The issue, in this case, is not so much proprietary forks of LLVM itself, which(as you note) are largely not worth the effort; but situations where a given architecture, say, is supported only in the form of a proprietary distribution of the LLVM core + a proprietary code generator for the given platform. You then end with a situation where the platoform vendor saved some money; but the platform is basically closed.
It's roughly analogous to being concerned about GPL compliance in companies that make routers and other little consumer electronics widgets: Such companies generally make no modifications of broad interest to the linux kernel, so their noncompliance imperils the development of the kernel not at all; but the modifications that they do make are generally vital to being able to build your own kernel for that particular device.
You aren't concerned because the WRT-54G firmware contained novel kernel innovations that will be of use to the linux kernel generally(indeed, the stock firmware was pretty unimpressive), you are concerned because you want to know what you need to know in order to build your own kernel for that device.
They rather stretch the definition of "standard"; but you can get some pretty beefy wall-warts.
The "ultra-small form factor" dell optiplexes have, historically, been powered by external power supplies(because hiding several pounds of the computer under the desk sure does make the case look smaller in the marketing glamor shots). This is a "wall wart" running a PC built with standard components(you could even get them with Prescot based Pentium Ds, back in the day, which were toasty bastards). The last such power supply I saw was rated at 220 watts, DC output, with a slightly higher draw at the wall.
Laptop chargers, for your basic cheap 15-17 inchers, are routinely in the 60-80 watt range. Your absurd desktop replacement units can push 150watts(you know that it isn't a good sign if your laptop's adapter has a 40mm fan in it...)
"s the days of GNU as the mainstream free and open development toolchain passé?"
I suspect that that will depend, in part, on how LLVM ends up being used. Since it is under a BSD-esque license, LLVM itself is definitely a candidate for being the "mainstream free and open development toolchain"; but only if the majority of real-world support scenarios don't involve proprietary actors taking advantage of that fact. In that case, it'll pretty much just end up being the core of a large number of binary, proprietary, BSPs and toolchains.
Given the good things that are said about its technical characteristics, I would hope that that doesn't happen; but the potential exists.
The fact that Bluefin are valuable has been responsible for the 80%-90% reduction in numbers; but also for the fact that people get real touchy about anything that threatens the last 10% or so.
The trouble here is that Bluefin like to go to the Gulf to spawn. If the delightful mixture of hydrocarbons and toxicologically troublesome dispersants turns out to poison eggs, sperm, or tiny juvenile fish, you could easily get an ecological impact equivalent to massive harvesting of the adult population; but without even the compensatory sushi.
Well, in this case, it is in humanity's self-interest, if nothing else, because bluefin tuna are legendarily tasty.
The ethical duties, if any, of environmental preservation are debatable. The fact that crashing the population of a species you like to eat is stupid and self-defeating isn't.
That and the fact that making boxes out of sheet metal is cheaper than making spheres out of sheet metal...
And PC-boards that aren't flat rectangles are very much special order items.(can you even get curved PCBs? There are those flexible plastic ribbon cables that sometimes have an IC or two soldered on; but I don't think I've ever seen a PCB of any nontrivial size that wasn't flat.)
If you want toolless, your best bet is actually soulless corporate drone-boxxen. They tend to be crushingly ugly; and slightly more expensive per unit spec than gamer homebuilds; but they are explicitly designed so that a monkey could replace pretty much FRU(with the possible exception of the motherboard, which is often screwed down) with its bare hands, and an experienced tech can replace all the FRUs in ~5 minutes.
Generic whitebox cases have gotten somewhat better in the toollessness department; but corporate drone-boxes have always been two steps ahead in that department.
With the increasing levels of integration(heck, you pretty much can't buy a motherboard without NIC and sound and scads of USB ports, and buying one without basic video isn't getting any easier), and the fact that we have all the lessons learned about cooling during the Prescott/space-heater era being applied to much cooler chips, the rise of mini-ITX seems like a obvious development. Multilayer PCBs aren't crazy expensive; but every square inch isn't free.
Well, Nature is a subsidiary of Macmillan, which is a subsidiary of Verlagsgruppe Georg von Holtzbrinck. I assume that it gets divided up among the shareholders in one or more of those entities.
