But there is a contradiction in the conventional wisdom:Lisp will make you a better
programmer,and yet you won't use it.
This is faulty logic. It is not a contradiction. There are lots of things that you can learn to become a better programmer. I too believe that Lisp is one of them. But this does not imply that Lisp is a good programming language. It implies that it makes a good learning tool. That's it.
Lisp is a neat language, but it's software engineering qualities are way behind the times. A friend and I actually did an interesting project comparing Lisp to Java. He was well skilled in Lisp, and I Java. We defined a project, and then I implemented it in Java, and my friend implemeneted it in Lisp. We then compared the results: the Java version took less time to develop, took less time to debug, contained slightly less lines of code, was more easily modified (we added an additional objective), and was more efficient. Keep in mind that my friend is a Lisp advocate, and a skilled developer.
Assuming, of course, that monetary motivations are the only ones that matter. That however has been proven wrong repeatedly, for instance by the huge amount of free software produced by enthusiastic volunteers.
I've really just started thinking about these ideas, so please continue the
debate, and point out where you think I'm mistaken.
I don't believe the open source gift economy is a counter example for
this. I believe it would be overly romantic to believe that if developers
stopped being paid that they would continue to enthusiastically donate
software. The bottom line is that every developer that contributes to open
source has either: (a) made lots of money already (b) is making good money, or
(c) has the promise of making good money in the future. In fact
contributing to open source is a good way to learn and ultimately make good
money.
If you took away the money, people might continue to contribute to open
source. In fact open source would probably be all that there was.
But people would no longer specialize in being a developer since they would need
to seek other means to support themselves. And ultimately, the quality of
code being contributed would greatly suffer in quality.
The open source gift economy is about "how much money is
enough". It's about donating your time and skills for worthy causes
instead of making more money. And the simple fact is that the gift
economy exists and thrives only because it is supported by a money economy.
The presumption that software developers would stop being paid is hard to
take seriously, and is just a mental exercise. But I can easily imagine a
not too distant future in which a musical artist works long and hard, investing
all of their talent and creativity, to ultimately produce a piece of
music. This music is then sold, but instead of the musician being rewarded
relative to how much their work is appreciated, only a small number of copies
are sold before it becomes widely available for free download. Sure, this
musician might then make money using alternative means, for example, giving a
tour. But what if the kind of music does not lend itself to touring?
Some of my favorite music is music I play in the background while I code or do
other work. I don't want to actually go to a concert given by these
musicians; I just want their music.
Musicians, then, failing to be paid for their skills, turn to other means for
income. They may still contribute some of their music for free, but the
bottom line is that they no longer specialize in being a musician, and the
result is that quality suffers.
There is a big difference between using a technology because you think the risks are worth the benefit, and using a technology because you're ignorant of its risks.
I recognize that this will likely be unpopular, but before you mod this down, consider it.
There seems to be two ends of this spectrum:
This article, "The Right To Read", portrays one end of the spectrum. It portrays a world in which information is overly expensive. Yes, a world in which all information (books, music, etc.) is overly expensive would be bad. It would financially lock us out of good ideas on which to build new better ideas.
The other end of the spectrum is a world in which all information is free. This would also be bad. What motivation would there be to provide new better ideas? None.
So these two ends of the spectum represent economics which are broken.
The current situation for the music industry is a good example of an economic system which has been broken. However much fun it is at the moment, and I'm certainly having a good time, the long term consequense of not rewarding musical artists for their work will be shitty music.
Currently, the music industry has little to no control over how much I pay for a piece of music. Typically, it is $0 because that's what I choose to pay because I can. But this is broken. The person doing the selling needs to set the price, not the person doing the "buying". This is just basic economics. A fair price is reached when both the seller and the buyer agree.
I honestly hope that we can ultimately devise methods which will allow authors to control digital information. I don't think it will be easy, in fact I don't think it is theoretically possible, but ultimately it is important to refix the economics that we have managed to break by at least making it very inconvenient. It would be short-sited to believe anything else, and would simply be restoring a condition which existed before the digital era.
At the same time, I'm also confident that the world envisioned by "The Right To Read" will never happen. That is also about broken economics, but in the opposite direction: the seller has all the power. But this is a condition which does not exist long term in a free market. If your information is too expensive, I'll just buy someone elses. And please don't tell me that you expect someone to own the viewing rights to something like the periodic table of elements.
I should also mention that anyone that has done system administration for a large university knows the value of the warning message quoted in "The Right To Read", and the actions that it describes. There are a million ways to abused student computer services, and students seem to find every one.
