Having had a parking break fail, your assessment is a little off. It didn't account for a non-owner parking it. The last driver may be considered to be in control of the car at the time of the crash. They are the one that turned the wheel and set the brake. If they did it wrong, it's not the owner's fault. If they did it right, then they can point to the owner for not maintaining the car, but they would still likely be the first point of interest.
The injured party was suing the driver, not the driver's insurance company. However the attorney was from the insurance company, it was not the driver's personal attorney.
You sue who you have the relationship with (the one who it you), regardless of who's liable (unless the person at issue is a minor, in which case you sue the parents). The insurance company's lawyer was there because the insurance company is the responsible party. They contractually took that role from the owner.
The point is that it's a solved problem, even if you personally refuse to help friends in need, the problem is common and solved. Smart-cars doesn't change that.
Likely it would be a case where the liability would be somewhat shared, and fault would be found to help with safety fixes, but not to assign liability.
Nope. My insurance accepts liability (up to a certain point, and for certain things). If I'm sued, the insurance company enjoins the suit, becoming a co-defendant, taking direct liability.
But then there's a difference between the legal definition of liable (the one responsible for paying the bill is legally liable, meeting your definition), and the dictionary definition (the one responsible for the action that caused the loss).
Wait, why should I pay insurance once I'm not in control of the vehicle?
If you own a car and let someone borrow it, you are still paying for the insurance on the vehicle while they are driving it. If it's an issue, why haven't you complained before now? There's no change. The owner is insured.
I don't have to have a license to own a house, but I do have a large civil liability when I purchase one. Having a civil liability enumerated doesn't imply licensing is required. I should just have a way to sue the person who sends an infected file to me. They caused me $0.05 in damages. Not much, but is is actual damage, so I should be able to collect. If I could, then people would secure or disconnect their computers.
I've done it in real life. Backwards as it is, building a single middleware system "close" to the final goal, then publishing it to the separate interface and back-end teams made for a much better result than the last middleware project where trying to get buy-in from all teams for every decision burned $100 million before the plug was pulled and years passed before they attempted again. Having the smaller teams make changes necessary to integrate is always easier than having the big piece accommodate the little ones.
If the path to the end doesn't change the effort or result, then try adding water to acid next time and see if the identical proportions act differently whether you add acid to water or water to acid.
When Wal-Mart wants a new interface with suppliers, they talk to them to get a general wish list, then do their own thing, that may or may not match suppliers' requirements. Wal-Mart then informs the suppliers of what they need to do to sell to Wal-Mart. And they comply, almost always. That's much simpler and easier to do (for both Wal-Mart and often the suppliers themselves), than cooperating on a mutually agreeable interface between Wal-Mart and thousands of suppliers. It works in real life in many situations, you just must be not very good.
And what if the computer will never hand over control until fully parked in a legal and safe location with no imminent threat?
I can imagine a number of edge cases where the computer would not be able to see a safe path, and would give back control, but could do so without any immediate action by the driver. The two I listed elsewhere are changing weather conditions where it's no longer safe to proceed, and something like an event parking lot, where the lot is unlined, mostly uncontrolled, and pedestrians and cars are roaming in all directions.
Which, to me, is kind of a fairly fundamental problem with self driving cars. It's all or nothing. And if *all* the cars on the road aren't autonomous, then the autonomous ones are mostly a traffic hazard with no clear liability.
How are the self-driving ones a hazard, and the self-driven ones not?
You presume the computer will hand over at impact point. The computer will hand over with a large buffer, and if the human doesn't respond quickly enough, the computer will park the car in a safe place. "problems" include parking in a parking lot with no lines, and people walking from all directions and cars traveling in unpredictable lines (temporary lots at events are often like this). There will be plenty of notice, and the car can find a sub-optimal location to pull over and stop, for the human to take over. Also, if there is an issue with certain circumstances, heavy snow in the day where the lines are being covered and traction is variable, while others aren't slowing enough, so that if you go a "safe speed" you are likely to be hit from behind, and if you go the prevailing speed, you are likely to lose control, you can still turn off the path to the destination, and park on a side street so the human has infinite time to take the handover. It would never attempt a handover with "imminent collision with deer detected, please regain control of car for liability reasons". That's an anti-car strawman. That case is easy to program anyway. Steer straight, stop (on turn, stay in lane, stop). If you can't stop in time, still try. Never swerve, even for a cow or other larger animal.
