*I* don't, but some of those people Linus was talking about probably do.
My point isn't that C or C++ have bad type safety (or good type safety for that matter) just that if type safety is a real concern for you, C++ isn't really going to leave you better off than C. In either one, the compiler won't save you if you make a type related error. Note that dynamic_cast adds a RUNtime test, the compiler still doesn't save you.
That's not necessarily true in C. If you use defines for the pushing and popping (queue/dequeue, link/unlink, etc) and build the chain pointers into the object, it will happily enforce the type based on the type of the head object. Once the pre-processor is done, the compiler will complain if you try to assign struct foo_t *next the address of a struct bar_t
There is, in fact, evidence that your concerns are well founded. I wish I had a reference for it but the research demonstrated an uncomfortable likelihood that a search against a large pool of DNA profiles would nearly inevitably return several false matches.
And that was before considering that few DNA samples are perfect in the first place.
A lot of people would have a lot less confidence in DNA matches if they knew that really only a tiny sampling of the DNA's characteristics are spot checked for a match. They definitely do not match them up codon by codon like many people imagine.
I wouldn't use it for most firmware either. Firmware often has size constraints that C++ is less than friendly about. It also often runs in a limited environment that may not be friendly to C++.
Actually, the kernel code is quite clear and understandable to any decent C programmer. I find it hard to consider it losing BADLY to anything.
Note that C++ isn't necessarily type safe either. It's not at all difficult to cast an object pointer to the wrong type and crash. It's no harder than it is in C.
I won't claim the IRS is angelic. In fact, I think we need to review what information they collect to be on the safe side.
But the NSA has clearly stepped over the line. It is no longer a possibility to worry about, it is a fact. They aren't just a potential enemy to be watched, they are an actual current enemy to be eliminated.
He worried because innocence doesn't necessarily result in a good outcome in criminal proceedings. If it actually goes to trial, bankruptcy is often the reward for the innocent. Other times, they end up convicted anyway and lose years of their lives (or simply lose their life in the case of capital murder).
So he was more like "OMG they think I brutally murdered someone, I'm going to prison and I didn't even do it!"
“I had lots of days sitting at the house with the dog,” he recalled in an interview, “wondering if these guys were going to use a battering ram to bust open the door and shoot my dog after he started barking at them.”
Sadly, we cannot claim his fear was unjustified. It certainly isn't without precedent.
I'm reminded of the 'serial killer' who raged across Europe for many years. It turned out that the DNA of the 'serial killer' was actually from the woman who packaged the cotton swabs.
It also turned out that they bought sterile swabs meant for medical use, but not swabs clear of DNA meant for forensic investigation.
The statistics are tricky there. The odds against a perfect false match to a particular bit of evidence are quite high. The odds of a false familial match against a degraded sample from ANY crime scene are quite different. I can't imagine why anyone would want to effectively sign up to be a go-to suspect.
Given news footage I have seen of someone's house after a search (which turned out to be based on an anonymous tip), I'd say TV has it right. They ripped sheet rock from the walls, destroyed furniture, and spread the insides of their mattresses over the floor. They didn't even apologize when it was over.
People would be more willing to cooperate if doing so wasn't so frequently punished even when you're innocent.
Law enforcement and our 'justice system' need to drill it into their skulls that unless they are very careful, the process of investigating and ultimately finding a person not guilty is itself an ordeal that amounts to punishment of the innocent.
It's no wonder people prefer to be left out of it rather than cooperate.
I'll bet that 'suspect' would have preferred not to be molested by the police over what was akin to a fishing expedition. They had no actual evidence that he committed the crime. It seems doubtful that it was actually probable cause.
They certainly owed him the courtesy of QUICKLY letting him know he had no need to worry about further suspicion hanging over him. Like doing their test IMMEDIATELY and letting him know within a couple days of taking the sample.
Sorry, no. As despised as the IRS is, it performs according to it's charter and the constitution (for the most part).
The NSA is an actual domestic enemy of the people. It's activities are illegal and it is actively damaging everyone's rights and security.
Nationalized health care would NOT give any government agent the ability to know exactly where I am most all of the time like the NSA illegally hacking my phone would. It would not let anyone know who I talk to, when that happens, or what was said like the illegal NSA hacking does.
Put a net under the bridge and they'll just run out in front of a bus or pull a knife on a cop instead. It simply isn't possible to eliminate even the most impulsive forms of death.
For it to have even a hint of effectiveness, they would actually have to nerf the world. If they're not going to do that, then they will have no effect at all.
