with sufficient power to travel between planets unless we discover some fancy new physics.
This is the crew return vehicle. Nothing really spectacular about that.
"I feel sorry for the crew who has to spend all that time in that shit box."
They won't. And you can really consider that capsule is more or less the escape pod from the real spaceship. Other way to think about it is the "shipping container for the crew and return samples".
I suppose most of the time the crew will have more spacious quarters, specially when en route to Mars. The capsule will also never get to the Martian surface - they will probably have a descent vehicle either with them since Earth or safely parked in Martian orbit as well as an ascent vehicle landed near their working site on Mars that's there since before they leave Earth.
There is really not enough data to attest Apollo spacecraft were much safer than the shuttles. There were less than two dozen Apollo manned launches with one nearly (because the crew got really, really lucky) catastrophic accident and more than a hundred shuttle launches done by a small fleet that went to space a couple times each with two very serious mishaps.
The best one can do is to extrapolate on data from about a hundred Soyuz missions. Soyuz seems to be slightly safer than shuttle and has in common with the Orion both the 60's tech and the mostly expendable architecture (IIRC, some systems are transferred from a used Soyuz to a new one after being recertified).
There are ways it can be built that can allow for a reasonable amount of trust.
I worked in the Brazilian e-voting system and I can easily enumerate a couple ideas I had in 2002, during the project.
- There was a printer option that would be attached to the voting machine that would print the vote in human readable form. The paper votes would be deposited in a sealed ballot and random machines would be selected for examination. Since a tampered with software could detect the printer and behave well when it was attached and change votes when not, we could either to put printers in all of them or build "dummy printers" that print nothing but look like printers for the rest of the voting machine.
- A vote file cryptographically signed with the help of a TPM, tying the vote file to a given voting machine and making it really, really hard to inject a forged file in the vote counting process.
- Have the voting crews authorize voters with their own PINs.
And the mechanisms that have been in place since I was a kid:
- Making every citizen vote. This makes it rather easy to point out anyone trying to vote for people they know won't be there: If you go to vote and the machine says you already voted, call the police.
- Assign voters to specific machines - this way they can know how many votes each machine will receive. There are special situations like candidates allowed to vote anywhere, but you can deal with them.
- Having every voting machine assigned to three randomly-selected adults to act as a crew to validate each voter that has to vote there.
- Having police arrest any people campaigning within half a mile from a voting place.
- Mining vote patterns for discrepancies that trigger investigations. One that was detected was a series of votes with less than 30 second intervals between them that happened near closing time.
And a whole lot more measures that help.
It's not that hard. It may not be perfect, but, it is sure good enough.
"This isn't just Diebold. This is dozens of state, local, and federal agencies that abjectly failed in their duties to their constituents to protect the voting system"
Having a lot of people to hang should not deter you from starting. It's a huge undertaking, but it's also a very important one.
You know how it works: the earlier you start, the sooner you finish the job.
The people responsible for this, from the lowest negligent intern up to the board, as well as every government official who vouched for the system, should be in jail.
How do you refund the people for corrupt government elected through electronic vote manipulation?
"First off, learn to use the 'quote' tag, fucktard. It was really challenging trying to sort your ramblings out when they were intertwined with my erudite arguments."
I am sorry to provide exercise for your brain, but I don't really like to write in HTML. I really don't care about you that much as to create a minor nuisance for me.
As for Apple, what exactly could they deliver of OLPCs? An OS, a WebKit-based browser and an e-mail program? Oh... I had no idea they were so difficult to do or that they were not readily available on just about every platform. Applications? Which ones? iWork? Keynote? It's not a general purpose laptop - it's a learning tool for kids.
As for the portability of Linux, it covers far more platforms than OSX. Having distros maintained for as many platforms ensure you have a full operating environment on each and every one of them. Porting for OSX is more painful because of the lack (on default installs) of X. No X means no GUI, or, at least, no Unix-style GUI. It doesn't matter how much certified UNIX OSX is, it won't run any Unix software that expects to see X. And, BTW, being a certified UNIX is... well... not worth much these days unless you want to sell for the US government. For most of the world, it's utterly irrelevant.
The article you mentioned is about cross-compiling with different versions of glibc. This is usually hard and that's why good distros provide build farms. Want your package built for, say, Itanium? Drop it in the proper place and there it goes. Isn't it wonderful what you can achieve when you just know the resources you have? OTOH, if you think you are clever enough to do it on your own, you should at least RTFM and learn that different platforms may have different library versions, including the more important ones.
"And yet you provide not ONE single example... Hmmm..."
As for a good example of Linux on ARM since ever, you can check the availability on distro sites. ARM ports have been available since, at least, Debian 2.2. Since a major Debian release takes a lot of time, I assume 2.2 happened ages ago. AFAIK, this was before OSX was even running on x86s (of course, it ran on x86s as OPENSTEP, but that's another story you are probably too young to remember). Before that, it also ran on Motorola's 68K (just as Linux, BTW) family.
