It is not possible to make a general statement about which layer of
interpretation is the right one to be made public. Higher levels, closer to
the final results, are more likely to be reusable by other researchers.
However, higher levels of interpretation provide the least information for
someone attempting to confirm that the total analysis is valid.
You're wrong. It is perfectly clear what needs to be published openly: whatever is necessary for someone to confirm that the total analysis is valid.
That is the fundamental principle required for scientific progress. The fact that this statement is not specific enough to prescribe exactly what needs to be done in every scientific experiment is not a flaw. If in doubt, err on the side of caution, ie publish more than is strictly necessary to confirm that the total analysis is valid.
Yes, this may sometimes be costly, but so what? Some experiments are costly, and that often causes scientists to think trough variations that show the same effect in a cheaper way. Such calculations can and should factor in the cost of making the necessary data available. They are not orthogonal to the science per se, because ensuring repeatability and verifiability is at the heart of the scientific method.
So when designing an experiment, think about what it will take not just to convince the journal's referees, but also to prove repeatability against anyone who is willing to properly test it. Then implement the experiment and provide the proof, and publish the summary in a journal.
No, cryptographic hash functions have certain strong guarantees, but
all(*) hash functions are supposed to mimic independent, uniformly random, behaviour
of inputs. Since in the physical world, inputs often come from
processes, and processes tend to evolve continuously, the inputs to be
hashed by a computer system often have some amount of similarity if
they occur close together in time. Thus to transform consecutive
inputs into a pair of independent uniformly random hashes, it is
desirable that small changes in the input result in completely changed
output.
(*) There are exceptions, such as when devising algorithms for locality sensitive hashing, but they are few.
You're being stupid by displaying your ignorance. You assume, entirely baselessly, that a human being is worth more than a horse. This is true in today's western, industrial society but has not been so historically.
In pre-industrial societies horses are important, and have more value than an average human being.
A horse can carry warriors, and pull heavy equipment, which is more than a single human labourer can do. Moreover, horses are usually owned by important people, and in hierarchical societies the possessions of one's betters often have more value than oneself.
You don't even have to go to exotic historical examples, in the Wild West one hundred years ago, a horse thief would be hanged. That shows clearly that a horse had at least the value of a human being in America.
To take a modern analogy: there is a threat. You have an infantryman and an F22. Would you let the infantryman sacrifice himself to save the group, or should the F22 be destroyed instead?
Fitt's law is bullshit however. It's stupid to throw your mouse against the top of the screen to access a menu. It's a lot smarter to reserve a button on the mouse (maybe the fourth or fifth, eg to use with your thumb), and have the menu just appear where you are. It's faster and there's less focus switching overall. Basically you get a "menu" button on your mouse, and your eyes get to stay looking at the place on the screen where they were looking before.
Firstly, Pratchett's comment has nothing to do with a paradox something of the sort. It's a simple claim that scientists are bad at estimating very small probabilities, and typically get them wrong by a factor of hundreds of thousands. This is actually true and rather insightful in a the-emperor-has-no-clothes kind of way, and also not very deep at all.
The concept of the long tail is somewhat more interesting, but not that deep either. It's merely about realizing that many processes aren't Gaussian, unlike what students are lead to believe in highschool and various introductory courses which are not primarily about statistics.
However, your distinction between likely and unlikely events is confused. If you are going to label two events as likely and unlikely, then you are asserting that the likely event is to be observed with higher probability than the unlikely event.
This is always true by definition.
What you are trying to say is that, if you restrict yourself to a particular family of events and you compare the probability of occurrence of an unspecified member of the family with the probability of occurrence of a single specified member, then the former can be larger.
As an example, consider the family of events {the hour of your death is N}. It is fairly unlikely that I can predict the hour of your death (not being a serial killer myself), so if I specify the event {the hour of your death is 12am} then the probability of occurrence is small. But if I do not specify the event, by saying {the hour of your death is N, where N is some hour in the day}, then that event is certain. Of course I haven't said anything interesting *with certainty*, whereas in the case of 12am I have said something interesting *with low probability*.
The tragedy of statistics is that the great majority of things we know with high probability aren't interesting, and the majority of things that are interesting have low probabilities or cannot be estimated accurately.
Firstly, thank you for such a considered reply. I think you'll find, if you look through my comment history, that on my serious days whenever I comment on the BSD licence, I do appreciate the advantages of that license over the GPL for certain goals. While I personally always use the GPL for my projects, I don't advocate it for everything.