Driving is easy. Emotional people trying to "drive" while reading the newspaper is hard.
Because people don't think modular. A complex system is easier than a simple system with multiple I/O to other incompletely defined systems.
That, and they did it backwards. There should have been one portal per state. Whether the state or the feds built it doesn't matter. Then the fed one integrates the 50 states to give some generic information and direct signups to the state portal. If they had built 50 portals with a shared home page, they'd have done better. Then, the states that were working are integrated in the fed, and stand alone, user's choice. The states that declined to make their own get one made for them, likely similar to what they would have made.
one rule with 50 cases (a single 50-state site) is complex and doomed to failure. 50 rules with 1 case each is much easier.
People try to solve complex problems, when it's really a collection of simple problems. The problem isn't programming or development, but problem solving. Solve the problem, and the solution is easier.
A contractor dealing with the government knows what they are getting into, or is incompetent in dealing with the government. If they can't deliver, then that is their fault. Part of requirements gathering is inferring unstated requirements. If you are too incompetent to "find" all the requirements, you shouldn't be developing.
An ultra safe mode should brake less "quicker". "Safe" is leaving more distance, so driving safe will result in braking earlier, but less hard. At least, when done right, which so many seem to get wrong. Also, AI "safe" (programmed to be aware of the tragedy of the commons) will realize that faster is safer. So "safer" program should have speed as mostly irrelevant, but perhaps buzz the human sooner for unexpected conditions.
How is faster safer? Cars spend less time on the road, so they get off sooner, leaving more room for others. The less congestion, the safer (at least with human drivers), so go fast for safety.
45 would be an inefficient speed with self-driving cars. If everyone else programs for "best speed" then your best economy will likely be joining in a train of self-drivers for aerodynamic efficiency. Of course, if you insist on sleeping in the back of your van (illegal in most places today, unless you are sleeping in a seat, in which case you could sleep up front), when most other cars are Prius-like in cross-section, you may still have issues with trains.
For road efficiency, they should redirect "slow" cars to US1, rather than I5, or something like that, but you'd probably complain that the additional turns would disturb your sleep.
With everyone else at optimal speed, your best option for the self-driving car would be to sleep in the driver seat (10 p.m. or whenever you go to sleep normally) and program it with an arrival time at 9 a.m. Then, at 3 a.m. it'll turn on and drive itself, with you on board, rather than leaving early and blocking the road to go less efficiently in a manner than harms others by obstructing the road.
Had you not heard of Florida? The punch-out ballots (famous in 2000, and still around in 2004, no idea how many places still use them, if any) don't have any indication on them what the question was. Much like a scantron test sheet, you fill in bubbles and hope the answer key is correct, but you have no way of knowing how it is actually counted.
And the fill-in type (used in the Coleman-Franken race) are also ambiguous, even though the discussions on it I saw on the wikipedia page didn't indicate anyone ever asserted they were ambiguous. The reason the Florida-style punch outs gained popularity is that what to do with "stray marks" and improperly filled bubbles is troublesome, so the more restrictive the interface with the physical ballot, the lower the chance of marking error.
As far as voting systems go, if you can't come up with one that doesn't get insulted heavily, maybe you're just bad at devising voting systems.
Usually the comments aren't insults of the systems, that's not an insult. You can't insult an apple. But they are personal insults aimed at me personally for pointing out that the USA was founded on open voting, and it worked (better than today's system) for the first 100 years. Open voting breaks down when the governemnt does, and the Civil War brought the end to Open Voting for popular elections, but Open Voting is still very common for petitions and in legislatures (I don't know of any of either where anonymous is allowed). When I point out those things, people tend to insult me. Have you ever heard of the government taking a "legalize pot" petition (there have been many) and raiding everyone on the list? No? Then why would you think it would happen under open voting?
Absentee votes are (or ought to be) a small portion of the overall vote — whereas under your proposal all votes will be known to all making retaliation of all sorts immediately possible. Does not even have to be about kneecaps: Fire Department may choose not to rush to a house, whose owner voted to keep their salaries capped, for example.
So, would that be legal or illegal? I imagine that if such illegal retalliation was discovered, the vote for the funding of the department would decrease, rather than increase the next vote. And the department would be keeping a very large and complicated database. What would they do if the home is owned by someone who voted against them, and rented by someone who voted for them? Try to save the contents, but let the structure burn? How about vice versa?
