2) I'm really glad I live there, and wish everyone had the opportunity to live like that
The problem is that the people who do live like that only do so by other people not living like that. There is a reason why so much manufacturing and mining is done in developing countries with little or no restrictions, and it isn't just the labor cost.
3) Environmental destruction has not actually happened here because of environmental regulations, which means it is an abstract concept. I wish it was for the rest of the world to
As above, if those regulations were in effect globally, *fewer* people would live like that and it would cost disproportionately more.
1) And why is land scarce? Oh, because dictators own it all. Then you realise that everyone could fit in the southern USA at Los Angles pop densities. And yes, I do realise that they will need more land to provide all the stuff they are using
I do agree that hoarding by violent means is the bigger problem than any artificial scarcity and I have no inkling of how land *really* looks, but on the other extreme of 'LA density everywhere' is a silly concept, considering that LA would not be possible if not for food and other products being brought into the city. If LA were given an impenetrable dome suddenly, a great many would starve when the food stores went out for lack of farmland to cultivate new food.
3) Environmental destruction may be permanent in some instances. That's why we need to help developing countries move beyond that phase of construction
It's naive to presume that doing it safe and clean at the similar (really, you are hoping for lower) costs.
4) Because they can't afford to.
Because the price of restoring it when technically possible is higher than any value extracted. If you are going to pay the price to put things back the way they were, then you are committing to put more resources into restoration than you will extract in ruining it.
I guarantee you that most of the products you consume are imported from places that do significant harm to their environment. Our amenities come by exploiting willingness of other places to fubar their environment. The cost and sometimes simply the forbidden nature of extracting resources from the earth in a developed country combined with similar effects on processing those resources make developing countries the place to do it. If every square foot on the planet were as restricted as the 'developed world', then all the toys and necessities would be much more rare and much more expensive.
I'm not an uber-envorinmentalist or anything, but I have to acknowledge that my country's domestic rules and regulations are nice for our land, but bordering on hypocritical with the huge imports of goods made possible by other places not having such restrictions.
Frankly, there is no valid reason for starting a new program in C in this day and age.
One reason I can think of is incurring the libstdc++ dependency in an embedded or otherwise size critical context. Sometimes you still have precious little OS storage, and your convenience/cost compromises might land shy of C++. This is a matter of design tradeoff, some contexts demand nothing more than assembler, or a libc smaller than glibc, and some 'embedded' can do nearly whatever the hell it wants.
Lower cost of ownership - BS, too much time is spent hacking up config files to make crap work or work right
This perception problem is common with perl. You can run a modern linux system fairly naively, but the enthusiasts will make the 'fancy tricks' look like 'the' way of life. There are advanced capabilities that are possible with fancy trickery, but most people would do fine if they ignored those capabilities and just used their system. Same applies to a lot of systems with advanced capabilities, for a simple example look up source engine command parameters for adjusting multiplayer settings. If you naively use those games, you'll likely have no problem, but if you search for cl_updaterate and start reading, you would be given the impression that there are tons and tons of parameters that simply *must* be tweaked for an 'acceptable' experience.
CLI/scripting system that actually works - BS, anything you can write and make work in Linux, I can in Windows
There is what is technically in theory possible, and then there is how to do it. If you compare vbscript/cmdscript to bash/(python|perl|etc), the syntax is dramatically better. Now Powershell brings a fair amount of better syntax and has at least decent interactive startup time in Windows 7/28kr2 (before that it was a mess), but in WindowsPE, you are still stuck with cmd/vbscript as your tools. In my opinion, Powershell leans a little more toward perl/python in some ways that make things like bash a touch easier for typical shell scripting.
Most open source software runs on it - Show me anything worthwhile that doesn't run in Windows or have a better alternative there
Compiz. Window scaling with title search is incredibly awesome in terms of practical application whereas Aero is strictly just eye-candy. Most applications will run on either platform, but on a typical linux it's 'aptitude install openoffice' rather than trying to find individual websites to host the software.
Drivers for just about any piece of hardware ever built - BS, that's the primary thing most users have issues with, half baked drivers
All platforms are subject to crappy drivers from crappy vendors. I had to search all over the damn place about networking on a new Windows 7 going out every couple ours to find out that the manufacturer's driver would go tits up due to segmentation offload (turning it off involved navigating several dialogs graphically, in linux it would just be 'ethtool -K eth0 tso off'). Pick a reputable hardware vendor and you'll get a good driver experience regardless of platform. Pick a sketchy vendor, and your experience is likely to suck regardless of OS.
