There's a difference between not protecting your child every moment of every day and pure neglect. If you see that they're hurting themselves and will continue to do so until someone intervenes, and yet you do nothing...well, I hardly need to tell you which that is.
Isn't it a bloody obvious method for installing an operating system? Copy the image over, apply some hardware configuration scripts (and whatever else you need, such as default packages).
The most revolutionary part is giving you a legitimate method of creating other install images. And that's not terribly revolutionary, given nLiteOS.
There are three distributions listed that conform to LSB3.0: SUSE, RedHat, and Asianux. Why should I write for LSB?
FTFT (from the...tutorial): LSB has binary compatibility standards, so I can compile once and run anywhere. But if the application is GPL and nontrivial, it shouldn't be that hard to get it into the package repositories in question. Otherwise, it's probably in a scripting language, so the end user doesn't have to build or install it anyway.
This is really only important for commercial Linux software.
"if you worry about a little second hand smoke (generally upscale restaurant bars aren't as smokey at lunch) then you need to reconsider that diet coke you're drinking."
Meh. I'm not worried due to my health; cigarette smoke is the worst smell I've yet encountered. I'd walk a quarter mile to avoid it. Skipping out on a room full of drunks is a small cost.
But you can actually use this to develop an OSS emulator for ARM processors. You can use it like an ARM processor for compatibility testing.
As for the source, you can easily compare two codebases to catch potential infringement. Given that the source for each is open, there should be no worries at all; the developers could start by setting up the comparison tool. The real worry would be any patents that Microsoft acquired during the course of creating the emulator.
Cleanroom reverse engineering is a way of being absolutely certain that there's no code sharing when changing licenses; when the code is open for both projects, that becomes less of a concern since it's easy to check.
So a component is exactly analogous to a function?
"The difference with processes is that components can be combined at run-time as the result of computation."
The components themselves can be combined, or their return values? The latter can be done simply in most shells; the former can be done simply in most shells. Unless you mean that I can combine functions of one program with functions of another without touching a compiler. Which I can already do with scripting languages, and don't want to do unless the function is documented.
Now, you still need a scheduler to tell you which component to run, right? Your parallelizing operators could result in hanging programs when you have large components running with short ones, or crashes if there are multiple parallelizable components to deal with on a single processor core. You still need to schedule the operations.
I don't think this is a complete replacement for the current computational model.
Richard Stallman said: "The GNU Emacs General Public License said all modified versions must be under the GNU Emacs General Public License. The GCC General Public Licence said all modified versions must be under the GCC General Public License. What if you wanted to merge Emacs and GCC?"
It would be in line with the core philosophy of Emacs, I admit, but oh, the horrors.
Nah, they'll just get annoyed that their computer is slow. They don't link their behavior with the computer's state unless the relationship is clear and immediate. Programmers and techies are used to that sort of thinking, but then, they already use Firefox for the most part.
They currently have another problem. How do you read the data? The protein is colored and shaped differently if it's been activated by light; either we use some sort of biological reader that bonds to the rest state of the protein but no other, or we shine light on the disk to see what color each bit is--oh, looks like that file is just 0xEEEEEE and takes up the entire disk...
You say that, and I even agree to an extent. However, the vast majority of people want something that functions well with little or no maintenance and has features and an interface that they want. And the vast majority of people are entirely unfamiliar with Linux, so they believe their choice is between getting a new, expensive computer and staying with Windows.
Moreover, Linux is rather impractical for those with dialup Internet service or (gasp) no Internet connection at all. Last I checked, that's still a popular way to go in many areas.
I think that the current Linux geeks will end up spreading Linux to their families.
Windows is opt-out, not opt-in; I think that, most of the time, people would prefer not to use Windows and the attendant expensive software bundle. Linux is an alternative, especially with systems like Ubuntu that take the pain out of administration.
Basically, Windows isn't reaching new people; Linux is. Even if the market share for Linux in established markets grows very slowly, it has to beat Microsoft eventually. Not necessarily in my lifetime, but eventually.
With a company such as TiVO, it's simple enough to do. Otherwise, how can you disable a fast forward button on software that you don't control? You'd need to transmit a virus continually that can exploit weaknesses in all DVR systems and disable that feature.
Of course, this would be entirely illegal, and ABC would go down in flames. Are they simply mad, then? Will they attempt to stop me from changing channels during commercials, too?