There are issues if people have a real-time access to a full database with lookup by person, but restrict the database to manual-access-only, and almost all of those problems go away. All that's left is the fraud we have today with absentee votes.
I've never lived in a place with absentee votes that were restricted in any way. Unless the "absentee voting" is voting in person, it's still open to corruption. And 100% of the population "could" vote absentee, and no rules would prevent that, so there's nothing in the current system that would prevent widespread fraud that you fear, unless everywhere I haven't lived doesn't allow absentee.
Maybe, some way of publishing the votes anonymously can be found — the way students' grades are published... With some sort of digest of their name (instead of the name itself), for example — so everybody can verify their vote without knowing that of the others.
The problem with that I was pointed to is that someone must own the list. If the government manages and prints the list, then they could put the name with the number like students grades are done. Then someone would be able to leak it back to the fire department. Everyone must "trust" someone for their vote system to work. The real arguments come down to who do you trust, and how much.
Student grades are usually a subset of SSN, which isn't exactly "secret", and would allow trivial sharing of secrets for verification. A private per-vote secret would require a paper receipt for someone to be able to remember it, so it'd be easier for someone to demand to get proof of vote.
If the fixes are so simple for the current system, why haven't they been done in the 100+ years the system has been in place?
And I'm not proposing a "new" system. The system I'm proposing was used for the first 100 or so years in the USA. And it never saw disuse, as most legislatures have open voting for all laws, and petitions and such are open and verified voting. So we've used it continuously in some form since the founding of the USA. We just ended open voting temporarily for a little Civil War, and it was time to switch back about 1910 or so, and we are a little over due for return to the previous system that works better in a stable country than anonymous voting. Yes, it works worse in countries with an unstable government and open revolt. But we don't have that now, and haven't for many years.
At no point then, does it become possible to tamper with the ballot without it being noticed.
You focused on one point I didn't ever address. How does your system address booth tampering, ballot stuffing, and ballot loss? All three have been proven to happen in the US, and no measure has been implemented to prevent them. Well, aside from e-voting, that is considered the worst of all worlds.
After I discussed measures to make it difficult to commit fraud?
Your reply, "I do know that, given any suspicion, people will look carefully at my ballot and interpret it the way I made it out." indicates you think there is 0% probability that anyone could ever mis-interpret your vote. Given the famous results in Florida, we have proof that some votes will not be counted in the manner you think you voted. Reality trivially proves you wrong, and I pointed that out. Your reply was "nuh uh", and I considered that a denial of belief a vote could be mis-counted.
That would be a naive and incorrect position.
What I want is a system that can let a voter see how their vote was cast *after* it was cast, and can trace a ballot back to a human. Verifiability is 100% missing from today's anonymous system. Someone could change your vote, or miscount it, and you could never know (yes, I know you have "faith" that it will be counted correctly, but if it weren't you couldn't know). And someone could steal 1000 valid ballots and fill them out in a desired manner, and stuff them in a ballot box, and there's no way to detect it once stuffed (again, I know there are measures to avoid stuffing, but they are demonstrably failures, when there have been reports of boxes with more ballots in them than delivered to the polling place the box came from).
I'm not proposing a system. I've suggested specific verified voting systems, and they've been insulted many times here, so I've reduced my comments to pointing out the many faults in the anonymous system, and that no new vulnerability is introduced through verified voting.
Just because the Germans did them worse doesn't mean the US didn't run concentration camps in WWII and funnel civilian American citizens into them, based on racial profiling. Just because the USSR did a police state worse than the US doesn't mean the US doesn't have one. Just because there exists a worse example isn't a counterpoint. In the US you can be assaulted by the police and arrested for almost anything. Why doesn't that qualify?
The US wanted to permanently hold Vietnam in the sense they wanted to permanently prevent an unfriendly government from taking charge. That's permanent.
Woodrow Wilson saw that the War to End All Wars wouldn't. WWII is a direct result of the conservatives in the US blocking the League of Nations. That's not hindsight. That's ignored foresight.
The Bug described runs on Gasoline. Diesel requires a hot combustion chamber. Glow plugs and block heaters should keep up, if used appropriately. In some places, they leave the cars on all day, idling for long periods. So you'll find engine timers in Alaska on fleet vehicles, as that's a more accurate measure of use than distance for gaging maintenance.
As for lube, run slightly lighter oil than spec. Cold lubrication is better, and hot lubrication worse. In winter, if you aren't driving cross country, it shouldn't be an issue.