I suspect the plan price was enough to kill it off, though I have some shred of hope that humanity saw the ads that did nothing more than say "OMG you can facebook" and collectively shrugged. Of course, the devices looked bad and the interface in the ads looked even worse.
In terms of Verizon worshipping, the full story sounds more like some MS guy liked Sidekicks (which if they were thinking clearly would be recognized as a stopgap rather than an endgame). Then they acquired danger and decided to do it their own way while keeping sidekick barely alive. On top of targeting a rapidly evaporating market as smartphones came down, they made it horribly late by insisting upon essentially starting from scratch (reminds me of the hotmail acquisition and initial attempt to move to winnt servers). By the time they were ready to move, smartphones and smartphone plans had come down so much, the Kin was laughable.
Essentially, it's the story of MS' life in the mobile market. They don't get the concepts until someone else has already gone. I don't know if/when MS will finally give up. The handset vendors have everything to gain from a platform like Android and nothing really to lose. I personally like WebOS, but I don't think there is room for more than one Apple-like entity in the market, and Android is really out-MSing MS in the space. MS made share gains by all the OEM support, and HTC, Samsung, et al are more in charge of their own destinies than MS ever would allow.
You made a mistake.. To avoid confusion, you designated a *specific* one of the children as being the child described as being a boy. this achieves the same effect as prescribing a specific ordering. It ignores the probability that the unnamed 'Boy' is the child you called Peter, which would make the probability that the language describes the first and third set twice as high as second or fourth (those cases have twice as many 'Peter' candidates, and you ruled one of those candidates out arbitrarily). This is really hard to think through, and we intuitively want to make the boy/boy orderings unambiguous by picking which of the boys is *the* boy, however, that assumes data not given in the original statement. In a random distribution, 75% of the permutations can be described as "one is a boy", and 2/3 of that set is mixed gender. however, only 50% can be described as "the first one is a boy", bringing it back to 50%.
One way of doing that would be to split the above list in half:
bb, bg | gb gg
and ask "At least one of my children is a boy..." in the first case and "At least one of my children is a girl..." in the second case.
I don't see how you think you've changed the problem. If you are grouping 'bg|gb' as one group on par with bb and gg, first that's 3/4ths of the group count and the middle group is twice the size of the other groups. If you are saying that bg, bg is one group, and gb gg is another group, then "at least one of my children is a boy' would describe the entire first group and half of the second, making the grouping meaningless. If you are saying the goal is to pick which half of the | describes it, then using the unordered information provided, the probability of guessing the correct side can be 2/3rds.
You haven't changed the problem in any way whatsoever..
The difference is the 'naive' code you write to do things in the simplest manner *can* run on a CPU. For the GPU languages, you *must* make those optimizations. This is not to undercut the value of GPU (as Intel concedes, the gap is large), but it does serve to counteract the dramatic numbers tauted by nVidia.
nVidia compared expert tuned and optimized performance metrics on their product and compared against stock, generic benchmarks on intel products.
Magny-Cours is currently showing significant performance advantage over Intel's offerings while at the same time AMD's Evergreen *mostly* shows performance advantages over nVidia's Fermi despite making it to market ahead of Fermi.
AMD is currently providing the best tech on the market This will likely change, but at the moment, things look good for them.
Intel and nVidia are stuck in the mode of realistically needing one another and simultaneously downplaying the other's contribution.
AMD can use what's best for the task at hand/accurately portray the relative importance of their CPUs/GPUs without undermining their marketing message.
Evergreen had a *huge* lead over pre-Fermi nVidia chips, and still leads in 32-bit precision (and by extension most of what the mass market cares about), but 64-bit precision lags Fermi. Of course, Evergreen beat Fermi to market by a large large margin.
Everything that's flying/walking around your fruit trees should probably not be doing that.
You mean like the creatures that fly around and help your plants pollinate so they actually, you know, bear fruit? I guess the point could be made that this would only be used when frost was a risk and therefore no pollinating would be occurring anyway.
On top of being highly capable at massively parallel floating point math (the bread and butter of top500 and most all real world HPC applications), GPU chips benefit from economies of scale by having a much larger market to sell chips to. If Intel has an HPC-only processor, I don't see it really surviving. There have been numerous HPC only accelerators that provided huge boosts over cpus that flopped. GPUs growing into that capability is the first large scale phenomenon in hpc with legs.
Disagree. Very few people avoid preventative care in other insured areas.
