You can have your code under BSD license, because it's less restricting than GPL. But if anyone tries to use your application as a part of a commercial closed-source project, then they will be violating _GPL_ license of QT. Which, sort of, defeats the whole purpose of BSD license...
Except that they *can* buy a commercial license for QT, which makes everything legal. It does work, kind-of...
Well, what exactly is "the industry" if not a system put in place so artists can be rewarded for creating music? At least, that was the original idea
No, "the industry" is a system put in place so that "the industry" can make a lot of cash. This rewarding creativity thing is a post hoc rationalisation for PR reasons.
That article is interesting, but by concentrating on either very large sites or very small sites, it misses one point where Java does excel: fitting large numbers of concurrent connections on a single server. A Java servlet architecture allows session or application scope objects that remain in-memory for long periods of time. This is useful for information cacheing (thus reducing the number of database queries the application will need to perform) and for maintaining user session information. Techniques for persisting objects between queries in PHP rely on serializing them at the end of one query and deserializing them for the next, which is an unnecessary overhead.
Did anyone notice these chips are using the 65nm process? At what point do the shear quantum affects overcome the deterministic EE rules that are used to design the chips? I don't know, but wikipedia defines a nanoparticle as one with at least one dimension less than 100nm. http://en.wikipedia.org/wiki/Nanoparticle Given that definition every transistor's source, drain and gate are nanoparticles. And we expect them to behave classically why?
Because experimental evidence suggests that they do. I think current theory is that transistors will behave deterministically down to at least 20nm. I'm really not sure what the relevance of nanoparticles is; the cut off there is entirely arbitrary and probably corresponds to the point that people generally start measuring in nanometres rather than fractions of microns. There's no particular physical effect that suddenly starts occuring at 100nm.
My understanding is that these problems allow someone who has the ability to execute code on the system to escalate privileges (get root).
This is true.
Single-user systems would not be affected.
This is not. Ther are huge numbers of issues here, with a wide variety of potential effects.
For instance: "General Protection (#GP) Fault May Not Be Signaled on Data Segment Limit Violation above 4-G Limit". This could result in the situation you describe in your first sentence. Also, software may rely on it to prevent misbehaviours of all other kinds: remote buffer overflows for instance, might be caught in some software by setting a data segment limit on the segment in use and assuming it will be caught. Other software might plausibly run in a loop until the violation occurs thus indicating it has run out of data to process in memory; such software may crash on a Core 2, and because no details are given about when this occurs ("May Not Be Signmaled"), we don't know that it couldn't be an intermittent failure.
"(E)CX May Get Incorrectly Updated When Performing Fast String REP MOVS or Fast String REP STOS With Large Data Structures" -- This could almost certainly cause buffer overflows in some software. Much more plausibly than the last one. Apparently a fix is available for this issue, but clearly not for open source users.
"Incorrect Address Computed For Last Byte of FXSAVE/FXRSTOR Image Leads to Partial Memory Update" -- This could be a more serious calculation accuracy problem than the Pentium FDIV bug was for some programs.
"Cache Data Access Request from One Core Hitting a Modified Line in the L1 Data Cache of the Other Core May Cause Unpredictable System Behavior" or "Concurrent Multi-processor Writes to Non-dirty Page May Result in Unpredictable Behavior" -- Multithreaded programs that access the same data in multiple threads can make anything get screwed up.
"PREFETCH Instruction Execution under Some Conditions May Lead to Processor Livelock", "REP Store Instructions in a Specific Situation may cause the Processor to Hang" -- I don't think I need comment further on these.
The number of cores is expected to double every 18 months for the next decade and reach 256 in a decade.
Right. Not sure I'm with you there. 256 cores is a lot, and I doubt that the infrastructure of (e.g.) memory bandwidth and power supply would be able to keep up with such demands.
Clock rates of commodity processors have stopped improving since mid-2003. This followed several decades in which clock rates have doubled every 18 months.
