I know of people that DONT have a TV and DONT watch BBC but because he owns a laptop he is forced to pay the "tax".
Only if he watches live streams. I don't own a TV and don't watch live streaming, therefore I don't need a license. I can still watch the iPlayer on-demand stuff (I could watch things like 4 on demand, but it's full of ads, so I don't).
Well when you have a massive debt, everyone has to give up something.. and that includes (unfortunately) scientists
Science is not a cost, it's an investment. You don't fix the economy by stopping spending money on things that will give a return.
Maybe those researching "blue skies" projects that have gone no where should be cut.
The departments where people only do research that is guaranteed to work are usually the weaker ones. Good research addresses problems where the solution isn't known, where there are only some approximate ideas about what it may be, and where failure is likely. A big problem in academia today is exactly the attitude in your post - that people who do research that may fail should be penalised.
They have security vulnerabilities fairly regularly. Ones that are remotely exploitable are rarer, but the cited one from the grandparent was first known in 2004, not fixed until 2006, and allowed someone to anyone who could make you display an image (e.g. in a web page) run arbitrary code in your kernel. It gets cited a lot because it's a perfect case study in stunningly incompetent security.
Jazelle is not publicly documented. The technology basically reserves a few registers for VM state and the rest for the stack of the current program. You use a special branch-to-Jazelle-mode instruction and then it starts executing Java bytecodes, trapping into the emulator for complex ones. There are two problems with this. The first is that the lack of documentation and the requirement to pay a patent license fee if you do use it even with the (expensive) documentation means that there is no open source implementation. The second is that it executes JVM bytecodes, not Dalvik ones.
The myth that more the eyeball, more successful the ad is just that, a myth. Pushed ad nauseum by mindless bots like you.
It's not totally untrue. Brand awareness does help sales. If people are looking for a product of a certain type, they are more likely to buy one from a company that they have heard of than one that they haven't.
In the case of Apple, however, it's a pointless goal. Everyone who hasn't been living under a rock for the past decade has heard of Apple. They've also probably heard of Nokia, Samsung, and Dell. An advert for a company in this situation needs to say more than 'we exist and are in this market!' it needs to say 'our products are good for these reasons.'
Shoot spectators? There are loads of them, so if you lock them all in you could probably draw it out for the duration of the event...
Note for British police, politicians, and other idiots: The above is a (weak) joke, and I am not seriously advocating shooting olympic attendees. Removing the lead before making them into soylent green adds too much to the processing costs.
I didn't understand this in the marketing material. Time Machine has allowed you to back up to different devices for ages (since its introduction?). What is the new bit?
When developing proprietary software for a client, I've often wished some software had BSD style license, so I could just take the code and use it in a proprietary solution
And if you did, and you found a bug, would you report it? Would you provide a useful test case? Would you provide a patch? If you needed to add a new feature, would you push it upstream?
If you answered yes to any of the above questions, then you'd have made a useful contribution to the BSD ecosystem while writing proprietary software. On contrast, if you implement a proprietary replacement for your client then you've contributed nothing to the Free Software world. And that's the real advantage of the BSDL - that it doesn't push people away from Free Software, who might otherwise become contributors.
There is BSD code in use by Microsoft and Apple that has been extended, closed and made unavailable to the community
There is. There is also a huge amount of BSD licensed code that Apple has released publicly. A lot in LLVM, and some in FreeBSD, both in the kernel and libc. Most Apple-originated code these days is Apache 2.0 licensed, but they've been willing in the past to relicense code that we (FreeBSD) want to adopt. In short, we've benefitted from their use of our code. We've also seen fairly significant contributions recently from companies like Juniper, who traditionally maintained their own fork of FreeBSD. They've learned that it's cheaper to keep their changes as small as possible: people buy hardware from them, not an OS.
That sort of makes BSD code long-term unsustainable.
Quite the reverse, it provides a transition mechanism, which is something that RMS quite often ignores the need for. You can't reach a world where all software is free without going via a world where some is free, then most is free. BSD-style licenses make it easy for software companies to incrementally transition from selling a product to selling development services. The GPL is an all-or-nothing approach. We've seen a lot of companies when faced with the GPL as the only existing option decide to start from scratch and write a proprietary replacement. In contrast, if a BSDL version exists, they'll usually take it, improve it, and push changes upstream. Upstreaming their changes gives two important benefits. The first is that it lets other people fix bugs in their code (and even do code review). The second is that it means someone else won't implement the feature they need in a different - and incompatible - way and make it very hard for them to pull in upstream changes. And, importantly, doing this with BSDL code doesn't have any implications for the rest of their stack. Once they've seen the cost savings for doing this with externally originating code, some companies then start to do it with the rest. Most software is written in-house by companies that need to use it, not by software companies, and so being able to outsource some of their development effort (often to their competitors) is a good strategy. With a BSDL approach, they can do it in small steps and see if it really does make sense for them.