Checking your tire pressure is free and the much more likely consequence of not doing so is having to buy new tires and/or rims at your expense. If you know you have an electrical problem, you'll get it taken care of, but if everything seems fine, you aren't paying 200 bucks every year for some inspector to look over your house for signs of trouble. You could have used oil change as an example to be closer, but that's to avoid a near certain catastrophic failure.
Even with insurance carriers frequently reducing any copay to zero to encourage visits, people still tend not to go out of inconvenience (in my case, my insurance pays me to go). These exams are done when you think nothing is wrong, and bloodwork is done and asymptomatic conditions can be detected that require minimal treatment or just simple lifestyle changes that avert severe conditions.
Not only health insurance does this, but most car insurance policies will 100% cover a chip or cracked windshield repair to avoid paying for windshield replacement when it grows beyond the point of hope.
I'm already dealing with UEFI systems with base configuration utilities that REQUIRE graphics and simply do not work with serial console redirection. I implore all serious vendors that will have to provide any interface via UEFI to support the text out methods. As any site familiar with Linux/Unix, or hell, even Windows Powershell/cmd users is aware, a GUI is not inherently superior to CLI just because it is shiny and there are many many scenarios where disdain for CLI is uncalled for.
To 'root' my pre on the first day involved only downloading the official development platform from Palm for Linux. I didn't have to go to Windows or OSX or wait for someone in the community to 'jailbreak'. Meanwhile, Android phones from most manufacturers take a few weeks for the community to jailbreak before the fun begins. I'd rather go with a platform where the manufacturer blatantly allows the users the power Palm does. I find it ironic as the base platform is more closed in theory, but in practice is a bit more amenable to hacking.
Though I'm personally not enthused about their HTML5/Javascript 'premiere' approach to applications, I do like the simplicity of SDL/GL/C code to develop other apps.
As a user, I find WebOS' current interface a bit slicker on the multitasking front.
Of course, all this said I don't think I'll ever be interested in a tablet. It's in a useless spot for me of not being as useful as a laptop yet not as convenient as my 'phone'.
Canonical has nothing to really gain unless their leadership wants to ditch Ubuntu for SuSE/SLES, which I just don't see happening. I don't see Canonical benefiting from maintaining two very distinct distros. If thinking they'll buy the company and force migrate to Ubuntu, that has more potential to drive them away from Canonical than trying to displace that market without buying it. If they bought and 'forced' customers to change (Enterprise customers are very change-averse), the customers may comply and change to a vendor that didn't force the issue.
Similarly, I see RHAT in the same position.
I could in theory see Oracle, depending on if they want to be more 'differentiated' than Unbreakable lets them be today. I might even see Dell (the only major hardware player without any OS to call 'their own', Oracle has Solaris/Unbreakable, HP has HP-UX and WebOS, IBM has AIX...). IBM seems in theory a potential fit, but they seem to regard Linux cautiously and despite having the Linux investment and capability to deliver a standalone distro to date with instant credibility in the enterprise market, they haven't made any effort. HP could also be interested in an enterprise-credible distribution. Of course, among hardware vendors, putting themselves in competition with RHAT, a current partner would be tricky.
Generally though, I absolutely can't imagine a company already solidly in the linux distro business have anything to gain from acquiring SuSE product/responsibilities.
Are you saying that a kernel module written against 2.6.18-8.el5 without any knowledge of what will happen up to 199 would compile without issue, because I call BS. I have dealt with RHEL5.2->5.3 and other such transitions and had to get new vendor source code as some include files changed in the kernel. Could that new source code compile against the old tree, sure, but it's easy to make the code have the right #ifdefs after the fact.
I'm not saying ignoring the new drivers is good, but 'pretending' that it even resembles '2.6.18' with all the backport is just that, pretending. I could understand if RH didn't change subsystem-wide aspects and managed to get drivers to work, but they do change certain things. I wouldn't begrudge them saying the nature of linux forces jumps in the version number to reasonably provide value, but the '2.6.18' is misleading.
I have to agree with the sentiment of the above, even if 'honesty' isn't the right word. They backport stuff from 2.6.32 to 2.6.18. They break the kernel module interfaces (drivers have had to change their source code to follow the 5.x series). The resultant thing tends to exhibit some of the worse of both worlds.
Slashdot Mirror
NetApp invented WAFL before ZFS. They claim ZFS itself infringes on WAFL technology.
NetApp uses WAFL on their NAS. A filesystem they invented before ZFS. They claim ZFS violates patents of WAFL.