Right. You know, I'm sure the fastest desktop processor you could buy in June 2003 had a clock speed of about 3GHz. Clearly I'm imagining the availability of 4GHz chips on the market today. Yeah, sure, it's slowed down. It hasn't stopped, though. I'm also clearly imagining that Core2 chips achieve more calculations per core-second than Pentium 4 chips running at significantly faster clock speeds.
Basically, the entire thesis here is that improvement in individual processor core performance has been halted for the last 4 years. This blantantly is not the case.
[Vishkin] opined that a mathematical model, called PRAM for Parallel Random-Access Model (or Machine), would be a proper framework. The PRAM is a simple extension to the standard RAM (for random access machine) model used to teach serial algorithms in every standard Computer Science curriculum.
Can somebody help me out here. I've never heard of this "random access machine" model. Are we talking about a von Neumann machine, or something else?
In his PhD thesis, Vishkin proposed a simple "work-depth" methodology for designing parallel algorithms: Formulate your algorithm in the form of rounds, where each round can include any number of operations that could all be performed concurrently had there been enough hardware to execute them. For performance, the design objective should be to minimize two parameters: (i) work - the total number of operations over all rounds, and (ii) depth - the number of rounds. A simple example for such a parallel algorithm follows.
Well, duh. Thanks for enlightening us. To improve performance, minimize the number of steps that must be taken sequentially, and perform as many as possible in parallel, but don't make too many parallel ones either. Clearly revolutionary thinking, there.
PRAM algorithms often allow an "arbitrary concurrent write" resolution, where several concurrent attempts to write into the same memory location result in one of these writes, but we don't know in advance which one. Note also that such "semantics" cannot even be expressed in any of the common serial programming languages.
...
Err, OK. Me, I thought starting several threads and making them all write to the same location would result in an unpredictable choice of the values written being stored in that location in standard languages. But then I don't have a PhD in parallel programming techniques, just ten years of industry experience writing multithreaded software, so what do I know?
All successful general-purpose computers since the 1940s rely on the so-called Von- Neumann apparatus. Is there a way to upgrade, rather than completely replace, this successful apparatus to handle parallelism?
Yes. You place multiple von neumann machines with access to the same memory, and provide wtructures for sending control signals between them so that a thread on one processor can start a new thread on another one. This is generally called SMP, and has been used extensively since a long time before this paper was written, so why are you even asking the question? There are of course alternative approaches (e.g. NUMA) that can provide better efficiency in some cases, but the basic question is answered already.
To be fair, if this crowd had a version of their chip implemented on 65nm silicon, it would probably outperform the Cell in several key areas. For a start, it has a maximum parallelism of 64 simultaneous instructions -- I believe the Cell can only reach 10 (?). Of course, writing a real program that takes advantage of that much parallelism is a little tricky...
Anyone remember the hype of the i860? Great on paper, but not so great in reality. I really hope this works though, von Neuman architecture was always supposed to be a stop-gap (even vN said so I think).
As far as I can tell, there's no really significant departure from von neumann architecture here. They have a processor capable of executing 64 concurrent threads, 'fork' and 'join' instructions, and a version of C that has been extended to be able to use them. I'm not sure I really see what's so revolutionary here -- I've been reading about prototypes of similar ideas to this since the late 90s.
I have no idea how exactly the metal roof is destroying the signal, whether it is causing severe multipath or simply absorbing it completely, but it does it quite effectively.
It'll be the former. It can't absorb signals that aren't sent in its direction (which typically the ones you want won't be), but what it can do is reflect back lots of slightly-out-of-phase signals from different points that confuse your receivers. Stick up a metal plate about a metre behind your TV antenna and see what happens to the signal. I'm not sure whether ensuring it's earthed would help here -- it might cause it to absorb more of the signal and reflect less, which could be enough to make your network work.
The real problem is that the frequency is substantially higher than mobile phone frequencies. 2.4GHz has substantially less capacity to penetrate stuff that might be in the way than 900MHz.