They make one-handed keyboards just for people like you. They're a little pricey, but you have to admit that they can't exactly spread out their development costs the same way that other companies can with more popular models.
Until the designers start to realise that we're not all 6' tall
Actually, dude, I'm 6'3", and the world is not designed for me at all. Chairs, cars, desk height, door ways, door knobs, etc. I personally think the perfect height is 5'11": tall enough to be "tall" and short enough to fit everything well.
Of course you experience less movement with Dvorak. QWERTY was designed to foil the typists who "fig jammed" the old mechanical typewriters by typing to fast.
It's true that Qwerty was designed to reduce jams. However, they did this by designing the key layout so that it would maximize the amount of alternation between hands. As a small example, the keys used to type the word "flake" alternate between the left and right side perfectly. Most words, of course, do not alternate perfectly, but Qwerty does alternate better than Dvorak. The fact that Qwerty was designed to alternate is one of the reasons that it is not such a bad keyboard layout since alternation also allows more efficient typing.
The Dvorak was the original design.
This is just false. Qwerty was designed in 1872. Dvorak was designed in 1932.
The Kinesis "Evolution" is not an evolutionary step for Kinesis. It's a step backwards.
I have two Kinesis Ergo keyboards, and love them. The design of these keyboards is completely new and very ergonomic. But I believe Kinesis has a hard time selling them since it takes a while to get used to them because they are so different.
Microsoft, and many other companies, took the low road: make a keyboard that is a little more ergonomic, but very easy to use if you are already accustomed to a normal keyboard. These keyboards sell very well, mainly because there is very little retraining time required.
I really repected Kinesis when it came out with the Ergo since it really is an evolutionary step in keyboard design. And my hands rejoice in its use. Every other keyboard makes my hands hurt after a full day of coding.
But the "Evolution" is basically the same keyboard design as a standard keyboard, with the exception that you can move the seperate halves around as desired. This *might* be beneficial if your cheif problem is back and neck pain, but it's likely a step in the wrong direction if your problem is with your hands.
The Qwerty keyboard layout was designed in 1872. The Dvorak keyboard layout was designed in 1932. Since then, many people have sat around and argued which is better. But it's not clear that either one is truly the best.
I think they are both pretty good. But I'm confident that today we can do better. And this is a project that I've just started to look into.
I'm currently an SDE, but most of my graduate work was in theory. There are two reasons why I think we can do better today:
A) We have developed many algorithms that can solve or closely approximate hard optimization problems.
B) We have computers.
I've broken the problem of determining a good keyboard layout into the following steps:
1) Determine what people actually type. I'm not quite sure how to best do this. Ideally, everyone would send me a file which captured the key strokes that they've made over the past decade. Realistically, I was thinking of simply gathering data from the web, and using word frequency data.
2) Determine how to effectively model typing. There are some interesting experiments that show that, for example, your index finger is faster and stronger than your pinky. Also, you can type more efficiently if the keys you need to type alternate hands. Here are some interesting reports on typing micro-benchmarks and such. The goal of this model would be to tell you how efficient a particular keyboard layout is. Until I read otherwise, I'm assuming that maximizing efficiency also maximizes ergonomics.
3) Develop an algorithm. This is the fun part. Develop an alorithm which searches the very large search space of possible keyboard layouts and find one which is as close to optimal as possible as defined by the efficiency of the model. I may find that this optimization problem is best expressed at a semidefinite program, or such, but it will be interesting to see where the model takes me.
4) Real world test. I have no illusions of gradeur. I don't expect people to come rushing to try out the keyboard layout I design. But I'm hoping that I can convince a few people to try it out and give it a test drive.
I would like to post this as an "ask slashdot" after I make more progress and can give more specifics, but does anyone have some good suggestions at this point? I'm especially interested in information concerning (1) and (2). Can anyone point me to other keyboard research?
BTW, I currently have two Kinesis Ergo keyboards (one for work and one for home) which can switch between qwerty and dvorak. I highly recommend them. There are many reasons I prefer them over a normal keyboard, but the biggest reason is the thumb keys. When the standard computer keyboard was designed, they just took the qwerty keyboard layout and added function keys around the edges. Thus, on a standard keyboard, the pinky (the weakest finger) is completely overworked while the thumbs (the stongest "finger") hardly break a sweat.
One the one hand, Microsoft doesn't deserve the benefit of the doubt when trying to figure out their motives behind any action that they perform. They deserve our skepticism.
