Not in the least bit, observe, just verified with ldd that Xorg and firefox have libz dynamically linked in on my system, which means on program restart, it will pick up the code from the shared library at runtime. It's the whole point of a dynamicly linked library.
Now once upon a time, a lot of distributions (and open source projects out of the box even) would just static link in libz for some reason or another, but after some security issues in the past that caused massive headaches for package maintainers, that practice has largely ceased.
I do have to disagree a bit about MS and IBM not being innovative. Don't know enough about Oracle to say one way or another.
MS does have some neat research going on, but 99% of it never goes near their mainstream product line, and instead, they spend effort copying up and coming ideas that are proving themselves in the field and making them fear being left out. The really innovative stuff from their employees largely sits on the back burner, being considered not worth the risk to bring to the forefront, until, of course, competitors start embracing a similar idea, then they may use it.
IBM is similar, but occasionally brings stuff out. A recent, very successful example is Blue Gene. A system with remarkable communications architecture connecting low-power (compute and literal sense) processing elements at incredible scale. They are guilty, however, or slowing the rate of innovation to a crawl once a product shows degrees of success. Once the formula has appeared to work, they don't mess with it, again, considered too risky to do so. They'll make evolutionary changes to keep pace with their perceived competition, but little else.
While not personally familiar with how the builds in Debian happen, a cross compiling build system is how a lot of vendors deal with this very issue.
I know of vendors that support equipment they don't have a single sample of. Of course, they warn their customers and typically have one or a very small number of early-adopter customers who maybe get a good price break or simply want new features enough to explicitly desire bleeding edge to serve as testing for their releases.
It has to walk, but nothing says the speed of the device being controlled has to be exactly the speed of walking. Walking could translate to 100 MPH movement if you wanted. Cockroaches don't exactly have the capacity to learn how to manipulate something other than what they are born knowing, so it makes sense.
Besides, the point of this is supposed to be at least somewhat artistic. Also, a kind of experiment to see how a cockroach compares to traditional methods of using programmed microcontrollers to acheive the same thing.
But his point is that is a not so common need, and that you can already have multiple IPv6 addresses to a host (hell, even with IPv4) you can do it. It doesn't make sense to eat up huge amounts of address space for the sake of an uncommon need that already has a solution. On the other hand, every host is on a network, and every reasonable network has multiple hosts, so the address specifying both makes sense, as it is used 100% of the time.
Not so much as betting on Linux as trying to be loud that they embrace it.
Now that they sold the PC division, the lowest-end they go is workstation, which they support linux in just like servers (well, technically, point of sale equipment is more low end, but that mostly runs linux too...) From zSeries down to Intellistation, it is tested, but to some degree feels like a second hand citizen to Windows where applicable (except in special circumstances, like x86 based cluster solutions).
To their credit, the pSeries servers at least seem to treat Linux very well and not treat it worse than AIX in terms of support...
I even have a SPARCbook 2 deep in my closet somewhere. Ahh, good old laptop, 32-bit sparc, 540 MB hard drive... AUI ethernet (woo), 640x480 LCD, and you could beat the *crap* out of someone if you wanted, or deflect bullets to medium-sized ICBMs if you wanted.
cel://syncorbit/Sol%3AEarth/2005-06-25T17%3A49%3A0 8.00000?x=sbsIqCwAe9e+DA&y=DhVf7tVWGQ&z=ZSnBOVkXLQ UQ&ow=0.249240&ox=-0.207134&oy=-0.047122&oz=-0.944 857&select=Sol:Earth&fov=32.524258&ts=0.000000<d =0&rf=71571&lm=2 This is a view from Earth's surface (US East coast) of the three planets. As always, beware of spaces/line breaks in the URL (alt-g to open arbitrary URL).
I meant ctrl-g, not shift g, a really cool app, for those playing around, hitting 'h' selects sol, then you can hit g to get back to familiar territory.
Not specifically for this event, but a really awesome simulator for astronomy stuff in general is celestia. http://www.shatters.net/celestia/
Click on the earth, hit g (actually celestia starts you at earth), you can hit o to mark orbits to make it easier to find the planets (hit p to show labels to identify the planets themselves.) I hit shift-g to go to the surface of earth and then find the planets. Can give you a good idea where to look really. Being on the surface isn't particularly important at first, don't want the planet being between you and your destination until you are sure it wouldn't be.