Wifi should be a supplement for larger areas with changing configurations (meeting rooms, conference rooms, etc) not as a replacement for a typical wired setup. Unless you plan on constantly rearranging your cubes, there's no reason to do that.
Except cash. Installing cabling can be quite expensive. A 5 person office could easily cost £500 to cable up, including health & safety inspections (which may or may not be mandatory depending on your locality, but are almost always a good idea when installing cables around employees). Wireless networking hardware for that same office might come in at £300. Savings almost certainly scale up with larger offices.
There are other reasons, too... I visited an IBM office a few years back where the employees were expected to "hot desk": they didn't have an assigned place to work, and when turning up just found a free desk anywhere and hooked in to the wireless network. They had a mobile phone base unit, so their mobile phones turned into office phones while they were in the building, etc. Management ensured there were slightly fewer convenient places to work than there were employees in the building on an average day -- encourages people to turn up early.
Yes, I've had the same idea in the past. Kind of like a difference disk image without the emulated hardware, just done at an API level. Install software anywhere just by copying its files. Protect conflicting apps from each other by having them in separate filesystems. It's a promising idea...
I don't understand the logic here. A company has the right to use any service or media codec they want. So why do FOSSies whine so much when a company chooses Windows Media Player or WMV?
The BBC isn't simply "a company": it's a nationalised organisation that exists by governmental charter. Part of its charter is that it should be free from commercial influence. We're worried that the decision to use MS formats (and hence excluding a not insignificant proportion of the potential users of the service) has been based less on technical suitability for the purpose than it has on Microsoft's commercial influence.
Since the MacroLife concept isn't widely spoken of, let me elucidate: 1) You build a space-based factory. 2) You build a colony nearby to manage it. 3) People get comfortable living in the colony, and enlarge it, and make it self-sufficient. 4) There's a political dispute. 5) People living in the colony attach an engine, and depart slowly for "elsewhere". 6) You don't want a tremendously high speed, because you collect materials along the way.
Interesting. Have you read "The Galactic Whirlpool" by David Gerrold (bad title, but really rather good book)? It follows a similar hypothesis. Except for point 6.
The physical limits of the current mechanism of the trend (i.e., the reduction in feature size of transistors) are probably about 15-20 years away. After that, different mechanisms of improvement will be necessary, but as Kurzweil has pointed out every time we've reached the limits of some particular technology in the past, a new one has become available just as development is tailing off which extends the trend even further. Will this happen when we reach the limit of silicon? I can't be sure, but it seems a little short-sighted just to assume it won't.
a human [...] averages over 100 watts if they expend 2500kcal in a day.
2500kcal / 1 day =~ 10,000KJ / 1 day = 10,000KWs / 86,400s = 116W.
That's not far off, plus 2500kcal is the recommended energy intake for an adult male. Others have lower recommended energy intakes. 100W average is probably about right.
Show me an AI which consumes less than 100 Watts of energy and costs less than 1 dollar per day and I'll start to believe in "The Singularity".
Can you suggest any reason why such a thing is inplausible? Yes, current generation computation devices are power-hungry and expensive to run. But this is probably an artifact of current technological limitations, rather than a law of nature. There is a known minimal thermal energy consumption per non-reversible calculation. We're nowhere near that level at the moment.
I see no reason to assume that the human brain is unique in its capacity to reach these levels of efficiency, and little reason to believe that it cannot be greatly surpassed.
How about suing for specific action in compliance with the agreement? EULAs, for example, contain a clause that the copyright owner can force you to *destroy* your media. Suing to enforce the agreement is just as valid
In general, you can't do it without evidence of agreement. As it's perfectly possible to not agree to the GPL and then commit copyright infringement by redistributing, saying this is what happened is always an option if a violator really wanted to hold onto their code.
It isn't, except doing so kind of defeats the purpose of not dynamically linking, because it's just as hard to make it work. It also makes the obfuscation the OP was asking about in his next question even less useful than it was to start with.
