You know, there ought to be some status in law, where a person's opinion is no longer regarded as more meaningful than purely entropic noise. Lets call it cum granum salis. A person having this status could neither be sole plaintiff nor sole witness in a legal case. If they wanted to bring suit, they'd have to have a guarantor willing to swear to their truthfulness - who would risk having their own opinions taken cum granum salis if the case were subsequently thrown out for lack of merit.
The first recipients should probably be Jack Thompson, and SCO.
Not sure if Lisp is really the tool for the task. Typed-data, untyped-variable languages are intrinsically slow. They have to re-check their assumptions nearly every time they touch storage. Even advanced Lisp compilers don't improve this much.
My own recommendation would be a mix of asm (low-level optimized stuff), Ocaml(engine) and lua (control and configuration).
You're right about one thing though: C++ is a fucking blight. It's used because "it's the standard, it's fast, and you can hire developers". To which the obvious answer is: none of the above matters if you can't use it to deliver the goods! And you can't.
Me, I'm a libertarian anarchist, which is to say about as utterly cold-blooded purist free-trader as it gets. And frankly, to me, the Chinese have the moral high ground.
"How dare you make a product competing with our government granted monopoly" - is not the sentiment of a free trader.
Of course now the EU will get in on the act, announce a product named "framboise", with the contract parceled out in a rigged open tender to a semi-nationalized French firm which just happens to have a bunch of EC bureaucrats owning shares...
Taking them out of the core - or in any other way causing them to languish - is as bad as deleting them. They're useful because they're kept current, and kept in sync with each other and with the core.
I value their large repository for its well-organized dependency tracking and the fact that packages have all been tested for mutual compatibility.
I don't care how often "stable" releases. I track "testing" with frequent dist-upgrades on my desktop machines, and on servers I'd not worry if "stable" was a bit long in the tooth.
Throwing away the packages to get a rapid release cycle would be a bad bargain for me.
I have cogent and realistic proposal for the software patent system, namely shut it down. Entirely, immediately, and without compensation. Or perhaps you meant "a wimpy compromise that papers over the cracks and appeases vested interests"? Software patents are conceptually broken and no half-solution will improve them.
A dog that sick, the only mercy is to shoot it quickly.
...that turning on you computer's networking and opening a hole in the firewall are two seperate operations.
I can see the conceptual theory, but still, it's bad design. Typical programmer-invented UI. Their mistake is "lets force the user to make the coder's job easy" and "implementation detail leaking out of the abstraction". We implemented TCP and access control seperately, so we configure them seperately.
A better way to do it would be to make the networking switch-on implicit: if you open a hole for TCP in the access file, this causes TCP to be enabled on the server. Therefore configuration follows interface ("these networking modes are supported, make it happen") not implementation.
It's not necessarily the case that companies opposing patents should also avoid using them. It's also a consistent, non-hypocritical position to:
1) use them, because turning down business advantage is stupid, and because your arm will be twisted with other people's patents unless you have enough of your own to twist back
2) oppose them, because it's a whole lot of effort spent basically trying to manoeuver the government to twist each other's arms, and business would be so much simpler and more profitable if nobody had to do that.
...every regulation creates a cartel of businesses who can afford to comply, a bureaucracy who decide when to apply it (usually "captured" by the cartel), and a lobby of vested interests. All of whom have far more personal incentive to defend a harmful regulation than anyone else has to repeal it. All of whom have selfish incentives to make the regulatory regime more complex and burdensome.
There is no such thing as a good or harmless regulation.
It seems to me that the utility range for embedded linux is narrow. If you're running on a full PC platform and presenting a UI on a monitor, or doing a task for which an off-the-peg solution exists, you'll want and chances are you'll have no other sensible choice but to use windows (eg: cash machines, point-of-sale). If you're doing hard realtime or running on a very small device, Linux won't cut it. Palmtops are suited to running Linux, but the users probably don't want it. Embedded Linux seems to me to have staked out a niche in dedicated but soft-realtime devices that are sub-PC but still pretty beefy. Examples: DVRs, MP3 players, high-end phones.