It doesn't always much like things really do, especially with atmospheric simulation not being particularly complex in celestia, but it is cool none the less. Particularly to click on one of the three planets close together, hit g, then look at everything from their perspective.
The key here is that in the commodity PC business, there are no margins, like you say, and IBM is just too topheavy to compete effectively in such a market.
IBM is still in the professional workstation and server market, where there is a high margin and that can still fly.
Apple is a high-margin style seller, whose products are priced at a premium and aim to be a status symbol of sorts, with some emphasis on a tightly controlled platform enabling a solid desktop user experience. Their strategy is very much related to mercedes and the like with somewhat improved quality, a lot of status, and high margins to compensate for the low marketshare that approach gives.
To say things are underway is an understatement, several projects already resolve the issue and are running primetime. This is part of the problem
esound, arts, and dmix all work around this. All have working solutions to the problem, and each has different people putting them forward and to date the platform as a whole has not converged on a working solution. Here the notion of flexibility of choice backfires, without one clear path to point someone in the direction of.
Apple has little to gain from the Linux kernel, and a lot to lose. The mach kernel is BSD licensed, and you can bet your ass that they will have some modifications to try to avoid booting on generic PC clones that they will not want to open up. Sure you could do this with linux, but they already have a kernel that works for them, their developers are familiar with, and they don't need much else.
I seriously doubt Apple would go down that road, though the hack seems likely, at least to run it under another os (mac on linux for example).
Apple's interest is in the profit margin on their complete solution, hardware, os, etc. They maintain relatively low volumes, but do what they can to justify a large profit margin. OSX by itself to a small number of people is enough to make them buy the high-profit margin solution.
Now let's say they released OSX for arbitrary x86 systems. Those small number of people who would have bought the whole solution for the sake of the OS now just have to buy the OS, which by itself has a hard timee a) not being pirated and b) having a profit amount per unit anywhere near a complete system. Volumes may increase at the reduced barrier to entry, but let's face it, MS still dominates and the vast majority have a solution already and OSX isn't that attractive.
But typically you can acheive 'insanely great' without sacrificing backwards compatibility to the extent Apple has, essentially saying screw you to their existing customer base.
MS has baggage mainly relating to permissions and security, but the codebase of XP isn't horribly crippled feature-wise and maintains good backward compatibility. Not perfect, but good. (I prefer linux, but the track record for nice binary compatibility is perhaps the worst of all considered cases)
OS9 to OSX was painful, but relatively well-executed, and probably necessary with the drastic change in memory management and multitasking model. Kinda like the win3.x to win95 and win98 to winxp leaps, just a little rougher.
m68k to something else was an obvious and needed jump, unfortunately. It was not reasonably avoidable, at the time no one could have said x86 would succeed as it did and that m68k would be left adrift as time marched on.
However, the move to PPC was baffling. x86 had the track record and the proliferation of that platform was obvious. Despite Motorola's problems in delivering good m68k designs, they went largely with Motorola again, though admittedly IBM being in there was probably reassuring. They had a choice at that point of the proven and ubiquitous x86 platform, or an equivalent platform performance wise that may or may not succeed in the personal computing marketplace, and it pretty much failed, despite years of trying and tons of marketing to try to cover up the shortcomings of their chosen platform. Now they have to backpedal, essentially admit they have been BSing, but promise that this time they are right.
Slashdot Mirror
She must just be trying get rid of some of that pesky hot grits slashdotters keep pouring down her pants..
Not in the least bit, observe, just verified with ldd that Xorg and firefox have libz dynamically linked in on my system, which means on program restart, it will pick up the code from the shared library at runtime. It's the whole point of a dynamicly linked library.
Now once upon a time, a lot of distributions (and open source projects out of the box even) would just static link in libz for some reason or another, but after some security issues in the past that caused massive headaches for package maintainers, that practice has largely ceased.
http://www.snopes.com/business/misxlate/nova.asp
Favorite example:
Would you refuse to buy a dinette set called 'Notable' because, obviously, it says there is "no table"?
