> Parental responsibility has to have boundaries, and the parents seem to have taken reasonable steps.
Not having actually read TFA, I'm not so sure. Telling someone to do something is good, but is useless if there are no checks to verify that what you say has been followed. So I still think it's negligent - they apparently neglected to follow up on their initial bit of good parenting.
Yup, after posting, I popped off to their site to download the package and see what's inside, and indeed it's different from the ethernet/bt/dsp/mpeg/etc. drivers that I've worked with in the past. The size of it means that it's not easy to work out exactly what's going on from a quick inspection. The objects without source indeed didn't even look like they were primarily static const u8 fw[]={/*hex data*/}; as it looked like they contained lots of large wasteful lookup tables (so "data", rather than "code") but I got bored before paging down too many megabytes of it.
This distribution method in some ways makes me laugh, it reminds me of some of the alcohol sales laws we have here. I can buy a bottle of beer fresh from the tap at the shop, but they cannot put a lid on it for me. I have to cap it myself. (They only have a licence to sell open containers.) They provide the caps for me, it's just that I have to do the final putting-together stage myself. And in many ways that's just dumb. If the workaround for not being able to ship it fully put together is just to ship in 2 parts, and have the recipient put it together, then nothing has really been prevented. But both sides will never back down, more's the shame.
Indeed, and Adams had even already worked on a Sci-Fi/Fantasy program which featured universal - straight to the mind - translation: Doctor Who (the episode where this plot-line is first introduced certainly post-dates Star Trek, but the fact that he wrote for the program after that point gives him no plausible deniability). He borrowed an awful lot.
And - OT - apart from Martin Freeman, the recent movie was PANTS!
In my experience, about 90% of the time I'm left confused. The output is almost always meaningless. For it to be coherently insulting would require a bizarre fluke of clarity.
As a concrete example, Google will translate the local "no" word, which has no alternative meanings, into "yes", "no", and "maybe". Bizarrely, it usually also wants to translate "Estonia" or "Finland" into "England". This is so broken that it's worse than useless. Google should be ashamed of it.
I'm not an expert on the nvidia case itself, but I do know a bit about some of the other "blob" drivers. In those cases, the real drivers (the smarts) are firmware that is uploaded onto a separate core (sometimes on the same die as the CPU(s), but sometimes off-die) that does not run linux. And therefore is not governed by the kernel's GPL, any more than a bitmap image pulled off the boot partition and flashed up on screen at boot time would be. That's payload, that's not control, it's input and output for the kernel, it's not part of the kernel. The actual linux kernel module is usually just a firmware uploading wrapper, some initial configuration of io parameters (dma areas, buses, clocks, etc.) and after that some simple shims.
If your premises are all true, your logic is good, and your conclusions are sound.
It is indeed the nature of "B" that is important. Does it borrow from Linux's implementation, which would make it a derived work, or is it merely sharing the linux API, a scenario explicitly permitted?
RTS's description of what "B" is is highly ambiguous - to such an extent that it *looks* evasive, and deliberately deceiving. Just seeing what they said invites you to probe further, one cannot just take it at face value, it looks like we're being lied to. However, the lawyers are in contact with each other, I doubt that any new revelations will all be made publicly.
Well, as the FSF wrote the licence, I think their interpretation of what the licence applies to should carry some weight, and they say "If the program dynamically links the plug-in [...] they form a single program, which must be treated as an extension of both the main program and the plugin.".
If one doesn't like their licence, and doesn't agree to it, then one has no right to distribute the licenced software. So if they don't agree that the kernel plus a scsi driver is a derivative work of the linux kernel that is covered by the GPL, they can't distribute the linux kernel. This is the stance that the original complainant, Alan Cox, and others have expressed on LKML, so is a perfectly usual interpretation. And Alan Cox should know, as he's had to work with these things very closely in the last half decade, he's probably one of the kernel's most-equipped guys in this field. (But of course he immediately sics the vicious attack lawyers at the target, his skill is knowing when to do that.)