But on the other hand, and I believe this is your point, if the slashdot community doesn't really analyze Microsoft's actions, and just complains about everything they do, the slashdot community will lose respect. Many of us are becoming a Turning machine that just rejects. There's nothing intellegent about that.
Here's a recent
article from the Economist on Microsoft which asks the question, "has Microsoft changed its ways?"
I respect the Economist quite a bit because they have built up a solid reputation over a long period of time.
I can't say the same for Microsoft.
I'm not suggesting that I agree with putting down Microsoft at every opportunity, but they don't deserve the benefit of the doubt when trying to figure out their motives behind an action.
It sounds like you are saying that because you specialize in something "useful", you make the world a better place, but gymnasts, and in general, people that specialize in sports, do not, and that they should be ashamed for their lack of usefulness.
But perhaps gymnasts do perform a useful function. I think it has to do with the brief period of happiness brought to the people that watch them flip about. At least that seems to make my mom very happy. I don't think gymnasts make me as happy as my mother, but as a scientist, I'm generaly curious about limits of human potenial. And gymnasts provide an interesting look at the fundamental abilities of the human body.
I understand your reaction, however. I think that many people over-value professional sports. And, please correct me if I'm wrong, but I think your response is to under-value them. But be careful. I don't believe that over-compensating is the best way to point out someone's error since you're trying to point out an error with an error.
How long has Gymnastics been around? Relatively speaking, the video gamming era has just begun. What you consider "good" at Quake today, will likely be below average tomorrow. It's not harder to be a gymnast than a gamer. It's just harder to reach what the world generally considers to be a "good" gymnast than it is to be a "good" gamer. But what we consider "good" will change over time, and we will raise our standards as gamers get better and better.
Imagine someone that trains their whole life to play Quake (or some other video game). They work out their fingers (or whatever does the controlling) doing special exercises each morning. They cross-train and practice their response times and general hand-eye coordination playing other games. They eat a special diet which helps them focus mentally for long periods of time.
The interesting thing to see will be if professional video gamers will specialize in just one game. I don't think so since video gamming is so closely tied to technology. Baseball will likely not change that much over the couse of a person's life. But video games will always be changing.
there is no such thing as the "perfect optimizing compiler". To be verifiably optimal, as well as knowing everything there is to know about the machine's internal architecture, it would have to have complete knowledge about the dataset that the program to be compiled is to be run on
And this is one of the reasons why the Java VM technologies that are being developed along with HotSpot are so interesting since they perform profiling on the fly and restucture code for better preformance. Of course you need to pay for this profiling and restructuring during execution, but apparently Sun is betting that it's overall a performance win.
I've never seen a proof, but I suspect the perfect optimizing compiler is a travelling salesman-class problem. Does anyone have any proofs of my suspicions?
I'm not sure what you mean by "the perfect optimizing compiler". There are many things that a compiler can optimize, and some of them are at odds with each other. For example, an optimal register allocation will likely not coincide with the optimal instruction scheduling.
Perhaps you just mean, "whatever makes it fastest", but of course that's undecidable.
If by "travelling salesman-class problem" you mean that some of the optimization problems for compilers are NP-hard, then you are right. Register allocation (minimizing the spill cost), for example, is NP-hard. Here's a paper on a really nice register allocation algormithm, along with a proof that register allocation is NP-hard. He also presents a non-exponential time approximation algorithm that performs very well (heads up compiler writers!).
Of course, compilers still don't produce code as good as hand-coded assembly language and please don't quote me the "myth of the magic compiler" that is supposed to produce code better than humans
As hardware get more and more complex, don't you believe that the number of things that a assembly writer needs to be concerned with increases to the point that it will be unreasonible? For most programmers, has that point not already been reached?
And even if a human can write assembly better than a compiler, is it worth the cost? For the majority of us, the answer is clearly no. What percentage of the develpers you know actually even know the assembly language for just one architecture, let alone know how to code it really well?
I think this discussion would definitely benefit from some guidance from a professional pychologist. I am not a professional pychologist, so perhaps someone can sharpen some of these issues.
For clarity (of both expression and thought), it might be helpful to think a second about the different kinds of personalities we are discussing. There seem to be two somewhat orthogonal continuums:
1) How competent someone is.
2) How curious someone is.
I personally know developers that are not particularly curious, but are highly competent. I would catagorize these developers as "professionals", and their contributions are without question very important. They tend to make strong managers since their personalities foster confidence, both from their team and from management, and they tend to be very organized. "professionals" tend to be motivated because they wish to have a succesful career.
