I made this mistake, and its contents (i.e., Windows 8 on portables might consolidate mobile and PC use cases) could have been summed up in half the length of the slashdot summary, which in turn was just astroturfing. Then, as usual with Microsoft propaganda trying to outdumb readers, the slashdot summary does NOT say: "Windows mobile version xxx is pretty good / growing / will be the good for whatever", but, adapting the message to the slashdot crowd, first it tactically discredits MS just to buy our agreement, then slip on the real message in parentheses, almost subliminally, which is the assumed 5% market share. It's not even near that except in certain Seattle and Espoo cafeterias.
The article itself is full of cheap dramatization like "When it was pointed out he forgot Microsoft, Schmidt said this wasn’t a mistake. (then in a standalone paragraph) I believe it is." and patronizing passages like "Seriously stop and think about this for a minute. [...] give it 12 months and try again."
It's almost as if someone asked what equipment he needs for performing bypass surgery. First requirement would be a bookshelf (to be filled with books for one's study), and second is a cabinet for keeping degrees, diplomas and continuing education. I'd love to know what motivated the question to begin with, it sounds almost eerie, esp. the professional part. Maybe a PHB who wants to monitor his engineers' purchase requests or a lottery winner with his dreams.
Well, it's a glaring error coming from someone who teaches kids, period. You're right, we're not in class, but perhaps said teacher is eager to learn about and correct this basic issue. In addition, naming himself "VVrath" can be seen as solicitation for pointing out improvement opportunities, and who are you to interfere with that?
I realize that many slashdotters don't give a flying rat's ass about grammar (mixing their with there, it's with its), so feel free to mod this to oblivion while modding up the equally uninteresting grammar denialist ("shit like this", "we're not in class", wow this IS novel and insightful). I like the fact that some people have the spine to express what is worth expressing despite the wrath of the plebs.
I don't know where you get your numbers from and of what unit of measure. It's clear that thermal insulation is night and day between parallel panes with a distance in the teens (mm) filled with gas (i.e. standard fare double-pane window) vs. non-parallel pane windows which are filled with air (other gas would escape and would be too expensive at >10x the volume to begin with). Have you ever held your hand near a single-pane window in wintertime, as a more accessible experiment? The air falls down quite fast. Maybe it's not 5-6mph but between non-parallel panes air would rotate at several turns a minute.
If economical window sound insulation is needed, stick to parallel panes as the main window, filled with gas, and use glasses of unequal thickness, the heavier the better. An ex-relative of mine lives in a very busy city road and their three-pane window (getting more prevalent in Europe) does wonders without exotic things like non-parallel panes that insulate poorly and you can't open them. On a visit, it was quite an eerie feeling when they opened their window at once and all that noise came in.
Correct, since the libraries are either-or and 3.whatever was spottily supported the last time around (12 months ago), there are effectively two languages, Python2 (you need for those libraries) and Python3 (you'd rather use). Python's attractiveness diminished so much so that neither of these two languages were chosen.
| I rent. If I want, a simple phone call to one person ends my relationship and I can move.
In the US it was customary to lease for a minimum of one year, and upon renewal, it typically needs to be renewed for one year. Now the problem is, if you rent, it is usually because you have a relatively mobile lifestyle in that you always move where there is a good job opportunity. Now it is statistically unlikely that your wish to move for a job will coincide with the lease cycle, unless you swim in job offers or have children and need to adhere to a school year. Though I'd be OK with a minimum lease term, it is a rip-off to have to renew for another year as it either adds a constraint to your life or adds cost when you violate that constraint.
Or did the rental market become more reasonable since the mortgage crisis?
Don't consult acoustics with a constructor. Generally, constructors know how to sell what they can do, and they will happily construct your idea or in the absence of it (or despite of it) their own. But they don't need to live with the consequences that their work is ineffective or suboptimal. Rather, obtain knowledge from an acoustician, from building acoustics books/education or Slashdot, with the obvious tradeoff between accuracy and cost. Use constructors to construct.
Yes, read it again, that's the point; if the air gap is too large, then the air has an easier way of circulating between the panes, leading to worse insulation.