Spammers operate because of the enormous economies of scale that exist with email. You can send out literally millions of emails for practically nothing.
Speaking as a web developer who once had a client send out a batch of spam, I can tell you that having his web site (which supplied only to UK residents) hit simultaneously by tens of thousands of chinese people really did cost him a fair amount.
Slashdot Mirror
You can have your code under BSD license, because it's less restricting than GPL. But if anyone tries to use your application as a part of a commercial closed-source project, then they will be violating _GPL_ license of QT. Which, sort of, defeats the whole purpose of BSD license...
Except that they *can* buy a commercial license for QT, which makes everything legal. It does work, kind-of...
Well, what exactly is "the industry" if not a system put in place so artists can be rewarded for creating music? At least, that was the original idea
No, "the industry" is a system put in place so that "the industry" can make a lot of cash. This rewarding creativity thing is a post hoc rationalisation for PR reasons.
That article is interesting, but by concentrating on either very large sites or very small sites, it misses one point where Java does excel: fitting large numbers of concurrent connections on a single server. A Java servlet architecture allows session or application scope objects that remain in-memory for long periods of time. This is useful for information cacheing (thus reducing the number of database queries the application will need to perform) and for maintaining user session information. Techniques for persisting objects between queries in PHP rely on serializing them at the end of one query and deserializing them for the next, which is an unnecessary overhead.
You sure? (OK, so it needs an eZ80, not a Z80, and its for direct OS-less embedding not CP/M, but still...)
Did anyone notice these chips are using the 65nm process?
At what point do the shear quantum affects overcome the deterministic EE rules that are used to design the chips? I don't know, but wikipedia defines a nanoparticle as one with at least one dimension less than 100nm. http://en.wikipedia.org/wiki/Nanoparticle
Given that definition every transistor's source, drain and gate are nanoparticles. And we expect them to behave classically why?
Because experimental evidence suggests that they do. I think current theory is that transistors will behave deterministically down to at least 20nm. I'm really not sure what the relevance of nanoparticles is; the cut off there is entirely arbitrary and probably corresponds to the point that people generally start measuring in nanometres rather than fractions of microns. There's no particular physical effect that suddenly starts occuring at 100nm.
My understanding is that these problems allow someone who has the ability to execute code on the system to escalate privileges (get root).
This is true.
Single-user systems would not be affected.
This is not. Ther are huge numbers of issues here, with a wide variety of potential effects.
For instance: "General Protection (#GP) Fault May Not Be Signaled on Data Segment Limit Violation above 4-G Limit". This could result in the situation you describe in your first sentence. Also, software may rely on it to prevent misbehaviours of all other kinds: remote buffer overflows for instance, might be caught in some software by setting a data segment limit on the segment in use and assuming it will be caught. Other software might plausibly run in a loop until the violation occurs thus indicating it has run out of data to process in memory; such software may crash on a Core 2, and because no details are given about when this occurs ("May Not Be Signmaled"), we don't know that it couldn't be an intermittent failure.
"(E)CX May Get Incorrectly Updated When Performing Fast String REP MOVS or Fast String REP STOS With Large Data Structures" -- This could almost certainly cause buffer overflows in some software. Much more plausibly than the last one. Apparently a fix is available for this issue, but clearly not for open source users.
"Incorrect Address Computed For Last Byte of FXSAVE/FXRSTOR Image Leads to Partial Memory Update" -- This could be a more serious calculation accuracy problem than the Pentium FDIV bug was for some programs.
"Cache Data Access Request from One Core Hitting a Modified Line in the L1 Data Cache of the Other Core May Cause Unpredictable System Behavior" or "Concurrent Multi-processor Writes to Non-dirty Page May Result in Unpredictable Behavior" -- Multithreaded programs that access the same data in multiple threads can make anything get screwed up.
"PREFETCH Instruction Execution under Some Conditions May Lead to Processor Livelock", "REP Store Instructions in a Specific Situation may cause the Processor to Hang" -- I don't think I need comment further on these.