I do have to disagree a bit about MS and IBM not being innovative. Don't know enough about Oracle to say one way or another.
MS does have some neat research going on, but 99% of it never goes near their mainstream product line, and instead, they spend effort copying up and coming ideas that are proving themselves in the field and making them fear being left out. The really innovative stuff from their employees largely sits on the back burner, being considered not worth the risk to bring to the forefront, until, of course, competitors start embracing a similar idea, then they may use it.
IBM is similar, but occasionally brings stuff out. A recent, very successful example is Blue Gene. A system with remarkable communications architecture connecting low-power (compute and literal sense) processing elements at incredible scale. They are guilty, however, or slowing the rate of innovation to a crawl once a product shows degrees of success. Once the formula has appeared to work, they don't mess with it, again, considered too risky to do so. They'll make evolutionary changes to keep pace with their perceived competition, but little else.
Everyone knows they did it to screw with horoscopes.
"...regular one is too hard."
Only if longer than 4 hours, then you need to see a physician...
Long before today Apple knew about the low watt and dual core roadmaps. Long before that announcement.
This wasn't a 'whip the line in shape, we are losing to intel'.
While not personally familiar with how the builds in Debian happen, a cross compiling build system is how a lot of vendors deal with this very issue.
I know of vendors that support equipment they don't have a single sample of. Of course, they warn their customers and typically have one or a very small number of early-adopter customers who maybe get a good price break or simply want new features enough to explicitly desire bleeding edge to serve as testing for their releases.
It has to walk, but nothing says the speed of the device being controlled has to be exactly the speed of walking. Walking could translate to 100 MPH movement if you wanted. Cockroaches don't exactly have the capacity to learn how to manipulate something other than what they are born knowing, so it makes sense.
Besides, the point of this is supposed to be at least somewhat artistic. Also, a kind of experiment to see how a cockroach compares to traditional methods of using programmed microcontrollers to acheive the same thing.
But his point is that is a not so common need, and that you can already have multiple IPv6 addresses to a host (hell, even with IPv4) you can do it. It doesn't make sense to eat up huge amounts of address space for the sake of an uncommon need that already has a solution. On the other hand, every host is on a network, and every reasonable network has multiple hosts, so the address specifying both makes sense, as it is used 100% of the time.
Not so much as betting on Linux as trying to be loud that they embrace it.
Now that they sold the PC division, the lowest-end they go is workstation, which they support linux in just like servers (well, technically, point of sale equipment is more low end, but that mostly runs linux too...) From zSeries down to Intellistation, it is tested, but to some degree feels like a second hand citizen to Windows where applicable (except in special circumstances, like x86 based cluster solutions).
To their credit, the pSeries servers at least seem to treat Linux very well and not treat it worse than AIX in terms of support...
I even have a SPARCbook 2 deep in my closet somewhere. Ahh, good old laptop, 32-bit sparc, 540 MB hard drive... AUI ethernet (woo), 640x480 LCD, and you could beat the *crap* out of someone if you wanted, or deflect bullets to medium-sized ICBMs if you wanted.
cel://syncorbit/Sol%3AEarth/2005-06-25T17%3A49%3A0 8.00000?x=sbsIqCwAe9e+DA&y=DhVf7tVWGQ&z=ZSnBOVkXLQ UQ&ow=0.249240&ox=-0.207134&oy=-0.047122&oz=-0.944 857&select=Sol:Earth&fov=32.524258&ts=0.000000<d =0&rf=71571&lm=2
This is a view from Earth's surface (US East coast) of the three planets. As always, beware of spaces/line breaks in the URL (alt-g to open arbitrary URL).
I meant ctrl-g, not shift g, a really cool app, for those playing around, hitting 'h' selects sol, then you can hit g to get back to familiar territory.
Not specifically for this event, but a really awesome simulator for astronomy stuff in general is celestia. http://www.shatters.net/celestia/
Click on the earth, hit g (actually celestia starts you at earth), you can hit o to mark orbits to make it easier to find the planets (hit p to show labels to identify the planets themselves.)
I hit shift-g to go to the surface of earth and then find the planets. Can give you a good idea where to look really. Being on the surface isn't particularly important at first, don't want the planet being between you and your destination until you are sure it wouldn't be.
