You seem to assume that the scientists whom have worked on the GPS project are unable to compensate for satelite movement, curvature of the earth, etc.
No, I'm assuming the accuracy required is beyond (way beyond would be my intuition) what is possible. We're talking about being to be able to measure velocities of *1* mms-1 (in a second), to accuracy of order 1 to 3% for it to be useful..
To be honest your argument has little to no foundation - seems more like a comment of a first year physics student than someone who knows about GPS systems.
I've worked out some numbers. I admit I don't know much about GPS systems today, but I don't think they're capable of the accuracy above. Show me they're wrong, or show me real-world systems which are capable of those accuracies.
FWIW, I am aware that GPS compensates for doppler shift for signals, and many other things (even special relativity). That isn't what I was disputing with the original poster though, try reading. I believe they were claiming GPS/directly/ calculates velocity of a vehicle from the doppler shift of the signals.
You have to measure 1 *millimetre* per second speed through doppler shift, with some accuracy, in order to deduce the speed of a fast-moving car to be approx 126km/h. With an accuracy of 10%, you need to measure to within ±0.1mms-1 to deduce a speed of 126km/h ±13km/h.
First, when the GPS unit itself calculates the speed, it records your instantaneous velocity, not an average. It calculates this using the doppler shift present in the GPS signals picked up by the unit, not from how far the unit has travelled.
Let's go over some basics:
a) There is no such thing as "instantaneous velocity" - as velocity is a function of time.
Corrolary: You can/approach/ t=0, but the closer to "near-instantaneous velocity" you try to measure, the more accurate your measurements must be - alternatively the higher the margin of error will be.
And the problem with the radar/lasar guns is indeed that, because they try calculate "near-instantaneous velocity" they are very *very* susceptible to error, particularly at the ranges the police often try use them at (hundreds of metres).
b) Noticing a doppler shift in waves from a (relatively) stationary source would require that you have a non-zero velocity relative to the source (ie the distance between you and the source change). I'm reasonably sure this velocity would be immeasurable from a consumer car in a GPS over a short period of time and, further, that any measurable doppler would be due far more to the/satellite/ moving, not the car..
I.e. I havn't done the calculations (it's not just linear, cause any doppler will be induced by the curvature of the earth, not directly by the car's speed), but you're talking about measuring doppler due to/millimetres/ of movement (curvature of the earth), as a car moves perhaps a tennes of few metres. It's beyond believable we could measure that with any useful accuracy in a car.
So I call bullshit, unless you show me the numbers to prove otherwise.
I thought the user accounts were lame (they were introduced what, 6 to 8 months after/. started, anyone remember?). But gave in after about a month or two (I can't remember what made me do so, some new feature must have been added for users I guess, e.g. comment history or somesuch?).
Hmm. If we delve into the gun-lobby links you gave, the worldnetdaily one is spinning this data from Universiteit Leiden. The data doesn't seem that clear-cut, indeed it seems rates are the same. I'm not sure how to interpret this data though (there's not much of it). I'm not sure either one of us should draw any strong conclusions from it.
Air-rifles: Yes, I did note in another comment they can be lethal. As the article notes though, over half the recorded "firearms" incidents were of 'minor wounds' (i.e. kids/teens shooting each other with pellet guns quite likely).
The one clear-statistic available though is that the USA has the highest rates (a few times over) of gun-shot fatalaties in the developed world...
Got any figures on that, out of curiosity? (The 4k figure in that article includes over 2k 'minor' injuries from air rifles / pellet guns. Hardly violent crime - though dangerous and irresponsible - but a shot from a pellet gun usually only stings a bit).
The UK btw has gotten pretty weird about crime statistics, cause of changes in how they reward police officers (bonuses depend on the number of crime 'detections', etc..) over the course of the labour government.
Go find out the definition of monopoly. For/practical/ purposes, the only definition which matters is legal/regulatory - and guess what, MS has been found to engage in monopolistic practices by the judicial systems in both the EU and the USA (and I think one Asian country too?).
But go ahead, give me whatever dictionary definition supports your view - it has no bearing.
I sure as hell doubt MS is going to license every line of their code to some random company,
They can be made to do so. (They're a monopoly, antithetical to free market economies, so they no longer deserve the freedom to run their company for their own benefit).
BTW, bear in mind that in the UK and Ireland, the general legal definition of "firearms" includes things like pellet-guns and air-rifles. A high-profile "gun crime" a few years ago involved the fatal shooting of a toddler with an air-rifle in Scotland.