Guess what? We have a Google Browser now. It comes in two flavors, neither of which is overtly branded; these are Opera and Firefox. And Google doesn't even have to develop either of them themselves.
Yes, but on my ten-year-old hardware, Linux autodetects and Windows still doesn't.
There are good reasons for this (licensing and such), but none that the end users actually sympathize with. It's the same with Ubuntu's non-support of proprietary media (mp3's etc).
Okay, so we make the tools illegal. But what do we do about ECHELON? It was illegal since its inception, and still it happened.
Should we disband all US intelligence agencies? The NSA's existence was classified originally, and remained so for years. If we disband the agencies, we'll just get secret ones in turn.
Should we stop electing career politicians? The issue seems to be with the position of President--after that, you're not allowed to hold political office, so you may as well risk everything; and you have enough power to infringe on citizens' rights significantly. Congresspeople, on the other hand, can be reelected indefinitely, so there's something at stake for them until they retire. Also, they have peers, not just underlings; committees, not unilateral control.
There's another issue. Every four or eight years, the US gets a new ruler with a clean slate that can abuse the people further, a ruler that feels a need to prove himself internationally via war.
Solutions? Don't have one President; have five or fifteen. The term length can stay the same; just divide the country into regions to elect a President, with new ones taking power on a rotating basis. Perhaps a fifth of the Presidency gets reelected each year; that would be manageable. It's much easier for dirty secrets to get leaked if you've got three new auditors per year, and three conspirators retiring each year with nothing more to lose.
A permanent member of the UN Security Council started the invasion. Another permanent member was also involved. The only Arab nation on the Security Council is Qatar.
Of the other permanent members, there's Russia, France, and China. France issued reproof as a nation; Russia and China seem to be adopting more isolationist policies than the others.
So, you'd have two definite vetoes if you introduced the resolution. Submitting such a proposal would be a waste of political clout.
It was a joke. The two are reasonably close if you're muttering or slurring.
Latin 'r' was most likely a tap or a trill; Spanish 'r' or 'rr' would be closest. At any rate, English 'r' is different and has a significant effect on the preceding vowel that would not be present in Latin. Thus, a native Spanish pronunciation of 'or' would work for the last syllable.
For the first, though, my English pronunciation barely deviates from my Latin. I'm told I speak English with an odd accent, though. (I'm a native speaker. Go figure.)
And I'm sure your accent is better. Hey, let's find a native Latin speaker from the classical period to compare.
What would you suggest to replace processes, scheduling, and a filesystem?
The process/scheduling/filesystem paradigm is essentially a conversion of the problem of running an arbitrary number of virtual Turing machines on a single Turing machine. It's a straightforward reduction from the simplest abstraction of a general computer. If you can come up with a better system, you'll get a PhD and a Turing prize pretty much immediately.
But let's look at this in more detail.
We have processes: that's what is currently trying to get access to the CPU. In order for you to run anything on your computer, you need some list of entities that will access the CPU. We have a scheduler: that determines what currently has access to the CPU. If you don't include this, you can only run one process at a time. That means every application has to include an operating system. We have a filesystem: that's all the data you have. How the hell can you replace this? You can create a different form, but it'd still be a data repository of some sort. Perhaps you object to the files/folders abstraction? The only possible replacement is files/metadata, or just files. The latter is disorganized and thus undesirable--you don't want to sift through all your OS files to find the document you were just editing. And the files/metadata system taking over requires very fast searching, which we simply don't have yet.
So, this 70-year-old system (Turing machines were first discussed in 1936) is still the best fit for our current hardware. That's partly because we have the model, so we develop OSes that abstract hardware to the model, so we get hardware that makes it easier to abstract to that model; it's also because the model works extremely well in general. I don't think we'll replace it any time soon.
Perhaps SkyOS would benefit from adopting the BSD driver model. Then they'd be able to support a lot of hardware for little work, and continue to support new hardware without any added effort.
I would have suggested using the Linux driver model, but the viral licensing issue would kill the company.
Slashdot Mirror
Because SourceForge seems to be dying--no CVS services for several weeks tends to kill the faith people have in you.
On the other hand, I'd have said the same about AOL, too.
Not that I'm arguing with the decision--it's good to have multiple repositories in case one goes down.
There's a difference between not protecting your child every moment of every day and pure neglect. If you see that they're hurting themselves and will continue to do so until someone intervenes, and yet you do nothing...well, I hardly need to tell you which that is.