"But when it comes to LEGAL LIABILITY (which is what matters to organizations such as the OLPC project), WHO CAN BE MADE RESPONSIBLE?"
You gotta be kidding, right?
If you hire someone to develop for you, you write a contract. It's that simple. The FLOSS advantage is that you and your contractor can piggyback on the work of others (like Apple does with WebKit, GNU compilers and a whole lot of OSX stuff) while still being able to hire people to do specific jobs with that very rich codebase.
Even if Apple offered OSX at no cost for OLPC, it would still belong to Apple. That means OLPC would not be able to change OSX in ways Apple didn't want to. If, in any given time, Apple decided it would no longer offer it, OLPC would die instantly.
And that is something you really don't want to happen: have another company decide your fate. If they wanted that, they could easily have gone for Microsoft from day one.
So, again, why ensure failure over time if you can at least have a shot going the other direction?
Sorry to tell you, but you must have heard this before: you are little more than an annoying Mac fanboy.
Let's do it this way: you search, let's say, the Debian repositories for ARM architecture, and then you let us know if you can't find a given program you need that much.
Usually, package maintainers do a very nice job of porting programs. I would not be surprised if you could run Tuxracer on an IBM zSeries mainframe.
And you don't need to be dismissive on the users of OLPCs. People have the ability to learn, given the incentives.
"Here's my point: OLPC can actually CONTRACT with Apple. How do they MAKE "Linux" (whatever THAT is!) perform ON SCHEDULE, TO SPEC, *OR ELSE*?"
That's hard to do when you accept Apple's offer is to do it for free. How can you exactly force Apple to do something for free? If you have a ton of money, you can hire them, but how do you justify them doing that instead of fostering open-source development that you can own in the end? Would OLPC own the ARM port of OSX? Of course not.
"Second, IT'S NOT JUST THE OS, STUPID!" yada yada yada "not only does "Linux" (whatever THAT is!) have to be ported; but also a gazillion other OSS applications."
Linux already runs on ARM. As do a lot of Linux applications. On ARM, on MIPS, on POWER/PowerPC, SPARC and others. Most of the time, all you need is to recompile the program. This is, you know, the beauty of real Unix-like OSs.
"STARTING with Apple (and its ARM experience, solid, platform-agnostic OS, and application portfolio) sure makes a LOT of sense to the OLPC people."
I bet you can find a dozen other companies more experienced in ARM processors than Apple. They just started it the other year. Linux is running on ARM architectures (from the very simple embedded ones to netbook-class OMAP ones) since like ever.
"And besides, none of this PRECLUDES the" F/OSS community" (whatever THAT is!) from offering one, or even a hundred, "Linux" (whatever THAT is!) distros, does it?"
No. But why would you hire Apple when you could just pick one and cultivate it to the stage you want? You can hire Apple just as much you can hire Canonical or Red Hat. If you can pay, you can have some very talented people.
"So, what was your point, again, exactly?"
Sorry, but your argument crumbled. Can you provide a good one the next time?
Slashdot Mirror
But can Debian folks bundle ZFS according to Debian's own rules?
There are not many new Xbox 1's and the x86 inside them does not compare well in terms or raw horsepower even to current ARM processors that draw a tiny fraction of the power they use to heat their surroundings.
They can't build a decent website. I sincerely doubt they will be able to build a decent console.
Pity, because the idea is good - a small inexpensive machine with reasonable graphics performance that could be used for just about anything. Not necessarily a console, a computer or a set-top-box but, depending on the user, a combination of the three.
I would love such a device.
I can imagine that during a crisis of some sort you may need to function with less than ideal number or representatives, but I see no such crisis at hand and said representatives should not be allowed to declare it by themselves. Also, there should be a penalty for invoking such a condition - like not being able to run for any election office or to work in any government position for a given period.
IIRC, in one of the Rama books, there is a species that considers participating in a war is a crime punished by death. It doesn't prevent one from happening, but it sure makes one unwilling to declare war unless for a very, very good reason.
A whole lot of wars would be prevented by laws much gentler than that.
There is one good reason to send humans to live outside Earth, but it is still very, very expensive. At these prices, we won't even consider it seriously until we are confronted with a very real threat of extinction.
You should not forget the Soyuz vehicles were improved since the early 60s while the shuttle had, IIRC, 20 years less time for design changes. The latest Soyuz is many generations beyond the two first-generation units that actually killed astronauts and while some mishaps do happen, nobody got killed in their latest design.
Also, the design is much simpler. When reliability is at stake, I go for the simpler design. Shuttles are impressive machines, but there is a job being done and a 737-sized fragile beast is not the best tool for sending people to orbit.
