Frankly, the "thousands of years of trial and error" are largely irrelevant. Pretty much any large collection of plant extracts will include many compounds with pharmacological effects on humans, so there's a high probability that some of them will be 1) previously unknown to medical science (because medical science takes time, and it hasn't had a lot of it yet), and 2) useful.
There's no magic to traditional blah blah whatever nonsense. You assemble a great big grab-bag of random junk, and there's a good chance of something interesting being in among the rubbish.
In what way is this sudden? People have been fighting over the rhetorical effect of fonts in soft-copy documents pretty much as long as we've had raster displays, and the serious academic study of the visual-rhetorical effects of typography in digital texts goes back at least to the early '90s. (The earliest essay on the subject in Handa's Visual Rhetoric in a Digital World collection is from '93, for example.)
Fucking kids. This thing that you think is new is not new.
That was part of the joke. (And the metrics are really not the most important feature of haiku anyway; complaining about the meter is rather missing the point.) But, hey, thanks for playing.
It cost more like $600 to get an entire 20' x 60' x 14' (front) x 20' (back) building insulated with 4" thick spray-fill insulation.
That's not "a good job insulating a home". That's a good job (if it's done right) of insulating the wall cavities.
I've owned three old houses (still own two of them), and have lived in others. There's a hell of a lot more heat loss through and around old windows and doors than there is through the walls. Even my old Craftsman house in Nebraska that had uninsulated wall cavities lost a lot more heat around the windows.
And envelope gaps like that cause drafts, which contribute disproportionately to subjective temperature - people feel colder in a relatively warm room with a cold draft than they do in a well-sealed colder room. So drafts encourage people to turn the heat up significantly higher to compensate, and if the home has forced-air heating, as is common with older houses in the US, they're just pressurizing the conditioned space and wasting more conditioned air to the outside.
And yes, particularly when you're talking about old homes, decent windows are very expensive. When I get replacement windows I don't go for triple glazing or Argon fill - independent testing shows those provide little incremental benefit. But I do get wood cladding because they're going in a historic home, and more importantly every window is a custom build because none of the openings are a standard size. So I average around $1000 per unit.
I was an OS/2 developer and user from 1989 through 1998, and worked extensively with all versions from 1.1 EE on. Besides being a key platform for the commercial software package I worked on, OS/2 was my client platform of choice until it became impossible to get decent video drivers for the laptops (Thinkpads, ironically) I was issued by my employer, sometime around 1996. Even after that I continued to work on commercial software for OS/2; and since that package was also available on a host of other platforms, from MS-DOS to CICS, I had plenty to compare it with.
The SIQ was definitely a problem, both in Presentation Manager and in WPS. Yes, ultimately the blame goes to misbehaving applications; but applications do misbehave. And, true, issues with it are not common, but they're common enough to be a problem. Perhaps once a month I'd have to remotely kill[1] some OS/2 process that was hanging the GUI.
That's often enough to be a concern for some users, particularly people developing GUI software, who might accidentally introduce such a bug.
I'm glad you folks have resurrected commercial OS/2. I might even buy a copy myself, for old time's sake. (Really, $130 is a drop in the bucket; I routinely spend more than that at the hardware store.) But I think we're all better off if we can be honest about its limitations. Then people can work to alleviate them.
[1] One of the advantages of working on a middleware / application engine product was I could do pretty much anything I wanted remotely. Around 1990 I used it to build a client/server version of RCS for OS/2 - note this was around the time CVS was invented, so distributed version control systems weren't widely known on small-system platforms. Then I ported the client to the AS/400, to give us RCS on OS/400. Good times.
What professor are you talking about? Seba is a lecturer. And barely that - someone above said he was an adjunct, but from his self-promotion website it looks like he's just taught some workshops in Stanford's Continuing Ed program.
I'm closer to a fucking professor than he is - I've at least taught full-semester credit courses in actual degree programs.
Oh, and he's a "serial entrepreneur". Can't forget that.
For long-distance driving, you have the "peleton" approach, where you could easily have a a truck providing power to re-charge a series vehicle traveling together on the highway.
Yes, that's exactly what I want to do - travel down the highway coupled to a truck for half an hour. Of course the existing roads are utterly unsuited to this anyway, and most of my long-distance driving is on roads where I rarely encounter any other vehicles traveling in my direction.
