RMS doesn't assume that humanity has a hacker complex: he neither insists nor expects that most users of Free software do anything other than use the binaries.
RMS doesn't think using an application that someone else hosts is illegal: he suggests Free software projects use Savannah, a Free web app.
In fact, RMS wholeheardedly supports commercialisation of Free software. After leaving his university position, he made his first bunch of cash by selling binaries of Free software.
SMTP servers that you use, he believes, should use software that is Free so, should you decide, you can buy some hardware and bandwidth and host your own email server with alterarions you decide.
RMS doesn't believe open source "is as evil as any other software". Indeed he doesn't care about open source as a movement. He can't hate the software since most of it is technically Free software too. He just doesn't like the philosophy of software freedom being confused with the engineering method of open source.
Obviously I'm not RMS and can't speak for the man himself, but it seems to me that you're being rather unfair on him.
Everyone knows that debugging is twice as hard as writing a program in the first place. So if you're as clever as you can be when you write it, how will you ever debug it?
* "The Elements of Programming Style", 2nd edition, chapter 2
Firstly, you can file many items in the same label at once far easier. It's just an extra step per mail, rather than an extra "2" in Yahoo mail.
Secondly, drag and drop is hardly "one step". At best it's the equivalent of two clicks and I'd probably argue that it's more like 2.5 click in terms of effort. Why? Aim+Press, Aim+release. And you've got the fact that you have to exert effort to keep the button pressed while aiming the second time.
I would suggest that Gmail could be made more convenient by perhaps allowing an option to automatically archive when a label is given, but definitely only an option since there are those who like to file mails with multiple labels.
How about the various form of twinning that occur, which in rare cases leads to one twin actually becoming part of the other, and needing to be removed so that the fully grown twin can live? That other twin (which cannot survive in any scenario) is human, and it is its own entity
There are two entirely different scenarios being posed. In the situation above, you state that there are two human lives at risk - how do you balance between them. It's the same as saying "You're wife and child are dangling from different cliffs. Both could fall at any moment. You have time to save one - which do you choose?" It's a moral dilemma, a no-win situation - whichever way you choose, a human dies, and your choice will be based upon this knowledge.
They're actually quite similar in the respects that matter. I'll alter your analogy to demonstrate. So they're dangling from different cliffs, fine. You have two ropes, one in each hand. On the left side is one life, on the right another. But the difference is that while both ropes are breaking, one rope is *way* stronger than the other. Two "rational" choices:
1. Do nothing; Rationale: We can't make a decision that would kill a life. Outcome: Both ropes break eventually and they both die.
2. Let go of the weaker rope and pull on the stronger rope to bring it to safety. Rationale: Save the one with most chance of surviving, even when you have to kill the other to do so. Outcome: You probably save a life.
I'd go for number 2 every time, though as the chances of survival for either become more similar (i.e. the weaker rope becomes more similar in strength to the stronger rope), the case for action becomes weaker.
See the current front-page interview with Chavez on aljazeera.net.
He explicitly states he is not anti-US, he has good friends who are US citizens and he has nothing against the US people. Indeed Venezuela has actually double the amount of fuel it donates to poor US citizens over the winter months when they're own government is quite happy to let them freeze.
He *is* anti-Bush which due to your political system of confusing the head of state with the executive leader means that some of your citizens will take them to mean the same thing. But as the statisticians have shown Bush wasn't actually duely elected; he and his cronies threw it. Chavez was put in charge by his population.
For the Slashdot editors to publish on the front-page "Chavez is Anti-US" is a new low for Slashdot.
Actually, I've just finished a PhD, whereby I had to compare the performance between a Pentium 4 HT @ 2.8Ghz, an AMD Athlon XP 2200 and an Intel Pentium-M 1.7GHz. The nature of the tests was audio signal processing.
Though the Athlon performed better than the Pentium-M, it was quite comparable, and both far outperformed the Pentium 4. This is despite the far lower power/cooling requirements of the Pentium-M.
I imagine if you search for general performance characteristics of the Pentium-M you might find it is better than you think.
OK, first off, with a GPL'd client, you _cannot_ assume the client is secure. There is nothing stopping a would-be cheater analysing the data-transportation code
(i.e. client-server protocol), and then telnet'ing to the server, logging on and telling the server he has load of money/items/ex-points/whatever the server will believe him saying. You cannot assume the client will be trust worthy. Period.
However, there are ways to burden the server with less info - you just have to make sure that the client can only request to do 'legal' things. If you think about a cpu that has two modes - OS kernel mode and app space mode, where in app space mode it can only do 'legel' things (i.e. stuff that doesn't mess other stuff up), then the same can be applied to the game. Of course this all depends on how they've designed the games upto this point anyway. If they cannot break the operations/requests to validatable stuff then this could all be irrelevent anyway.