2) I'm really glad I live there, and wish everyone had the opportunity to live like that
The problem is that the people who do live like that only do so by other people not living like that. There is a reason why so much manufacturing and mining is done in developing countries with little or no restrictions, and it isn't just the labor cost.
3) Environmental destruction has not actually happened here because of environmental regulations, which means it is an abstract concept. I wish it was for the rest of the world to
As above, if those regulations were in effect globally, *fewer* people would live like that and it would cost disproportionately more.
1) And why is land scarce? Oh, because dictators own it all. Then you realise that everyone could fit in the southern USA at Los Angles pop densities. And yes, I do realise that they will need more land to provide all the stuff they are using
I do agree that hoarding by violent means is the bigger problem than any artificial scarcity and I have no inkling of how land *really* looks, but on the other extreme of 'LA density everywhere' is a silly concept, considering that LA would not be possible if not for food and other products being brought into the city. If LA were given an impenetrable dome suddenly, a great many would starve when the food stores went out for lack of farmland to cultivate new food.
3) Environmental destruction may be permanent in some instances. That's why we need to help developing countries move beyond that phase of construction
It's naive to presume that doing it safe and clean at the similar (really, you are hoping for lower) costs.
4) Because they can't afford to.
Because the price of restoring it when technically possible is higher than any value extracted. If you are going to pay the price to put things back the way they were, then you are committing to put more resources into restoration than you will extract in ruining it.
I guarantee you that most of the products you consume are imported from places that do significant harm to their environment. Our amenities come by exploiting willingness of other places to fubar their environment. The cost and sometimes simply the forbidden nature of extracting resources from the earth in a developed country combined with similar effects on processing those resources make developing countries the place to do it. If every square foot on the planet were as restricted as the 'developed world', then all the toys and necessities would be much more rare and much more expensive.
I'm not an uber-envorinmentalist or anything, but I have to acknowledge that my country's domestic rules and regulations are nice for our land, but bordering on hypocritical with the huge imports of goods made possible by other places not having such restrictions.
Frankly, there is no valid reason for starting a new program in C in this day and age.
One reason I can think of is incurring the libstdc++ dependency in an embedded or otherwise size critical context. Sometimes you still have precious little OS storage, and your convenience/cost compromises might land shy of C++. This is a matter of design tradeoff, some contexts demand nothing more than assembler, or a libc smaller than glibc, and some 'embedded' can do nearly whatever the hell it wants.
Lower cost of ownership - BS, too much time is spent hacking up config files to make crap work or work right
This perception problem is common with perl. You can run a modern linux system fairly naively, but the enthusiasts will make the 'fancy tricks' look like 'the' way of life. There are advanced capabilities that are possible with fancy trickery, but most people would do fine if they ignored those capabilities and just used their system. Same applies to a lot of systems with advanced capabilities, for a simple example look up source engine command parameters for adjusting multiplayer settings. If you naively use those games, you'll likely have no problem, but if you search for cl_updaterate and start reading, you would be given the impression that there are tons and tons of parameters that simply *must* be tweaked for an 'acceptable' experience.
CLI/scripting system that actually works - BS, anything you can write and make work in Linux, I can in Windows
There is what is technically in theory possible, and then there is how to do it. If you compare vbscript/cmdscript to bash/(python|perl|etc), the syntax is dramatically better. Now Powershell brings a fair amount of better syntax and has at least decent interactive startup time in Windows 7/28kr2 (before that it was a mess), but in WindowsPE, you are still stuck with cmd/vbscript as your tools. In my opinion, Powershell leans a little more toward perl/python in some ways that make things like bash a touch easier for typical shell scripting.
Most open source software runs on it - Show me anything worthwhile that doesn't run in Windows or have a better alternative there
Compiz. Window scaling with title search is incredibly awesome in terms of practical application whereas Aero is strictly just eye-candy. Most applications will run on either platform, but on a typical linux it's 'aptitude install openoffice' rather than trying to find individual websites to host the software.
Drivers for just about any piece of hardware ever built - BS, that's the primary thing most users have issues with, half baked drivers
All platforms are subject to crappy drivers from crappy vendors. I had to search all over the damn place about networking on a new Windows 7 going out every couple ours to find out that the manufacturer's driver would go tits up due to segmentation offload (turning it off involved navigating several dialogs graphically, in linux it would just be 'ethtool -K eth0 tso off'). Pick a reputable hardware vendor and you'll get a good driver experience regardless of platform. Pick a sketchy vendor, and your experience is likely to suck regardless of OS.