Very few people drive more than 600 miles in a day non-stop.
And for those of us who do, or who drive long distances in rural areas, this is still a non-starter.
Not soon. Several times a year I drive across half the country (Eastern time zone into Mountain). Much of that is on rural roads where you're likely to see more animals than other vehicles on the road, and you can go fifty miles between human-built structures. No one is putting a charging station there any time soon.
Yes, it's likely that eventually there will be charging stations in most of the little towns that now have gas stations. But recharging time versus refueling time will still be a huge impediment to this sort of travel.
I don't expect to ever own a fully-electric vehicle. I would be hugely surprised if it becomes impractical in my lifetime to use an ICE vehicle. And if it does, I'll put a fucking generator and big propane tanks in the back of an EV pickup, and have myself a propane-electric vehicle. Screw giant battery packs and recharge stations.
Of course, I still have a hard-wired landline phone, too. Because it continues to function in prolonged power outages, after the cell tower batteries are drained. How long have experts been predicting the demise of POTS?
You think a factor of 3.7 in cost per unit distance is compelling? You're out of your mind. Here in the US there's so little market for that degree of savings that companies don't even bother selling highly efficient vehicles.
Even in some unicorn-and-fairy near-future world where electric vehicles are dominant, at 3.7x base cost ICEs would be well-suited to serve as middle-class Veblen goods, just as big SUVs and higher trim lines are now.
Essentially it means AMT will be disabled. "Provisioning" means "making AMT available for use"; unprovisioning is the the opposite.
If the Intel tool[1] shows your system is provisioned, go ahead and unprovision it. Regardless of what you might be using AMT for in your organization, it's not worth the risk.
The Intel tools[2] include some documentation which, while pretty bad, does help clarify some things.
[1] Or some alternative, if you're not running Windows; unfortunately I have no idea what the alternative might be, because Intel's communication in this issue is abysmal. Their tools are crap and their instructions are crap, even for Windows, and if you're running anything else then ha ha fuck you. Really this is an appallingly poor response by a vendor, and Intel deserve every bit of bad press and more.
[2] They can be downloaded from Intel here. Of course that page requires Javascript because it's impossible to provide downloads without Javascript. No, wait, it's because the people in charge of Intel's website are idiots.
I worked on such a system around that time (mainframe migration)
Platform Solutions?
and the Itaniums were pretty quirky compared to x86 servers.
Quirky, yes. I've worked with Itanium HP-UX systems. They're kind of a pain to debug in assembly when you don't have symbols; the "make the compiler do the work" philosophy behind VLIW does not make for happy times when staring at disassembled instructions and a list of register values.
Itanium is also not forgiving of Undefined Behavior, which is good in that it helps enforce better coding practices, but bad when it produces weird heisenbugs that take forever to track down.
An example: Itanium registers have a trap representation, a "not-a-value" value that, when encountered, generates a CPU trap. Registers are often set to this trap representation so you'll know if buggy code tries to read an uninitialized register.
The Itanium ABI puts the return value from a subroutine call in a register (as is typical). If the subroutine doesn't set a return value, that register will have whatever it had before, which may be the trap representation.
Now, say you have a piece of ancient C code that calls an external function without a declaration. Such a call implicitly has a return type int. What if the function actually is defined with no return type (a void function), but the calling code tries to examine its return value anyway? You'll get whatever old value was in the return-value register, which sometimes, depending on call path and phase of the moon, will be the trap representation. Attempting to read that will cause a CPU trap.
In HP-UX, the kernel intercepts that CPU trap and raises SIGILL. So sometimes your program will get a SIGILL, in kernel code. Nothing about those symptoms says "you called a void function without a prototype, doofus". It is, shall we say, inobvious.
A static analyzer like lint should catch that sort of thing, but if that's typical of the code you're dealing with, your analyzer will probably find a million suspect constructs. Sure, in an ideal world you'd fix them all, but it'll be a while before you correct the cause of that intermittent SIGILL - and you still won't know why.
Not that I'm bitter. Really, a trap representation in a register is basically a good idea. It's just HP-UX handles it poorly. The AS/400 did a much better job with this sort of thing.