Other than having very thin clients (read: dumb) and a super-server then security is a hard aspect of open software to solve. There may be another way out though.
Suppose you have three totally random and arbitrarily chosen clients A, B and C. Now suppose A wants to do action X (like move to some point in space). You cant trust A to tell you whether it's legal of him/her to do X themselves - they might lie in order to cheat. But get this: you coudl ask B and C togther to see if A could do X - if and only if they both say yes should A be allowed. Otherwise (if they both say no) A should be told off or (if they disagree) B&C should bew told off and rechosen. This allows some wieght to be taken from the server and processing distributed. It does need B and C to know all about A (in order for them to make the decision properly). In addition to this it is considerably more secure that assuming the clients will always tell the truth and only do what they're allowed.
The other way cheats can break the game is through snooping on other people's data streams. This again does not need security through obscurity - a good encryption scheme should not rely on secrecy of the algorithm (cipher) only of the key. For the ultimate security, a one-time-pad would be a good idea, which is implemented in HardEncrypt (there's a/. article a while back on this). Any thing else could lead to successful snooping.
OK two things about your post that I don't think are particularely coherent.
1: You have not given any specific proof or explicit examples that show free-speech/open-source/whatever is _genuinely opposing corporatism/capitalism: Marx's quote is quite right - _between_ the states True and False there is nothing - no comprimise. These are genuinely opposing quantities; their definition is based as the opposite of the other. However, I would challenge you to prove that commerce/etc. is fundamentally based as the opposite of free-speech/open-source/etc.
2: Again, possibly some mis-use of Marx: Suppose the world was made of two types of people (or entity); one who fundamentally supported free-speech/etc. and the other who fundamentally opposed it. Fine - but there are many, many such entities. The world exists not as one big True and another big False, but as lots of (relatively) little 'True's and 'False's. When combined into a macro-model, these 'appear' as smooth gradients. For example, suppose you have a picture made of discrete elements - 'dots' - each one can be either black or white. Now lets distribute them randomly throughout: Look close, and there are lots of black dots, and lots of white dots. Look from afar and take in the whole picture, and you no longer see the dots, but smooth greys. And sure, for some simple pictures with low-resolutions, you may see the individual dots, but this is the _world_ - see it whole and it all combines into one cosmopolitan blur.
One the other hand, suppose the world (as I believe it is) is made of entities (people/organisations/etc.) that have no rock-solid well-defined fundamentals (as you seem to be quoting Marx to tell us all that it is), but who exist in the world as analogue tones between the two (non-basically-opposing) extremes. Where does that leave your argument?
Sure - you have you're opinion, just as Brian does, but don't pretend to give us a rock-solid proof quoting a well-respected historical figure any more than the bloke before you.
Disclaimer: I know only a modest amount about neuropsychology/theory of computation...
OK, I've thought about this for a while and here goes:
As far as I can tell, there is very little we actually *know* about the brain. It cannot be likened in any definate way to a (theoretically) lossless accurate finite state machine/automaton, so there is no hard and fast way to answer this question. Get rid of that 122 meg estimate - he's obviously talking about something else;).
Basically, from what I can tell, the brain works _as a whole_ to store any one given 'item of information'. This is unlike a conventional computer that has a specific location to store a specific bit of data.
Any storage device on a conventional computer (as far as I know) uses some sort of definate addressing mechanism to access a particular peice of data.
On the other hand the brain "stores" pretty much everything you experienced, whether you remember it or not. The problem is how it is/addressed/. In the brain's case, there may be loads of 'data' stored, its just you cant remember the 'links' needed to actually get to it - ever been caught out by that word you know, but you just cant quite remember? (I know I have;).
The way the brain 'records' experiences is by changing (m)any neorones' receptibility to neighbouring cells as well as those neorones' internal chemical (in)balance. This would give a "random" (hey - never use the 'r' word without the quotes) yet analogue data storage that clearly cannot be enumerated in any definative way to bytes.
I believe it's entirely possible that the brain could even 'record' data in more outlandish methods such as small inductive/capacitive fields and chaotic electrical interference. What is definately known is that the way the brain stores data is probably more than the "just count the number of cells and multiply by ten" way shown in one of the replies. Such a way would be inmature, inaccurate and an insult to the more thoughtful amongst us.
Now, since I've said all that about just how impossible it is to measure the maximum capacity of the brain, what is probably a more practical answer is how much the brain will store on average. This I dont know, though I guess that that would be a question for the psychologists out there - and maybe thats where the 122 meg guess comes into it.
Slashdot Mirror
And where, specifically does he say it's ok to buy access to someone else's SMTP server when said server uses non-Free software?
As for Google, agm is exactly right: he shouldn't use it (or any of its non-Free services) according to his own philosophy.