Slashdot Mirror
Having had a parking break fail, your assessment is a little off. It didn't account for a non-owner parking it. The last driver may be considered to be in control of the car at the time of the crash. They are the one that turned the wheel and set the brake. If they did it wrong, it's not the owner's fault. If they did it right, then they can point to the owner for not maintaining the car, but they would still likely be the first point of interest.
There's no hardware lockout on aircraft. If you run over maintenance time, you can keep flying until the FAA grounds you.
You stated that everyone who can't develop all of the requirements perfectly is incompetent. This places you in a truly rare group.
I never used the word "perfect(ly)". That you don't understand plain English may be your issue.
The injured party was suing the driver, not the driver's insurance company. However the attorney was from the insurance company, it was not the driver's personal attorney.
You sue who you have the relationship with (the one who it you), regardless of who's liable (unless the person at issue is a minor, in which case you sue the parents). The insurance company's lawyer was there because the insurance company is the responsible party. They contractually took that role from the owner.
The point is that it's a solved problem, even if you personally refuse to help friends in need, the problem is common and solved. Smart-cars doesn't change that.
Likely it would be a case where the liability would be somewhat shared, and fault would be found to help with safety fixes, but not to assign liability.
Nope. My insurance accepts liability (up to a certain point, and for certain things). If I'm sued, the insurance company enjoins the suit, becoming a co-defendant, taking direct liability.
But then there's a difference between the legal definition of liable (the one responsible for paying the bill is legally liable, meeting your definition), and the dictionary definition (the one responsible for the action that caused the loss).
Wait, why should I pay insurance once I'm not in control of the vehicle?
If you own a car and let someone borrow it, you are still paying for the insurance on the vehicle while they are driving it. If it's an issue, why haven't you complained before now? There's no change. The owner is insured.
I don't have to have a license to own a house, but I do have a large civil liability when I purchase one. Having a civil liability enumerated doesn't imply licensing is required. I should just have a way to sue the person who sends an infected file to me. They caused me $0.05 in damages. Not much, but is is actual damage, so I should be able to collect. If I could, then people would secure or disconnect their computers.
I've done it in real life. Backwards as it is, building a single middleware system "close" to the final goal, then publishing it to the separate interface and back-end teams made for a much better result than the last middleware project where trying to get buy-in from all teams for every decision burned $100 million before the plug was pulled and years passed before they attempted again. Having the smaller teams make changes necessary to integrate is always easier than having the big piece accommodate the little ones.
If the path to the end doesn't change the effort or result, then try adding water to acid next time and see if the identical proportions act differently whether you add acid to water or water to acid.
When Wal-Mart wants a new interface with suppliers, they talk to them to get a general wish list, then do their own thing, that may or may not match suppliers' requirements. Wal-Mart then informs the suppliers of what they need to do to sell to Wal-Mart. And they comply, almost always. That's much simpler and easier to do (for both Wal-Mart and often the suppliers themselves), than cooperating on a mutually agreeable interface between Wal-Mart and thousands of suppliers. It works in real life in many situations, you just must be not very good.
I can imagine a number of edge cases where the computer would not be able to see a safe path, and would give back control, but could do so without any immediate action by the driver. The two I listed elsewhere are changing weather conditions where it's no longer safe to proceed, and something like an event parking lot, where the lot is unlined, mostly uncontrolled, and pedestrians and cars are roaming in all directions.
Which, to me, is kind of a fairly fundamental problem with self driving cars. It's all or nothing. And if *all* the cars on the road aren't autonomous, then the autonomous ones are mostly a traffic hazard with no clear liability.
How are the self-driving ones a hazard, and the self-driven ones not?
You presume the computer will hand over at impact point. The computer will hand over with a large buffer, and if the human doesn't respond quickly enough, the computer will park the car in a safe place. "problems" include parking in a parking lot with no lines, and people walking from all directions and cars traveling in unpredictable lines (temporary lots at events are often like this). There will be plenty of notice, and the car can find a sub-optimal location to pull over and stop, for the human to take over. Also, if there is an issue with certain circumstances, heavy snow in the day where the lines are being covered and traction is variable, while others aren't slowing enough, so that if you go a "safe speed" you are likely to be hit from behind, and if you go the prevailing speed, you are likely to lose control, you can still turn off the path to the destination, and park on a side street so the human has infinite time to take the handover. It would never attempt a handover with "imminent collision with deer detected, please regain control of car for liability reasons". That's an anti-car strawman. That case is easy to program anyway. Steer straight, stop (on turn, stay in lane, stop). If you can't stop in time, still try. Never swerve, even for a cow or other larger animal.