Eh? USB 1.0 and 2.0 don't give the device DMA (for most protocols). The USB controller has DMA, and a vulnerability in the controller firmware could be exploited by a device - I'm pretty sure there are real-world examples. (There have been USB controller bugs that provided RCE.)
But it doesn't really matter. While external-bus-with-DMA is a whopping great security hole and a profoundly Bad Idea - and yet another example of why the people who design such things badly need remedial security education - even non-DMA interfaces like USB 1.0 and 2.0 provide a huge attack surface, because of how they're treated by the OS.
Common OSes helpfully provide things like automounting filesystems, and that's a big pile of code that's likely to have exploitable vulnerabilities. Or they treat devices that claim to be, say, keyboards (USB Rubber Ducky) or network cards (Fuller's attack) as completely trustworthy and assume that such devices are working for the benefit of the user.
The oft-cited dictum that physical control of a machine completely defeats security is naive and patently untrue - it's very difficult to defeat tamper-resistant hardware, for example. But It does provide a far more generous attack surface. And while external DMA-enabled ports are a sweet gift to an attacker, plain ol' USB is a pretty big hole too.
I don't care if telecoms lock the devices they subsidize. The problem for me is that I want a rather rare combination of hardware features, and it's hard to find phones that provide it. Usually the ones that do are old models customized for some telecom that were remaindered and dumped on the grey market.
That sort of phone is unsubsidized, so there's no reason they should be locked; but since they were originally intended to be sold subsidized they are.
My worry is that the selection of devices that are sold full-price, unlocked, and rootable will be limited, and I won't be able to get one with the feature set I want.
But I suspect in not too many years I'll be moving back to a feature phone anyway. Smartphones are steadily getting less palatable and Fuchsia looks like a big step in that direction.
Ajit should not be attacking free political speech, he is already in trouble killing Net Neutrality. He needs to apologise.
Dream on. He's been a tool of his paymasters since he was a Verizon lawyer. Dude puts the "lack" (as in "... of ethics", "... of character", "... of original thought") in "lackey".
Yes. And every case in which an FCC indecency fine is shot down is another blow to the foolish 1978 Pacifica decision, the one where SCOTUS decided that broadcast media are "uniquely intrusive" and so don't deserve full First Amendment protection.
That decision has been roundly criticized from the moment it was written, and the Roberts court has generally not viewed it kindly (particularly not when the government pulls out the "think of the children" argument, as in Entertainment Merchants Association). They voided FCC fines in 2012, in FCC v. Fox, and they might very well do so again.
There's no way to know what the author meant. After "the freezing Antarctic" (was it liquid recently? just letting us know it's cold there?), I think it's safe to say this is not someone you'd be looking to for any sort of technical accuracy. We're probably lucky the article doesn't claim elves will keep it cold. (And what they hell is the antecedent of the plural pronoun "their" in that sentence - "fifths"? Yes, the vast density of those fifths!)
In any case, it's really water's enthalpy of melting that matters here, as you say. My guess is the author of the article has never heard of enthalpy.
Plus the set of keywords is small and easy to learn anyway
Yes, and they have nothing to do with English grammar, or word inflections, or orthography (once you memorize that small set of keywords), or any of the other things that are important to learning English as a natural language.
The relevance of programming-language keywords to learning English is scarcely more the relevance of Greek letters in mathematical notation is to learning Greek.
Now, learning to read comments written in English in source code has some relevance to learning programming, since so much extant source has comments in English; and so some extent the same can be said for identifiers. But again there's a mighty gulf between having basic reading competency in a language and being able to speak or write it. I've had reading courses in French, but I never learned to speak it or even really understand any of the spoken language; and similarly the two years I had of Japanese made me much better at reading it than writing or speaking.
Slashdot Mirror
Frankly, the "thousands of years of trial and error" are largely irrelevant. Pretty much any large collection of plant extracts will include many compounds with pharmacological effects on humans, so there's a high probability that some of them will be 1) previously unknown to medical science (because medical science takes time, and it hasn't had a lot of it yet), and 2) useful.
There's no magic to traditional blah blah whatever nonsense. You assemble a great big grab-bag of random junk, and there's a good chance of something interesting being in among the rubbish.
In what way is this sudden? People have been fighting over the rhetorical effect of fonts in soft-copy documents pretty much as long as we've had raster displays, and the serious academic study of the visual-rhetorical effects of typography in digital texts goes back at least to the early '90s. (The earliest essay on the subject in Handa's Visual Rhetoric in a Digital World collection is from '93, for example.)