So you think they ship a linux driver that they're claiming is proprietary, and a linux kernel that is GPL, and they _aren't_ linking the two? In particular given that they've admitted "we only use Linux kernel symbols that are..." https://lkml.org/lkml/2012/11/8/533
> 2. There is at least some code that they are distributing which they do not own the copyrights to.
> If you are claiming 2, then I hope you have evidence
The product, according to RTS itself - but worded very ambiguously, seems to ship with either linux or a derivitive work of linux. They do not own the copyrights to linux.
"Yes, we ship GPL-ed linux; and yes, we ship a closed-source driver which was originally written as a linux kernel module; but no, that driver is never linked to the kernel that's running and runs stand-alone, and therefore is not in violation of linux's GPL.".
Not entirely on speculation. The device claims to run the GPL'ed linux kernel. Therefore the module they link to the kernel must also be GPL.
The comments from the subsystem maintainer (like about 70% of the comments on/.) indicate that he is ignorant of the meaning of the GPL. This *alone* is a fairly good reason to dismiss him from his role even if his company is innocent of the allegations.
Try telling the environmentalists that their geothermal energy is just waste heat from a huge fission reactor deep underground, and watch their heads explode!
Intresting. I know that american football teams have huge numbers of players with different specialities, so most are not on the patch at any one time. So what's the average amount of action that each *player* participates in, per game?
I'd be willing to bet you use and AND of a whole lot of NOTs quite a lot when playing sudoku. How else will you know that only one thing is permitted in a square unless you've got 8 NOTs all ANDed together?
> if a theory is so general it can not be falsified it is not science.
Yes, but "supersymmetry" isn't "a theory", it's not the science. It's a label that is applied to the whole family of putative theories that are trying to be science, and which share a common core feature. It's not "general", it's "several". I hate to stand up for supersymmetry, as none of its expressions show the elegance that I like in science (et gustibus non disputandem est), but thinking of it as one single target that can be shot down is in error.
Not long before Newton some other guy (Galileo?) proposed the acceleration of objects falling under gravity such that the speed was proportional to the disance already moved. Newton as we know modelled it differently. The other guy's theory fell down when it was realised that an object would never start to fall. So they shot it down, and Newton's took over. That didn't mean that "gravitational acceleration" was so general it couldn't be disproved and wasn't a science.
> it makes it less interesting science
In some ways, definitely - yours seems to overlap somewhat with my 'elegance' point of view. If there is enough room to be making many many different models, then it looks like there's more guesswork involved than insight. Anyone can roll their own supersymmetric theory - download the new SuSy model GUI-based wizard trial version, and generate your own model in only 10 clicks! First 20 models free!
I'm not sure the number of cores is reliable. One figure said a quarter of a million nvidea cores. That works out at 100 boards (2.5kcores/board). Another figure from an other article http://www.anandtech.com/show/6421/inside-the-titan-supercomputer-299k-amd-x86-cores-and-186k-nvidia-gpu-cores claimed Titan was 18k opteron with 18k nvideas, one each. The latter implies 50 million of nvidea cores.
As the latter has a photo of a rack, with 4 opterons (16 cores each) plus 4 nvideas (2688 cores each), I'm much more likely to believe Anand than top500.org.
Nope. Because you must always presume that the attacker can try things many many many times.
If the library they're trying to return into (google "return to libc" or "ret2libc" attack if you're not familiar with common attacks) can be at one of 4096 different addresses (assuming 12 bits in the address are randomised), and assuming a probe can come in (with a slightly different payload) every 20 seconds (chosen as it might be too low for a rate limiter to notice), then you'll be rooted in less than a day, as 86400>20*4096. If you don't have a rate limiter and there are 200 attacks per second with slightly different payloads, then you'll be rooted in 20s.
I have a cryptographic background, and "less broken", is still "broken". It's "inc-/dec-reasing the constant" in the algorithm, not "inc-/dec-reasing the complexity".