I won't count that. The Apollo 1 vehicle had a more than one atm pure Oxygen atmosphere that made everything inside it flammable and a hatch that was very hard to open (I think that was because the combustion lowered internal pressure below 1 atm exerting a force on the hatch that astronauts were not able to counter). After the accident and the design changes, the normal air atmosphere and explosives around the hatch made it a lot less unsafe to fly (nobody in his right mind would claim space travel is safe).
There is no reason not to attach an Apollo-like capsule to a interplanetary spacecraft that has a nuclear-thermal (or anything "N") engine strapped on it. Still, you are right - the eco-folks will make it nearly impossible to have a spacecraft with sufficient power to travel between planets unless we discover some fancy new physics.
This is the crew return vehicle. Nothing really spectacular about that.
Not until it becomes much cheaper.
For the current prices, it's probably cheaper to extract metals from thin air than it is to mine an asteroid for it.
"I feel sorry for the crew who has to spend all that time in that shit box."
They won't. And you can really consider that capsule is more or less the escape pod from the real spaceship. Other way to think about it is the "shipping container for the crew and return samples".
I suppose most of the time the crew will have more spacious quarters, specially when en route to Mars. The capsule will also never get to the Martian surface - they will probably have a descent vehicle either with them since Earth or safely parked in Martian orbit as well as an ascent vehicle landed near their working site on Mars that's there since before they leave Earth.
There is really not enough data to attest Apollo spacecraft were much safer than the shuttles. There were less than two dozen Apollo manned launches with one nearly (because the crew got really, really lucky) catastrophic accident and more than a hundred shuttle launches done by a small fleet that went to space a couple times each with two very serious mishaps.
The best one can do is to extrapolate on data from about a hundred Soyuz missions. Soyuz seems to be slightly safer than shuttle and has in common with the Orion both the 60's tech and the mostly expendable architecture (IIRC, some systems are transferred from a used Soyuz to a new one after being recertified).
So, the Orion/Unix analogy doesn't work.
Anyway, I was expecting something bigger ;-)
I only hope they pick up people who are claustrophiles.
You will still have to make the propellant.
Only your ray-guns should have problems. Any weapon based on a mechanical trigger that starts a chemical reaction should remain functional.
Depending on the event severity, you should be careful in case your railguns misfire.
There are ways it can be built that can allow for a reasonable amount of trust.
I worked in the Brazilian e-voting system and I can easily enumerate a couple ideas I had in 2002, during the project.
- There was a printer option that would be attached to the voting machine that would print the vote in human readable form. The paper votes would be deposited in a sealed ballot and random machines would be selected for examination. Since a tampered with software could detect the printer and behave well when it was attached and change votes when not, we could either to put printers in all of them or build "dummy printers" that print nothing but look like printers for the rest of the voting machine.
- A vote file cryptographically signed with the help of a TPM, tying the vote file to a given voting machine and making it really, really hard to inject a forged file in the vote counting process.
- Have the voting crews authorize voters with their own PINs.
And the mechanisms that have been in place since I was a kid:
- Making every citizen vote. This makes it rather easy to point out anyone trying to vote for people they know won't be there: If you go to vote and the machine says you already voted, call the police.
- Assign voters to specific machines - this way they can know how many votes each machine will receive. There are special situations like candidates allowed to vote anywhere, but you can deal with them.
- Having every voting machine assigned to three randomly-selected adults to act as a crew to validate each voter that has to vote there.
- Having police arrest any people campaigning within half a mile from a voting place.
- Mining vote patterns for discrepancies that trigger investigations. One that was detected was a series of votes with less than 30 second intervals between them that happened near closing time.
And a whole lot more measures that help.
It's not that hard. It may not be perfect, but, it is sure good enough.
"This isn't just Diebold. This is dozens of state, local, and federal agencies that abjectly failed in their duties to their constituents to protect the voting system"
Having a lot of people to hang should not deter you from starting. It's a huge undertaking, but it's also a very important one.
You know how it works: the earlier you start, the sooner you finish the job.
Just curious (non-US here)... How exactly it is defined?
Jury first, because it's a subtle one and not too many people will ever notice.
Ammo second, because you can always say it makes the world a safer place for children.
Then you are free to pick soap or ballot. Take one away and the other will take care of itself.
"refund"!
"refund"?!
The people responsible for this, from the lowest negligent intern up to the board, as well as every government official who vouched for the system, should be in jail.
How do you refund the people for corrupt government elected through electronic vote manipulation?
Since when MS-SQL users know anything else? They don't compare it with other databases: they rely on what they read on MS's Technet.
Not at all.
"First off, learn to use the 'quote' tag, fucktard. It was really challenging trying to sort your ramblings out when they were intertwined with my erudite arguments."