Between you and me, I'd pick the 12" concrete wall, or better yet, a 5" concrete wall and a 4" concrete wall with some air in between, or rockwool. The jeans solution didn't say what surrounds the cavity - chickenwire? drywall? And also, the old jeans are uneven; will rot, compact, get wet, catch fire, house mice or smell bad. While a mass-spring-mass system is better than the same mass with no spring between them, I'd argue for the merits of the 12" concrete wall; one could always attach another layer if needed (e.g. some rockwool and drywall that is not fixed to the concrete wall). There are all kinds of weird effects going on; it is possible to worsen the acoustic situation, for example a new mass layer highlights some frequency, or due to the way it is suspended, it conducts vibrations, turning them into giang speakers. Where's an acoustician when we need one?
Never heard of this one and while it sounds plausible, I find its effectiveness limited, due to the typically very small angles achievable (if the window should be openable). What I read in a building acoustics book is that that glass panes of different thickness are helpful, because they act as mass-spring-mass systems and their pass-through filter frequency will be different. Ideally there is an air gap between the panes, but multiple panes are useful too. For example, 4mm glass + 10mm air + 2x3mm glass can be a good window choice.
Also, insulation will be worse with planes that are not parallel; by necessity it means that on average you have a significantly larger air gap between the two planes, and the internal air (or noble gas mixture) is better able to circulate, reducing insulation. There is a reason why windows are the way they are, at least in Europe (I found the prevalence of single-glazed windows in the UK and the US appalling from an acoustic and environmental viewpoint, their insulation capability is a small fraction of that of double-glazed ones, often worsened by gaping holes).
It's a miracle to me how easy it is to forget about some key benefits of car ownership.
1. A lot of people own a car not because they have to but because they need to.
2. I would not be as comfortable in a car if it was used by the commons the same way public transportation is, and it would actually be worse: people are better at refraining from vandalism and disgusting things in the public, while they would be left to their own devices in a privately used car. Probably hardcore vandalism like knifed seats would be rare, due to renter identification, but there are a lot of bodily excretions that are even worse but harder to detect or penalize.
3. I can hop into my car at a moment's notice.
4. One can choose a car based on brand or performance preferences.
So there are lots of reasons for not burying the concept of car ownership just because of an upcoming invention. If you look at economy alone, why don't you put people in the rooms that you don't occupy? A large percent of rooms spend a large percent of time empty. Why don't you invite others to build on your lot, in the corner that is unused 99% of the time? Because man is not a pure economical animal.
And I don't personalize the interior of the car, don't clean it too often and don't drive a lot.
In fact there are arguments for increased car ownership with self-driving cars, too. Someone without a driver's license (for whatever reason) might own one. Someone who can't park near the house or the office might want one (or someone who can't park safely, or who is afraid to drive). Or anyone who uses public transportation so he can read along (or drink before heading back home) may become a buyer. I'm not suggesting that all these individual gains are good for the society, but some benefits of ownership will remain present, and there will be totally new benefits as well.
Harvest the mechanical energy of the keystroke. Either by using a manual typewriter, or using magnets and coils in the keys to generate tiny currents, which in turn charge a capacitor, which drive the mW level processor and LCD. If that does not work, make the kids exercise by taking turns on a stationary bicycle with a dynamo.
> You found 10ms to shave off a single calculation? 10 ms is practically infinity, even 10 ns would be important on most modern machines.
That's nothing, kiddos! Where I work, even sub-nanosecond time differences matter a lot on single calculations! And we use the fastest electrons we can get!
What's your interview technique? I'm just asking because an interview setting is very different from a work setting. To give an example, once I was interviewed for a position and questions about C/C++ came out, and I said, I haven't programmed in it, so I probably scored -2 on a scale of 1 to 5 (and didn't get hired). Then at work, soon a problem came up that required C programming, and guess what, with no experience in it whatsoever I was able to create working code (not a big task but still about a thousand lines, incl. GPGPU, OpenGL and use of template libraries). The development was kind of slow (maybe twice as long as of what I use daily) and obviously a good learning experience.