Opteron has 9 metal layers. I'm no EEE, but I know those layers aren't all available for interconnects, at least one layer will be dedicated to ground and power (IMLU).
If the traces generally *can* cross, how come *none* of the multi-core CPU plans show such core interconnects? They're/all/ through crossbars. It seems that running very high-speed interconnects in parallel, or across each other, in different layers, is more expensive and/or less effective than using transistors as a switching fabric (ie crossbar).
If you can't use birth control for religious reasons, don't have sex.
What an inane statement.
- It can be hard to use birth control even if you have/no/ religious problems using it.. - Even if you don't mind using birth control, it may not be easily available - There may be cultural biases against using certain forms of contraception (e.g. it may not be 'manly'), in such cases you need *active* programmes to encourage condom use.
Damn, even educated, prosperous, long married couples have 'accidents' - surely everyone knows of friends who have a child or sibling that is ten+ years younger than the next oldest...
"Don't have sex" - what a dumb thing to expect humanity to be capable of. You should become a priest.
the idea that a block of code that happens to make use of well-known hooks in the Linux kernel is a derivative work.
Ah.. Hmm, well:
a) Using well-known hooks, for some definition of 'well-known' that would constitute a functional interface (i.e. stable, documented, widely implemented and used) likely would NOT constitute a derived work. There is case-law on that in the USA (and in the UK too I think), and further there is an explicit statement in the Linux COPYING file to state that using its well-known hooks does not lead to derivation (those hooks are called 'syscalls').
I'm simplifying, so Google for "filter test altai" if you're curious about the details (applying in the USA anyway).
b) All other 'hooks' in Linux can be very fluid and often very Linux-specific. Whether their use is a copyright derivation 'barrier' would depend on the precise hooks, and it'd be very debatable. You'd need a judge and possibly a jury for a definitive answer, to carry out the filter test.
c) We know ESX is NOT using the hooks in a..
d) We don't know what hooks ESX IS using, cause the code concerned is proprietary. However see b) and e).
e) There is another kind of use that copyright can recognise, the wholesale dependence on a work. E.g. it seems to be alleged that ESX depends on Linux for IO, depends on SMB for file-transfering in userspace, etc. I.e., if true, that's a lot of GPL code that ESX depends on, which VMWare chose to depend on - no one held a gun to their head and forced them to implement their product on top of Linux...
The legal advice I've heard says that depending on GPL code (in the 'required' sense) in one's GPL-incompatible code is a risky position to take..
Finally, you may wish to look at my other comments in this thread, particularly my comparison to Xen.
You, on the other hand, told me (by extension) that what I said sucks. Thanks for that, it really helped. Why don't you go after my spelling and grammar next, so you can feel extra superior today?
I'm sorry your analogy sucked. That wasn't my fault;). In pointing out how it sucked, I did get you though to post a comment relating to the/actual/ subject - from which useful discussion can proceed (unlike the analogy;) ).
Discussion probably should be considered a journey toward truth, rather than a duel to win - one side or the other being superior is often quite transient..
NB: I speak only for myself here. Further, I have never used ESX and am giving argument based on descriptions of ESX given by the article and certain posters here. Hence my comments should be read as quite speculative.
Then why do I hear arguments every other day about whether to include binary-only proprietary drivers? By definition, they couldn't be if they weren't GPL according to your argument.
Uhm, by definition, if the legality of binary-only, proprietary drivers for GPLed OSes were obviously legal - there'd be NO NEED FOR ARGUMENT! You're hearing those arguments, cause they're in a grey area. Some drivers are not derived works - e.g. Linus thinks NVidias' are such a case - and some are.
It all depends on how the driver was developed, and what it depends on.
I did read the article, but the fucking argument was so convoluted - and then confused by the comments here as to what was really meant - that it took me an hour to figure it out - and I'm still called wrong by everybody.
The arguments here on/. are so convoluted because badly-informed people seem to feel empowered to make authorative statements on the matter. Those statements then too often tend to be wrong and need correction. With all due respect, you're one of those people - refrain from commenting if you don't understand something (better, phrase your understanding as a question).
(FWIW: We're all ill-informed in most areas, we can only ever be relatively well-informed in a very few subjects. So please don't feel insulted. You probably could me teach lots of stuff about plasma physics, or who knows what;) ).
Fuck it, it's all bullshit. Who gives a shit except some asshole named Hellwig?
Christ, I'm starting to get sick of this license horseshit. Public domain every fucking thing and everybody including Linus STFU and get back to coding.
