When you show the grandparent's comment to your overlords, please take the bolded section and print it in 72-point text. The new design is quite possibly the worst thing I've ever seen anyone try to do to Slashdot and it will kill the site. I'm not as active as I was a year ago due to lack of time, but in your statistics last year I was one of the 5 most active commenters for one or two quarters. I would not visit Slashdot if they switched to this abomination.
I never understood this either. For example, they ripped all of the locale support out of libc to remove bloat. Which then means that everyone who uses the NDK reimplements it in another library, which is not shared...
Some bits of the FreeBSD kernel make it into the BSD server in the XNU kernel. One of the big ones is the MAC framework (SEBSD), which is shared between FreeBSD and XNU and supports pluggable access control policies. This is used to implement the code signing logic on Juniper routers and the application sandboxing on iOS and OS X. There are some pretty big differences to the VM subsystems on both (they're both derived from Mach 2.5, but they've diverged hugely since then).
You might want to scan the Android stack for FreeBSD copyrights sometime. Most of the Android libc is a slimmed-down version of the FreeBSD one, and there are quite a few other bits of FreeBSD code in there too. In terms of lines of code, I think there is about as much FreeBSD code in Android as there is Linux code.
I'm not sure what your point is. In the House, all seats were up for reelection and the Republicans won on seats but did not get the majority of the popular vote, so this means that they represent the will of the people? In the Senate, not all seats were up for reelection, but in the 1/3 that were, the Republicans lost 2 seats overall and lost the popular vote, so this means that they represent the will of the people? In the Presidential election, there is only one seat and it was up for reelection, the Republicans lost both the popular vote and failed to win the seat, so this means that they represent the will of the people?
You can write off the Senate elections as only giving you the views of 1/3 of the country, but you're still left with two national elections where more people voted against the Republicans than voted for them (I'm not sure what the statistics are for third parties, so this may be true for Democrats too) and where more people voted for the Democrats than voted for the Republicans. And yet you still claim that the Republican majority in the House represents the will of the people?
So yes, by the rules of the system, which are arguably superior to the hypothetical rules you seem to wish we operated under, the House Majority Republicans *are* there due to the will of the people, and *do* represent it.
If by 'the people' you mean the commission that defines constituency boundaries, then I suppose you're right. If you mean 'the people who voted' then you are wrong. If you mean 'the people who are eligible to vote' then neither party can claim to be even approximately representing the will of the people.
The popular vote count in both elections is definitely relevant if you're claiming (as the grandparent was) that a narrow victory represents 'the will of the people'. Getting 49.6% of the popular vote does not mean that you represent the will of the people, it means that you represent the will of about half of the people. Having a majority with this amount of the vote means that constituency boundaries favoured you this time. It is most definitely not a mandate to push through your agenda without compromising, as the grandparent appeared to think.
You're an idiot, you know that? The House of Representatives represents the will of the people. The people want to get rid of Obamacare.
You realise that in the last election:
The Democrats won the Presidency.
In the Senate, of the 33 seats that were up for election, 23 were Democrat and the Republicans only needed to take 4 of them to have a majority. Instead, they lost 2.
In the House elections, the Democrats won the majority of the popular vote, got 3.4% more of the popular vote than last time (the Republicans got 4.8% less than last time), gained 8 seats, but still don't have an overall majority because of the way constituency boundaries are placed.
Or, to put it another way, in 2012:
In the Presidental election, 51.1% voted Democrat, 47.2% voted Republican.
In the Senate election, 53.7% voted Democrat, 42.1% voted Republican.
In the House election, 48.3% voted Democrat, 46.9% voted Republican.
So, in terms of popular vote, the Democrats go a (very slim) majority in two of the three elections, the Republicans didn't manage it in one. In the election where the Republicans did best (the Presidential race, 47.2%), they didn't get as much of the vote as the Democrats did in the election where they did the worst (House, 48.3%). Neither party got anywhere close to the percentage of the vote that enables someone who doesn't lie for a living to claim to have a mandate from the people.
Now, I realise the state of mathematical education is pretty poor in the USA, but being able to tell which of two numbers is bigger than the other is surely something that is covered.