An example that has a similar kind of idea, even though it's a bit unrelated, is that of Dan Bernstein's timing attacks on AES. AES - demonstrated, through all (known) mathematical and computational techniques to be secure - can be broken, if the enemy gets to probe you often enough (maybe many thousands of times, an attack that is actually practical, and has been performed in reality). Of course, in that case it was only particular (cache-sensitive, lookup table driven) implementations that suffered the attack, but it's concrete proof that you must never underestimate the power of multiple repeated queries from a malicious adversary.
Indeed, insults do not make my case, they simply decorate my case. If the slashdot audience is going to have to plough through your idiocy, they may as well at least have something that adds some amusement to the drudgery.
And you couldn't explain yourself out of a paper bag. You were totally and utterly in the wrong in the first place, and, once that was demonstrated to you and to the slashdot reading public, had the stupidity to re-iterate how wrong you were. That's either brave or stupid.
Fluidity is irrelevant. Intramolecular forces are relevant. Liquids have them. Gases pretty much don't.
For ideal gases, free of intermolecular attractions, what's really most relevant is molecular speed, but (mean) speed is inversely proportional to sqrt(molecular weight). Therefore molecular weight *is* important. And for most of the atmosphere by volume, density is so low that intra-molecular attractive forces are indeed negligible, and so in the largest part of the atmosphere the gases do indeed behave like ideal gases.
Why else do you think the outermost region of the earth's atmosphere is mostly hydrogen? And the layer below that is mostly helium. And has been for your billions of years.
So please do some research before posting nonsense on slashdot. Here's a good place to start: http://www.shodor.org/os411/courses/411b/module02/unit01/images/heterosphere.gif The atmosphere's *way* thicker than just the blue thing that has the fluffy floaty rainy things in it, you know.
Liquids and gasses are very different beasts. One has enough intermolecular forces to bind the molecules into an effectively incompressible mass, the other hasn't, and has components that only interact with each other through random collisions. You're comparing apples to class III orange stars.
You appear to have a problem with the concept of direction. The problem being addressed was that of stuff with the evil bit set coming *in* from the internet. Moving a PC relative to the windows does not change that at all, for example. Most of your defences seem to *presume* there's a man on the inside, but if you've got a man on the inside, there's no defence if there are *any* input devices at all. And a computer with no input devices can not receive or process any information. And therefore isn't a computer any more. You're just trying to show off that you can think of lots of ways of getting data off a supposedly secure system, which is pointless willy waving, anyone can do that. Why did you mention destroying guests' phones, but not the pad of paper they had in their pocket? Or their memories?
I do understand the difference, and you are a fucking lazy idiot. If you'd have followed the link I mentioned, you'd have seen the following: """ In addition to the Floatingsheep blog we have a number of academic papers in various stages of review, revision and published form.... The Professional Geographer 64(4): 602-617... Journal of Urban Technology 18(1): 115-132.... Journal of Urban Technology 18(1): 93-114... Tijdschrift voor Economische en Sociale Geografie 101(4): 422â€"436... World Health and Medical Policy 2(2): 7-33.... Environment and Planning A 42(4): 763-764.... GeoForum 40(4): 523-534... GeoForum 38(6): 1322-1343... Environment and Planning B 34(3): 466-482.... Societies and Cities in the Age of Instant Access. 231-244 """
Good. Fuck off then. You clearly do not have enough intelligence to understand the simple word "insinuation", for example, so I can't expect to see much in the way of intelligent debate from you.
Slashdot Mirror
> Parental responsibility has to have boundaries, and the parents seem to have taken reasonable steps.
Not having actually read TFA, I'm not so sure. Telling someone to do something is good, but is useless if there are no checks to verify that what you say has been followed. So I still think it's negligent - they apparently neglected to follow up on their initial bit of good parenting.