It doesn't always much like things really do, especially with atmospheric simulation not being particularly complex in celestia, but it is cool none the less. Particularly to click on one of the three planets close together, hit g, then look at everything from their perspective.
Argument still stands, they have to use an off-the-shelf video chipset. Or at least there is no reason they wouldn't so far.
The key here is that in the commodity PC business, there are no margins, like you say, and IBM is just too topheavy to compete effectively in such a market.
IBM is still in the professional workstation and server market, where there is a high margin and that can still fly.
Apple is a high-margin style seller, whose products are priced at a premium and aim to be a status symbol of sorts, with some emphasis on a tightly controlled platform enabling a solid desktop user experience. Their strategy is very much related to mercedes and the like with somewhat improved quality, a lot of status, and high margins to compensate for the low marketshare that approach gives.
To say things are underway is an understatement, several projects already resolve the issue and are running primetime. This is part of the problem
esound, arts, and dmix all work around this. All have working solutions to the problem, and each has different people putting them forward and to date the platform as a whole has not converged on a working solution. Here the notion of flexibility of choice backfires, without one clear path to point someone in the direction of.
I see two key points, one is that you've ditched your experience years ago, which means a lot has changed.
Second, SuSE at least is not such a crap distro anymore. RH has a lot of issues to date, but SuSE has pulled things off nicely.
They announced the dev kit would be P4 3.6 GHz, the spec for building applications for it indicates two platforms, PPC and ia32 by name.
Security through obscurity, indeed.
The problem with such a scheme is that it is useless unless relatively standard, and if relatively standard easily overcome.
Encrypted mail is a simple, much more secure way.
Don't know why people feel the need to try to get fancy and end up with something crappy when solid solutions exist.
Apple has little to gain from the Linux kernel, and a lot to lose. The mach kernel is BSD licensed, and you can bet your ass that they will have some modifications to try to avoid booting on generic PC clones that they will not want to open up. Sure you could do this with linux, but they already have a kernel that works for them, their developers are familiar with, and they don't need much else.
I seriously doubt Apple would go down that road, though the hack seems likely, at least to run it under another os (mac on linux for example).
Apple's interest is in the profit margin on their complete solution, hardware, os, etc. They maintain relatively low volumes, but do what they can to justify a large profit margin. OSX by itself to a small number of people is enough to make them buy the high-profit margin solution.
Now let's say they released OSX for arbitrary x86 systems. Those small number of people who would have bought the whole solution for the sake of the OS now just have to buy the OS, which by itself has a hard timee a) not being pirated and b) having a profit amount per unit anywhere near a complete system. Volumes may increase at the reduced barrier to entry, but let's face it, MS still dominates and the vast majority have a solution already and OSX isn't that attractive.
There is a reason why it appears that way, they ship with an ink cartridge that is far from full.
But typically you can acheive 'insanely great' without sacrificing backwards compatibility to the extent Apple has, essentially saying screw you to their existing customer base.
MS has baggage mainly relating to permissions and security, but the codebase of XP isn't horribly crippled feature-wise and maintains good backward compatibility. Not perfect, but good. (I prefer linux, but the track record for nice binary compatibility is perhaps the worst of all considered cases)
OS9 to OSX was painful, but relatively well-executed, and probably necessary with the drastic change in memory management and multitasking model. Kinda like the win3.x to win95 and win98 to winxp leaps, just a little rougher.
m68k to something else was an obvious and needed jump, unfortunately. It was not reasonably avoidable, at the time no one could have said x86 would succeed as it did and that m68k would be left adrift as time marched on.
However, the move to PPC was baffling. x86 had the track record and the proliferation of that platform was obvious. Despite Motorola's problems in delivering good m68k designs, they went largely with Motorola again, though admittedly IBM being in there was probably reassuring. They had a choice at that point of the proven and ubiquitous x86 platform, or an equivalent platform performance wise that may or may not succeed in the personal computing marketplace, and it pretty much failed, despite years of trying and tons of marketing to try to cover up the shortcomings of their chosen platform. Now they have to backpedal, essentially admit they have been BSing, but promise that this time they are right.