That's not usually the case. If you keep the fact that you're gay secret from your wife, then there's something that you can be blackmailed over. That makes you a security risk. The security clearance process that I've been through (in the UK) didn't care if you were gay or had a string of mistresses, as long as these weren't things that you wanted to keep secret.
There's a big difference between using technology to aid in parenting and teaching and using technology as a substitute for parenting and teaching. This was true of television, of home computers, and is now true of tablets and smartphones.
The study was published a year or two ago, so I'm not sure why it's appearing on Slashdot now - normally I'd blame the 'editors' for this kind of late submission, but apparently they found a 'news' site with a slow news day that decided that old research papers now count as news. It's a pretty neat concept though and more recent work has improved the accuracy a lot by feeding other sensors into the mix.
Okay, since you've obviously never even visited a research lab, let alone worked in one, let me give you a brief overview of how it works:
There are funded projects and there are resources. Some resources are allocated to a specific project and some are not. Typically, the ones allocated to a project can be used for other things if that project doesn't need them, but it gets priority. There's usually a fair amount of unused infrastructure that can be used for unofficial projects if no one else needs it.
In a well-run research lab (of which Bell Labs and Xerox PARC in this era are archetypes), there is no strict accounting of time to projects. People are expected to work on some things as part of big ongoing projects, but they have a lot of free time to devote to other things that they consider fun and interesting. This is done because the people running the lab know that these spare-time projects are how you get the seeds of the next iteration of big projects. The same is true in most research labs, including most universities, which is why I suspect that you've never been to university, as you'd have encountered this concept before.
UNIX was such a project. It was not part of any funded project at Bell Labs and was done by a few guys for fun. It then grew and was used in some funded projects (troff was the product of one such project), but wasn't officially backed by AT&T until long after it was created.
There is a distinction between funding a project and paying an individual's salary. The small group that worked on UNIX (well, UNICS back then, in one of Peter's characteristic puns - it was renamed UNIX later when it had multi-user support) had their salaries paid by Bell Labs, but they were being paid to work on other things. Their work on UNIX was not backed by management and was not funded. They had no resources allocated to the UNIX project, they used whatever they could scrounge. I've done similar things with machines a few generations old to build infrastructure for fun projects and occasionally these go on to become funded projects.
MINIX 3 is a pretty conventional microkernel. Take a look, for example, at some of the virtual memory stuff done in L4/HURD. This can be seen as a continuation of the external pager support in Mach, but with a much simpler and cleaner design and a lot of direct application to current workloads.
Wow, really high standard of debate you have there. You appear to have no idea of how things work in research labs (industrial or academic). Did you attend university?
I'm neither cold fjord nor the AC, but you are entirely wrong. The current project that I work on is headed by Peter Neumann (one of the architects of MULTICS), and so I have a fairly good understanding of the environment that led to the creation of UNIX from conversations with someone who was there at the time. Pretty much everything that cold fjord said is correct. They got an old computer that no one else wanted and hacked up an OS in their own spare time. They didn't aim to take over the world - they just wanted something that had the bits of MULTICS that they liked (Peter has some frank opinions on their choices here) that they could use.
And they've been using it to explore some quite interesting ideas in kernel design. The fine-grained compartmentalism that a microkernel provides (at the expense of some performance) is starting to look more attractive in a world where computers run in very hostile environments and yet even a 50% slower kernel would have a negligible impact on user-perceived performance (or battery life).
I was 7 when I learned to program. We had one lesson a week taught by the school's headmaster on whatever he thought was interesting, and so he taught some programming in BASIC[1] on the BBC Model B. He also taught some geometry using Logo on the same machine. It was connected to a big TV (which, by modern standards, is a small TV), and he'd ask the class to describe the program and he'd type it. After school and at lunch and break times there were a few of these machines that we could use, and I learned a bit more. I asked my father to teach me a real language, and he taught me PL/M86 (which I still miss sometimes), and I then moved on to C[2].
When I got to university, I discovered how much of the theoretical side I was missing. The main problem with teaching programming at an early age is that it really needs to be accompanied by teaching logic and then game and graph theory. I've seen classes that do this well for under-10s, but they're very rare.
[1] The Dijkstra comment that teaching BASIC should be a criminal offence doesn't really apply to BBC BASIC, which had full support for structured programming, an integrated assembler, and direct access to memory-mapped hardware.