But what about the opposite approach: don't even try to enter those markets. Why should a company always try to become bigger even in areas that are not its strength?
Because the world changes. Remember SGI? They completely owned the graphics workstation market. Then a few of their engineers realised that they could make something that wasn't as good as their workstation products, but could be sold cheaply enough to be a mass-market product. They took it to management, who said that they didn't want to enter a new market and that they were happy with their place in their existing one. So the engineers left and founded nVidia. Now, if you want to buy a 3D workstation, you typically by a commodity computer and a high-end nVidia GPU. Within a few years, nVidia was outspending SGI on R&D, because although their per-unit profits were much lower their volume was orders of magnitude higher.
When I was a student I worked at a not-very-reputable computer shop, which did the replacement trick in the original post quite often. At the time, it was more common for drives to fail because of the control circuitry than because of the spinning rust bit. Replacing the controller board worked in a lot of cases.
With hard driver repair, it's often the case that it can either be repaired easily, or it can't be repaired at all without a significant investment in time and equipment. If the platter is damaged, you're basically screwed: if you've got the right setup then you can read the undamaged bits after disassembling the drive in a clean room, but you probably won't get much useful data. If the head or motor has died, but not damaged the disk, then a cleanroom disassembly and remounting of the platters will work, but it requires specialised equipment and is likely to be more expensive than the value of the data (you have backups, right?). If it's a failure on the controller board, then just swapping them over is easy, and a repair shop will probably have replacement boards in stock and can do it for you easily.
However, in terms of the thought experiment... we know that they would agree to it, since we've had ARM devices produced by small firms with locked down OS configurations
Not the same thing. Manufacturers have locked down the OS because they want to, although that's rare - a lot of Chinese-manufactured ARM devices triple boot WinCE, Android and Ubuntu. Much more commonly, they've done so because their customers (i.e. the mobile networks) have required it. Doing something because your customer requires it is very different from doing it because your supplier required it. The former makes a sale for you, the latter may lose sales for you.
And, increasingly, these are replaced by conference calls. If the meeting is quick enough that the time saving would be worthwhile, then it's typically better to use telecommunication. If it's a meeting where face to face time is more important than speed, then typically saving a couple of hours in travelling for a multi-day visit isn't a noticeable benefit, and the number of people for whom it is is not large enough to cover the cost of maintaining the aircraft.
The problem is that flight time is often a fairly small component of long trips. For me to get to the USA from the UK typically involves a couple of hours of getting to the airport, two hours of check in, and close to an hour of getting past immigration and picking up luggage at the far end, followed by however long it takes to get to my final destination (which may be via a connecting flight). It's basically a day of travelling. Even if you could do the transatlantic bit in an hour, it would still be most of a day. This is a big part of the reason why the Eurostar is a huge win for travelling from the UK to Belgium or France: shorter check-in times, and your luggage goes on with you so no need to wait for it to be loaded or unloaded, so even though the train is significantly slower you can be almost at your destination before someone taking a plane would even be on the runway.
WW2 I might agree with, but WW1? What exactly was the cause there? The soldiers on both sides had far more in common with each other than with their leaders. I suppose that reducing excess population in the lower classes counts as a cause, but it's not one that I'd encourage people to be willing to die for...
One of the common examples of anticompetitive behaviour that antitrust cases cover is using a monopoly in one area to gain one in another. The fact that Microsoft doesn't have a monopoly in ARM operating systems does not help them defend against this sort of complaint, because the argument would be that they are using their monopoly on x86 operating systems to gain one on ARM operating systems.
To test whether this is a valid complaint, the court would propose the thought experiment: imagine that Microsoft is a new startup producing their first OS and makes the same requirement. Would manufacturers agree to it? If not, then Microsoft has violated antitrust regulations. I would imagine not, because being able to install a different OS after market would minimise buyers' risk (see any discussion of HP TouchPads and count the people saying 'well, if I don't like WebOS I can always install Android...').
This is something DirectX gets for free from COM. When you request a handle to DirectX, it's via a versioned COM interface (basically, an array of function pointers). You don't need an explicit compatibility tab, because each new version of DirectX comes with a new set of COM interfaces, but you always request the one you were compiled with. That doesn't protect you from bugs, from dependence on driver issues (one good one is that some early drivers happened to enumerate texture formats in a particular order and a load of games depend on this, so break with newer drivers), and so on.