Right. Not sure I'm with you there. 256 cores is a lot, and I doubt that the infrastructure of (e.g.) memory bandwidth and power supply would be able to keep up with such demands.
Right. You know, I'm sure the fastest desktop processor you could buy in June 2003 had a clock speed of about 3GHz. Clearly I'm imagining the availability of 4GHz chips on the market today. Yeah, sure, it's slowed down. It hasn't stopped, though. I'm also clearly imagining that Core2 chips achieve more calculations per core-second than Pentium 4 chips running at significantly faster clock speeds.
Basically, the entire thesis here is that improvement in individual processor core performance has been halted for the last 4 years. This blantantly is not the case.
Can somebody help me out here. I've never heard of this "random access machine" model. Are we talking about a von Neumann machine, or something else?
Well, duh. Thanks for enlightening us. To improve performance, minimize the number of steps that must be taken sequentially, and perform as many as possible in parallel, but don't make too many parallel ones either. Clearly revolutionary thinking, there.
Err, OK. Me, I thought starting several threads and making them all write to the same location would result in an unpredictable choice of the values written being stored in that location in standard languages. But then I don't have a PhD in parallel programming techniques, just ten years of industry experience writing multithreaded software, so what do I know?
All successful general-purpose computers since the 1940s rely on the so-called Von-
Neumann apparatus. Is there a way to upgrade, rather than completely replace, this
successful apparatus to handle parallelism?
Yes. You place multiple von neumann machines with access to the same memory, and provide wtructures for sending control signals between them so that a thread on one processor can start a new thread on another one. This is generally called SMP, and has been used extensively since a long time before this paper was written, so why are you even asking the question? There are of course alternative approaches (e.g. NUMA) that can provide better efficiency in some cases, but the basic question is answered already.
To be fair, if this crowd had a version of their chip implemented on 65nm silicon, it would probably outperform the Cell in several key areas. For a start, it has a maximum parallelism of 64 simultaneous instructions -- I believe the Cell can only reach 10 (?). Of course, writing a real program that takes advantage of that much parallelism is a little tricky...
Anyone remember the hype of the i860? Great on paper, but not so great in reality. I really hope this works though, von Neuman architecture was always supposed to be a stop-gap (even vN said so I think).
As far as I can tell, there's no really significant departure from von neumann architecture here. They have a processor capable of executing 64 concurrent threads, 'fork' and 'join' instructions, and a version of C that has been extended to be able to use them. I'm not sure I really see what's so revolutionary here -- I've been reading about prototypes of similar ideas to this since the late 90s.
I have no idea how exactly the metal roof is destroying the signal, whether it is causing severe multipath or simply absorbing it completely, but it does it quite effectively.
It'll be the former. It can't absorb signals that aren't sent in its direction (which typically the ones you want won't be), but what it can do is reflect back lots of slightly-out-of-phase signals from different points that confuse your receivers. Stick up a metal plate about a metre behind your TV antenna and see what happens to the signal. I'm not sure whether ensuring it's earthed would help here -- it might cause it to absorb more of the signal and reflect less, which could be enough to make your network work.
The real problem is that the frequency is substantially higher than mobile phone frequencies. 2.4GHz has substantially less capacity to penetrate stuff that might be in the way than 900MHz.
Wifi should be a supplement for larger areas with changing configurations (meeting rooms, conference rooms, etc) not as a replacement for a typical wired setup. Unless you plan on constantly rearranging your cubes, there's no reason to do that.
Except cash. Installing cabling can be quite expensive. A 5 person office could easily cost £500 to cable up, including health & safety inspections (which may or may not be mandatory depending on your locality, but are almost always a good idea when installing cables around employees). Wireless networking hardware for that same office might come in at £300. Savings almost certainly scale up with larger offices.