[2] Back then, you really needed makefiles because there was no equivalent to a modern compiler driver. Compilation, assembly, and linking were all separate, manual, steps.
The problem is that there's no well-supported way, for example in the DNS, for a server to say which certificate it will use. If HTTPS required two certificates, then that would just mean that you'd need to compromise two CAs (or one CA and get them to issue two certs), which given what we already know about the NSA program, has already happened. This is something that people like Ben Laurie at Google are working on with Certificate Transparency: trying to ensure that there is a recorded and verifiable chain showing that a certificate was issued to the real owner of the domain and that it is not being MITM'd.
The point of using control-alt-delete is that it's a key combination that can not be caught by any userspace process that does not have a special permission. This means that it's impossible to spoof the login screen on Windows without already having compromised the kernel. It doesn't matter what the key combination is, as long as it's one that is not delivered by the normal keypress event delivery mechanisms. Control-alt-delete is a reasonable choice, because no application author is likely to complain that they can't use this shortcut combination.
That's right, the difference is syntax and you've picked a terrible example. A better example would be method invocation. The differences between calling a C function pointer, a C++ virtual function, a Java method or an Objective-C method are significant.
That makes sense. I get a couple of 'call for papers' emails a day from dubious journals, often with such broad titles as 'The Journal of Modern Research', so it would be completely impossible for them to rate articles. The research establishment in the UK has tried quite hard over the last decade to counter this 'publish loads of crap' incentive. The old Research Assessment Exercise and the new Research Excellence Framework by which departments are assessed requires a small number (about one per year) of 'research outputs'. These can be high-impact papers, books, and so on, and in computer science can include things like published open source software (which counts as technology transfer if you can point to people using it).
And it's a completely fucked-up policy, because the hoops that a US company needs to jump through to hire an Iranian national are insane. So you end up educating a load of people, then telling them that they're second-class people and sending them back home. Guess how favourably disposed they are to the USA after that...
No it isn't. NaCl is a great proof of concept. It shows that you can sandbox x86 apps using some static analysis of the binaries and a few other constraints (it also showed that segmentation support on modern x86 chips is pretty poor and terrible on Atom). The problem is that it only works on x86 binaries. What proportion of Web use these days is (ARM-based) phones and tablets? 20%? If you make something that only works for 80% (and falling) of your customers, then that's a problem.
PNaCl is promising, but it's currently in early draft stage. It hard-codes some things into the ABI too early and misses other important things (e.g. no mechanism for exceptions, and they're very difficult to implement correctly in a PNaCl model). And, unlike NaCl, PNaCl relies on a complex compiler being bug free for security, and we all know how well that worked out for Java...
Slashdot Mirror
When you show the grandparent's comment to your overlords, please take the bolded section and print it in 72-point text. The new design is quite possibly the worst thing I've ever seen anyone try to do to Slashdot and it will kill the site. I'm not as active as I was a year ago due to lack of time, but in your statistics last year I was one of the 5 most active commenters for one or two quarters. I would not visit Slashdot if they switched to this abomination.
I never understood this either. For example, they ripped all of the locale support out of libc to remove bloat. Which then means that everyone who uses the NDK reimplements it in another library, which is not shared...
Some bits of the FreeBSD kernel make it into the BSD server in the XNU kernel. One of the big ones is the MAC framework (SEBSD), which is shared between FreeBSD and XNU and supports pluggable access control policies. This is used to implement the code signing logic on Juniper routers and the application sandboxing on iOS and OS X. There are some pretty big differences to the VM subsystems on both (they're both derived from Mach 2.5, but they've diverged hugely since then).
You might want to scan the Android stack for FreeBSD copyrights sometime. Most of the Android libc is a slimmed-down version of the FreeBSD one, and there are quite a few other bits of FreeBSD code in there too. In terms of lines of code, I think there is about as much FreeBSD code in Android as there is Linux code.
I'm not sure what your point is. In the House, all seats were up for reelection and the Republicans won on seats but did not get the majority of the popular vote, so this means that they represent the will of the people? In the Senate, not all seats were up for reelection, but in the 1/3 that were, the Republicans lost 2 seats overall and lost the popular vote, so this means that they represent the will of the people? In the Presidential election, there is only one seat and it was up for reelection, the Republicans lost both the popular vote and failed to win the seat, so this means that they represent the will of the people?