Yup, after posting, I popped off to their site to download the package and see what's inside, and indeed it's different from the ethernet/bt/dsp/mpeg/etc. drivers that I've worked with in the past. The size of it means that it's not easy to work out exactly what's going on from a quick inspection. The objects without source indeed didn't even look like they were primarily
static const u8 fw[]={/*hex data*/};
as it looked like they contained lots of large wasteful lookup tables (so "data", rather than "code") but I got bored before paging down too many megabytes of it.
This distribution method in some ways makes me laugh, it reminds me of some of the alcohol sales laws we have here. I can buy a bottle of beer fresh from the tap at the shop, but they cannot put a lid on it for me. I have to cap it myself. (They only have a licence to sell open containers.) They provide the caps for me, it's just that I have to do the final putting-together stage myself. And in many ways that's just dumb. If the workaround for not being able to ship it fully put together is just to ship in 2 parts, and have the recipient put it together, then nothing has really been prevented. But both sides will never back down, more's the shame.
Indeed, and Adams had even already worked on a Sci-Fi/Fantasy program which featured universal - straight to the mind - translation: Doctor Who (the episode where this plot-line is first introduced certainly post-dates Star Trek, but the fact that he wrote for the program after that point gives him no plausible deniability). He borrowed an awful lot.
And - OT - apart from Martin Freeman, the recent movie was PANTS!
In my experience, about 90% of the time I'm left confused. The output is almost always meaningless. For it to be coherently insulting would require a bizarre fluke of clarity.
As a concrete example, Google will translate the local "no" word, which has no alternative meanings, into "yes", "no", and "maybe". Bizarrely, it usually also wants to translate "Estonia" or "Finland" into "England". This is so broken that it's worse than useless. Google should be ashamed of it.
I'm not an expert on the nvidia case itself, but I do know a bit about some of the other "blob" drivers. In those cases, the real drivers (the smarts) are firmware that is uploaded onto a separate core (sometimes on the same die as the CPU(s), but sometimes off-die) that does not run linux. And therefore is not governed by the kernel's GPL, any more than a bitmap image pulled off the boot partition and flashed up on screen at boot time would be. That's payload, that's not control, it's input and output for the kernel, it's not part of the kernel. The actual linux kernel module is usually just a firmware uploading wrapper, some initial configuration of io parameters (dma areas, buses, clocks, etc.) and after that some simple shims.
If your premises are all true, your logic is good, and your conclusions are sound.
It is indeed the nature of "B" that is important. Does it borrow from Linux's
implementation, which would make it a derived work, or is it merely sharing the linux API, a scenario explicitly permitted?
RTS's description of what "B" is is highly ambiguous - to such an extent that it *looks* evasive, and deliberately deceiving. Just seeing what they said invites you to probe further, one cannot just take it at face value, it looks like we're being lied to. However, the lawyers are in contact with each other, I doubt that any new revelations will all be made publicly.
Well, as the FSF wrote the licence, I think their interpretation of what the licence applies to should carry some weight, and they say "If the program dynamically links the plug-in [...] they form a single program, which must be treated as an extension of both the main program and the plugin.".
If one doesn't like their licence, and doesn't agree to it, then one has no right to distribute the licenced software. So if they don't agree that the kernel plus a scsi driver is a derivative work of the linux kernel that is covered by the GPL, they can't distribute the linux kernel. This is the stance that the original complainant, Alan Cox, and others have expressed on LKML, so is a perfectly usual interpretation. And Alan Cox should know, as he's had to work with these things very closely in the last half decade, he's probably one of the kernel's most-equipped guys in this field. (But of course he immediately sics the vicious attack lawyers at the target, his skill is knowing when to do that.)
So you think they ship a linux driver that they're claiming is proprietary, and a linux kernel that is GPL, and they _aren't_ linking the two? In particular given that they've admitted "we only use Linux kernel symbols that are ..." https://lkml.org/lkml/2012/11/8/533
Wanna buy a bridge?
Licences are not contracts, according to the people who wrote the GPL:
http://www.gnu.org/philosophy/enforcing-gpl.html
Don't believe them? Would a lawyer saying "A license is not a contract" swing things?
http://www.groklaw.net/article.php?story=20031214210634851
> 2. There is at least some code that they are distributing which they do not own the copyrights to.