The credit default swaps and some other financial instruments are moderately hard to explain
Stephanie Flanders and Robert Preston have both done a good job in their BBC blogs. And they usually structure their articles with some definitions and background at the top, then the news part after the first subheading. If you're familiar with the background (i.e. you read their last few articles on related subjects) then you can skip the first few paragraphs. I usually now start in the middle, and if I found something I didn't understand go back to the top and read the bit where they explain it. It's a shame that they don't make more use of hypertext: it's pretty trivial to just link jargon terms to a definition on first use...
And define it! People read news to become more educated. BBC articles quite often have a little sidebar describing what the terms mean (often in an incredibly patronising way, but you can't have everything). If you're familiar with the subject, then you get to read a more information-dense article. If you're not, then the sidebar teaches you the basic terminology, which helps you the next time you try to read something technical in the same field.
If you want to be pulled over cycling in Cambridge, go through a red light by Parker's Piece. There's a police station right next to it and there's often someone just leaving or arriving who is very happy to have a little chat with dangerous cyclists.
27" seems to be close to the sweet spot in terms of physical size. I put mine above my laptop, so I can move things onto it when I'm working on them and leave my laptop screen cluttered. I would really love to have a 300dpi 27" monitor though...
No, but that's an idiotic comparison. I don't put things on the other side of my window that need to be in my locus of attention, so there is no chance that not being able to keep everything outside in my field of will result in lower productivity.
There's no such thing as 'enough', let alone 'too much' screen space.
I disagree. The optimal amount of screen space just fills my field of vision. Add any more, and spacial memory starts to work less well. I spent a while working with dual 30" monitors, and it was really easy to lose windows and have to spend some time finding them. A 27" display seems about right, maybe slightly small, but it's possible to put it where you don't need to turn your head to see everything.
That depends on who you are. As an individual, sure. As a company with 10,000 employees, things look quite different. Companies like Red Hat or iX Systems will happily let you pay for a fork to be maintained on your behalf.
Slashdot Mirror
I know of people that DONT have a TV and DONT watch BBC but because he owns a laptop he is forced to pay the "tax".
Only if he watches live streams. I don't own a TV and don't watch live streaming, therefore I don't need a license. I can still watch the iPlayer on-demand stuff (I could watch things like 4 on demand, but it's full of ads, so I don't).
Well when you have a massive debt, everyone has to give up something.. and that includes (unfortunately) scientists
Science is not a cost, it's an investment. You don't fix the economy by stopping spending money on things that will give a return.
Maybe those researching "blue skies" projects that have gone no where should be cut.
The departments where people only do research that is guaranteed to work are usually the weaker ones. Good research addresses problems where the solution isn't known, where there are only some approximate ideas about what it may be, and where failure is likely. A big problem in academia today is exactly the attitude in your post - that people who do research that may fail should be penalised.
They have security vulnerabilities fairly regularly. Ones that are remotely exploitable are rarer, but the cited one from the grandparent was first known in 2004, not fixed until 2006, and allowed someone to anyone who could make you display an image (e.g. in a web page) run arbitrary code in your kernel. It gets cited a lot because it's a perfect case study in stunningly incompetent security.
Jazelle is not publicly documented. The technology basically reserves a few registers for VM state and the rest for the stack of the current program. You use a special branch-to-Jazelle-mode instruction and then it starts executing Java bytecodes, trapping into the emulator for complex ones. There are two problems with this. The first is that the lack of documentation and the requirement to pay a patent license fee if you do use it even with the (expensive) documentation means that there is no open source implementation. The second is that it executes JVM bytecodes, not Dalvik ones.
The myth that more the eyeball, more successful the ad is just that, a myth. Pushed ad nauseum by mindless bots like you.
It's not totally untrue. Brand awareness does help sales. If people are looking for a product of a certain type, they are more likely to buy one from a company that they have heard of than one that they haven't.
In the case of Apple, however, it's a pointless goal. Everyone who hasn't been living under a rock for the past decade has heard of Apple. They've also probably heard of Nokia, Samsung, and Dell. An advert for a company in this situation needs to say more than 'we exist and are in this market!' it needs to say 'our products are good for these reasons.'
Shoot spectators? There are loads of them, so if you lock them all in you could probably draw it out for the duration of the event...