A developer that is both competent and naturally curious tends to be natually motivated to learn about new advances or just simply hone their skills by playing around. They tend to be natually enthusiastic about their work, and that tends to rub off on the whole environment and makes others more enthusiastic. These guys tend to be "geeks".
A developer that is curious, but not particularly competent tends to be "cute". These guys are good for doing staight forward jobs, and also for general entertainment purposes assuming that they are not also cocky.
Most companies have a good mix of these personality types, and I would argue that that's a good thing. Perhaps some of the interactions would be as follows: The geeks play around and come up with ideas. The professionals consider the ideas and make refinements. The cute guys (please, no gay jokes) implement the easy stuff and generally help the professionals and geeks feel good about themselves and motivate them by reminding them that they are indeed talented. The professionals keep the geeks in line and prevent them from goofing off too much, and the geeks keep the professionals on their toes and motivated to keep up with new advances.
Yes, but that's feeling sadness for the user of the toaster.
I don't understand what you are saying.
If I had to cast the toaster analogy into somthing that makes sense concerning the matter at hand, it would be that all of the perfectly operational parts of a toaster get thrown out when one little part breaks down.
I don't think that I will ever consider AI to be a life form.
This really all comes down to your worldview. Do you hold a supernaturalist worldview, or a naturalist worldview?
If you hold a naturalist worldview (which I do) then you don't believe in God (or Gods), and you don't believe that you have a soul, or some part of your mind which continues to exist after death. It's a pretty depressing worldview, but one that makes a lot of sense to me based on my experience. In this worldview, it's rather easy to believe that you are nothing more than a machine. If nature, the tireless experimenter, can do it, then why not us. It may takes us a while, but it's likely a matter of time, and definitely a possibility.
If, on the other hand, you are a supernaturalist, then you most likely believe that you possess something that a "machine" could never have, for example, a soul. I believe the notion of AI as a life form is currently inconsistent with most religions.
Competition is part of life. Failure is definitely a feature. I agree that you shouldn't feel bad for something that fails.
I don't feel bad for Be. But I do feel bad for the developers. I think the developers have done a good job.
I don't understand the relevance of your toaster analogy. Here's an analogy for you: if you did a great job at work, but got laid off because your company went south, should we not feel bad for you?
Are you so sure? The difference between an ant and a human is huge. But the difference between us, and some other technologically advanced species may not be that large. The main difference might simply be technology.
To put this in perspective, do you think it's possible that we could achieve those technologies within 2500 years? If so, then how different do you think that humans will be in 2500 years? To help answer this, consider looking backwards: do you think that a conversation with Plato would be interesting?
I would be afraid of reading a news site which *only* displayed news articles that the site guessed I would like based on the virtual identity that this site constructed for me.
A virtual identity, just like a real person, should be treated as something which is dynamic. Perhaps a good way of thinking about how to establish and maintain a virtual identity is using a standard AI technique: hill climbing. A news site could choose, say, 90% of the articles that it will first display to you from a list of articles that it thinks you will like. The remaining 10% could be choosen completely at random. Thus, this system would simultaneously (1) discover what other topics you may like, or have recently grown to like, and (2) expose you to articles that you would not have normally been exposed to based on your current virtual identity.
Typically, hill climbing is used to find an "optimal" solution to a fixed problem. But in this case, it would be used to continually probe you to see how you may have changed, while at the same time keeping you connected to "common knowledge".
I'm still not convinced you'll get the global minimum via the method described, though - it's not something I'll just buy with a handwave.
Agreed.
I guess what's most interesting about this example is that it challenges us to think about a "computer" in a way that we normally do not. Makes me wonder if someday my system will have an add-on NP-complete card for handling a specific NP-complete problem.
Unfortunately, you need to infer the definition of the problem from his description of the solution. I originally thought it was an MST as well, but as SLi points out it sounds like it is a Steiner tree problem. To be exact, a Euclidean Steiner tree problem, which is in NP-hard.
you've got a solution to a particular version of the minimum spanning tree problem (the points are all located in 2-D eucludian space, which is not necessarily true for general problem).
So it would seem that this is not an MST problem. Additionally, the fact that the graph is embeddable in a 2D Euclidian space does not make a difference. MST is in P, regardless of the weights assigned to the edges.
Slashdot Mirror
But there is a contradiction in the conventional wisdom:Lisp will make you a better programmer,and yet you won't use it.
This is faulty logic. It is not a contradiction. There are lots of things that you can learn to become a better programmer. I too believe that Lisp is one of them. But this does not imply that Lisp is a good programming language. It implies that it makes a good learning tool. That's it.