I understand the motivation, but there are simply far too many ways to die if you want to. Even if we nerf the whole world and baby proof everything, we won't stop an adult or teen that wants to die from doing so. But we will make the world a worse place to live.
There is a significant chance that we will just force people to choose a horrific way to die or (perhaps worse) a way that is as likely to leave them in a horrific but not dead state as it is to kill them.
Except it suggests that we are filling a reservoir and that once full we may expect the rate of warming to increase. It still suggests that we had best cut our emissions before we get into serious problems. It could also be serious trouble if the reservoir is carbonic acid formation in the ocean.
I can say that the model would necessarily have a parameter for CO2 level and a function to determine it's level as the model evolved. If the real world doesn't match the evolution of the CO2 in the model, all bets are off and we can't really say what (if anything) the model is telling us.
Did you read the part about how the necessary conditions the model was based on weren't met so it can only be a thermal or CO2 sink? Or the part about how the temperature rise after 2000 is and remains non-zero?
I wake up one morning feeling dizzy. I know that disorientation training for the space program causes dizziness. Conclusion, I joined the space program in my sleep. The data fits perfectly!
Or perhaps since I am still in bed I should consider a different explanation. Perhaps I should take my temperature and consider seeing a doctor.
It is worth looking at why model C seems to track. It isn't being takes seriously as-is because we know the conditions it assumed did not happen. CFCs were sharply cut and some effort has been put towards methane and NOx, but there hasn't been a net reduction in CO2.
It is most probable that there is a CO2 or thermal sink that hasn't been accounted for. It could be huge or it could be just about full.
Part of the problem is that the deniers give the impression (to the scientists) that they will sit on their ass till the sea rises enough to float their chair unless the prediction is unmitigated disaster. So there is a tendency to report the worst case scenario.
In fact, I suspect that as long as there's money to be made by the deniers, Jesus could appear before them personally and explain it and they wouldn't believe him either.
In between are a bunch of people not sure who to believe.
Slashdot Mirror
*I* don't, but some of those people Linus was talking about probably do.
My point isn't that C or C++ have bad type safety (or good type safety for that matter) just that if type safety is a real concern for you, C++ isn't really going to leave you better off than C. In either one, the compiler won't save you if you make a type related error. Note that dynamic_cast adds a RUNtime test, the compiler still doesn't save you.
I've done that. It's not that out there.
Others use per-type link/unlink functions declared inline.
Yes, you SHOULD use that for casting, but you don't HAVE to.
In C, you CAN define a base struct that includes an enum specifying the correct type but you certainly don't have to.
That is, C++ offers a convenient way to do it right but doesn't require it, so it remains easy to cast incorrectly and do a crash and burn.
That's not necessarily true in C. If you use defines for the pushing and popping (queue/dequeue, link/unlink, etc) and build the chain pointers into the object, it will happily enforce the type based on the type of the head object. Once the pre-processor is done, the compiler will complain if you try to assign struct foo_t *next the address of a struct bar_t
Exactly, so why the claims that C++ is so much safer?
So how does that differ at all from C? Shouldn't != can't
There is, in fact, evidence that your concerns are well founded. I wish I had a reference for it but the research demonstrated an uncomfortable likelihood that a search against a large pool of DNA profiles would nearly inevitably return several false matches.
And that was before considering that few DNA samples are perfect in the first place.
A lot of people would have a lot less confidence in DNA matches if they knew that really only a tiny sampling of the DNA's characteristics are spot checked for a match. They definitely do not match them up codon by codon like many people imagine.
I wouldn't use it for most firmware either. Firmware often has size constraints that C++ is less than friendly about. It also often runs in a limited environment that may not be friendly to C++.
Actually, the kernel code is quite clear and understandable to any decent C programmer. I find it hard to consider it losing BADLY to anything.
Note that C++ isn't necessarily type safe either. It's not at all difficult to cast an object pointer to the wrong type and crash. It's no harder than it is in C.
I won't claim the IRS is angelic. In fact, I think we need to review what information they collect to be on the safe side.
But the NSA has clearly stepped over the line. It is no longer a possibility to worry about, it is a fact. They aren't just a potential enemy to be watched, they are an actual current enemy to be eliminated.
He worried because innocence doesn't necessarily result in a good outcome in criminal proceedings. If it actually goes to trial, bankruptcy is often the reward for the innocent. Other times, they end up convicted anyway and lose years of their lives (or simply lose their life in the case of capital murder).
So he was more like "OMG they think I brutally murdered someone, I'm going to prison and I didn't even do it!"
From TFA:
“I had lots of days sitting at the house with the dog,” he recalled in an interview, “wondering if these guys were going to use a battering ram to bust open the door and shoot my dog after he started barking at them.”