Fucking kids. This thing that you think is new is not new.
That was part of the joke. (And the metrics are really not the most important feature of haiku anyway; complaining about the meter is rather missing the point.) But, hey, thanks for playing.
Which IT jobs will last forever?
None. Things change. Hell, most of 'em will be gone by the time the sun finishes boiling off the oceans.
(And, yes, the article is a pile of rubbish, but since the premise is idiotic, that hardly matters.)
What is your obsession with hyphenating "asshole"? Are you... are you from the past?
God damned time travelers taking our flame-war jobs.
Man, people will try to pass anything off as haiku.
It cost more like $600 to get an entire 20' x 60' x 14' (front) x 20' (back) building insulated with 4" thick spray-fill insulation.
That's not "a good job insulating a home". That's a good job (if it's done right) of insulating the wall cavities.
I've owned three old houses (still own two of them), and have lived in others. There's a hell of a lot more heat loss through and around old windows and doors than there is through the walls. Even my old Craftsman house in Nebraska that had uninsulated wall cavities lost a lot more heat around the windows.
And envelope gaps like that cause drafts, which contribute disproportionately to subjective temperature - people feel colder in a relatively warm room with a cold draft than they do in a well-sealed colder room. So drafts encourage people to turn the heat up significantly higher to compensate, and if the home has forced-air heating, as is common with older houses in the US, they're just pressurizing the conditioned space and wasting more conditioned air to the outside.
And yes, particularly when you're talking about old homes, decent windows are very expensive. When I get replacement windows I don't go for triple glazing or Argon fill - independent testing shows those provide little incremental benefit. But I do get wood cladding because they're going in a historic home, and more importantly every window is a custom build because none of the openings are a standard size. So I average around $1000 per unit.
Then I install them myself, because I'm cheap.
There never was an SIQ problem in OS/2.
I'm afraid I have to disagree.
I was an OS/2 developer and user from 1989 through 1998, and worked extensively with all versions from 1.1 EE on. Besides being a key platform for the commercial software package I worked on, OS/2 was my client platform of choice until it became impossible to get decent video drivers for the laptops (Thinkpads, ironically) I was issued by my employer, sometime around 1996. Even after that I continued to work on commercial software for OS/2; and since that package was also available on a host of other platforms, from MS-DOS to CICS, I had plenty to compare it with.
The SIQ was definitely a problem, both in Presentation Manager and in WPS. Yes, ultimately the blame goes to misbehaving applications; but applications do misbehave. And, true, issues with it are not common, but they're common enough to be a problem. Perhaps once a month I'd have to remotely kill[1] some OS/2 process that was hanging the GUI.
That's often enough to be a concern for some users, particularly people developing GUI software, who might accidentally introduce such a bug.
I'm glad you folks have resurrected commercial OS/2. I might even buy a copy myself, for old time's sake. (Really, $130 is a drop in the bucket; I routinely spend more than that at the hardware store.) But I think we're all better off if we can be honest about its limitations. Then people can work to alleviate them.
[1] One of the advantages of working on a middleware / application engine product was I could do pretty much anything I wanted remotely. Around 1990 I used it to build a client/server version of RCS for OS/2 - note this was around the time CVS was invented, so distributed version control systems weren't widely known on small-system platforms. Then I ported the client to the AS/400, to give us RCS on OS/400. Good times.
Slashdot headlines suddenly become misleading drivel!
Perhaps "suddenly" is now used to mean "gradually", just as "literally" is used for "figuratively".
What professor are you talking about? Seba is a lecturer. And barely that - someone above said he was an adjunct, but from his self-promotion website it looks like he's just taught some workshops in Stanford's Continuing Ed program.
I'm closer to a fucking professor than he is - I've at least taught full-semester credit courses in actual degree programs.
Oh, and he's a "serial entrepreneur". Can't forget that.
I know, I know - research is hard.
For long-distance driving, you have the "peleton" approach, where you could easily have a a truck providing power to re-charge a series vehicle traveling together on the highway.
Yes, that's exactly what I want to do - travel down the highway coupled to a truck for half an hour. Of course the existing roads are utterly unsuited to this anyway, and most of my long-distance driving is on roads where I rarely encounter any other vehicles traveling in my direction.