Note for British police, politicians, and other idiots: The above is a (weak) joke, and I am not seriously advocating shooting olympic attendees. Removing the lead before making them into soylent green adds too much to the processing costs.
I didn't understand this in the marketing material. Time Machine has allowed you to back up to different devices for ages (since its introduction?). What is the new bit?
When developing proprietary software for a client, I've often wished some software had BSD style license, so I could just take the code and use it in a proprietary solution
And if you did, and you found a bug, would you report it? Would you provide a useful test case? Would you provide a patch? If you needed to add a new feature, would you push it upstream?
If you answered yes to any of the above questions, then you'd have made a useful contribution to the BSD ecosystem while writing proprietary software. On contrast, if you implement a proprietary replacement for your client then you've contributed nothing to the Free Software world. And that's the real advantage of the BSDL - that it doesn't push people away from Free Software, who might otherwise become contributors.
There is BSD code in use by Microsoft and Apple that has been extended, closed and made unavailable to the community
There is. There is also a huge amount of BSD licensed code that Apple has released publicly. A lot in LLVM, and some in FreeBSD, both in the kernel and libc. Most Apple-originated code these days is Apache 2.0 licensed, but they've been willing in the past to relicense code that we (FreeBSD) want to adopt. In short, we've benefitted from their use of our code. We've also seen fairly significant contributions recently from companies like Juniper, who traditionally maintained their own fork of FreeBSD. They've learned that it's cheaper to keep their changes as small as possible: people buy hardware from them, not an OS.
That sort of makes BSD code long-term unsustainable.
Quite the reverse, it provides a transition mechanism, which is something that RMS quite often ignores the need for. You can't reach a world where all software is free without going via a world where some is free, then most is free. BSD-style licenses make it easy for software companies to incrementally transition from selling a product to selling development services. The GPL is an all-or-nothing approach. We've seen a lot of companies when faced with the GPL as the only existing option decide to start from scratch and write a proprietary replacement. In contrast, if a BSDL version exists, they'll usually take it, improve it, and push changes upstream. Upstreaming their changes gives two important benefits. The first is that it lets other people fix bugs in their code (and even do code review). The second is that it means someone else won't implement the feature they need in a different - and incompatible - way and make it very hard for them to pull in upstream changes. And, importantly, doing this with BSDL code doesn't have any implications for the rest of their stack. Once they've seen the cost savings for doing this with externally originating code, some companies then start to do it with the rest. Most software is written in-house by companies that need to use it, not by software companies, and so being able to outsource some of their development effort (often to their competitors) is a good strategy. With a BSDL approach, they can do it in small steps and see if it really does make sense for them.
But what about the opposite approach: don't even try to enter those markets. Why should a company always try to become bigger even in areas that are not its strength?
Because the world changes. Remember SGI? They completely owned the graphics workstation market. Then a few of their engineers realised that they could make something that wasn't as good as their workstation products, but could be sold cheaply enough to be a mass-market product. They took it to management, who said that they didn't want to enter a new market and that they were happy with their place in their existing one. So the engineers left and founded nVidia. Now, if you want to buy a 3D workstation, you typically by a commodity computer and a high-end nVidia GPU. Within a few years, nVidia was outspending SGI on R&D, because although their per-unit profits were much lower their volume was orders of magnitude higher.
When I was a student I worked at a not-very-reputable computer shop, which did the replacement trick in the original post quite often. At the time, it was more common for drives to fail because of the control circuitry than because of the spinning rust bit. Replacing the controller board worked in a lot of cases.
With hard driver repair, it's often the case that it can either be repaired easily, or it can't be repaired at all without a significant investment in time and equipment. If the platter is damaged, you're basically screwed: if you've got the right setup then you can read the undamaged bits after disassembling the drive in a clean room, but you probably won't get much useful data. If the head or motor has died, but not damaged the disk, then a cleanroom disassembly and remounting of the platters will work, but it requires specialised equipment and is likely to be more expensive than the value of the data (you have backups, right?). If it's a failure on the controller board, then just swapping them over is easy, and a repair shop will probably have replacement boards in stock and can do it for you easily.
However, in terms of the thought experiment ... we know that they would agree to it, since we've had ARM devices produced by small firms with locked down OS configurations
Not the same thing. Manufacturers have locked down the OS because they want to, although that's rare - a lot of Chinese-manufactured ARM devices triple boot WinCE, Android and Ubuntu. Much more commonly, they've done so because their customers (i.e. the mobile networks) have required it. Doing something because your customer requires it is very different from doing it because your supplier required it. The former makes a sale for you, the latter may lose sales for you.