Nine-tenths of embedded anything will now and forever be running hand-hacked C or assembler over bare metal, or an "OS" that amounts to little more than a runloop and a deterministic thread-scheduler.
The "GNU/Linux" thing has always stuck me as self-centered. I mean, sure it's both GNU and Linux, but... A name is not a chemical formula out of which a thing's structure can be parsed. Otherwise my OS would be something more like OnceKnoppix SortaDebianTesting Gnu/Linux/bitsOfBSD/Xorg/KDE/SunJava/OpenOffice.or g... and it wouldn't stop there. Heck I could just dump a list of my apt packages, their repository and version, but even that would be incomplete as it would miss out historical influences since purged.
At which point does a name come to encompass the totality of elapsed events since absolute tick zero?
I'll continue calling it "Linux" or maybe "Debian testing", because that's good enough and does nobody any special favours.
The trouble is that Java is literally the best tool for the task. It occupies a sort of "local best" in the phase-space of programming langauge features, when considered for use as a cross-platform userspace network stack.
- Perl, Python: nice library, too slow, come standard on unix but requires installing a seperate and hairy interpreter on windows. Difficult to use for large programs or large teams.
- C, C++: impossible to make secure and difficult to debug, insufficient abstraction, hairy library, barely modular, cross platform only through the messy hack of conditional compilation.
- D, Ocaml,Ada, Haskell: secure, obscure, no library, not really cross platform enough.
- Java: huge standard library and zillions of free drop-in modules, comfortable level of abstraction, modular and suitable for large programs, cross platform enough, fast enough, simple enough, popular enough, secure enough, requires an interpreter but most everyone already has it.
In this new Freenet, network connections only pass through a select few friends, but the routing layer hides this - files are globally available, as they used to be. You've misunderstood the protocol design.
Also, you've even misunderstood the "select few friends" thing. It's not that you can exclude people. It's that you have to actively include people - and you have to have their permission first.
An analogy would be: passing messages between people by telling a trusted friend, he tells his trusted friend, and so on until it reaches the destination.
Not to mention, KDE's window manager has active focus-stealing prevention. Linux apps can misbehave, but the WM will slap them down and refuse to comply.
It really does matter if you "own" the OS or merely "rent" it.
It's not just naked human beings. It's guys cumming on womens faces saying, "Take that bitch, want some more?" This kind of material can be very harmful to kids.
This idea that people (not just kids) sponge up bad attitudes from role models in the media (and that includes porn) is really pervasive throughout society, but I've seen no evidence for it. In fact I'd call the counter-evidence pretty strong. How many efforts to wield this "power" (for good, "good", and plain evil) have failed? Remember Eminem's "mosh"? He rocked the other guy's vote. Not an isolated incident.
Truth is, I just don't believe that any human learns bad attitudes from the media accidentally. It's a matter of (instinctual) intent, and not at all passive. Kids deliberately pick up and adopt the pervasive mores of their culture - and in most cases, just as deliberately reject what society sees as taboo, quickly picking up which media messages slot into which category.
I think where folks get confused is: kids learn actions as actions, words as words. So if you have a culture that talks taboo about sex but doesn't act taboo, the kids will copy both halves. Thus the wierd American culture of the prurient priss rolls on into another generation.
The core problem of funding is that a government space program - NASA - doesn't get its' money back. For a repeat performance, it has to tax people all over again, and there are political and economic limits.
A private, commercial space program is predicated on profit. It may be a bit slow starting, but if properly run, it makes back more than the money it invests. The next time, it can go further. There is no theoretical ceiling on the funding of a for-profit space program, so long as it continues turning a profit.
NASA is not merely underfunded. It's impossible to fund. Only private spaceflight will get us off this rock.
Would i ever buy a PDF? Maybe. Would I ever buy a DRM'd book? Not if my life depended on it. A book with worse limitations than paper is not useful, now or ever.
Would you hear yourself! "We wuz robbed" my left foot. You were defeated by a small but solid margin in an election no more corrupt than normal - probably less so, given the high turnout and unusually close scrutiny.