Lisp is a neat language, but it's software engineering qualities are way behind the times. A friend and I actually did an interesting project comparing Lisp to Java. He was well skilled in Lisp, and I Java. We defined a project, and then I implemented it in Java, and my friend implemeneted it in Lisp. We then compared the results: the Java version took less time to develop, took less time to debug, contained slightly less lines of code, was more easily modified (we added an additional objective), and was more efficient. Keep in mind that my friend is a Lisp advocate, and a skilled developer.
Assuming, of course, that monetary motivations are the only ones that matter. That however has been proven wrong repeatedly, for instance by the huge amount of free software produced by enthusiastic volunteers.
I've really just started thinking about these ideas, so please continue the debate, and point out where you think I'm mistaken.
I don't believe the open source gift economy is a counter example for this. I believe it would be overly romantic to believe that if developers stopped being paid that they would continue to enthusiastically donate software. The bottom line is that every developer that contributes to open source has either: (a) made lots of money already (b) is making good money, or (c) has the promise of making good money in the future. In fact contributing to open source is a good way to learn and ultimately make good money.
If you took away the money, people might continue to contribute to open source. In fact open source would probably be all that there was. But people would no longer specialize in being a developer since they would need to seek other means to support themselves. And ultimately, the quality of code being contributed would greatly suffer in quality.
The open source gift economy is about "how much money is enough". It's about donating your time and skills for worthy causes instead of making more money. And the simple fact is that the gift economy exists and thrives only because it is supported by a money economy.
The presumption that software developers would stop being paid is hard to take seriously, and is just a mental exercise. But I can easily imagine a not too distant future in which a musical artist works long and hard, investing all of their talent and creativity, to ultimately produce a piece of music. This music is then sold, but instead of the musician being rewarded relative to how much their work is appreciated, only a small number of copies are sold before it becomes widely available for free download. Sure, this musician might then make money using alternative means, for example, giving a tour. But what if the kind of music does not lend itself to touring? Some of my favorite music is music I play in the background while I code or do other work. I don't want to actually go to a concert given by these musicians; I just want their music.
Musicians, then, failing to be paid for their skills, turn to other means for income. They may still contribute some of their music for free, but the bottom line is that they no longer specialize in being a musician, and the result is that quality suffers.
How do you solve this problem?
There is a big difference between using a technology because you think the risks are worth the benefit, and using a technology because you're ignorant of its risks.
Hey, whatever works for you, dude.
I recognize that this will likely be unpopular, but before you mod this down, consider it.
There seems to be two ends of this spectrum:
This article, "The Right To Read", portrays one end of the spectrum. It portrays a world in which information is overly expensive. Yes, a world in which all information (books, music, etc.) is overly expensive would be bad. It would financially lock us out of good ideas on which to build new better ideas.
The other end of the spectrum is a world in which all information is free. This would also be bad. What motivation would there be to provide new better ideas? None.
So these two ends of the spectum represent economics which are broken.
The current situation for the music industry is a good example of an economic system which has been broken. However much fun it is at the moment, and I'm certainly having a good time, the long term consequense of not rewarding musical artists for their work will be shitty music.
Currently, the music industry has little to no control over how much I pay for a piece of music. Typically, it is $0 because that's what I choose to pay because I can. But this is broken. The person doing the selling needs to set the price, not the person doing the "buying". This is just basic economics. A fair price is reached when both the seller and the buyer agree.
I honestly hope that we can ultimately devise methods which will allow authors to control digital information. I don't think it will be easy, in fact I don't think it is theoretically possible, but ultimately it is important to refix the economics that we have managed to break by at least making it very inconvenient. It would be short-sited to believe anything else, and would simply be restoring a condition which existed before the digital era.
At the same time, I'm also confident that the world envisioned by "The Right To Read" will never happen. That is also about broken economics, but in the opposite direction: the seller has all the power. But this is a condition which does not exist long term in a free market. If your information is too expensive, I'll just buy someone elses. And please don't tell me that you expect someone to own the viewing rights to something like the periodic table of elements.
I should also mention that anyone that has done system administration for a large university knows the value of the warning message quoted in "The Right To Read", and the actions that it describes. There are a million ways to abused student computer services, and students seem to find every one.
The logitech keyboard uses wireless RF. Have there been any studies on the affect of surrounding ourselves with RF devices for most of the day?
I'm just concerned (paraniod?) that instead of getting RSI, now we'll get cancer. I would much rather have RSI.
What you want is a Maltron.