However, I think it is clear that the currency of the GPL community are ideas and implementations. These are the membership dues you pay to be part of, and considered to be part of, the community. It makes no sense to argue against this, as if it was perhaps a minor flaw in the license that can or should be remedied for the purpose of keeping commercial players in the club. Yes, it is a goal of the GPL community to grow, people such as RMS are very clear on that, but membership dues are required from all members - that is what limits the growth and so it should be.
I come from an academic background, and the history of science suggests strongly that progress requires full disclosure of ideas and methods. In the 17th century and before, scientists hid their methods, and hid their results. These were competitive advantages. People like Cardano kept jealously the secret of solving cubic equations - it was literally worth money, as there were circus competitions where people solved equations in front of an audience. Science made very little progress, and many ideas were lost, and many ideas were improperly evaluated and wrongly praised. Today, Science requires full disclosure, and scientists often disclose their work in progress years before it is fully completed. The rate of discoveries has skyrocketed, and the rate at which scientific advances are incorporated into society has too. No serious scientist would argue for hiding *anything* as a competitive advantage - if anything they would prefer to share trivial advances on the off chance that they can be viewed as having invented them. The lesson has been learned.
I view open source in this light. There are parts of the community which have learned the lesson and espouse full sharing of all "competitive advantages", and there are more primitive parts which do not agree. I call them primitive in the same way that I view 17th century scientists as primitive. They didn't know any better, and that was tragic. Many of their ideas and methods can't be used any more, but we remember them for their historical importance, at least sometimes.
I don't feel the need to convince you to choose the GPL, I choose it because any other license is just a waste. And I do expect those who join the community to pay their dues, because that is what grows the community's value, and allows it to thrive.
I actually read slashdot in w3m. I do it because it's fast, I don't
care about images and design, and most importantly it integrates very
well with emacs for writing comments. Finally, it's fast, really fast.
For example, if I drill down into a comment and go back up, there's no delay
while the browser has to re-render the page, refresh the content or run some javascript hooks, etc. It's just the little things that I notice every time I accidentally use a graphical browser to view slashdot.
Considering that the request states "I am the rights holder and they don't
have rights to distribute", step 3 is trivial.
That's probably the *least* trivial step. Deciding who is the rights holder is what lawyers are needed for, whenever there is a dispute. The world is full of business partners who, after years of working together, part ways *both* claiming the rights to their work. How is an ISP to know at a glance that this takedown isn't part of some ongoing legal dispute where nobody yet knows who has what rights? It might not be, and then again it might be.
On the contrary, that point is exactly *why* we GPL advocates advocate it. We don't want to enable companies which have no intention whatsoever to be part of the community. They're free to do everything themselves, and good luck to them. But giving them a leg up to get off the ground just so they can be selfish assholes with their ideas? Why should we? And yes, sharing your *ideas* and *implementations* is what we mean by being part of the community. If this isn't for you, then don't let the door hit you on the way out, thanks.
Using google glasses... good. It won't provide any more information than the
officer already has access to, or that can't be mined off a conventional
camera's video. It may just provide the info quicker, when the officer needs
it.
The timely arrival of information can interfere with the correct application of decisions. Suppose a cop sees a minor infraction, like crossing the street 50m away from a pedestrian crossing. The choice to go and give that person a ticket or let it go is a function of the traffic conditions, ie how dangerous this behaviour is at the time. It shouldn't be about *who* the person is. Now suppose the glasses come up with a bunch of internet accusations against the jaywalker about beating his wife. So the cop decides to go talk to the guy and give him a ticket anyway.
By giving agents *more* information than necessary, it makes it harder or impossible for them to make correct decisions. It's not unlike hiring decisions, say. If every resume has race, religion, and age right next to the name, that's going to influence decisions for the worse. It's extra information, but it should *not* be available to the decision maker.
The internet consists of hardware and software and things worth stealing. The first has long development cycles, and is more difficult to modify than the second. The second is extremely varied and full of vulnerabilties that are often easy to patch one instance at a time, but hard to patch simultaneously and comprehensively across the network. The third are things that shouldn't be accessible from the Internet in the first place, like our real names just so we can have a Google account, our credit card numbers just so merchants don't have to ask us when they want to charge us, our activity records just so we can be manipulated through ads, etc.
We can't change the first two without destroying the Internet, but there's no reason why computers should contain so much valuable information to steal.
On the other hand, the profits from existing drugs helps pay for research
into new drugs to some degree, so could India be similar to a big company
that uses Open Source but doesn't release any of their own code?
Research, like many other activities, is influenced by fashions. Arguably, by making a particular drug widely available very cheaply, the resulting enhanced interest will stimulate more funding for similar drugs, by virtue of the increased popularity and availability. This increased funding within the industry could be many times more valuable than the amount a single company can derive from the consequences of patent protection.