Driving is easy. Emotional people trying to "drive" while reading the newspaper is hard.
Because people don't think modular. A complex system is easier than a simple system with multiple I/O to other incompletely defined systems.
That, and they did it backwards. There should have been one portal per state. Whether the state or the feds built it doesn't matter. Then the fed one integrates the 50 states to give some generic information and direct signups to the state portal. If they had built 50 portals with a shared home page, they'd have done better. Then, the states that were working are integrated in the fed, and stand alone, user's choice. The states that declined to make their own get one made for them, likely similar to what they would have made.
one rule with 50 cases (a single 50-state site) is complex and doomed to failure. 50 rules with 1 case each is much easier.
People try to solve complex problems, when it's really a collection of simple problems. The problem isn't programming or development, but problem solving. Solve the problem, and the solution is easier.
A contractor dealing with the government knows what they are getting into, or is incompetent in dealing with the government. If they can't deliver, then that is their fault. Part of requirements gathering is inferring unstated requirements. If you are too incompetent to "find" all the requirements, you shouldn't be developing.
An ultra safe mode should brake less "quicker". "Safe" is leaving more distance, so driving safe will result in braking earlier, but less hard. At least, when done right, which so many seem to get wrong. Also, AI "safe" (programmed to be aware of the tragedy of the commons) will realize that faster is safer. So "safer" program should have speed as mostly irrelevant, but perhaps buzz the human sooner for unexpected conditions.
How is faster safer? Cars spend less time on the road, so they get off sooner, leaving more room for others. The less congestion, the safer (at least with human drivers), so go fast for safety.
45 would be an inefficient speed with self-driving cars. If everyone else programs for "best speed" then your best economy will likely be joining in a train of self-drivers for aerodynamic efficiency. Of course, if you insist on sleeping in the back of your van (illegal in most places today, unless you are sleeping in a seat, in which case you could sleep up front), when most other cars are Prius-like in cross-section, you may still have issues with trains.
For road efficiency, they should redirect "slow" cars to US1, rather than I5, or something like that, but you'd probably complain that the additional turns would disturb your sleep.
With everyone else at optimal speed, your best option for the self-driving car would be to sleep in the driver seat (10 p.m. or whenever you go to sleep normally) and program it with an arrival time at 9 a.m. Then, at 3 a.m. it'll turn on and drive itself, with you on board, rather than leaving early and blocking the road to go less efficiently in a manner than harms others by obstructing the road.
And the fill-in type (used in the Coleman-Franken race) are also ambiguous, even though the discussions on it I saw on the wikipedia page didn't indicate anyone ever asserted they were ambiguous. The reason the Florida-style punch outs gained popularity is that what to do with "stray marks" and improperly filled bubbles is troublesome, so the more restrictive the interface with the physical ballot, the lower the chance of marking error.
As far as voting systems go, if you can't come up with one that doesn't get insulted heavily, maybe you're just bad at devising voting systems.
Usually the comments aren't insults of the systems, that's not an insult. You can't insult an apple. But they are personal insults aimed at me personally for pointing out that the USA was founded on open voting, and it worked (better than today's system) for the first 100 years. Open voting breaks down when the governemnt does, and the Civil War brought the end to Open Voting for popular elections, but Open Voting is still very common for petitions and in legislatures (I don't know of any of either where anonymous is allowed). When I point out those things, people tend to insult me. Have you ever heard of the government taking a "legalize pot" petition (there have been many) and raiding everyone on the list? No? Then why would you think it would happen under open voting?
Absentee votes are (or ought to be) a small portion of the overall vote — whereas under your proposal all votes will be known to all making retaliation of all sorts immediately possible. Does not even have to be about kneecaps: Fire Department may choose not to rush to a house, whose owner voted to keep their salaries capped, for example.
So, would that be legal or illegal? I imagine that if such illegal retalliation was discovered, the vote for the funding of the department would decrease, rather than increase the next vote. And the department would be keeping a very large and complicated database. What would they do if the home is owned by someone who voted against them, and rented by someone who voted for them? Try to save the contents, but let the structure burn? How about vice versa?