They make one-handed keyboards just for people like you. They're a little pricey, but you have to admit that they can't exactly spread out their development costs the same way that other companies can with more popular models.
Until the designers start to realise that we're not all 6' tall
Actually, dude, I'm 6'3", and the world is not designed for me at all. Chairs, cars, desk height, door ways, door knobs, etc. I personally think the perfect height is 5'11": tall enough to be "tall" and short enough to fit everything well.
You make two statements. Both of them are false.
Of course you experience less movement with Dvorak. QWERTY was designed to foil the typists who "fig jammed" the old mechanical typewriters by typing to fast.
It's true that Qwerty was designed to reduce jams. However, they did this by designing the key layout so that it would maximize the amount of alternation between hands. As a small example, the keys used to type the word "flake" alternate between the left and right side perfectly. Most words, of course, do not alternate perfectly, but Qwerty does alternate better than Dvorak. The fact that Qwerty was designed to alternate is one of the reasons that it is not such a bad keyboard layout since alternation also allows more efficient typing.
The Dvorak was the original design.
This is just false. Qwerty was designed in 1872. Dvorak was designed in 1932.
Thanks for playing.
The Kinesis "Evolution" is not an evolutionary step for Kinesis. It's a step backwards.
I have two Kinesis Ergo keyboards, and love them. The design of these keyboards is completely new and very ergonomic. But I believe Kinesis has a hard time selling them since it takes a while to get used to them because they are so different.
Microsoft, and many other companies, took the low road: make a keyboard that is a little more ergonomic, but very easy to use if you are already accustomed to a normal keyboard. These keyboards sell very well, mainly because there is very little retraining time required.
I really repected Kinesis when it came out with the Ergo since it really is an evolutionary step in keyboard design. And my hands rejoice in its use. Every other keyboard makes my hands hurt after a full day of coding.
But the "Evolution" is basically the same keyboard design as a standard keyboard, with the exception that you can move the seperate halves around as desired. This *might* be beneficial if your cheif problem is back and neck pain, but it's likely a step in the wrong direction if your problem is with your hands.
The Qwerty keyboard layout was designed in 1872. The Dvorak keyboard layout was designed in 1932. Since then, many people have sat around and argued which is better. But it's not clear that either one is truly the best.
I think they are both pretty good. But I'm confident that today we can do better. And this is a project that I've just started to look into.
I'm currently an SDE, but most of my graduate work was in theory. There are two reasons why I think we can do better today:
A) We have developed many algorithms that can solve or closely approximate hard optimization problems.
B) We have computers.
I've broken the problem of determining a good keyboard layout into the following steps:
1) Determine what people actually type. I'm not quite sure how to best do this. Ideally, everyone would send me a file which captured the key strokes that they've made over the past decade. Realistically, I was thinking of simply gathering data from the web, and using word frequency data.
2) Determine how to effectively model typing. There are some interesting experiments that show that, for example, your index finger is faster and stronger than your pinky. Also, you can type more efficiently if the keys you need to type alternate hands. Here are some interesting reports on typing micro-benchmarks and such. The goal of this model would be to tell you how efficient a particular keyboard layout is. Until I read otherwise, I'm assuming that maximizing efficiency also maximizes ergonomics.
3) Develop an algorithm. This is the fun part. Develop an alorithm which searches the very large search space of possible keyboard layouts and find one which is as close to optimal as possible as defined by the efficiency of the model. I may find that this optimization problem is best expressed at a semidefinite program, or such, but it will be interesting to see where the model takes me.
4) Real world test. I have no illusions of gradeur. I don't expect people to come rushing to try out the keyboard layout I design. But I'm hoping that I can convince a few people to try it out and give it a test drive.
I would like to post this as an "ask slashdot" after I make more progress and can give more specifics, but does anyone have some good suggestions at this point? I'm especially interested in information concerning (1) and (2). Can anyone point me to other keyboard research?
BTW, I currently have two Kinesis Ergo keyboards (one for work and one for home) which can switch between qwerty and dvorak. I highly recommend them. There are many reasons I prefer them over a normal keyboard, but the biggest reason is the thumb keys. When the standard computer keyboard was designed, they just took the qwerty keyboard layout and added function keys around the edges. Thus, on a standard keyboard, the pinky (the weakest finger) is completely overworked while the thumbs (the stongest "finger") hardly break a sweat.
Opps. I didn't finish what I was trying to say.
One the one hand, Microsoft doesn't deserve the benefit of the doubt when trying to figure out their motives behind any action that they perform. They deserve our skepticism.