I am sorry to provide exercise for your brain, but I don't really like to write in HTML. I really don't care about you that much as to create a minor nuisance for me.
As for Apple, what exactly could they deliver of OLPCs? An OS, a WebKit-based browser and an e-mail program? Oh... I had no idea they were so difficult to do or that they were not readily available on just about every platform. Applications? Which ones? iWork? Keynote? It's not a general purpose laptop - it's a learning tool for kids.
As for the portability of Linux, it covers far more platforms than OSX. Having distros maintained for as many platforms ensure you have a full operating environment on each and every one of them. Porting for OSX is more painful because of the lack (on default installs) of X. No X means no GUI, or, at least, no Unix-style GUI. It doesn't matter how much certified UNIX OSX is, it won't run any Unix software that expects to see X. And, BTW, being a certified UNIX is... well... not worth much these days unless you want to sell for the US government. For most of the world, it's utterly irrelevant.
The article you mentioned is about cross-compiling with different versions of glibc. This is usually hard and that's why good distros provide build farms. Want your package built for, say, Itanium? Drop it in the proper place and there it goes. Isn't it wonderful what you can achieve when you just know the resources you have? OTOH, if you think you are clever enough to do it on your own, you should at least RTFM and learn that different platforms may have different library versions, including the more important ones.
"And yet you provide not ONE single example... Hmmm..."
As for a good example of Linux on ARM since ever, you can check the availability on distro sites. ARM ports have been available since, at least, Debian 2.2. Since a major Debian release takes a lot of time, I assume 2.2 happened ages ago. AFAIK, this was before OSX was even running on x86s (of course, it ran on x86s as OPENSTEP, but that's another story you are probably too young to remember). Before that, it also ran on Motorola's 68K (just as Linux, BTW) family.
"But when it comes to LEGAL LIABILITY (which is what matters to organizations such as the OLPC project), WHO CAN BE MADE RESPONSIBLE?"
You gotta be kidding, right?
If you hire someone to develop for you, you write a contract. It's that simple. The FLOSS advantage is that you and your contractor can piggyback on the work of others (like Apple does with WebKit, GNU compilers and a whole lot of OSX stuff) while still being able to hire people to do specific jobs with that very rich codebase.
Even if Apple offered OSX at no cost for OLPC, it would still belong to Apple. That means OLPC would not be able to change OSX in ways Apple didn't want to. If, in any given time, Apple decided it would no longer offer it, OLPC would die instantly.
And that is something you really don't want to happen: have another company decide your fate. If they wanted that, they could easily have gone for Microsoft from day one.
So, again, why ensure failure over time if you can at least have a shot going the other direction?
Sorry to tell you, but you must have heard this before: you are little more than an annoying Mac fanboy.
Let's do it this way: you search, let's say, the Debian repositories for ARM architecture, and then you let us know if you can't find a given program you need that much.
Usually, package maintainers do a very nice job of porting programs. I would not be surprised if you could run Tuxracer on an IBM zSeries mainframe.
And you don't need to be dismissive on the users of OLPCs. People have the ability to learn, given the incentives.
"Here's my point: OLPC can actually CONTRACT with Apple. How do they MAKE "Linux" (whatever THAT is!) perform ON SCHEDULE, TO SPEC, *OR ELSE*?"
That's hard to do when you accept Apple's offer is to do it for free. How can you exactly force Apple to do something for free? If you have a ton of money, you can hire them, but how do you justify them doing that instead of fostering open-source development that you can own in the end? Would OLPC own the ARM port of OSX? Of course not.
"Second, IT'S NOT JUST THE OS, STUPID!" yada yada yada "not only does "Linux" (whatever THAT is!) have to be ported; but also a gazillion other OSS applications."
Linux already runs on ARM. As do a lot of Linux applications. On ARM, on MIPS, on POWER/PowerPC, SPARC and others. Most of the time, all you need is to recompile the program. This is, you know, the beauty of real Unix-like OSs.
"STARTING with Apple (and its ARM experience, solid, platform-agnostic OS, and application portfolio) sure makes a LOT of sense to the OLPC people."
I bet you can find a dozen other companies more experienced in ARM processors than Apple. They just started it the other year. Linux is running on ARM architectures (from the very simple embedded ones to netbook-class OMAP ones) since like ever.
"And besides, none of this PRECLUDES the" F/OSS community" (whatever THAT is!) from offering one, or even a hundred, "Linux" (whatever THAT is!) distros, does it?"
No. But why would you hire Apple when you could just pick one and cultivate it to the stage you want? You can hire Apple just as much you can hire Canonical or Red Hat. If you can pay, you can have some very talented people.
"So, what was your point, again, exactly?"
Sorry, but your argument crumbled. Can you provide a good one the next time?
Care to explain why?
I am asking because I could not do it without resorting to some type of circular reasoning.