The point isn't really to make publishing "free", because that isn't possible. Even your kid brother's myspace page costs something to keep on the web. The internet and desktop publishing software makes it cheaper; but can't entirely eradicate the cost.
The advantage, though, is that by tweaking the payment model, you can massively increase the accessibility of the research(which is arguably an ethical imperative when it is publicly funded, and a nice perk in all events), cut down on the ability of parasitic middlemen to skim money off the top, and not actually change the distribution of the cost burden all that much.
If journal subscriptions cost money; but publishing an article is free, any research institution will have to have a larger library budget(so the researchers have access to the literature) and a slightly smaller research budget(since the researchers don't pay publishing expenses), the public either gets nothing, or pays nontrivial fees for access.
If the subscription costs nothing; but publishing carries a fee, research institutions will have smaller library budgets(since they won't have to buy the journal) and slightly larger research budgets(since publishing is now a "research" expense). The public at large, and people at poor institutions, get access. The middlemen are reduced to working at cost, rather than extracting rent.
The problem is not with people taking risks(well, that bothers the nanny-staters almost as much as the source of the stem cells bothers the godbots; but that isn't a big deal); but with how the sellers are representing the risks. Competent individuals choosing to take risks, or not, is freedom. Hucksters misrepresenting risks to desperate sick people is somewhere between fraud and manslaughter, depending on how it goes.
How many of these various offshore stem-cell shops fall into the following categories?
1. Scientists/research MDs whose interpretation of risk/reward tradeoffs differs from that of the FDA. In this category I would put more or less orthodox researchers who are of the position that the risks of stem cell use(cancer, infection, immune responses, etc.) are either just not that serious compared to the potential benefits and/or are the individual's choice to make.
2. Sincere cranks. In this category would go the various flavors of nutter who have gone straight off the deep end in terms of actual research about what stem cells are capable of, and how to make them do it; but are fully sincere in their belief that stem cells are the magic bullet against autism or aging, or whatever they are selling them as.
3. Cynical hucksters: All the research seriousness of the above; but without the slightly wild-eyed sincerity. However, they know that lying to desperate sick people is both easy and lucrative.
I once had a user request training when their keyboard was replaced.
To be fair, the old keyboard was a basic 101 key model, and the new one had some media control(start, stop, pause, little volume knob) buttons on it.
The user was informed, as politely as shock allowed, that the function of each key on the new keyboard was the same as that of the old, save for the additional keys, whose use was optional, and not required for the performance of any job-related function.
I'm not sure why anybody would listen to Gartner..
on
Time To Dump XP?
·
· Score: 1
Gartner have a well deserved reputation for authoritative delivery of a mixture of the blatantly wrong and the painfully obvious; but they seem to be veering largely toward the second camp with this one.
Even extended enterprise support for XP isn't going to last forever, and whatever legacy crap people are worried about isn't going to become any more compatible as time passes. As for "training", home users' access to XP has been(barring active effort on their part) largely cut off for some time now, so the ones that aren't mac users at home will be getting exposed anyway.
Unless you have some scheme to drop Microsoft, it seems pretty blatantly obvious that planning for their latest flavor is your only choice at this point.
It tries; but the other devices(those that survive) complain that the 802.11 compliance of a $50 1.2 kilowatt cavity magnetron leaves something to be desired...
Anxiety has reasonably well studied chemical markers. We don't typically bother to test them; because people running to their shrink to lie about being anxious just isn't a huge problem(and, to the degree that it does happen, a lot of first line anxiolytics(lorazipam, diazapam, clonazapam, beer, etc. are cheap and not wildly harmful. Who cares if some people who want them get their hands on them?)
Panic attacks might be fakeable by a good actor(I don't know if manipulating your pulse and andrenaline levels is something they teach you in acting school; but I won't dismiss the notion out of hand); but, unless he is actually a master manipulator who has been fooling everybody for years, this guy is not a good actor.
I'd go with the "embarrassment/reprisal" hypothesis, myself.
The DoD's networks are supposed to be all secure and advanced and stuff. Getting hacked by a single sad-case foreign national, acting without support, makes them look pathetic.