> If you are claiming 2, then I hope you have evidence
The product, according to RTS itself - but worded very ambiguously, seems to ship with either linux or a derivitive work of linux. They do not own the copyrights to linux.
https://lkml.org/lkml/2012/11/9/97
This is nothing to do with contract law.
*And* when the device in question *runs linux*.
"Yes, we ship GPL-ed linux; and yes, we ship a closed-source driver which was originally written as a linux kernel module; but no, that driver is never linked to the kernel that's running and runs stand-alone, and therefore is not in violation of linux's GPL.".
Yeah, right.
Not entirely on speculation. The device claims to run the GPL'ed linux kernel. Therefore the module they link to the kernel must also be GPL.
/.) indicate that he is ignorant of the meaning of the GPL. This *alone* is a fairly good reason to dismiss him from his role even if his company is innocent of the allegations.
The comments from the subsystem maintainer (like about 70% of the comments on
diff --git a/MAINTAINERS b/MAINTAINERS
index fdc0119..5859f2a 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -6735,7 +6735,6 @@ F: fs/sysv/
F: include/linux/sysv_fs.h
TARGET SUBSYSTEM
-M: Nicholas A. Bellinger <nab@linux-iscsi.org>
L: linux-scsi@vger.kernel.org
L: target-devel@vger.kernel.org
L: http://groups.google.com/group/linux-iscsi-target-dev
Try telling the environmentalists that their geothermal energy is just waste heat from a huge fission reactor deep underground, and watch their heads explode!
Intresting. I know that american football teams have huge numbers of players with different specialities, so most are not on the patch at any one time. So what's the average amount of action that each *player* participates in, per game?
I'd be willing to bet you use and AND of a whole lot of NOTs quite a lot when playing sudoku. How else will you know that only one thing is permitted in a square unless you've got 8 NOTs all ANDed together?
> if a theory is so general it can not be falsified it is not science.
Yes, but "supersymmetry" isn't "a theory", it's not the science. It's a label that is applied to the whole family of putative theories that are trying to be science, and which share a common core feature. It's not "general", it's "several". I hate to stand up for supersymmetry, as none of its expressions show the elegance that I like in science (et gustibus non disputandem est), but thinking of it as one single target that can be shot down is in error.
Not long before Newton some other guy (Galileo?) proposed the acceleration of objects falling under gravity such that the speed was proportional to the disance already moved. Newton as we know modelled it differently. The other guy's theory fell down when it was realised that an object would never start to fall. So they shot it down, and Newton's took over. That didn't mean that "gravitational acceleration" was so general it couldn't be disproved and wasn't a science.
> it makes it less interesting science
In some ways, definitely - yours seems to overlap somewhat with my 'elegance' point of view. If there is enough room to be making many many different models, then it looks like there's more guesswork involved than insight. Anyone can roll their own supersymmetric theory - download the new SuSy model GUI-based wizard trial version, and generate your own model in only 10 clicks! First 20 models free!
I'm not sure the number of cores is reliable. One figure said a quarter of a million nvidea cores. That works out at 100 boards (2.5kcores/board). Another figure from an other article
http://www.anandtech.com/show/6421/inside-the-titan-supercomputer-299k-amd-x86-cores-and-186k-nvidia-gpu-cores
claimed Titan was 18k opteron with 18k nvideas, one each. The latter implies 50 million of nvidea cores.
As the latter has a photo of a rack, with 4 opterons (16 cores each) plus 4 nvideas (2688 cores each), I'm much more likely to believe Anand than top500.org.
Nope. Because you must always presume that the attacker can try things many many many times.