I suspect the plan price was enough to kill it off, though I have some shred of hope that humanity saw the ads that did nothing more than say "OMG you can facebook" and collectively shrugged. Of course, the devices looked bad and the interface in the ads looked even worse.
In terms of Verizon worshipping, the full story sounds more like some MS guy liked Sidekicks (which if they were thinking clearly would be recognized as a stopgap rather than an endgame). Then they acquired danger and decided to do it their own way while keeping sidekick barely alive. On top of targeting a rapidly evaporating market as smartphones came down, they made it horribly late by insisting upon essentially starting from scratch (reminds me of the hotmail acquisition and initial attempt to move to winnt servers). By the time they were ready to move, smartphones and smartphone plans had come down so much, the Kin was laughable.
Essentially, it's the story of MS' life in the mobile market. They don't get the concepts until someone else has already gone. I don't know if/when MS will finally give up. The handset vendors have everything to gain from a platform like Android and nothing really to lose. I personally like WebOS, but I don't think there is room for more than one Apple-like entity in the market, and Android is really out-MSing MS in the space. MS made share gains by all the OEM support, and HTC, Samsung, et al are more in charge of their own destinies than MS ever would allow.
I generated 1000000 pairs of children at random
Holy crap you were busy
You made a mistake.. To avoid confusion, you designated a *specific* one of the children as being the child described as being a boy. this achieves the same effect as prescribing a specific ordering. It ignores the probability that the unnamed 'Boy' is the child you called Peter, which would make the probability that the language describes the first and third set twice as high as second or fourth (those cases have twice as many 'Peter' candidates, and you ruled one of those candidates out arbitrarily). This is really hard to think through, and we intuitively want to make the boy/boy orderings unambiguous by picking which of the boys is *the* boy, however, that assumes data not given in the original statement. In a random distribution, 75% of the permutations can be described as "one is a boy", and 2/3 of that set is mixed gender. however, only 50% can be described as "the first one is a boy", bringing it back to 50%.
One way of doing that would be to split the above list in half:
bb, bg | gb gg
and ask "At least one of my children is a boy..." in the first case and "At least one of my children is a girl..." in the second case.
I don't see how you think you've changed the problem. If you are grouping 'bg|gb' as one group on par with bb and gg, first that's 3/4ths of the group count and the middle group is twice the size of the other groups. If you are saying that bg, bg is one group, and gb gg is another group, then "at least one of my children is a boy' would describe the entire first group and half of the second, making the grouping meaningless. If you are saying the goal is to pick which half of the | describes it, then using the unordered information provided, the probability of guessing the correct side can be 2/3rds.
You haven't changed the problem in any way whatsoever..
The difference is the 'naive' code you write to do things in the simplest manner *can* run on a CPU. For the GPU languages, you *must* make those optimizations. This is not to undercut the value of GPU (as Intel concedes, the gap is large), but it does serve to counteract the dramatic numbers tauted by nVidia.
nVidia compared expert tuned and optimized performance metrics on their product and compared against stock, generic benchmarks on intel products.
Magny-Cours is currently showing significant performance advantage over Intel's offerings while at the same time AMD's Evergreen *mostly* shows performance advantages over nVidia's Fermi despite making it to market ahead of Fermi.
AMD is currently providing the best tech on the market This will likely change, but at the moment, things look good for them.
AMD is the most advantaged on this front...
Intel and nVidia are stuck in the mode of realistically needing one another and simultaneously downplaying the other's contribution.
AMD can use what's best for the task at hand/accurately portray the relative importance of their CPUs/GPUs without undermining their marketing message.
Evergreen had a *huge* lead over pre-Fermi nVidia chips, and still leads in 32-bit precision (and by extension most of what the mass market cares about), but 64-bit precision lags Fermi. Of course, Evergreen beat Fermi to market by a large large margin.
Everything that's flying/walking around your fruit trees should probably not be doing that.
You mean like the creatures that fly around and help your plants pollinate so they actually, you know, bear fruit? I guess the point could be made that this would only be used when frost was a risk and therefore no pollinating would be occurring anyway.
On top of being highly capable at massively parallel floating point math (the bread and butter of top500 and most all real world HPC applications), GPU chips benefit from economies of scale by having a much larger market to sell chips to. If Intel has an HPC-only processor, I don't see it really surviving. There have been numerous HPC only accelerators that provided huge boosts over cpus that flopped. GPUs growing into that capability is the first large scale phenomenon in hpc with legs.
Disagree. Very few people avoid preventative care in other insured areas.