But on the other hand, and I believe this is your point, if the slashdot community doesn't really analyze Microsoft's actions, and just complains about everything they do, the slashdot community will lose respect. Many of us are becoming a Turning machine that just rejects. There's nothing intellegent about that.
Here's a recent article from the Economist on Microsoft which asks the question, "has Microsoft changed its ways?"
I respect the Economist quite a bit because they have built up a solid reputation over a long period of time.
I can't say the same for Microsoft.
I'm not suggesting that I agree with putting down Microsoft at every opportunity, but they don't deserve the benefit of the doubt when trying to figure out their motives behind an action.
Ouch. You really don't like gymnasts, do you?
It sounds like you are saying that because you specialize in something "useful", you make the world a better place, but gymnasts, and in general, people that specialize in sports, do not, and that they should be ashamed for their lack of usefulness.
But perhaps gymnasts do perform a useful function. I think it has to do with the brief period of happiness brought to the people that watch them flip about. At least that seems to make my mom very happy. I don't think gymnasts make me as happy as my mother, but as a scientist, I'm generaly curious about limits of human potenial. And gymnasts provide an interesting look at the fundamental abilities of the human body.
I understand your reaction, however. I think that many people over-value professional sports. And, please correct me if I'm wrong, but I think your response is to under-value them. But be careful. I don't believe that over-compensating is the best way to point out someone's error since you're trying to point out an error with an error.
How long has Gymnastics been around? Relatively speaking, the video gamming era has just begun. What you consider "good" at Quake today, will likely be below average tomorrow. It's not harder to be a gymnast than a gamer. It's just harder to reach what the world generally considers to be a "good" gymnast than it is to be a "good" gamer. But what we consider "good" will change over time, and we will raise our standards as gamers get better and better.
Imagine someone that trains their whole life to play Quake (or some other video game). They work out their fingers (or whatever does the controlling) doing special exercises each morning. They cross-train and practice their response times and general hand-eye coordination playing other games. They eat a special diet which helps them focus mentally for long periods of time.
The interesting thing to see will be if professional video gamers will specialize in just one game. I don't think so since video gamming is so closely tied to technology. Baseball will likely not change that much over the couse of a person's life. But video games will always be changing.
there is no such thing as the "perfect optimizing compiler". To be verifiably optimal, as well as knowing everything there is to know about the machine's internal architecture, it would have to have complete knowledge about the dataset that the program to be compiled is to be run on
And this is one of the reasons why the Java VM technologies that are being developed along with HotSpot are so interesting since they perform profiling on the fly and restucture code for better preformance. Of course you need to pay for this profiling and restructuring during execution, but apparently Sun is betting that it's overall a performance win.
I've never seen a proof, but I suspect the perfect optimizing compiler is a travelling salesman-class problem. Does anyone have any proofs of my suspicions?
I'm not sure what you mean by "the perfect optimizing compiler". There are many things that a compiler can optimize, and some of them are at odds with each other. For example, an optimal register allocation will likely not coincide with the optimal instruction scheduling.
Perhaps you just mean, "whatever makes it fastest", but of course that's undecidable.
If by "travelling salesman-class problem" you mean that some of the optimization problems for compilers are NP-hard, then you are right. Register allocation (minimizing the spill cost), for example, is NP-hard. Here's a paper on a really nice register allocation algormithm, along with a proof that register allocation is NP-hard. He also presents a non-exponential time approximation algorithm that performs very well (heads up compiler writers!).
Of course, compilers still don't produce code as good as hand-coded assembly language and please don't quote me the "myth of the magic compiler" that is supposed to produce code better than humans
As hardware get more and more complex, don't you believe that the number of things that a assembly writer needs to be concerned with increases to the point that it will be unreasonible? For most programmers, has that point not already been reached?
And even if a human can write assembly better than a compiler, is it worth the cost? For the majority of us, the answer is clearly no. What percentage of the develpers you know actually even know the assembly language for just one architecture, let alone know how to code it really well?
I think this discussion would definitely benefit from some guidance from a professional pychologist. I am not a professional pychologist, so perhaps someone can sharpen some of these issues.
For clarity (of both expression and thought), it might be helpful to think a second about the different kinds of personalities we are discussing. There seem to be two somewhat orthogonal continuums:
1) How competent someone is.
2) How curious someone is.
I personally know developers that are not particularly curious, but are highly competent. I would catagorize these developers as "professionals", and their contributions are without question very important. They tend to make strong managers since their personalities foster confidence, both from their team and from management, and they tend to be very organized. "professionals" tend to be motivated because they wish to have a succesful career.