Sure, if you have twelve gigs of RAM.
Isn't it a bloody obvious method for installing an operating system? Copy the image over, apply some hardware configuration scripts (and whatever else you need, such as default packages).
The most revolutionary part is giving you a legitimate method of creating other install images. And that's not terribly revolutionary, given nLiteOS.
Does it matter if they're stealing ideas or not? They still have the features in the end.
There are three distributions listed that conform to LSB3.0: SUSE, RedHat, and Asianux. Why should I write for LSB?
FTFT (from the...tutorial): LSB has binary compatibility standards, so I can compile once and run anywhere. But if the application is GPL and nontrivial, it shouldn't be that hard to get it into the package repositories in question. Otherwise, it's probably in a scripting language, so the end user doesn't have to build or install it anyway.
This is really only important for commercial Linux software.
"if you worry about a little second hand smoke (generally upscale restaurant bars aren't as smokey at lunch) then you need to reconsider that diet coke you're drinking."
Meh. I'm not worried due to my health; cigarette smoke is the worst smell I've yet encountered. I'd walk a quarter mile to avoid it. Skipping out on a room full of drunks is a small cost.
But you can actually use this to develop an OSS emulator for ARM processors. You can use it like an ARM processor for compatibility testing.
As for the source, you can easily compare two codebases to catch potential infringement. Given that the source for each is open, there should be no worries at all; the developers could start by setting up the comparison tool. The real worry would be any patents that Microsoft acquired during the course of creating the emulator.
Cleanroom reverse engineering is a way of being absolutely certain that there's no code sharing when changing licenses; when the code is open for both projects, that becomes less of a concern since it's easy to check.
So a component is exactly analogous to a function?
"The difference with processes is that components can be combined at run-time as the result of computation."
The components themselves can be combined, or their return values? The latter can be done simply in most shells; the former can be done simply in most shells. Unless you mean that I can combine functions of one program with functions of another without touching a compiler. Which I can already do with scripting languages, and don't want to do unless the function is documented.
Now, you still need a scheduler to tell you which component to run, right? Your parallelizing operators could result in hanging programs when you have large components running with short ones, or crashes if there are multiple parallelizable components to deal with on a single processor core. You still need to schedule the operations.
I don't think this is a complete replacement for the current computational model.
Richard Stallman said: "The GNU Emacs General Public License said all modified versions must be under the GNU Emacs General Public License. The GCC General Public Licence said all modified versions must be under the GCC General Public License. What if you wanted to merge Emacs and GCC?"
It would be in line with the core philosophy of Emacs, I admit, but oh, the horrors.
Nah, they'll just get annoyed that their computer is slow. They don't link their behavior with the computer's state unless the relationship is clear and immediate. Programmers and techies are used to that sort of thinking, but then, they already use Firefox for the most part.
They currently have another problem. How do you read the data? The protein is colored and shaped differently if it's been activated by light; either we use some sort of biological reader that bonds to the rest state of the protein but no other, or we shine light on the disk to see what color each bit is--oh, looks like that file is just 0xEEEEEE and takes up the entire disk...
Do you actually WANT to interview a job applicant who can't spell 20 words in a 150-word email?
You say that, and I even agree to an extent. However, the vast majority of people want something that functions well with little or no maintenance and has features and an interface that they want. And the vast majority of people are entirely unfamiliar with Linux, so they believe their choice is between getting a new, expensive computer and staying with Windows.
Moreover, Linux is rather impractical for those with dialup Internet service or (gasp) no Internet connection at all. Last I checked, that's still a popular way to go in many areas.
I think that the current Linux geeks will end up spreading Linux to their families.
Windows is opt-out, not opt-in; I think that, most of the time, people would prefer not to use Windows and the attendant expensive software bundle. Linux is an alternative, especially with systems like Ubuntu that take the pain out of administration.
Basically, Windows isn't reaching new people; Linux is. Even if the market share for Linux in established markets grows very slowly, it has to beat Microsoft eventually. Not necessarily in my lifetime, but eventually.
But combine it with Tekken, and then...!
Offensive lineman comes up, bodyslam, punch, helicopter kick, then HADOKEN!
With a company such as TiVO, it's simple enough to do. Otherwise, how can you disable a fast forward button on software that you don't control? You'd need to transmit a virus continually that can exploit weaknesses in all DVR systems and disable that feature.