Checking your tire pressure is free and the much more likely consequence of not doing so is having to buy new tires and/or rims at your expense. If you know you have an electrical problem, you'll get it taken care of, but if everything seems fine, you aren't paying 200 bucks every year for some inspector to look over your house for signs of trouble. You could have used oil change as an example to be closer, but that's to avoid a near certain catastrophic failure.
Even with insurance carriers frequently reducing any copay to zero to encourage visits, people still tend not to go out of inconvenience (in my case, my insurance pays me to go). These exams are done when you think nothing is wrong, and bloodwork is done and asymptomatic conditions can be detected that require minimal treatment or just simple lifestyle changes that avert severe conditions.
Not only health insurance does this, but most car insurance policies will 100% cover a chip or cracked windshield repair to avoid paying for windshield replacement when it grows beyond the point of hope.
I cringe every time I hear the word 'tweet'.
I'm already dealing with UEFI systems with base configuration utilities that REQUIRE graphics and simply do not work with serial console redirection. I implore all serious vendors that will have to provide any interface via UEFI to support the text out methods. As any site familiar with Linux/Unix, or hell, even Windows Powershell/cmd users is aware, a GUI is not inherently superior to CLI just because it is shiny and there are many many scenarios where disdain for CLI is uncalled for.
To 'root' my pre on the first day involved only downloading the official development platform from Palm for Linux. I didn't have to go to Windows or OSX or wait for someone in the community to 'jailbreak'. Meanwhile, Android phones from most manufacturers take a few weeks for the community to jailbreak before the fun begins. I'd rather go with a platform where the manufacturer blatantly allows the users the power Palm does. I find it ironic as the base platform is more closed in theory, but in practice is a bit more amenable to hacking.
Though I'm personally not enthused about their HTML5/Javascript 'premiere' approach to applications, I do like the simplicity of SDL/GL/C code to develop other apps.
As a user, I find WebOS' current interface a bit slicker on the multitasking front.
Of course, all this said I don't think I'll ever be interested in a tablet. It's in a useless spot for me of not being as useful as a laptop yet not as convenient as my 'phone'.
Canonical has nothing to really gain unless their leadership wants to ditch Ubuntu for SuSE/SLES, which I just don't see happening. I don't see Canonical benefiting from maintaining two very distinct distros. If thinking they'll buy the company and force migrate to Ubuntu, that has more potential to drive them away from Canonical than trying to displace that market without buying it. If they bought and 'forced' customers to change (Enterprise customers are very change-averse), the customers may comply and change to a vendor that didn't force the issue.
Similarly, I see RHAT in the same position.
I could in theory see Oracle, depending on if they want to be more 'differentiated' than Unbreakable lets them be today. I might even see Dell (the only major hardware player without any OS to call 'their own', Oracle has Solaris/Unbreakable, HP has HP-UX and WebOS, IBM has AIX...). IBM seems in theory a potential fit, but they seem to regard Linux cautiously and despite having the Linux investment and capability to deliver a standalone distro to date with instant credibility in the enterprise market, they haven't made any effort. HP could also be interested in an enterprise-credible distribution. Of course, among hardware vendors, putting themselves in competition with RHAT, a current partner would be tricky.
Generally though, I absolutely can't imagine a company already solidly in the linux distro business have anything to gain from acquiring SuSE product/responsibilities.
Are you saying that a kernel module written against 2.6.18-8.el5 without any knowledge of what will happen up to 199 would compile without issue, because I call BS. I have dealt with RHEL5.2->5.3 and other such transitions and had to get new vendor source code as some include files changed in the kernel. Could that new source code compile against the old tree, sure, but it's easy to make the code have the right #ifdefs after the fact.
I'm not saying ignoring the new drivers is good, but 'pretending' that it even resembles '2.6.18' with all the backport is just that, pretending. I could understand if RH didn't change subsystem-wide aspects and managed to get drivers to work, but they do change certain things. I wouldn't begrudge them saying the nature of linux forces jumps in the version number to reasonably provide value, but the '2.6.18' is misleading.
I don't think the kernel factors into the application ABI/API overmuch. I have binaries from very old kernels with no issue.
'ABI' compatibility in the kernel is not preserved. The kernel header files changed so that not only did drivers need recompile, they needed recoding.
I have to agree with the sentiment of the above, even if 'honesty' isn't the right word. They backport stuff from 2.6.32 to 2.6.18. They break the kernel module interfaces (drivers have had to change their source code to follow the 5.x series). The resultant thing tends to exhibit some of the worse of both worlds.