If the library they're trying to return into (google "return to libc" or "ret2libc" attack if you're not familiar with common attacks) can be at one of 4096 different addresses (assuming 12 bits in the address are randomised), and assuming a probe can come in (with a slightly different payload) every 20 seconds (chosen as it might be too low for a rate limiter to notice), then you'll be rooted in less than a day, as 86400>20*4096. If you don't have a rate limiter and there are 200 attacks per second with slightly different payloads, then you'll be rooted in 20s.
I have a cryptographic background, and "less broken", is still "broken". It's "inc-/dec-reasing the constant" in the algorithm, not "inc-/dec-reasing the complexity".
An example that has a similar kind of idea, even though it's a bit unrelated, is that of Dan Bernstein's timing attacks on AES. AES - demonstrated, through all (known) mathematical and computational techniques to be secure - can be broken, if the enemy gets to probe you often enough (maybe many thousands of times, an attack that is actually practical, and has been performed in reality). Of course, in that case it was only particular (cache-sensitive, lookup table driven) implementations that suffered the attack, but it's concrete proof that you must never underestimate the power of multiple repeated queries from a malicious adversary.
Indeed, insults do not make my case, they simply decorate my case. If the slashdot audience is going to have to plough through your idiocy, they may as well at least have something that adds some amusement to the drudgery.
And you couldn't explain yourself out of a paper bag. You were totally and utterly in the wrong in the first place, and, once that was demonstrated to you and to the slashdot reading public, had the stupidity to re-iterate how wrong you were. That's either brave or stupid.
I think I know which.
Fluidity is irrelevant. Intramolecular forces are relevant. Liquids have them. Gases pretty much don't.
For ideal gases, free of intermolecular attractions, what's really most relevant is molecular speed, but (mean) speed is inversely proportional to sqrt(molecular weight). Therefore molecular weight *is* important. And for most of the atmosphere by volume, density is so low that intra-molecular attractive forces are indeed negligible, and so in the largest part of the atmosphere the gases do indeed behave like ideal gases.
Why else do you think the outermost region of the earth's atmosphere is mostly hydrogen? And the layer below that is mostly helium. And has been for your billions of years.
So please do some research before posting nonsense on slashdot. Here's a good place to start:
http://www.shodor.org/os411/courses/411b/module02/unit01/images/heterosphere.gif
The atmosphere's *way* thicker than just the blue thing that has the fluffy floaty rainy things in it, you know.
Liquids and gasses are very different beasts. One has enough intermolecular forces to bind the molecules into an effectively incompressible mass, the other hasn't, and has components that only interact with each other through random collisions. You're comparing apples to class III orange stars.
You appear to have a problem with the concept of direction. The problem being addressed was that of stuff with the evil bit set coming *in* from the internet. Moving a PC relative to the windows does not change that at all, for example. Most of your defences seem to *presume* there's a man on the inside, but if you've got a man on the inside, there's no defence if there are *any* input devices at all. And a computer with no input devices can not receive or process any information. And therefore isn't a computer any more. You're just trying to show off that you can think of lots of ways of getting data off a supposedly secure system, which is pointless willy waving, anyone can do that. Why did you mention destroying guests' phones, but not the pad of paper they had in their pocket? Or their memories?
I do understand the difference, and you are a fucking lazy idiot. If you'd have followed the link I mentioned, you'd have seen the following: ... The Professional Geographer 64(4): 602-617 ... Journal of Urban Technology 18(1): 115-132. ... Journal of Urban Technology 18(1): 93-114 ... Tijdschrift voor Economische en Sociale Geografie 101(4): 422â€"436 ... World Health and Medical Policy 2(2): 7-33. ... Environment and Planning A 42(4): 763-764. ... GeoForum 40(4): 523-534 ... GeoForum 38(6): 1322-1343 ... Environment and Planning B 34(3): 466-482. ... Societies and Cities in the Age of Instant Access. 231-244
"""
In addition to the Floatingsheep blog we have a number of academic papers in various stages of review, revision and published form.
"""
Good. Fuck off then. You clearly do not have enough intelligence to understand the simple word "insinuation", for example, so I can't expect to see much in the way of intelligent debate from you.