Learn the truth: politics has always been a scam at the expense of the regular people since the days of ancient Rome and beyond, in way that goes far beyond party affiliation. Power simply makes hanging onto power easier.
You want less dishonesty in politics? Work to reduce the scope and scale the state. If excess power's there to wield, it will be abused, count on it.
Quantum Crypto was never about "oh no, the alarm tripped, someone's snooping". It was always about "assume some bits are snooped, some aren't, find and discard the snooped bits and just use the secret ones". It's a way of sharing key or one-time-pad material, not of passing meaningful data.
Paraphrasing from what someone said about Fortran: I don't know how people will store data in fifty years time, but it will involve UTF8, XML and PDF. They are all three category killers for a particular task.
It's a good tool, but for a different task. DjVu is an image compression format for images that can be cleanly defined as "important foreground, unimportant background". Therefore, it's ideal for scanned documents. PDF can contain scanned pages as images, but it's inefficient. However, most PDFs aren't full-page images, they're vector graphics and vector text. PDFs like that are very efficient, more than DjVu, but necessarily of digital origin. DjVu is really an archival tool for eg: scanned historical manuscripts, where you want the text pin-sharp, but the blotchy paper can safely be lo-rez.
...this is a government fishing expedition, and it's not even fishing for crime, it's fishing for data from which they suppose they might be able to theorize harm, and legislate a new crime.
In reality, they just want to believe that harm is somehow being done. They aren't after evidence or scientific proof. They're after data that can be munged to confirm their biases and those of their constituents.
For the record: in my own opinion images of sex, even wild and kinky sex, do not harm kids - and probably don't even much interest them except for snigger-value. All this fuss is saying much more about the repressed prurience of the more nutty kind of "cultural conservative" than about any scientific reality.
One thing I've thought of that these things would be useful for would be as "couriers" for important documents that can't be sent electronically. Contracts, etc. Possibly, donor organs. Other small packages that currently get rushed on aircraft, but would be better yet if they arrived with almost the speed of an email.
That seems like the next step up after there-and-back-again tourist flights.
Slashdot Mirror
Q: How many biomechanical engineers does it take to replace a wafer-thin oled sheet?
A: Only one, but you'd have to put an awful lot of voltage across him to get him to glow properly.
You know, there ought to be some status in law, where a person's opinion is no longer regarded as more meaningful than purely entropic noise. Lets call it cum granum salis. A person having this status could neither be sole plaintiff nor sole witness in a legal case. If they wanted to bring suit, they'd have to have a guarantor willing to swear to their truthfulness - who would risk having their own opinions taken cum granum salis if the case were subsequently thrown out for lack of merit.
The first recipients should probably be Jack Thompson, and SCO.
Not sure if Lisp is really the tool for the task. Typed-data, untyped-variable languages are intrinsically slow. They have to re-check their assumptions nearly every time they touch storage. Even advanced Lisp compilers don't improve this much.
My own recommendation would be a mix of asm (low-level optimized stuff), Ocaml(engine) and lua (control and configuration).
You're right about one thing though: C++ is a fucking blight. It's used because "it's the standard, it's fast, and you can hire developers". To which the obvious answer is: none of the above matters if you can't use it to deliver the goods! And you can't.
Me, I'm a libertarian anarchist, which is to say about as utterly cold-blooded purist free-trader as it gets. And frankly, to me, the Chinese have the moral high ground.
"How dare you make a product competing with our government granted monopoly" - is not the sentiment of a free trader.
Of course now the EU will get in on the act, announce a product named "framboise", with the contract parceled out in a rigged open tender to a semi-nationalized French firm which just happens to have a bunch of EC bureaucrats owning shares...
Taking them out of the core - or in any other way causing them to languish - is as bad as deleting them. They're useful because they're kept current, and kept in sync with each other and with the core.
I value their large repository for its well-organized dependency tracking and the fact that packages have all been tested for mutual compatibility.
I don't care how often "stable" releases. I track "testing" with frequent dist-upgrades on my desktop machines, and on servers I'd not worry if "stable" was a bit long in the tooth.
Throwing away the packages to get a rapid release cycle would be a bad bargain for me.