When made to look pathetic, those with power generally seek reprisal against their enemies.
Frankly, the DoD was lucky to have been hacked by him. He is largely harmless, and watching how he got in was probably instructive, to some degree. They really ought to spend less time hounding him, and more time thinking about the fact that certain other hackers are much less harmless, and substantially less likely to be turned over for a stay in PMITA prison by their host governments...
That seems logical enough(and, indeed, my HID-reporting UPS shows up in GNOME panel as a battery device, complete with historical charge statistics and stuff, upon the installation of an appropriate driver).
My point was just that, as in the case of USB_HID, a standard that is too flexible ends up not defining enough to save you from hardware-specific drivers. Outside of the (mostly) safe realm of mice, keyboards, and basic gamepads, "USB_HID" isn't much more of an assurance of "no 3rd party driver needed" than "PCIe device" is.
Unfortunately, the industry of "BPA free" products sprang up at pretty much exactly the same time that the industry of BPA free products did. It turns out that printing new labels is much easier than actually reformulating your products.
Also, the "Contains BPA; but nobody except professional toxicologists studying the subject and hardcore supply chain wonks knows that" industries have been largely unaffected.
Pretty much as theory would predict, the areas closest to ideal markets with zero barriers to entry and equally informed participants achieved something close to a free market solution. The areas that deviated from those assumptions, whether by fraud, subterfuge, imperfect information, or existence of externalities did not.
To a large extent, that is basically what does happen. Virtually all peripherals and add-on cards have onboard processors and firmware of substantial complexity presenting some sort of abstracted interface to the host. There are some minor hitches(Flash adds to BOM cost, so cheapskates build devices that depend on a host driver to feed them a firmware blob on initialization) and some more major ones.
The big, perhaps biggest, kicker is that "standard set of interface protocols" is inevitably either a slow moving mass(against which manufacturers who "just want one more little feature to enable use case X" will forever be rebelling and demanding nonstandard extensions) or a spec that is sufficiently versatile to describe almost anything, and thus describes almost nothing.
One example that springs to mind is USB_HID. For the basics, mouse, keyboard, gamepads, it works just fine. However, there are large numbers of devices that are technically "USB_HID"; but are sufficiently unlike the common use cases that "standard" USB_HID drivers make them do nothing useful. Specialized input devices, like SmartBoards, are guilty, as are such oddities as UPSes(don't ask me how a UPS is a "human interface device"; but that's what it reported itself as... USB_AVC has, to my knowledge, endured fewer outright oddities; but serious firmware bugs that break on some "standard" implementations are quite common.
It also creates a perverse incentive toward increasing the complexity(and typically the bugginess and per-manufacturer weirdness) of devices because of the need to shoehorn new sorts of peripherals into the set of "standard interfaces" understood by existing OSes.
I agree that the present state of drivers is pretty grim; but there is no easy way out.
Also, unless my inflation adjustments are wrong, 1.5 trillion in 2009 dollars is Four Times the value in 1921 dollars of the war reparations imposed on Germany by the Treaty of Versailles.
Yup, being a third-party facilitator to some file-sharing is four times as evil as WWI...
TFA is largely worthless; but EFI is actually a pretty big deal. In essence, it keeps the really hair and complex bits of the BIOS(y hello thar, ACPI, I am talking to you) and adds a giant heap of further complexity. Haven't you always wanted a BIOS that needs its own FAT32(or HFS+ in Apple's freaky nonconformant implementation) partition in order to store its own device drivers?
As with most Mac "firsts", it is and it isn't. The Gateway 610 Media Center came out with an EFI-based motheboard firmware in 2003, 3 years before Apple started shipping units with EFI. It offered no particularly compelling advantages over legacy BIOS, so there was no great rush among other manufacturers to do the same.
Apple's "first" was not doing it; but doing it exclusively across all their models.
Slashdot Mirror
The issue, in this case, is not so much proprietary forks of LLVM itself, which(as you note) are largely not worth the effort; but situations where a given architecture, say, is supported only in the form of a proprietary distribution of the LLVM core + a proprietary code generator for the given platform. You then end with a situation where the platoform vendor saved some money; but the platform is basically closed.