There are issues if people have a real-time access to a full database with lookup by person, but restrict the database to manual-access-only, and almost all of those problems go away. All that's left is the fraud we have today with absentee votes.
I've never lived in a place with absentee votes that were restricted in any way. Unless the "absentee voting" is voting in person, it's still open to corruption. And 100% of the population "could" vote absentee, and no rules would prevent that, so there's nothing in the current system that would prevent widespread fraud that you fear, unless everywhere I haven't lived doesn't allow absentee.
Maybe, some way of publishing the votes anonymously can be found — the way students' grades are published... With some sort of digest of their name (instead of the name itself), for example — so everybody can verify their vote without knowing that of the others.
The problem with that I was pointed to is that someone must own the list. If the government manages and prints the list, then they could put the name with the number like students grades are done. Then someone would be able to leak it back to the fire department. Everyone must "trust" someone for their vote system to work. The real arguments come down to who do you trust, and how much.
Student grades are usually a subset of SSN, which isn't exactly "secret", and would allow trivial sharing of secrets for verification. A private per-vote secret would require a paper receipt for someone to be able to remember it, so it'd be easier for someone to demand to get proof of vote.
And I'm not proposing a "new" system. The system I'm proposing was used for the first 100 or so years in the USA. And it never saw disuse, as most legislatures have open voting for all laws, and petitions and such are open and verified voting. So we've used it continuously in some form since the founding of the USA. We just ended open voting temporarily for a little Civil War, and it was time to switch back about 1910 or so, and we are a little over due for return to the previous system that works better in a stable country than anonymous voting. Yes, it works worse in countries with an unstable government and open revolt. But we don't have that now, and haven't for many years.
At no point then, does it become possible to tamper with the ballot without it being noticed.
You focused on one point I didn't ever address. How does your system address booth tampering, ballot stuffing, and ballot loss? All three have been proven to happen in the US, and no measure has been implemented to prevent them. Well, aside from e-voting, that is considered the worst of all worlds.
After I discussed measures to make it difficult to commit fraud?
Your reply, "I do know that, given any suspicion, people will look carefully at my ballot and interpret it the way I made it out." indicates you think there is 0% probability that anyone could ever mis-interpret your vote. Given the famous results in Florida, we have proof that some votes will not be counted in the manner you think you voted. Reality trivially proves you wrong, and I pointed that out. Your reply was "nuh uh", and I considered that a denial of belief a vote could be mis-counted.
That would be a naive and incorrect position.
What I want is a system that can let a voter see how their vote was cast *after* it was cast, and can trace a ballot back to a human. Verifiability is 100% missing from today's anonymous system. Someone could change your vote, or miscount it, and you could never know (yes, I know you have "faith" that it will be counted correctly, but if it weren't you couldn't know). And someone could steal 1000 valid ballots and fill them out in a desired manner, and stuff them in a ballot box, and there's no way to detect it once stuffed (again, I know there are measures to avoid stuffing, but they are demonstrably failures, when there have been reports of boxes with more ballots in them than delivered to the polling place the box came from).
I'm not proposing a system. I've suggested specific verified voting systems, and they've been insulted many times here, so I've reduced my comments to pointing out the many faults in the anonymous system, and that no new vulnerability is introduced through verified voting.
Just because the Germans did them worse doesn't mean the US didn't run concentration camps in WWII and funnel civilian American citizens into them, based on racial profiling. Just because the USSR did a police state worse than the US doesn't mean the US doesn't have one. Just because there exists a worse example isn't a counterpoint. In the US you can be assaulted by the police and arrested for almost anything. Why doesn't that qualify?
Good, then I can buy games in AU for my US Wii? I bought it after 2006, so it can't be an issue, right?
The US wanted to permanently hold Vietnam in the sense they wanted to permanently prevent an unfriendly government from taking charge. That's permanent.
Woodrow Wilson saw that the War to End All Wars wouldn't. WWII is a direct result of the conservatives in the US blocking the League of Nations. That's not hindsight. That's ignored foresight.
The Bug described runs on Gasoline. Diesel requires a hot combustion chamber. Glow plugs and block heaters should keep up, if used appropriately. In some places, they leave the cars on all day, idling for long periods. So you'll find engine timers in Alaska on fleet vehicles, as that's a more accurate measure of use than distance for gaging maintenance.
As for lube, run slightly lighter oil than spec. Cold lubrication is better, and hot lubrication worse. In winter, if you aren't driving cross country, it shouldn't be an issue.