Uhm, those who write the code decide the licence. And they inherently must care about that licence (e.g. those in the BSD community care very strongly that their code be used without much care).
If you don't like the licence, don't use the code. Linux is not the only 'FreeNix', try FreeBSD or one of the OpenSolaris distros.
The OP was using examples of all the console OS stuff - that was NOT the issue that I saw mentioned in the original article.
Uhm: The console OS *IS* Linux, which it seems is required in order for ESX to function, hence Linux (the "console" OS) IS THE ISSUE.
According to the OP, the "console" OS is actually providing the drivers to do IO - and hence is a bit more than just a "console" (ESX's claimed hardware support is "Whatever Linux supports..", according to the OP).
Apparently the source to the Linux bits/is/ available. It's the "depend on Linux for hardware IO" that Hellwig (and others?) object to as a licence violation.
To me, that's like telling me if I install an aftermarket stereo (module) in my Buick (kernel) using the wiring harness that came with the car (module hooks), my stereo is now considered to have been a derivative work of my car.
See... the thing is that that installing hardware is not covered by copyright law. So your analogy just doesn't work.
The problem with argument by analogy is that often the analogies just suck. And even if they don't, they'll eventually fail. And you spend time arguing about how the analogy related to the *actual* subject you're discussing, rather than the actual subject.
Slashdot Mirror
You seem to assume that the scientists whom have worked on the GPS project are unable to compensate for satelite movement, curvature of the earth, etc.
/directly/ calculates velocity of a vehicle from the doppler shift of the signals.
No, I'm assuming the accuracy required is beyond (way beyond would be my intuition) what is possible. We're talking about being to be able to measure velocities of *1* mms-1 (in a second), to accuracy of order 1 to 3% for it to be useful..
To be honest your argument has little to no foundation - seems more like a comment of a first year physics student than someone who knows about GPS systems.
I've worked out some numbers. I admit I don't know much about GPS systems today, but I don't think they're capable of the accuracy above. Show me they're wrong, or show me real-world systems which are capable of those accuracies.
FWIW, I am aware that GPS compensates for doppler shift for signals, and many other things (even special relativity). That isn't what I was disputing with the original poster though, try reading. I believe they were claiming GPS
It can be directly measured using the Doppler shift of light bouncing off a moving object
Doppler shift can not be measured without reference to time...
Your point about the problem with radar guns in nonsense, you clearly have no idea how they work.
You clearly have no idea how they are operated.
And yes, IAAP.
Not a very good one.
Have problems with calculus, eh?
;).
Not really. But you seem to have a problem with physics
Oh, just to clarify why I mention the speed...
:)
You have to measure 1 *millimetre* per second speed through doppler shift, with some accuracy, in order to deduce the speed of a fast-moving car to be approx 126km/h. With an accuracy of 10%, you need to measure to within ±0.1mms-1 to deduce a speed of 126km/h ±13km/h.
It's just not feasible
you're talking about measuring doppler due to /millimetres/ of movement (curvature of the earth), as a car moves a tennes of few metres.
Ah, that's roughly about 1mm for 35m of forward movement btw. For reference, 35ms-1 is about 126km/h (78mph).
First, when the GPS unit itself calculates the speed, it records your instantaneous velocity, not an average. It calculates this using the doppler shift present in the GPS signals picked up by the unit, not from how far the unit has travelled.
/approach/ t=0, but the closer to "near-instantaneous velocity" you try to measure, the more accurate your measurements must be - alternatively the higher the margin of error will be.
/satellite/ moving, not the car..
/millimetres/ of movement (curvature of the earth), as a car moves perhaps a tennes of few metres. It's beyond believable we could measure that with any useful accuracy in a car.
Let's go over some basics:
a) There is no such thing as "instantaneous velocity" - as velocity is a function of time.
Corrolary: You can
And the problem with the radar/lasar guns is indeed that, because they try calculate "near-instantaneous velocity" they are very *very* susceptible to error, particularly at the ranges the police often try use them at (hundreds of metres).
b) Noticing a doppler shift in waves from a (relatively) stationary source would require that you have a non-zero velocity relative to the source (ie the distance between you and the source change). I'm reasonably sure this velocity would be immeasurable from a consumer car in a GPS over a short period of time and, further, that any measurable doppler would be due far more to the
I.e. I havn't done the calculations (it's not just linear, cause any doppler will be induced by the curvature of the earth, not directly by the car's speed), but you're talking about measuring doppler due to
So I call bullshit, unless you show me the numbers to prove otherwise.