And for what it's worth I owe you no explanation; I'm just trying to do my bit to kurb ignorant ranting by presenting the facts.
RMS doesn't assume that humanity has a hacker complex: he neither insists nor expects that most users of Free software do anything other than use the binaries.
RMS doesn't think using an application that someone else hosts is illegal: he suggests Free software projects use Savannah, a Free web app.
In fact, RMS wholeheardedly supports commercialisation of Free software. After leaving his university position, he made his first bunch of cash by selling binaries of Free software.
SMTP servers that you use, he believes, should use software that is Free so, should you decide, you can buy some hardware and bandwidth and host your own email server with alterarions you decide.
RMS doesn't believe open source "is as evil as any other software". Indeed he doesn't care about open source as a movement. He can't hate the software since most of it is technically Free software too. He just doesn't like the philosophy of software freedom being confused with the engineering method of open source.
Obviously I'm not RMS and can't speak for the man himself, but it seems to me that you're being rather unfair on him.
Gav.
It was Brian Kernighan:
Everyone knows that debugging is twice as hard as writing a program in the first place. So if you're as clever as you can be when you write it, how will you ever debug it?
* "The Elements of Programming Style", 2nd edition, chapter 2
That's not very fair.
Firstly, you can file many items in the same label at once far easier. It's just an extra step per mail, rather than an extra "2" in Yahoo mail.
Secondly, drag and drop is hardly "one step". At best it's the equivalent of two clicks and I'd probably argue that it's more like 2.5 click in terms of effort. Why? Aim+Press, Aim+release. And you've got the fact that you have to exert effort to keep the button pressed while aiming the second time.
I would suggest that Gmail could be made more convenient by perhaps allowing an option to automatically archive when a label is given, but definitely only an option since there are those who like to file mails with multiple labels.
How about the various form of twinning that occur, which in rare cases leads to one twin actually becoming part of the other, and needing to be removed so that the fully grown twin can live? That other twin (which cannot survive in any scenario) is human, and it is its own entity
There are two entirely different scenarios being posed. In the situation above, you state that there are two human lives at risk - how do you balance between them. It's the same as saying "You're wife and child are dangling from different cliffs. Both could fall at any moment. You have time to save one - which do you choose?" It's a moral dilemma, a no-win situation - whichever way you choose, a human dies, and your choice will be based upon this knowledge.
They're actually quite similar in the respects that matter. I'll alter your analogy to demonstrate. So they're dangling from different cliffs, fine. You have two ropes, one in each hand. On the left side is one life, on the right another. But the difference is that while both ropes are breaking, one rope is *way* stronger than the other. Two "rational" choices:
1. Do nothing; Rationale: We can't make a decision that would kill a life. Outcome: Both ropes break eventually and they both die.
2. Let go of the weaker rope and pull on the stronger rope to bring it to safety. Rationale: Save the one with most chance of surviving, even when you have to kill the other to do so. Outcome: You probably save a life.
I'd go for number 2 every time, though as the chances of survival for either become more similar (i.e. the weaker rope becomes more similar in strength to the stronger rope), the case for action becomes weaker.
See the current front-page interview with Chavez on aljazeera.net.
He explicitly states he is not anti-US, he has good friends who are US citizens and he has nothing against the US people. Indeed Venezuela has actually double the amount of fuel it donates to poor US citizens over the winter months when they're own government is quite happy to let them freeze.
He *is* anti-Bush which due to your political system of confusing the head of state with the executive leader means that some of your citizens will take them to mean the same thing. But as the statisticians have shown Bush wasn't actually duely elected; he and his cronies threw it. Chavez was put in charge by his population.
For the Slashdot editors to publish on the front-page "Chavez is Anti-US" is a new low for Slashdot.
Actually, I've just finished a PhD, whereby I had to compare the performance between a Pentium 4 HT @ 2.8Ghz, an AMD Athlon XP 2200 and an Intel Pentium-M 1.7GHz. The nature of the tests was audio signal processing.
Though the Athlon performed better than the Pentium-M, it was quite comparable, and both far outperformed the Pentium 4. This is despite the far lower power/cooling requirements of the Pentium-M.
I imagine if you search for general performance characteristics of the Pentium-M you might find it is better than you think.
OK, first off, with a GPL'd client, you _cannot_ assume the client is secure. There is nothing stopping a would-be cheater analysing the data-transportation code
/. article a while back on this). Any thing else could lead to successful snooping.
(i.e. client-server protocol), and then telnet'ing to the server, logging on and telling the server he has load of money/items/ex-points/whatever the server will believe him saying. You cannot assume the client will be trust worthy. Period.