It's roughly analogous to being concerned about GPL compliance in companies that make routers and other little consumer electronics widgets: Such companies generally make no modifications of broad interest to the linux kernel, so their noncompliance imperils the development of the kernel not at all; but the modifications that they do make are generally vital to being able to build your own kernel for that particular device.
You aren't concerned because the WRT-54G firmware contained novel kernel innovations that will be of use to the linux kernel generally(indeed, the stock firmware was pretty unimpressive), you are concerned because you want to know what you need to know in order to build your own kernel for that device.
They rather stretch the definition of "standard"; but you can get some pretty beefy wall-warts.
The "ultra-small form factor" dell optiplexes have, historically, been powered by external power supplies(because hiding several pounds of the computer under the desk sure does make the case look smaller in the marketing glamor shots). This is a "wall wart" running a PC built with standard components(you could even get them with Prescot based Pentium Ds, back in the day, which were toasty bastards). The last such power supply I saw was rated at 220 watts, DC output, with a slightly higher draw at the wall.
Laptop chargers, for your basic cheap 15-17 inchers, are routinely in the 60-80 watt range. Your absurd desktop replacement units can push 150watts(you know that it isn't a good sign if your laptop's adapter has a 40mm fan in it...)
"s the days of GNU as the mainstream free and open development toolchain passé?"
I suspect that that will depend, in part, on how LLVM ends up being used. Since it is under a BSD-esque license, LLVM itself is definitely a candidate for being the "mainstream free and open development toolchain"; but only if the majority of real-world support scenarios don't involve proprietary actors taking advantage of that fact. In that case, it'll pretty much just end up being the core of a large number of binary, proprietary, BSPs and toolchains.
Given the good things that are said about its technical characteristics, I would hope that that doesn't happen; but the potential exists.
The fact that Bluefin are valuable has been responsible for the 80%-90% reduction in numbers; but also for the fact that people get real touchy about anything that threatens the last 10% or so.
The trouble here is that Bluefin like to go to the Gulf to spawn. If the delightful mixture of hydrocarbons and toxicologically troublesome dispersants turns out to poison eggs, sperm, or tiny juvenile fish, you could easily get an ecological impact equivalent to massive harvesting of the adult population; but without even the compensatory sushi.
Well, in this case, it is in humanity's self-interest, if nothing else, because bluefin tuna are legendarily tasty.
The ethical duties, if any, of environmental preservation are debatable. The fact that crashing the population of a species you like to eat is stupid and self-defeating isn't.
That and the fact that making boxes out of sheet metal is cheaper than making spheres out of sheet metal...
And PC-boards that aren't flat rectangles are very much special order items.(can you even get curved PCBs? There are those flexible plastic ribbon cables that sometimes have an IC or two soldered on; but I don't think I've ever seen a PCB of any nontrivial size that wasn't flat.)
If you want toolless, your best bet is actually soulless corporate drone-boxxen. They tend to be crushingly ugly; and slightly more expensive per unit spec than gamer homebuilds; but they are explicitly designed so that a monkey could replace pretty much FRU(with the possible exception of the motherboard, which is often screwed down) with its bare hands, and an experienced tech can replace all the FRUs in ~5 minutes.
Generic whitebox cases have gotten somewhat better in the toollessness department; but corporate drone-boxes have always been two steps ahead in that department.
With the increasing levels of integration(heck, you pretty much can't buy a motherboard without NIC and sound and scads of USB ports, and buying one without basic video isn't getting any easier), and the fact that we have all the lessons learned about cooling during the Prescott/space-heater era being applied to much cooler chips, the rise of mini-ITX seems like a obvious development. Multilayer PCBs aren't crazy expensive; but every square inch isn't free.
Well, Nature is a subsidiary of Macmillan, which is a subsidiary of Verlagsgruppe Georg von Holtzbrinck. I assume that it gets divided up among the shareholders in one or more of those entities.
The point isn't really to make publishing "free", because that isn't possible. Even your kid brother's myspace page costs something to keep on the web. The internet and desktop publishing software makes it cheaper; but can't entirely eradicate the cost.