Very few people drive more than 600 miles in a day non-stop.
And for those of us who do, or who drive long distances in rural areas, this is still a non-starter.
Not soon. Several times a year I drive across half the country (Eastern time zone into Mountain). Much of that is on rural roads where you're likely to see more animals than other vehicles on the road, and you can go fifty miles between human-built structures. No one is putting a charging station there any time soon.
Yes, it's likely that eventually there will be charging stations in most of the little towns that now have gas stations. But recharging time versus refueling time will still be a huge impediment to this sort of travel.
I don't expect to ever own a fully-electric vehicle. I would be hugely surprised if it becomes impractical in my lifetime to use an ICE vehicle. And if it does, I'll put a fucking generator and big propane tanks in the back of an EV pickup, and have myself a propane-electric vehicle. Screw giant battery packs and recharge stations.
Of course, I still have a hard-wired landline phone, too. Because it continues to function in prolonged power outages, after the cell tower batteries are drained. How long have experts been predicting the demise of POTS?
If tenor stops, what will we do with our vehicles?
You think a factor of 3.7 in cost per unit distance is compelling? You're out of your mind. Here in the US there's so little market for that degree of savings that companies don't even bother selling highly efficient vehicles.
Even in some unicorn-and-fairy near-future world where electric vehicles are dominant, at 3.7x base cost ICEs would be well-suited to serve as middle-class Veblen goods, just as big SUVs and higher trim lines are now.
Essentially it means AMT will be disabled. "Provisioning" means "making AMT available for use"; unprovisioning is the the opposite.
If the Intel tool[1] shows your system is provisioned, go ahead and unprovision it. Regardless of what you might be using AMT for in your organization, it's not worth the risk.
The Intel tools[2] include some documentation which, while pretty bad, does help clarify some things.
[1] Or some alternative, if you're not running Windows; unfortunately I have no idea what the alternative might be, because Intel's communication in this issue is abysmal. Their tools are crap and their instructions are crap, even for Windows, and if you're running anything else then ha ha fuck you. Really this is an appallingly poor response by a vendor, and Intel deserve every bit of bad press and more.
[2] They can be downloaded from Intel here. Of course that page requires Javascript because it's impossible to provide downloads without Javascript. No, wait, it's because the people in charge of Intel's website are idiots.
A fine post. If I had mod points at the moment, I'd probably discard my posts on this article just to mod it up.
I worked on such a system around that time (mainframe migration)
Platform Solutions?
and the Itaniums were pretty quirky compared to x86 servers.
Quirky, yes. I've worked with Itanium HP-UX systems. They're kind of a pain to debug in assembly when you don't have symbols; the "make the compiler do the work" philosophy behind VLIW does not make for happy times when staring at disassembled instructions and a list of register values.
Itanium is also not forgiving of Undefined Behavior, which is good in that it helps enforce better coding practices, but bad when it produces weird heisenbugs that take forever to track down.
An example: Itanium registers have a trap representation, a "not-a-value" value that, when encountered, generates a CPU trap. Registers are often set to this trap representation so you'll know if buggy code tries to read an uninitialized register.
The Itanium ABI puts the return value from a subroutine call in a register (as is typical). If the subroutine doesn't set a return value, that register will have whatever it had before, which may be the trap representation.
Now, say you have a piece of ancient C code that calls an external function without a declaration. Such a call implicitly has a return type int. What if the function actually is defined with no return type (a void function), but the calling code tries to examine its return value anyway? You'll get whatever old value was in the return-value register, which sometimes, depending on call path and phase of the moon, will be the trap representation. Attempting to read that will cause a CPU trap.
In HP-UX, the kernel intercepts that CPU trap and raises SIGILL. So sometimes your program will get a SIGILL, in kernel code. Nothing about those symptoms says "you called a void function without a prototype, doofus". It is, shall we say, inobvious.
A static analyzer like lint should catch that sort of thing, but if that's typical of the code you're dealing with, your analyzer will probably find a million suspect constructs. Sure, in an ideal world you'd fix them all, but it'll be a while before you correct the cause of that intermittent SIGILL - and you still won't know why.