Sadly, we cannot claim his fear was unjustified. It certainly isn't without precedent.
I'm reminded of the 'serial killer' who raged across Europe for many years. It turned out that the DNA of the 'serial killer' was actually from the woman who packaged the cotton swabs.
It also turned out that they bought sterile swabs meant for medical use, but not swabs clear of DNA meant for forensic investigation.
The statistics are tricky there. The odds against a perfect false match to a particular bit of evidence are quite high. The odds of a false familial match against a degraded sample from ANY crime scene are quite different. I can't imagine why anyone would want to effectively sign up to be a go-to suspect.
Given news footage I have seen of someone's house after a search (which turned out to be based on an anonymous tip), I'd say TV has it right. They ripped sheet rock from the walls, destroyed furniture, and spread the insides of their mattresses over the floor. They didn't even apologize when it was over.
People would be more willing to cooperate if doing so wasn't so frequently punished even when you're innocent.
Law enforcement and our 'justice system' need to drill it into their skulls that unless they are very careful, the process of investigating and ultimately finding a person not guilty is itself an ordeal that amounts to punishment of the innocent.
It's no wonder people prefer to be left out of it rather than cooperate.
I'll bet that 'suspect' would have preferred not to be molested by the police over what was akin to a fishing expedition. They had no actual evidence that he committed the crime. It seems doubtful that it was actually probable cause.
They certainly owed him the courtesy of QUICKLY letting him know he had no need to worry about further suspicion hanging over him. Like doing their test IMMEDIATELY and letting him know within a couple days of taking the sample.
Sorry, no. As despised as the IRS is, it performs according to it's charter and the constitution (for the most part).
The NSA is an actual domestic enemy of the people. It's activities are illegal and it is actively damaging everyone's rights and security.
Nationalized health care would NOT give any government agent the ability to know exactly where I am most all of the time like the NSA illegally hacking my phone would. It would not let anyone know who I talk to, when that happens, or what was said like the illegal NSA hacking does.
Put a net under the bridge and they'll just run out in front of a bus or pull a knife on a cop instead. It simply isn't possible to eliminate even the most impulsive forms of death.
For it to have even a hint of effectiveness, they would actually have to nerf the world. If they're not going to do that, then they will have no effect at all.
From TFA:
Instead of treating individual risk, means restriction entails modifying the environment by removing the means by which people usually die by suicide
So they really are talking about nerfing the world rather than making sure counselling is readily available.
In Japan, the hand can be used like a knife. But this doesn't work on a tomato.
I understand the motivation, but there are simply far too many ways to die if you want to. Even if we nerf the whole world and baby proof everything, we won't stop an adult or teen that wants to die from doing so. But we will make the world a worse place to live.
There is a significant chance that we will just force people to choose a horrific way to die or (perhaps worse) a way that is as likely to leave them in a horrific but not dead state as it is to kill them.
Besides, how will we cook without knives?
Except it suggests that we are filling a reservoir and that once full we may expect the rate of warming to increase. It still suggests that we had best cut our emissions before we get into serious problems. It could also be serious trouble if the reservoir is carbonic acid formation in the ocean.
I can say that the model would necessarily have a parameter for CO2 level and a function to determine it's level as the model evolved. If the real world doesn't match the evolution of the CO2 in the model, all bets are off and we can't really say what (if anything) the model is telling us.
Did you read the part about how the necessary conditions the model was based on weren't met so it can only be a thermal or CO2 sink? Or the part about how the temperature rise after 2000 is and remains non-zero?
I wake up one morning feeling dizzy. I know that disorientation training for the space program causes dizziness. Conclusion, I joined the space program in my sleep. The data fits perfectly!
Or perhaps since I am still in bed I should consider a different explanation. Perhaps I should take my temperature and consider seeing a doctor.
It is worth looking at why model C seems to track. It isn't being takes seriously as-is because we know the conditions it assumed did not happen. CFCs were sharply cut and some effort has been put towards methane and NOx, but there hasn't been a net reduction in CO2.
It is most probable that there is a CO2 or thermal sink that hasn't been accounted for. It could be huge or it could be just about full.
Part of the problem is that the deniers give the impression (to the scientists) that they will sit on their ass till the sea rises enough to float their chair unless the prediction is unmitigated disaster. So there is a tendency to report the worst case scenario.
In fact, I suspect that as long as there's money to be made by the deniers, Jesus could appear before them personally and explain it and they wouldn't believe him either.
In between are a bunch of people not sure who to believe.