The advantage, though, is that by tweaking the payment model, you can massively increase the accessibility of the research(which is arguably an ethical imperative when it is publicly funded, and a nice perk in all events), cut down on the ability of parasitic middlemen to skim money off the top, and not actually change the distribution of the cost burden all that much.
If journal subscriptions cost money; but publishing an article is free, any research institution will have to have a larger library budget(so the researchers have access to the literature) and a slightly smaller research budget(since the researchers don't pay publishing expenses), the public either gets nothing, or pays nontrivial fees for access.
If the subscription costs nothing; but publishing carries a fee, research institutions will have smaller library budgets(since they won't have to buy the journal) and slightly larger research budgets(since publishing is now a "research" expense). The public at large, and people at poor institutions, get access. The middlemen are reduced to working at cost, rather than extracting rent.
Step 1. Scientists do research(paid for largely by a mixture of tax money, and skimming from undergrads)
Step 2. Scientists write paper, submit to journal.
Step 3. Journal has other scientists(paid for by their respective universities) peer review paper for free.
Step 4. If journal decides to publish, they frequently demand copyright on paper.
Step 5. University library shells out nontrivial dead presidents so that scientists can read the papers they and their colleagues wrote.
They poison parasites, right?
The problem is not with people taking risks(well, that bothers the nanny-staters almost as much as the source of the stem cells bothers the godbots; but that isn't a big deal); but with how the sellers are representing the risks. Competent individuals choosing to take risks, or not, is freedom. Hucksters misrepresenting risks to desperate sick people is somewhere between fraud and manslaughter, depending on how it goes.
How many of these various offshore stem-cell shops fall into the following categories?
1. Scientists/research MDs whose interpretation of risk/reward tradeoffs differs from that of the FDA. In this category I would put more or less orthodox researchers who are of the position that the risks of stem cell use(cancer, infection, immune responses, etc.) are either just not that serious compared to the potential benefits and/or are the individual's choice to make.
2. Sincere cranks. In this category would go the various flavors of nutter who have gone straight off the deep end in terms of actual research about what stem cells are capable of, and how to make them do it; but are fully sincere in their belief that stem cells are the magic bullet against autism or aging, or whatever they are selling them as.
3. Cynical hucksters: All the research seriousness of the above; but without the slightly wild-eyed sincerity. However, they know that lying to desperate sick people is both easy and lucrative.
I once had a user request training when their keyboard was replaced.
To be fair, the old keyboard was a basic 101 key model, and the new one had some media control(start, stop, pause, little volume knob) buttons on it.
The user was informed, as politely as shock allowed, that the function of each key on the new keyboard was the same as that of the old, save for the additional keys, whose use was optional, and not required for the performance of any job-related function.
Gartner have a well deserved reputation for authoritative delivery of a mixture of the blatantly wrong and the painfully obvious; but they seem to be veering largely toward the second camp with this one.
Even extended enterprise support for XP isn't going to last forever, and whatever legacy crap people are worried about isn't going to become any more compatible as time passes. As for "training", home users' access to XP has been(barring active effort on their part) largely cut off for some time now, so the ones that aren't mac users at home will be getting exposed anyway.
Unless you have some scheme to drop Microsoft, it seems pretty blatantly obvious that planning for their latest flavor is your only choice at this point.
It tries; but the other devices(those that survive) complain that the 802.11 compliance of a $50 1.2 kilowatt cavity magnetron leaves something to be desired...
Anxiety has reasonably well studied chemical markers. We don't typically bother to test them; because people running to their shrink to lie about being anxious just isn't a huge problem(and, to the degree that it does happen, a lot of first line anxiolytics(lorazipam, diazapam, clonazapam, beer, etc. are cheap and not wildly harmful. Who cares if some people who want them get their hands on them?)
Panic attacks might be fakeable by a good actor(I don't know if manipulating your pulse and andrenaline levels is something they teach you in acting school; but I won't dismiss the notion out of hand); but, unless he is actually a master manipulator who has been fooling everybody for years, this guy is not a good actor.