Not that I'm bitter. Really, a trap representation in a register is basically a good idea. It's just HP-UX handles it poorly. The AS/400 did a much better job with this sort of thing.
Eh? USB 1.0 and 2.0 don't give the device DMA (for most protocols). The USB controller has DMA, and a vulnerability in the controller firmware could be exploited by a device - I'm pretty sure there are real-world examples. (There have been USB controller bugs that provided RCE.)
But it doesn't really matter. While external-bus-with-DMA is a whopping great security hole and a profoundly Bad Idea - and yet another example of why the people who design such things badly need remedial security education - even non-DMA interfaces like USB 1.0 and 2.0 provide a huge attack surface, because of how they're treated by the OS.
Common OSes helpfully provide things like automounting filesystems, and that's a big pile of code that's likely to have exploitable vulnerabilities. Or they treat devices that claim to be, say, keyboards (USB Rubber Ducky) or network cards (Fuller's attack) as completely trustworthy and assume that such devices are working for the benefit of the user.
The oft-cited dictum that physical control of a machine completely defeats security is naive and patently untrue - it's very difficult to defeat tamper-resistant hardware, for example. But It does provide a far more generous attack surface. And while external DMA-enabled ports are a sweet gift to an attacker, plain ol' USB is a pretty big hole too.
I don't care if telecoms lock the devices they subsidize. The problem for me is that I want a rather rare combination of hardware features, and it's hard to find phones that provide it. Usually the ones that do are old models customized for some telecom that were remaindered and dumped on the grey market.
That sort of phone is unsubsidized, so there's no reason they should be locked; but since they were originally intended to be sold subsidized they are.
My worry is that the selection of devices that are sold full-price, unlocked, and rootable will be limited, and I won't be able to get one with the feature set I want.
But I suspect in not too many years I'll be moving back to a feature phone anyway. Smartphones are steadily getting less palatable and Fuchsia looks like a big step in that direction.
the crappy 30 fps that people seem to think is "good enough."
Hell, as far as I'm concerned, 30 characters per second is good enough. Why the fuck would I need my phone to do 60+ frames per second?
On the other hand, it's pretty clear I can ignore Fuchsia forever. All the evidence so far suggests it has absolutely nothing I want.
Ajit should not be attacking free political speech, he is already in trouble killing Net Neutrality. He needs to apologise.
Dream on. He's been a tool of his paymasters since he was a Verizon lawyer. Dude puts the "lack" (as in "... of ethics", "... of character", "... of original thought") in "lackey".
Yes. And every case in which an FCC indecency fine is shot down is another blow to the foolish 1978 Pacifica decision, the one where SCOTUS decided that broadcast media are "uniquely intrusive" and so don't deserve full First Amendment protection.
That decision has been roundly criticized from the moment it was written, and the Roberts court has generally not viewed it kindly (particularly not when the government pulls out the "think of the children" argument, as in Entertainment Merchants Association). They voided FCC fines in 2012, in FCC v. Fox, and they might very well do so again.
That and/or melting enthalpy.
There's no way to know what the author meant. After "the freezing Antarctic" (was it liquid recently? just letting us know it's cold there?), I think it's safe to say this is not someone you'd be looking to for any sort of technical accuracy. We're probably lucky the article doesn't claim elves will keep it cold. (And what they hell is the antecedent of the plural pronoun "their" in that sentence - "fifths"? Yes, the vast density of those fifths!)
In any case, it's really water's enthalpy of melting that matters here, as you say. My guess is the author of the article has never heard of enthalpy.
Yes, but Exchange everything is atrocious, so users quickly become accustomed to it.
Exchange: Managing user expectations since 1993.
Plus the set of keywords is small and easy to learn anyway
Yes, and they have nothing to do with English grammar, or word inflections, or orthography (once you memorize that small set of keywords), or any of the other things that are important to learning English as a natural language.
The relevance of programming-language keywords to learning English is scarcely more the relevance of Greek letters in mathematical notation is to learning Greek.
Now, learning to read comments written in English in source code has some relevance to learning programming, since so much extant source has comments in English; and so some extent the same can be said for identifiers. But again there's a mighty gulf between having basic reading competency in a language and being able to speak or write it. I've had reading courses in French, but I never learned to speak it or even really understand any of the spoken language; and similarly the two years I had of Japanese made me much better at reading it than writing or speaking.