You can write off the Senate elections as only giving you the views of 1/3 of the country, but you're still left with two national elections where more people voted against the Republicans than voted for them (I'm not sure what the statistics are for third parties, so this may be true for Democrats too) and where more people voted for the Democrats than voted for the Republicans. And yet you still claim that the Republican majority in the House represents the will of the people?
So yes, by the rules of the system, which are arguably superior to the hypothetical rules you seem to wish we operated under, the House Majority Republicans *are* there due to the will of the people, and *do* represent it.
If by 'the people' you mean the commission that defines constituency boundaries, then I suppose you're right. If you mean 'the people who voted' then you are wrong. If you mean 'the people who are eligible to vote' then neither party can claim to be even approximately representing the will of the people.
The popular vote count in both elections is definitely relevant if you're claiming (as the grandparent was) that a narrow victory represents 'the will of the people'. Getting 49.6% of the popular vote does not mean that you represent the will of the people, it means that you represent the will of about half of the people. Having a majority with this amount of the vote means that constituency boundaries favoured you this time. It is most definitely not a mandate to push through your agenda without compromising, as the grandparent appeared to think.
You're an idiot, you know that? The House of Representatives represents the will of the people. The people want to get rid of Obamacare.
You realise that in the last election:
Or, to put it another way, in 2012:
So, in terms of popular vote, the Democrats go a (very slim) majority in two of the three elections, the Republicans didn't manage it in one. In the election where the Republicans did best (the Presidential race, 47.2%), they didn't get as much of the vote as the Democrats did in the election where they did the worst (House, 48.3%). Neither party got anywhere close to the percentage of the vote that enables someone who doesn't lie for a living to claim to have a mandate from the people.
Now, I realise the state of mathematical education is pretty poor in the USA, but being able to tell which of two numbers is bigger than the other is surely something that is covered.
That depends on who else is sitting on the limb, and what is underneath...
That's not usually the case. If you keep the fact that you're gay secret from your wife, then there's something that you can be blackmailed over. That makes you a security risk. The security clearance process that I've been through (in the UK) didn't care if you were gay or had a string of mistresses, as long as these weren't things that you wanted to keep secret.
There's a big difference between using technology to aid in parenting and teaching and using technology as a substitute for parenting and teaching. This was true of television, of home computers, and is now true of tablets and smartphones.
The study was published a year or two ago, so I'm not sure why it's appearing on Slashdot now - normally I'd blame the 'editors' for this kind of late submission, but apparently they found a 'news' site with a slow news day that decided that old research papers now count as news. It's a pretty neat concept though and more recent work has improved the accuracy a lot by feeding other sensors into the mix.
There are funded projects and there are resources. Some resources are allocated to a specific project and some are not. Typically, the ones allocated to a project can be used for other things if that project doesn't need them, but it gets priority. There's usually a fair amount of unused infrastructure that can be used for unofficial projects if no one else needs it.
In a well-run research lab (of which Bell Labs and Xerox PARC in this era are archetypes), there is no strict accounting of time to projects. People are expected to work on some things as part of big ongoing projects, but they have a lot of free time to devote to other things that they consider fun and interesting. This is done because the people running the lab know that these spare-time projects are how you get the seeds of the next iteration of big projects. The same is true in most research labs, including most universities, which is why I suspect that you've never been to university, as you'd have encountered this concept before.
UNIX was such a project. It was not part of any funded project at Bell Labs and was done by a few guys for fun. It then grew and was used in some funded projects (troff was the product of one such project), but wasn't officially backed by AT&T until long after it was created.
There is a distinction between funding a project and paying an individual's salary. The small group that worked on UNIX (well, UNICS back then, in one of Peter's characteristic puns - it was renamed UNIX later when it had multi-user support) had their salaries paid by Bell Labs, but they were being paid to work on other things. Their work on UNIX was not backed by management and was not funded. They had no resources allocated to the UNIX project, they used whatever they could scrounge. I've done similar things with machines a few generations old to build infrastructure for fun projects and occasionally these go on to become funded projects.