I have cogent and realistic proposal for the software patent system, namely shut it down. Entirely, immediately, and without compensation. Or perhaps you meant "a wimpy compromise that papers over the cracks and appeases vested interests"? Software patents are conceptually broken and no half-solution will improve them.
A dog that sick, the only mercy is to shoot it quickly.
...that turning on you computer's networking and opening a hole in the firewall are two seperate operations.
I can see the conceptual theory, but still, it's bad design. Typical programmer-invented UI. Their mistake is "lets force the user to make the coder's job easy" and "implementation detail leaking out of the abstraction". We implemented TCP and access control seperately, so we configure them seperately.
A better way to do it would be to make the networking switch-on implicit: if you open a hole for TCP in the access file, this causes TCP to be enabled on the server. Therefore configuration follows interface ("these networking modes are supported, make it happen") not implementation.
It's not necessarily the case that companies opposing patents should also avoid using them. It's also a consistent, non-hypocritical position to:
1) use them, because turning down business advantage is stupid, and because your arm will be twisted with other people's patents unless you have enough of your own to twist back
2) oppose them, because it's a whole lot of effort spent basically trying to manoeuver the government to twist each other's arms, and business would be so much simpler and more profitable if nobody had to do that.
...every regulation creates a cartel of businesses who can afford to comply, a bureaucracy who decide when to apply it (usually "captured" by the cartel), and a lobby of vested interests. All of whom have far more personal incentive to defend a harmful regulation than anyone else has to repeal it. All of whom have selfish incentives to make the regulatory regime more complex and burdensome.
There is no such thing as a good or harmless regulation.
It seems to me that the utility range for embedded linux is narrow. If you're running on a full PC platform and presenting a UI on a monitor, or doing a task for which an off-the-peg solution exists, you'll want and chances are you'll have no other sensible choice but to use windows (eg: cash machines, point-of-sale). If you're doing hard realtime or running on a very small device, Linux won't cut it. Palmtops are suited to running Linux, but the users probably don't want it. Embedded Linux seems to me to have staked out a niche in dedicated but soft-realtime devices that are sub-PC but still pretty beefy. Examples: DVRs, MP3 players, high-end phones.
Nine-tenths of embedded anything will now and forever be running hand-hacked C or assembler over bare metal, or an "OS" that amounts to little more than a runloop and a deterministic thread-scheduler.
The "GNU/Linux" thing has always stuck me as self-centered. I mean, sure it's both GNU and Linux, but... A name is not a chemical formula out of which a thing's structure can be parsed. Otherwise my OS would be something more like OnceKnoppix SortaDebianTesting Gnu/Linux/bitsOfBSD/Xorg/KDE/SunJava/OpenOffice.or g... and it wouldn't stop there. Heck I could just dump a list of my apt packages, their repository and version, but even that would be incomplete as it would miss out historical influences since purged.
At which point does a name come to encompass the totality of elapsed events since absolute tick zero?
I'll continue calling it "Linux" or maybe "Debian testing", because that's good enough and does nobody any special favours.
The trouble is that Java is literally the best tool for the task. It occupies a sort of "local best" in the phase-space of programming langauge features, when considered for use as a cross-platform userspace network stack.
- Perl, Python: nice library, too slow, come standard on unix but requires installing a seperate and hairy interpreter on windows. Difficult to use for large programs or large teams.
- C, C++: impossible to make secure and difficult to debug, insufficient abstraction, hairy library, barely modular, cross platform only through the messy hack of conditional compilation.
- D, Ocaml,Ada, Haskell: secure, obscure, no library, not really cross platform enough.
- Java: huge standard library and zillions of free drop-in modules, comfortable level of abstraction, modular and suitable for large programs, cross platform enough, fast enough, simple enough, popular enough, secure enough, requires an interpreter but most everyone already has it.
In this new Freenet, network connections only pass through a select few friends, but the routing layer hides this - files are globally available, as they used to be. You've misunderstood the protocol design.
Also, you've even misunderstood the "select few friends" thing. It's not that you can exclude people. It's that you have to actively include people - and you have to have their permission first.