There are other reasons, too... I visited an IBM office a few years back where the employees were expected to "hot desk": they didn't have an assigned place to work, and when turning up just found a free desk anywhere and hooked in to the wireless network. They had a mobile phone base unit, so their mobile phones turned into office phones while they were in the building, etc. Management ensured there were slightly fewer convenient places to work than there were employees in the building on an average day -- encourages people to turn up early.
Yes, I've had the same idea in the past. Kind of like a difference disk image without the emulated hardware, just done at an API level. Install software anywhere just by copying its files. Protect conflicting apps from each other by having them in separate filesystems. It's a promising idea...
I don't understand the logic here. A company has the right to use any service or media codec they want. So why do FOSSies whine so much when a company chooses Windows Media Player or WMV?
The BBC isn't simply "a company": it's a nationalised organisation that exists by governmental charter. Part of its charter is that it should be free from commercial influence. We're worried that the decision to use MS formats (and hence excluding a not insignificant proportion of the potential users of the service) has been based less on technical suitability for the purpose than it has on Microsoft's commercial influence.
ROTFL.
I take it you haven't read any of his books, then? Common themes being nanotech, AI and space travel...
Since the MacroLife concept isn't widely spoken of, let me elucidate:
1) You build a space-based factory.
2) You build a colony nearby to manage it.
3) People get comfortable living in the colony, and enlarge it, and make it self-sufficient.
4) There's a political dispute.
5) People living in the colony attach an engine, and depart slowly for "elsewhere".
6) You don't want a tremendously high speed, because you collect materials along the way.
Interesting. Have you read "The Galactic Whirlpool" by David Gerrold (bad title, but really rather good book)? It follows a similar hypothesis. Except for point 6.
Yes, but when?
The physical limits of the current mechanism of the trend (i.e., the reduction in feature size of transistors) are probably about 15-20 years away. After that, different mechanisms of improvement will be necessary, but as Kurzweil has pointed out every time we've reached the limits of some particular technology in the past, a new one has become available just as development is tailing off which extends the trend even further. Will this happen when we reach the limit of silicon? I can't be sure, but it seems a little short-sighted just to assume it won't.
a human [...] averages over 100 watts if they expend 2500kcal in a day.
2500kcal / 1 day =~ 10,000KJ / 1 day = 10,000KWs / 86,400s = 116W.
That's not far off, plus 2500kcal is the recommended energy intake for an adult male. Others have lower recommended energy intakes. 100W average is probably about right.
Show me an AI which consumes less than 100 Watts of energy and costs less than 1 dollar per day and I'll start to believe in "The Singularity".
Can you suggest any reason why such a thing is inplausible? Yes, current generation computation devices are power-hungry and expensive to run. But this is probably an artifact of current technological limitations, rather than a law of nature. There is a known minimal thermal energy consumption per non-reversible calculation. We're nowhere near that level at the moment.
I see no reason to assume that the human brain is unique in its capacity to reach these levels of efficiency, and little reason to believe that it cannot be greatly surpassed.
How about suing for specific action in compliance with the agreement? EULAs, for example, contain a clause that the copyright owner can force you to *destroy* your media. Suing to enforce the agreement is just as valid
In general, you can't do it without evidence of agreement. As it's perfectly possible to not agree to the GPL and then commit copyright infringement by redistributing, saying this is what happened is always an option if a violator really wanted to hold onto their code.
And this one suggests it still was as recently as late '05, so it seems to really be a recent change.
Hmm. It seems to have changed recently (for some value of recently).
See this mailing list post for a copy of it as it was a few years ago.
Did you read my entire post before you posted your reply?
It isn't, except doing so kind of defeats the purpose of not dynamically linking, because it's just as hard to make it work. It also makes the obfuscation the OP was asking about in his next question even less useful than it was to start with.
Spammers operate because of the enormous economies of scale that exist with email. You can send out literally millions of emails for practically nothing.
Speaking as a web developer who once had a client send out a batch of spam, I can tell you that having his web site (which supplied only to UK residents) hit simultaneously by tens of thousands of chinese people really did cost him a fair amount.