I'd go with the "embarrassment/reprisal" hypothesis, myself.
The DoD's networks are supposed to be all secure and advanced and stuff. Getting hacked by a single sad-case foreign national, acting without support, makes them look pathetic.
When made to look pathetic, those with power generally seek reprisal against their enemies.
Frankly, the DoD was lucky to have been hacked by him. He is largely harmless, and watching how he got in was probably instructive, to some degree. They really ought to spend less time hounding him, and more time thinking about the fact that certain other hackers are much less harmless, and substantially less likely to be turned over for a stay in PMITA prison by their host governments...
That seems logical enough(and, indeed, my HID-reporting UPS shows up in GNOME panel as a battery device, complete with historical charge statistics and stuff, upon the installation of an appropriate driver).
My point was just that, as in the case of USB_HID, a standard that is too flexible ends up not defining enough to save you from hardware-specific drivers. Outside of the (mostly) safe realm of mice, keyboards, and basic gamepads, "USB_HID" isn't much more of an assurance of "no 3rd party driver needed" than "PCIe device" is.
Unfortunately, the industry of "BPA free" products sprang up at pretty much exactly the same time that the industry of BPA free products did. It turns out that printing new labels is much easier than actually reformulating your products.
Also, the "Contains BPA; but nobody except professional toxicologists studying the subject and hardcore supply chain wonks knows that" industries have been largely unaffected.
Pretty much as theory would predict, the areas closest to ideal markets with zero barriers to entry and equally informed participants achieved something close to a free market solution. The areas that deviated from those assumptions, whether by fraud, subterfuge, imperfect information, or existence of externalities did not.
To a large extent, that is basically what does happen. Virtually all peripherals and add-on cards have onboard processors and firmware of substantial complexity presenting some sort of abstracted interface to the host. There are some minor hitches(Flash adds to BOM cost, so cheapskates build devices that depend on a host driver to feed them a firmware blob on initialization) and some more major ones.
The big, perhaps biggest, kicker is that "standard set of interface protocols" is inevitably either a slow moving mass(against which manufacturers who "just want one more little feature to enable use case X" will forever be rebelling and demanding nonstandard extensions) or a spec that is sufficiently versatile to describe almost anything, and thus describes almost nothing.
One example that springs to mind is USB_HID. For the basics, mouse, keyboard, gamepads, it works just fine. However, there are large numbers of devices that are technically "USB_HID"; but are sufficiently unlike the common use cases that "standard" USB_HID drivers make them do nothing useful. Specialized input devices, like SmartBoards, are guilty, as are such oddities as UPSes(don't ask me how a UPS is a "human interface device"; but that's what it reported itself as... USB_AVC has, to my knowledge, endured fewer outright oddities; but serious firmware bugs that break on some "standard" implementations are quite common.
It also creates a perverse incentive toward increasing the complexity(and typically the bugginess and per-manufacturer weirdness) of devices because of the need to shoehorn new sorts of peripherals into the set of "standard interfaces" understood by existing OSes.
I agree that the present state of drivers is pretty grim; but there is no easy way out.
The sheer number of repetitions of "trusted computing" and "trusted platform" in that document make the hair on the back of my neck rise.
Also, unless my inflation adjustments are wrong, 1.5 trillion in 2009 dollars is Four Times the value in 1921 dollars of the war reparations imposed on Germany by the Treaty of Versailles.
Yup, being a third-party facilitator to some file-sharing is four times as evil as WWI...
TFA is largely worthless; but EFI is actually a pretty big deal. In essence, it keeps the really hair and complex bits of the BIOS(y hello thar, ACPI, I am talking to you) and adds a giant heap of further complexity. Haven't you always wanted a BIOS that needs its own FAT32(or HFS+ in Apple's freaky nonconformant implementation) partition in order to store its own device drivers?
As with most Mac "firsts", it is and it isn't. The Gateway 610 Media Center came out with an EFI-based motheboard firmware in 2003, 3 years before Apple started shipping units with EFI. It offered no particularly compelling advantages over legacy BIOS, so there was no great rush among other manufacturers to do the same.
Apple's "first" was not doing it; but doing it exclusively across all their models.