However, there are ways to burden the server with less info - you just have to make sure that the client can only request to do 'legal' things. If you think about a cpu that has two modes - OS kernel mode and app space mode, where in app space mode it can only do 'legel' things (i.e. stuff that doesn't mess other stuff up), then the same can be applied to the game. Of course this all depends on how they've designed the games upto this point anyway. If they cannot break the operations/requests to validatable stuff then this could all be irrelevent anyway.
Other than having very thin clients (read: dumb) and a super-server then security is a hard aspect of open software to solve. There may be another way out though.
Suppose you have three totally random and arbitrarily chosen clients A, B and C. Now suppose A wants to do action X (like move to some point in space). You cant trust A to tell you whether it's legal of him/her to do X themselves - they might lie in order to cheat. But get this: you coudl ask B and C togther to see if A could do X - if and only if they both say yes should A be allowed. Otherwise (if they both say no) A should be told off or (if they disagree) B&C should bew told off and rechosen. This allows some wieght to be taken from the server and processing distributed. It does need B and C to know all about A (in order for them to make the decision properly). In addition to this it is considerably more secure that assuming the clients will always tell the truth and only do what they're allowed.
The other way cheats can break the game is through snooping on other people's data streams. This again does not need security through obscurity - a good encryption scheme should not rely on secrecy of the algorithm (cipher) only of the key. For the ultimate security, a one-time-pad would be a good idea, which is implemented in HardEncrypt (there's a
Gav
hey - it's funny.
get a sense of humour as well as your daily shot of m$ propaganda
OK two things about your post that I don't think are particularely coherent.
1: You have not given any specific proof or explicit examples that show free-speech/open-source/whatever is _genuinely opposing corporatism/capitalism: Marx's quote is quite right - _between_ the states True and False there is nothing - no comprimise. These are genuinely opposing quantities; their definition is based as the opposite of the other. However, I would challenge you to prove that commerce/etc. is fundamentally based as the opposite of free-speech/open-source/etc.
2: Again, possibly some mis-use of Marx:
Suppose the world was made of two types of people (or entity); one who fundamentally supported free-speech/etc. and the other who fundamentally opposed it. Fine - but there are many, many such entities. The world exists not as one big True and another big False, but as lots of (relatively) little 'True's and 'False's. When combined into a macro-model, these 'appear' as smooth gradients.
For example, suppose you have a picture made of discrete elements - 'dots' - each one can be either black or white. Now lets distribute them randomly throughout: Look close, and there are lots of black dots, and lots of white dots. Look from afar and take in the whole picture, and you no longer see the dots, but smooth greys. And sure, for some simple pictures with low-resolutions, you may see the individual dots, but this is the _world_ - see it whole and it all combines into one cosmopolitan blur.
One the other hand, suppose the world (as I believe it is) is made of entities (people/organisations/etc.) that have no rock-solid well-defined fundamentals (as you seem to be quoting Marx to tell us all that it is), but who exist in the world as analogue tones between the two (non-basically-opposing) extremes. Where does that leave your argument?
Sure - you have you're opinion, just as Brian does, but don't pretend to give us a rock-solid proof quoting a well-respected historical figure any more than the bloke before you.
Disclaimer: I know only a modest amount about neuropsychology/theory of computation...
;).
/addressed/. In the brain's case, there may be loads of 'data' stored, its just you cant remember the 'links' needed to actually get to it - ever been caught out by that word you know, but you just cant quite remember? (I know I have ;).
OK, I've thought about this for a while and here goes:
As far as I can tell, there is very little we actually *know* about the brain. It cannot be likened in any definate way to a (theoretically) lossless accurate finite state machine/automaton, so there is no hard and fast way to answer this question. Get rid of that 122 meg estimate - he's obviously talking about something else
Basically, from what I can tell, the brain works _as a whole_ to store any one given 'item of information'. This is unlike a conventional computer that has a specific location to store a specific bit of data.
Any storage device on a conventional computer (as far as I know) uses some sort of definate addressing mechanism to access a particular peice of data.
On the other hand the brain "stores" pretty much everything you experienced, whether you remember it or not. The problem is how it is
The way the brain 'records' experiences is by changing (m)any neorones' receptibility to neighbouring cells as well as those neorones' internal chemical (in)balance. This would give a "random" (hey - never use the 'r' word without the quotes) yet analogue data storage that clearly cannot be enumerated in any definative way to bytes.
I believe it's entirely possible that the brain could even 'record' data in more outlandish methods such as small inductive/capacitive fields and chaotic electrical interference. What is definately known is that the way the brain stores data is probably more than the "just count the number of cells and multiply by ten" way shown in one of the replies. Such a way would be inmature, inaccurate and an insult to the more thoughtful amongst us.
Now, since I've said all that about just how impossible it is to measure the maximum capacity of the brain, what is probably a more practical answer is how much the brain will store on average. This I dont know, though I guess that that would be a question for the psychologists out there - and maybe thats where the 122 meg guess comes into it.