And, increasingly, these are replaced by conference calls. If the meeting is quick enough that the time saving would be worthwhile, then it's typically better to use telecommunication. If it's a meeting where face to face time is more important than speed, then typically saving a couple of hours in travelling for a multi-day visit isn't a noticeable benefit, and the number of people for whom it is is not large enough to cover the cost of maintaining the aircraft.
The problem is that flight time is often a fairly small component of long trips. For me to get to the USA from the UK typically involves a couple of hours of getting to the airport, two hours of check in, and close to an hour of getting past immigration and picking up luggage at the far end, followed by however long it takes to get to my final destination (which may be via a connecting flight). It's basically a day of travelling. Even if you could do the transatlantic bit in an hour, it would still be most of a day. This is a big part of the reason why the Eurostar is a huge win for travelling from the UK to Belgium or France: shorter check-in times, and your luggage goes on with you so no need to wait for it to be loaded or unloaded, so even though the train is significantly slower you can be almost at your destination before someone taking a plane would even be on the runway.
WW2 I might agree with, but WW1? What exactly was the cause there? The soldiers on both sides had far more in common with each other than with their leaders. I suppose that reducing excess population in the lower classes counts as a cause, but it's not one that I'd encourage people to be willing to die for...
One of the common examples of anticompetitive behaviour that antitrust cases cover is using a monopoly in one area to gain one in another. The fact that Microsoft doesn't have a monopoly in ARM operating systems does not help them defend against this sort of complaint, because the argument would be that they are using their monopoly on x86 operating systems to gain one on ARM operating systems.
To test whether this is a valid complaint, the court would propose the thought experiment: imagine that Microsoft is a new startup producing their first OS and makes the same requirement. Would manufacturers agree to it? If not, then Microsoft has violated antitrust regulations. I would imagine not, because being able to install a different OS after market would minimise buyers' risk (see any discussion of HP TouchPads and count the people saying 'well, if I don't like WebOS I can always install Android...').
This is something DirectX gets for free from COM. When you request a handle to DirectX, it's via a versioned COM interface (basically, an array of function pointers). You don't need an explicit compatibility tab, because each new version of DirectX comes with a new set of COM interfaces, but you always request the one you were compiled with. That doesn't protect you from bugs, from dependence on driver issues (one good one is that some early drivers happened to enumerate texture formats in a particular order and a load of games depend on this, so break with newer drivers), and so on.
The credit default swaps and some other financial instruments are moderately hard to explain
Stephanie Flanders and Robert Preston have both done a good job in their BBC blogs. And they usually structure their articles with some definitions and background at the top, then the news part after the first subheading. If you're familiar with the background (i.e. you read their last few articles on related subjects) then you can skip the first few paragraphs. I usually now start in the middle, and if I found something I didn't understand go back to the top and read the bit where they explain it. It's a shame that they don't make more use of hypertext: it's pretty trivial to just link jargon terms to a definition on first use...
And define it! People read news to become more educated. BBC articles quite often have a little sidebar describing what the terms mean (often in an incredibly patronising way, but you can't have everything). If you're familiar with the subject, then you get to read a more information-dense article. If you're not, then the sidebar teaches you the basic terminology, which helps you the next time you try to read something technical in the same field.
If you want to be pulled over cycling in Cambridge, go through a red light by Parker's Piece. There's a police station right next to it and there's often someone just leaving or arriving who is very happy to have a little chat with dangerous cyclists.
and yet can do nothing about the fact that everyone's movements across and through London are routinely tracked
It is generally accepted even my the most liberal democracies that people - even visitors - lose some freedoms when in prison.
27" seems to be close to the sweet spot in terms of physical size. I put mine above my laptop, so I can move things onto it when I'm working on them and leave my laptop screen cluttered. I would really love to have a 300dpi 27" monitor though...
No, but that's an idiotic comparison. I don't put things on the other side of my window that need to be in my locus of attention, so there is no chance that not being able to keep everything outside in my field of will result in lower productivity.
There's no such thing as 'enough', let alone 'too much' screen space.
I disagree. The optimal amount of screen space just fills my field of vision. Add any more, and spacial memory starts to work less well. I spent a while working with dual 30" monitors, and it was really easy to lose windows and have to spend some time finding them. A 27" display seems about right, maybe slightly small, but it's possible to put it where you don't need to turn your head to see everything.
That depends on who you are. As an individual, sure. As a company with 10,000 employees, things look quite different. Companies like Red Hat or iX Systems will happily let you pay for a fork to be maintained on your behalf.