You young 'uns. Now, when ah wur a lad...
Capsaicin come out just as hot as it goes in.
Cause your not digesting it. Presumably because your gastro-intestinal tract isn't accustomed to it and is trying to expel it ASAP.
Keep eating chili and this stops happening pretty quickly, and you can start to enjoy the chili high.
(I don't think chili is an opiate though? Never heard that before, nor has wikipædia, AFAICT).
Me too! :)
/. started, anyone remember?). But gave in after about a month or two (I can't remember what made me do so, some new feature must have been added for users I guess, e.g. comment history or somesuch?).
I thought the user accounts were lame (they were introduced what, 6 to 8 months after
Hmm. If we delve into the gun-lobby links you gave, the worldnetdaily one is spinning this data from Universiteit Leiden. The data doesn't seem that clear-cut, indeed it seems rates are the same. I'm not sure how to interpret this data though (there's not much of it). I'm not sure either one of us should draw any strong conclusions from it.
Air-rifles: Yes, I did note in another comment they can be lethal. As the article notes though, over half the recorded "firearms" incidents were of 'minor wounds' (i.e. kids/teens shooting each other with pellet guns quite likely).
The one clear-statistic available though is that the USA has the highest rates (a few times over) of gun-shot fatalaties in the developed world...
We do look a lot better in plain violent crime.
Got any figures on that, out of curiosity? (The 4k figure in that article includes over 2k 'minor' injuries from air rifles / pellet guns. Hardly violent crime - though dangerous and irresponsible - but a shot from a pellet gun usually only stings a bit).
The UK btw has gotten pretty weird about crime statistics, cause of changes in how they reward police officers (bonuses depend on the number of crime 'detections', etc..) over the course of the labour government.
Go find out the definition of monopoly. For /practical/ purposes, the only definition which matters is legal/regulatory - and guess what, MS has been found to engage in monopolistic practices by the judicial systems in both the EU and the USA (and I think one Asian country too?).
But go ahead, give me whatever dictionary definition supports your view - it has no bearing.
I sure as hell doubt MS is going to license every line of their code to some random company,
They can be made to do so. (They're a monopoly, antithetical to free market economies, so they no longer deserve the freedom to run their company for their own benefit).
And from the article:
"In total, 3,995 people were shot, of whom about half (2,187) had minor wounds caused by air guns."
BTW, bear in mind that in the UK and Ireland, the general legal definition of "firearms" includes things like pellet-guns and air-rifles. A high-profile "gun crime" a few years ago involved the fatal shooting of a toddler with an air-rifle in Scotland.
Hilarious...
/far/ lower gun crime rate than the US.
The point is that most other countries (e.g. like the UK) have a
I'm not sure by what logic you can think that the fact that gun crime still exists in those other countries buttresses your position.
Opteron has 9 metal layers. I'm no EEE, but I know those layers aren't all available for interconnects, at least one layer will be dedicated to ground and power (IMLU).
/all/ through crossbars. It seems that running very high-speed interconnects in parallel, or across each other, in different layers, is more expensive and/or less effective than using transistors as a switching fabric (ie crossbar).
If the traces generally *can* cross, how come *none* of the multi-core CPU plans show such core interconnects? They're
??
Except it's not a general graph, it's constrained to vertices in a single 2D plane, which may not cross. (For reasons of cost).
Christ himself said to pay taxes that the government demands.
And Ireland demands taxes...
We just demand a smaller percentage, what's wrong with that?
If you can't use birth control for religious reasons, don't have sex.
/no/ religious problems using it..
What an inane statement.
- It can be hard to use birth control even if you have
- Even if you don't mind using birth control, it may not be easily available
- There may be cultural biases against using certain forms of contraception (e.g. it may not be 'manly'), in such cases you need *active* programmes to encourage condom use.
Damn, even educated, prosperous, long married couples have 'accidents' - surely everyone knows of friends who have a child or sibling that is ten+ years younger than the next oldest...
"Don't have sex" - what a dumb thing to expect humanity to be capable of. You should become a priest.
the idea that a block of code that happens to make use of well-known hooks in the Linux kernel is a derivative work.
;). In pointing out how it sucked, I did get you though to post a comment relating to the /actual/ subject - from which useful discussion can proceed (unlike the analogy ;) ).