No. First, there are benefits unique to audio interaction (immediacy, tonal emphasis, emotion); next up, benefits to physical presence (enhances interaction - e.g. pointing, gestures, a wider visual canvas), and finally you skipped this last bit with the pretty pictures...:)
"...code editors open in front of us, trying to demonstrate certain use-cases where visual coding is superior."
This is not very convincing. There are benefits unique to text representation (typefaces, size of text, color of text, capitals/lowercase, even the precise choice of idiom in languages such as perl where there are many ways to do something). These benefits arguably are more varied and precise than those offered by the spoken word. Pointing, gestures and visual canvas interactivity is the job of an editor or ide, as is changing the view of the information presented eg by folding away function bodies etc.
Furthermore, there are benefits that simply exist in text form and can never be duplicated in audio interaction, such as going back and annotating existing text, or skipping comments and going back and reading them later. The linearity of audio interaction makes hierarchical information structures essentially impossible.
We *already* have graphical code. What do you think the alphabet is? It's a modular, graphical coding system for concepts using icons. We use it to build intermediate high level graphical pictures called words, which are then interpreted according to ad hoc graphical rules (such as whether we read left to right, right to left, indent using space or braces etc) into graphical representations of logical processes known as programs.
Perhaps you imagine a more primitive 2D system where the low level structures are not built up from a small number of multi purpose alphabetic icons using elaborate juxtaposition rules, but rather exist individually in a huge library of predefined pictograms. We have that too, it's called Chinese.
Actually I blame the developers. Clearly they are modern "app" developers, as opposed to us old school application developers, from an era when "application" didn't mean a glorified web page drenched in javascript toys.
Any application developer worth anything knows about Model-View-Controller and can separate the underlying data model interaction protocols from the view presented to the user.
There is no excuse whatsoever for the loss of any existing features that are found in classic, nor is there any technical reason whatsoever why there has to be a migration to a single "new" site to keep up with the times. The slashdot website is just a view into the comment and stories database, and there should be many views for everybody to choose their preferred one at any time, including the "beta" one as just one of them. In fact, if slashdot published a reasonable API there would be plenty of low digit users who could whip up a sane interface before breakfast.
Quite true. Software should generally be Free. One of the benefits that we all gain from this is that, when software doesn't have to be locked up just to force people to pay for it, it can be distributed in source code form. And that helps prevent issues with embedded malware, and embedded exploitation of the users too.
It wouldn't be nicer. It would be the least they should possibly do.
Publishers like Elsevier are leaches sucking at the teat of scientific institutions, weakening their libraries, which are the cornerstone of humanity's research efforts. The sooner they FOAD the better.
The key is not to confuse fiction with reality which admittedly
many do.
And that is precisely the problem TFA highlights. Cameron confuses fiction with reality. To prevent prime ministers from acting on incorrect assumptions and faulty logic, it makes sense to oppose them doing things which lead them into confusing fiction with reality, such as watching cop dramas and discussing them publicly as justifications for draconian spying powers.
In other words, he deserves the crap he gets for allowing TV drama into policy decisions.
The NSA has backdoors into the major encryption systems, for example in RSA products. So recovery is basically trivial, if you rely on any products sold or provided by Microsoft, Apple, Google, etc. Every large US company in fact has too much to lose if they don't cooperate with the NSA, so pick any company where it makes sense for the NSA to put backdoors in. If that company still exists today, then you can conclude that it has a secret deal with the NSA to spy on its customers.
Slashdot Mirror
You're wrong. It is perfectly clear what needs to be published openly: whatever is necessary for someone to confirm that the total analysis is valid.
That is the fundamental principle required for scientific progress. The fact that this statement is not specific enough to prescribe exactly what needs to be done in every scientific experiment is not a flaw. If in doubt, err on the side of caution, ie publish more than is strictly necessary to confirm that the total analysis is valid.
Yes, this may sometimes be costly, but so what? Some experiments are costly, and that often causes scientists to think trough variations that show the same effect in a cheaper way. Such calculations can and should factor in the cost of making the necessary data available. They are not orthogonal to the science per se, because ensuring repeatability and verifiability is at the heart of the scientific method.
So when designing an experiment, think about what it will take not just to convince the journal's referees, but also to prove repeatability against anyone who is willing to properly test it. Then implement the experiment and provide the proof, and publish the summary in a journal.
(*) There are exceptions, such as when devising algorithms for locality sensitive hashing, but they are few.
No, but if you use your middle finger instead that counts as a win for the rest of us.