MINIX 3 is a pretty conventional microkernel. Take a look, for example, at some of the virtual memory stuff done in L4/HURD. This can be seen as a continuation of the external pager support in Mach, but with a much simpler and cleaner design and a lot of direct application to current workloads.
Wow, really high standard of debate you have there. You appear to have no idea of how things work in research labs (industrial or academic). Did you attend university?
Yup. And he's right...
I'm neither cold fjord nor the AC, but you are entirely wrong. The current project that I work on is headed by Peter Neumann (one of the architects of MULTICS), and so I have a fairly good understanding of the environment that led to the creation of UNIX from conversations with someone who was there at the time. Pretty much everything that cold fjord said is correct. They got an old computer that no one else wanted and hacked up an OS in their own spare time. They didn't aim to take over the world - they just wanted something that had the bits of MULTICS that they liked (Peter has some frank opinions on their choices here) that they could use.
And they've been using it to explore some quite interesting ideas in kernel design. The fine-grained compartmentalism that a microkernel provides (at the expense of some performance) is starting to look more attractive in a world where computers run in very hostile environments and yet even a 50% slower kernel would have a negligible impact on user-perceived performance (or battery life).
When I got to university, I discovered how much of the theoretical side I was missing. The main problem with teaching programming at an early age is that it really needs to be accompanied by teaching logic and then game and graph theory. I've seen classes that do this well for under-10s, but they're very rare.
[1] The Dijkstra comment that teaching BASIC should be a criminal offence doesn't really apply to BBC BASIC, which had full support for structured programming, an integrated assembler, and direct access to memory-mapped hardware.
[2] Back then, you really needed makefiles because there was no equivalent to a modern compiler driver. Compilation, assembly, and linking were all separate, manual, steps.
The problem is that there's no well-supported way, for example in the DNS, for a server to say which certificate it will use. If HTTPS required two certificates, then that would just mean that you'd need to compromise two CAs (or one CA and get them to issue two certs), which given what we already know about the NSA program, has already happened. This is something that people like Ben Laurie at Google are working on with Certificate Transparency: trying to ensure that there is a recorded and verifiable chain showing that a certificate was issued to the real owner of the domain and that it is not being MITM'd.
The point of using control-alt-delete is that it's a key combination that can not be caught by any userspace process that does not have a special permission. This means that it's impossible to spoof the login screen on Windows without already having compromised the kernel. It doesn't matter what the key combination is, as long as it's one that is not delivered by the normal keypress event delivery mechanisms. Control-alt-delete is a reasonable choice, because no application author is likely to complain that they can't use this shortcut combination.
Or
That's right, the difference is syntax and you've picked a terrible example. A better example would be method invocation. The differences between calling a C function pointer, a C++ virtual function, a Java method or an Objective-C method are significant.
That makes sense. I get a couple of 'call for papers' emails a day from dubious journals, often with such broad titles as 'The Journal of Modern Research', so it would be completely impossible for them to rate articles. The research establishment in the UK has tried quite hard over the last decade to counter this 'publish loads of crap' incentive. The old Research Assessment Exercise and the new Research Excellence Framework by which departments are assessed requires a small number (about one per year) of 'research outputs'. These can be high-impact papers, books, and so on, and in computer science can include things like published open source software (which counts as technology transfer if you can point to people using it).
And it's a completely fucked-up policy, because the hoops that a US company needs to jump through to hire an Iranian national are insane. So you end up educating a load of people, then telling them that they're second-class people and sending them back home. Guess how favourably disposed they are to the USA after that...
No it isn't. NaCl is a great proof of concept. It shows that you can sandbox x86 apps using some static analysis of the binaries and a few other constraints (it also showed that segmentation support on modern x86 chips is pretty poor and terrible on Atom). The problem is that it only works on x86 binaries. What proportion of Web use these days is (ARM-based) phones and tablets? 20%? If you make something that only works for 80% (and falling) of your customers, then that's a problem.
PNaCl is promising, but it's currently in early draft stage. It hard-codes some things into the ABI too early and misses other important things (e.g. no mechanism for exceptions, and they're very difficult to implement correctly in a PNaCl model). And, unlike NaCl, PNaCl relies on a complex compiler being bug free for security, and we all know how well that worked out for Java...