A developer that is both competent and naturally curious tends to be natually motivated to learn about new advances or just simply hone their skills by playing around. They tend to be natually enthusiastic about their work, and that tends to rub off on the whole environment and makes others more enthusiastic. These guys tend to be "geeks".
A developer that is curious, but not particularly competent tends to be "cute". These guys are good for doing staight forward jobs, and also for general entertainment purposes assuming that they are not also cocky.
Most companies have a good mix of these personality types, and I would argue that that's a good thing. Perhaps some of the interactions would be as follows: The geeks play around and come up with ideas. The professionals consider the ideas and make refinements. The cute guys (please, no gay jokes) implement the easy stuff and generally help the professionals and geeks feel good about themselves and motivate them by reminding them that they are indeed talented. The professionals keep the geeks in line and prevent them from goofing off too much, and the geeks keep the professionals on their toes and motivated to keep up with new advances.
Again, I'm no expert. What's your take?
Yes, but that's feeling sadness for the user of the toaster.
I don't understand what you are saying.
If I had to cast the toaster analogy into somthing that makes sense concerning the matter at hand, it would be that all of the perfectly operational parts of a toaster get thrown out when one little part breaks down.
You were one of these operational parts.
The next project for this group clearly should be a write-only file system: WOFS.
Or is that what MS frontpage already uses to back up your work?
I don't think that I will ever consider AI to be a life form.
This really all comes down to your worldview. Do you hold a supernaturalist worldview, or a naturalist worldview?
If you hold a naturalist worldview (which I do) then you don't believe in God (or Gods), and you don't believe that you have a soul, or some part of your mind which continues to exist after death. It's a pretty depressing worldview, but one that makes a lot of sense to me based on my experience. In this worldview, it's rather easy to believe that you are nothing more than a machine. If nature, the tireless experimenter, can do it, then why not us. It may takes us a while, but it's likely a matter of time, and definitely a possibility.
If, on the other hand, you are a supernaturalist, then you most likely believe that you possess something that a "machine" could never have, for example, a soul. I believe the notion of AI as a life form is currently inconsistent with most religions.
Competition is part of life. Failure is definitely a feature. I agree that you shouldn't feel bad for something that fails.
I don't feel bad for Be. But I do feel bad for the developers. I think the developers have done a good job.
I don't understand the relevance of your toaster analogy. Here's an analogy for you: if you did a great job at work, but got laid off because your company went south, should we not feel bad for you?
Are you so sure? The difference between an ant and a human is huge. But the difference between us, and some other technologically advanced species may not be that large. The main difference might simply be technology.
To put this in perspective, do you think it's possible that we could achieve those technologies within 2500 years? If so, then how different do you think that humans will be in 2500 years? To help answer this, consider looking backwards: do you think that a conversation with Plato would be interesting?
I would be afraid of reading a news site which *only* displayed news articles that the site guessed I would like based on the virtual identity that this site constructed for me.
A virtual identity, just like a real person, should be treated as something which is dynamic. Perhaps a good way of thinking about how to establish and maintain a virtual identity is using a standard AI technique: hill climbing. A news site could choose, say, 90% of the articles that it will first display to you from a list of articles that it thinks you will like. The remaining 10% could be choosen completely at random. Thus, this system would simultaneously (1) discover what other topics you may like, or have recently grown to like, and (2) expose you to articles that you would not have normally been exposed to based on your current virtual identity.
Typically, hill climbing is used to find an "optimal" solution to a fixed problem. But in this case, it would be used to continually probe you to see how you may have changed, while at the same time keeping you connected to "common knowledge".
Just a thought.
I'm still not convinced you'll get the global minimum via the method described, though - it's not something I'll just buy with a handwave.
Agreed.
I guess what's most interesting about this example is that it challenges us to think about a "computer" in a way that we normally do not. Makes me wonder if someday my system will have an add-on NP-complete card for handling a specific NP-complete problem.
Unfortunately, you need to infer the definition of the problem from his description of the solution. I originally thought it was an MST as well, but as SLi points out it sounds like it is a Steiner tree problem. To be exact, a Euclidean Steiner tree problem, which is in NP-hard.
you've got a solution to a particular version of the minimum spanning tree problem (the points are all located in 2-D eucludian space, which is not necessarily true for general problem).
So it would seem that this is not an MST problem. Additionally, the fact that the graph is embeddable in a 2D Euclidian space does not make a difference. MST is in P, regardless of the weights assigned to the edges.