In pre-industrial societies horses are important, and have more value than an average human being. A horse can carry warriors, and pull heavy equipment, which is more than a single human labourer can do. Moreover, horses are usually owned by important people, and in hierarchical societies the possessions of one's betters often have more value than oneself.
You don't even have to go to exotic historical examples, in the Wild West one hundred years ago, a horse thief would be hanged. That shows clearly that a horse had at least the value of a human being in America.
To take a modern analogy: there is a threat. You have an infantryman and an F22. Would you let the infantryman sacrifice himself to save the group, or should the F22 be destroyed instead?
Elmer Fudd using a drone to hunt Bugs Bunny? Dammit, I don't think I've seen that one yet.
Fitt's law is bullshit however. It's stupid to throw your mouse against the top of the screen to access a menu. It's a lot smarter to reserve a button on the mouse (maybe the fourth or fifth, eg to use with your thumb), and have the menu just appear where you are. It's faster and there's less focus switching overall. Basically you get a "menu" button on your mouse, and your eyes get to stay looking at the place on the screen where they were looking before.
Also, it was an American drone. It might have been commandeered by the NSA to kill a suspected terrorist.
Firstly, Pratchett's comment has nothing to do with a paradox something of the sort. It's a simple claim that scientists are bad at estimating very small probabilities, and typically get them wrong by a factor of hundreds of thousands. This is actually true and rather insightful in a the-emperor-has-no-clothes kind of way, and also not very deep at all.
The concept of the long tail is somewhat more interesting, but not that deep either. It's merely about realizing that many processes aren't Gaussian, unlike what students are lead to believe in highschool and various introductory courses which are not primarily about statistics.
However, your distinction between likely and unlikely events is confused. If you are going to label two events as likely and unlikely, then you are asserting that the likely event is to be observed with higher probability than the unlikely event. This is always true by definition.
What you are trying to say is that, if you restrict yourself to a particular family of events and you compare the probability of occurrence of an unspecified member of the family with the probability of occurrence of a single specified member, then the former can be larger.
As an example, consider the family of events {the hour of your death is N}. It is fairly unlikely that I can predict the hour of your death (not being a serial killer myself), so if I specify the event {the hour of your death is 12am} then the probability of occurrence is small. But if I do not specify the event, by saying {the hour of your death is N, where N is some hour in the day}, then that event is certain. Of course I haven't said anything interesting *with certainty*, whereas in the case of 12am I have said something interesting *with low probability*.
The tragedy of statistics is that the great majority of things we know with high probability aren't interesting, and the majority of things that are interesting have low probabilities or cannot be estimated accurately.
However, I think it is clear that the currency of the GPL community are ideas and implementations. These are the membership dues you pay to be part of, and considered to be part of, the community. It makes no sense to argue against this, as if it was perhaps a minor flaw in the license that can or should be remedied for the purpose of keeping commercial players in the club. Yes, it is a goal of the GPL community to grow, people such as RMS are very clear on that, but membership dues are required from all members - that is what limits the growth and so it should be.
I come from an academic background, and the history of science suggests strongly that progress requires full disclosure of ideas and methods. In the 17th century and before, scientists hid their methods, and hid their results. These were competitive advantages. People like Cardano kept jealously the secret of solving cubic equations - it was literally worth money, as there were circus competitions where people solved equations in front of an audience. Science made very little progress, and many ideas were lost, and many ideas were improperly evaluated and wrongly praised. Today, Science requires full disclosure, and scientists often disclose their work in progress years before it is fully completed. The rate of discoveries has skyrocketed, and the rate at which scientific advances are incorporated into society has too. No serious scientist would argue for hiding *anything* as a competitive advantage - if anything they would prefer to share trivial advances on the off chance that they can be viewed as having invented them. The lesson has been learned.
I view open source in this light. There are parts of the community which have learned the lesson and espouse full sharing of all "competitive advantages", and there are more primitive parts which do not agree. I call them primitive in the same way that I view 17th century scientists as primitive. They didn't know any better, and that was tragic. Many of their ideas and methods can't be used any more, but we remember them for their historical importance, at least sometimes.
I don't feel the need to convince you to choose the GPL, I choose it because any other license is just a waste. And I do expect those who join the community to pay their dues, because that is what grows the community's value, and allows it to thrive.
I actually read slashdot in w3m. I do it because it's fast, I don't care about images and design, and most importantly it integrates very well with emacs for writing comments. Finally, it's fast, really fast. For example, if I drill down into a comment and go back up, there's no delay while the browser has to re-render the page, refresh the content or run some javascript hooks, etc. It's just the little things that I notice every time I accidentally use a graphical browser to view slashdot.