Of course, this would be entirely illegal, and ABC would go down in flames. Are they simply mad, then? Will they attempt to stop me from changing channels during commercials, too?
Guess what? We have a Google Browser now. It comes in two flavors, neither of which is overtly branded; these are Opera and Firefox. And Google doesn't even have to develop either of them themselves.
Yes, but on my ten-year-old hardware, Linux autodetects and Windows still doesn't.
There are good reasons for this (licensing and such), but none that the end users actually sympathize with. It's the same with Ubuntu's non-support of proprietary media (mp3's etc).
Okay, so we make the tools illegal. But what do we do about ECHELON? It was illegal since its inception, and still it happened.
Should we disband all US intelligence agencies? The NSA's existence was classified originally, and remained so for years. If we disband the agencies, we'll just get secret ones in turn.
Should we stop electing career politicians? The issue seems to be with the position of President--after that, you're not allowed to hold political office, so you may as well risk everything; and you have enough power to infringe on citizens' rights significantly. Congresspeople, on the other hand, can be reelected indefinitely, so there's something at stake for them until they retire. Also, they have peers, not just underlings; committees, not unilateral control.
There's another issue. Every four or eight years, the US gets a new ruler with a clean slate that can abuse the people further, a ruler that feels a need to prove himself internationally via war.
Solutions? Don't have one President; have five or fifteen. The term length can stay the same; just divide the country into regions to elect a President, with new ones taking power on a rotating basis. Perhaps a fifth of the Presidency gets reelected each year; that would be manageable. It's much easier for dirty secrets to get leaked if you've got three new auditors per year, and three conspirators retiring each year with nothing more to lose.
A permanent member of the UN Security Council started the invasion. Another permanent member was also involved. The only Arab nation on the Security Council is Qatar.
Of the other permanent members, there's Russia, France, and China. France issued reproof as a nation; Russia and China seem to be adopting more isolationist policies than the others.
So, you'd have two definite vetoes if you introduced the resolution. Submitting such a proposal would be a waste of political clout.
It was a joke. The two are reasonably close if you're muttering or slurring.
Latin 'r' was most likely a tap or a trill; Spanish 'r' or 'rr' would be closest. At any rate, English 'r' is different and has a significant effect on the preceding vowel that would not be present in Latin. Thus, a native Spanish pronunciation of 'or' would work for the last syllable.
For the first, though, my English pronunciation barely deviates from my Latin. I'm told I speak English with an odd accent, though. (I'm a native speaker. Go figure.)
And I'm sure your accent is better. Hey, let's find a native Latin speaker from the classical period to compare.
What would you suggest to replace processes, scheduling, and a filesystem?
The process/scheduling/filesystem paradigm is essentially a conversion of the problem of running an arbitrary number of virtual Turing machines on a single Turing machine. It's a straightforward reduction from the simplest abstraction of a general computer. If you can come up with a better system, you'll get a PhD and a Turing prize pretty much immediately.
But let's look at this in more detail.
We have processes: that's what is currently trying to get access to the CPU. In order for you to run anything on your computer, you need some list of entities that will access the CPU.
We have a scheduler: that determines what currently has access to the CPU. If you don't include this, you can only run one process at a time. That means every application has to include an operating system.
We have a filesystem: that's all the data you have. How the hell can you replace this? You can create a different form, but it'd still be a data repository of some sort. Perhaps you object to the files/folders abstraction? The only possible replacement is files/metadata, or just files. The latter is disorganized and thus undesirable--you don't want to sift through all your OS files to find the document you were just editing. And the files/metadata system taking over requires very fast searching, which we simply don't have yet.
So, this 70-year-old system (Turing machines were first discussed in 1936) is still the best fit for our current hardware. That's partly because we have the model, so we develop OSes that abstract hardware to the model, so we get hardware that makes it easier to abstract to that model; it's also because the model works extremely well in general. I don't think we'll replace it any time soon.
Perhaps SkyOS would benefit from adopting the BSD driver model. Then they'd be able to support a lot of hardware for little work, and continue to support new hardware without any added effort.
I would have suggested using the Linux driver model, but the viral licensing issue would kill the company.
Depends on how much is built in. It looks like SkyOS will be great for my grandparents.
On the other hand, Ubuntu would be as well, provided I set it up beforehand.