An analogy would be: passing messages between people by telling a trusted friend, he tells his trusted friend, and so on until it reaches the destination.
Not to mention, KDE's window manager has active focus-stealing prevention. Linux apps can misbehave, but the WM will slap them down and refuse to comply.
It really does matter if you "own" the OS or merely "rent" it.
This idea that people (not just kids) sponge up bad attitudes from role models in the media (and that includes porn) is really pervasive throughout society, but I've seen no evidence for it. In fact I'd call the counter-evidence pretty strong. How many efforts to wield this "power" (for good, "good", and plain evil) have failed? Remember Eminem's "mosh"? He rocked the other guy's vote. Not an isolated incident.
Truth is, I just don't believe that any human learns bad attitudes from the media accidentally. It's a matter of (instinctual) intent, and not at all passive. Kids deliberately pick up and adopt the pervasive mores of their culture - and in most cases, just as deliberately reject what society sees as taboo, quickly picking up which media messages slot into which category.
I think where folks get confused is: kids learn actions as actions, words as words. So if you have a culture that talks taboo about sex but doesn't act taboo, the kids will copy both halves. Thus the wierd American culture of the prurient priss rolls on into another generation.
The core problem of funding is that a government space program - NASA - doesn't get its' money back. For a repeat performance, it has to tax people all over again, and there are political and economic limits.
A private, commercial space program is predicated on profit. It may be a bit slow starting, but if properly run, it makes back more than the money it invests. The next time, it can go further. There is no theoretical ceiling on the funding of a for-profit space program, so long as it continues turning a profit.
NASA is not merely underfunded. It's impossible to fund. Only private spaceflight will get us off this rock.
Baen if I'm nice and and amule if I'm nasty.
Would i ever buy a PDF? Maybe. Would I ever buy a DRM'd book? Not if my life depended on it. A book with worse limitations than paper is not useful, now or ever.
Would you hear yourself! "We wuz robbed" my left foot. You were defeated by a small but solid margin in an election no more corrupt than normal - probably less so, given the high turnout and unusually close scrutiny.
Learn the truth: politics has always been a scam at the expense of the regular people since the days of ancient Rome and beyond, in way that goes far beyond party affiliation. Power simply makes hanging onto power easier.
You want less dishonesty in politics? Work to reduce the scope and scale the state. If excess power's there to wield, it will be abused, count on it.
Quantum Crypto was never about "oh no, the alarm tripped, someone's snooping". It was always about "assume some bits are snooped, some aren't, find and discard the snooped bits and just use the secret ones". It's a way of sharing key or one-time-pad material, not of passing meaningful data.
IOW, the "attack" is meaningless.
Paraphrasing from what someone said about Fortran: I don't know how people will store data in fifty years time, but it will involve UTF8, XML and PDF. They are all three category killers for a particular task.
It's a good tool, but for a different task. DjVu is an image compression format for images that can be cleanly defined as "important foreground, unimportant background". Therefore, it's ideal for scanned documents. PDF can contain scanned pages as images, but it's inefficient. However, most PDFs aren't full-page images, they're vector graphics and vector text. PDFs like that are very efficient, more than DjVu, but necessarily of digital origin. DjVu is really an archival tool for eg: scanned historical manuscripts, where you want the text pin-sharp, but the blotchy paper can safely be lo-rez.
...this is a government fishing expedition, and it's not even fishing for crime, it's fishing for data from which they suppose they might be able to theorize harm, and legislate a new crime.
In reality, they just want to believe that harm is somehow being done. They aren't after evidence or scientific proof. They're after data that can be munged to confirm their biases and those of their constituents.
For the record: in my own opinion images of sex, even wild and kinky sex, do not harm kids - and probably don't even much interest them except for snigger-value. All this fuss is saying much more about the repressed prurience of the more nutty kind of "cultural conservative" than about any scientific reality.
One thing I've thought of that these things would be useful for would be as "couriers" for important documents that can't be sent electronically. Contracts, etc. Possibly, donor organs. Other small packages that currently get rushed on aircraft, but would be better yet if they arrived with almost the speed of an email.
That seems like the next step up after there-and-back-again tourist flights.