Agreed.
But he shouldn't make it too obvious, or we'll crucify him for being a coward who took the easy way out.
That's probably the *least* trivial step. Deciding who is the rights holder is what lawyers are needed for, whenever there is a dispute. The world is full of business partners who, after years of working together, part ways *both* claiming the rights to their work. How is an ISP to know at a glance that this takedown isn't part of some ongoing legal dispute where nobody yet knows who has what rights? It might not be, and then again it might be.
On the contrary, that point is exactly *why* we GPL advocates advocate it. We don't want to enable companies which have no intention whatsoever to be part of the community. They're free to do everything themselves, and good luck to them. But giving them a leg up to get off the ground just so they can be selfish assholes with their ideas? Why should we? And yes, sharing your *ideas* and *implementations* is what we mean by being part of the community. If this isn't for you, then don't let the door hit you on the way out, thanks.
The timely arrival of information can interfere with the correct application of decisions. Suppose a cop sees a minor infraction, like crossing the street 50m away from a pedestrian crossing. The choice to go and give that person a ticket or let it go is a function of the traffic conditions, ie how dangerous this behaviour is at the time. It shouldn't be about *who* the person is. Now suppose the glasses come up with a bunch of internet accusations against the jaywalker about beating his wife. So the cop decides to go talk to the guy and give him a ticket anyway.
By giving agents *more* information than necessary, it makes it harder or impossible for them to make correct decisions. It's not unlike hiring decisions, say. If every resume has race, religion, and age right next to the name, that's going to influence decisions for the worse. It's extra information, but it should *not* be available to the decision maker.
We can't change the first two without destroying the Internet, but there's no reason why computers should contain so much valuable information to steal.
Research, like many other activities, is influenced by fashions. Arguably, by making a particular drug widely available very cheaply, the resulting enhanced interest will stimulate more funding for similar drugs, by virtue of the increased popularity and availability. This increased funding within the industry could be many times more valuable than the amount a single company can derive from the consequences of patent protection.
This is not very convincing. There are benefits unique to text representation (typefaces, size of text, color of text, capitals/lowercase, even the precise choice of idiom in languages such as perl where there are many ways to do something). These benefits arguably are more varied and precise than those offered by the spoken word. Pointing, gestures and visual canvas interactivity is the job of an editor or ide, as is changing the view of the information presented eg by folding away function bodies etc.
Furthermore, there are benefits that simply exist in text form and can never be duplicated in audio interaction, such as going back and annotating existing text, or skipping comments and going back and reading them later. The linearity of audio interaction makes hierarchical information structures essentially impossible.
Perhaps you imagine a more primitive 2D system where the low level structures are not built up from a small number of multi purpose alphabetic icons using elaborate juxtaposition rules, but rather exist individually in a huge library of predefined pictograms. We have that too, it's called Chinese.
Any application developer worth anything knows about Model-View-Controller and can separate the underlying data model interaction protocols from the view presented to the user.
There is no excuse whatsoever for the loss of any existing features that are found in classic, nor is there any technical reason whatsoever why there has to be a migration to a single "new" site to keep up with the times. The slashdot website is just a view into the comment and stories database, and there should be many views for everybody to choose their preferred one at any time, including the "beta" one as just one of them. In fact, if slashdot published a reasonable API there would be plenty of low digit users who could whip up a sane interface before breakfast.
Agreed, especially since the Americans weren't critical to the German defeat [only the Japanese defeat in the Pacific].
Quite true. Software should generally be Free. One of the benefits that we all gain from this is that, when software doesn't have to be locked up just to force people to pay for it, it can be distributed in source code form. And that helps prevent issues with embedded malware, and embedded exploitation of the users too.
Publishers like Elsevier are leaches sucking at the teat of scientific institutions, weakening their libraries, which are the cornerstone of humanity's research efforts. The sooner they FOAD the better.
And that is precisely the problem TFA highlights. Cameron confuses fiction with reality. To prevent prime ministers from acting on incorrect assumptions and faulty logic, it makes sense to oppose them doing things which lead them into confusing fiction with reality, such as watching cop dramas and discussing them publicly as justifications for draconian spying powers.
In other words, he deserves the crap he gets for allowing TV drama into policy decisions.
The NSA has backdoors into the major encryption systems, for example in RSA products. So recovery is basically trivial, if you rely on any products sold or provided by Microsoft, Apple, Google, etc. Every large US company in fact has too much to lose if they don't cooperate with the NSA, so pick any company where it makes sense for the NSA to put backdoors in. If that company still exists today, then you can conclude that it has a secret deal with the NSA to spy on its customers.