Ah.. Hmm, well:
a) Using well-known hooks, for some definition of 'well-known' that would constitute a functional interface (i.e. stable, documented, widely implemented and used) likely would NOT constitute a derived work. There is case-law on that in the USA (and in the UK too I think), and further there is an explicit statement in the Linux COPYING file to state that using its well-known hooks does not lead to derivation (those hooks are called 'syscalls').
I'm simplifying, so Google for "filter test altai" if you're curious about the details (applying in the USA anyway).
b) All other 'hooks' in Linux can be very fluid and often very Linux-specific. Whether their use is a copyright derivation 'barrier' would depend on the precise hooks, and it'd be very debatable. You'd need a judge and possibly a jury for a definitive answer, to carry out the filter test.
c) We know ESX is NOT using the hooks in a..
d) We don't know what hooks ESX IS using, cause the code concerned is proprietary. However see b) and e).
e) There is another kind of use that copyright can recognise, the wholesale dependence on a work. E.g. it seems to be alleged that ESX depends on Linux for IO, depends on SMB for file-transfering in userspace, etc. I.e., if true, that's a lot of GPL code that ESX depends on, which VMWare chose to depend on - no one held a gun to their head and forced them to implement their product on top of Linux...
The legal advice I've heard says that depending on GPL code (in the 'required' sense) in one's GPL-incompatible code is a risky position to take..
Finally, you may wish to look at my other comments in this thread, particularly my comparison to Xen.
You, on the other hand, told me (by extension) that what I said sucks. Thanks for that, it really helped. Why don't you go after my spelling and grammar next, so you can feel extra superior today?
I'm sorry your analogy sucked. That wasn't my fault
Discussion probably should be considered a journey toward truth, rather than a duel to win - one side or the other being superior is often quite transient..
NB: I speak only for myself here. Further, I have never used ESX and am giving argument based on descriptions of ESX given by the article and certain posters here. Hence my comments should be read as quite speculative.
Then why do I hear arguments every other day about whether to include binary-only proprietary drivers? By definition, they couldn't be if they weren't GPL according to your argument.
/. are so convoluted because badly-informed people seem to feel empowered to make authorative statements on the matter. Those statements then too often tend to be wrong and need correction. With all due respect, you're one of those people - refrain from commenting if you don't understand something (better, phrase your understanding as a question).
;) ).
Uhm, by definition, if the legality of binary-only, proprietary drivers for GPLed OSes were obviously legal - there'd be NO NEED FOR ARGUMENT! You're hearing those arguments, cause they're in a grey area. Some drivers are not derived works - e.g. Linus thinks NVidias' are such a case - and some are.
It all depends on how the driver was developed, and what it depends on.
I did read the article, but the fucking argument was so convoluted - and then confused by the comments here as to what was really meant - that it took me an hour to figure it out - and I'm still called wrong by everybody.
The arguments here on
(FWIW: We're all ill-informed in most areas, we can only ever be relatively well-informed in a very few subjects. So please don't feel insulted. You probably could me teach lots of stuff about plasma physics, or who knows what
Fuck it, it's all bullshit. Who gives a shit except some asshole named Hellwig?
Christ, I'm starting to get sick of this license horseshit. Public domain every fucking thing and everybody including Linus STFU and get back to coding.
Uhm, those who write the code decide the licence. And they inherently must care about that licence (e.g. those in the BSD community care very strongly that their code be used without much care).
If you don't like the licence, don't use the code. Linux is not the only 'FreeNix', try FreeBSD or one of the OpenSolaris distros.
The OP was using examples of all the console OS stuff - that was NOT the issue that I saw mentioned in the original article.
Uhm: The console OS *IS* Linux, which it seems is required in order for ESX to function, hence Linux (the "console" OS) IS THE ISSUE.
According to the OP, the "console" OS is actually providing the drivers to do IO - and hence is a bit more than just a "console" (ESX's claimed hardware support is "Whatever Linux supports..", according to the OP).
Apparently the source to the Linux bits /is/ available. It's the "depend on Linux for hardware IO" that Hellwig (and others?) object to as a licence violation.
To me, that's like telling me if I install an aftermarket stereo (module) in my Buick (kernel) using the wiring harness that came with the car (module hooks), my stereo is now considered to have been a derivative work of my car.
See... the thing is that that installing hardware is not covered by copyright law. So your analogy just doesn't work.
The problem with argument by analogy is that often the analogies just suck. And even if they don't, they'll eventually fail. And you spend time arguing about how the analogy related to the *actual* subject you're discussing, rather than the actual subject.
Arguing by analogy sucks - don't do it.