However the code worked nicely and reliably; I know where it was awesome and where I cut some corners (being an alpha version); what the tradeoffs were (performance, memory, verbosity, portability etc.); and most of all, I had a full grasp of the functional model and the inner workings of the solution.
Whereas I have seen programmers who had used language X for many years and would have aced typical interview questions such as write a heap sort or what's the best data representation for something (e.g. doubly linked list) and would still make a worse programmer than me. On multiple occasions I inherited code in a language new to me and rewrote much of it of higher coherence.
A closing thought is that it turned out, on an interview I can even fail questions where I'm at home; in an interactive discussion maybe the best approach does not come to mind, and I can't program on paper (i.e. without a PC with quick lookups as needed). So depending on how you interview people, the question is how much bias you introduce for people who fit the interview style and against those who get the job done in good quality in a work setting.
With some current languages (khm.. Java... khm... C++) there is so much broilerplate and there are so many frameworks and libraries that unless you spend a lot of time with it, the most productive approach is to go by example. Also, consider that macros, library functions and template (generic) programming is programming by example at its core. It is no accident that frameworks and "solutions" come with an SDK, typically with examples.
On the other hand, if you have the time, you can start with more of a RISC-like language, and/or do programming expressly for the fun of it. Such languages: Lisp, Scheme, Logo, Forth or Python. If you do something long enough for the fun of it in an interactive, relaxed setting that allows you to get into the groove (i.e. give you the flow experience when hours just fly by) then you can build up significantly. For the flow experience, there are preconditions, such as don't set the bar too high or too low and make it real interactive with a series of quick modifications and tests rather than writing 20 lines of code then test. If you can't find fun in it, that's the first thing to reprogram inside. Think about whatever application interests you - control a robot, draw something interesting like a fractal (generative art), or program the simplest game you can think of. If you are interested.
Is this an attempt to prefer or glorify the job title "developer" over "programmer"? Let's not pretend. We should not create a false dichotomy which is programmer vs. developer. The layering continuum, if we go into that, is vast. If he writes code to solve financial problems, he might be a quant. Or he may be a researcher. Or relative to an organization, a team lead, a software architect or project manager.
Aside from the labeling, we can say for certainty that there are experienced programmers who should have picked another career (provable via the existence proof). Good programming definitely requires aptitude. But it's not true that there is a separate animal called "developer" which embodies these qualities; it just looks better on a resume. Programmer, developer, software engineer, code jedi, play with the words.
Shedding unnecessary layers is a step in the right direction. But to be really disruptive, it's not enough of a jump, because by the time it hits the markets, the current, expensive CPUs will be the low end. Instead, you have to work on eliminating components altogether. Currently, smartphones have a separate chip for baseband and CPU, although stronger and stronger CPUs, these days A9 cores, are integrated with baseband. The writing is on the wall (or in this PDF: http://tinyurl.com/9bjsgv7). To create a truly impressive performance gap, additional performance gains should be obtained. To benefit both the user and the cellular network, convert verbose HTML, CSS and JS directly into ARM Thumb instructions on the server side - it is a compressed instruction set, and the machine code could call natively coded, expressive functions on the phone so it could still result in denser code than e.g. Java or LLVM byte code. You need to look at what mobile technology will be real cheap, and target the solution to that. Probably it's going to be a SoC that integrates baseband, CPU, memory, storage or SD card controller, LCD driver and any sensors.
In the farther future, all these components will be integrated on the back of LCD panels, so the ultimate phone will have two components: an LCD panel with all necessary electronic components including touch, backlight, CPU and baseband, and the battery, which will be a polymer shaped to become the housing of the phone.There will always be low-end CPUs, because at every transition point, there is an economic incentive for smaller, cooler, less power hungry processors. To target that, you need to go native code, but in smarter ways than NaCl / PNaCl.
Yet another benefit with native code is that for things like games or immersive apps, Javascript JIT creates an unpredictable performance characteristic, and lags and stutters may result.
> The CPU and GPU will be about 4x the performance of the iPhone 4. Best guess is that the "5" has dual ARM Cortex A15 processors running at around 1200MHz
A Cortex A15 is an optimistic guess, there is some speculation on the web that the memory subsystem was improved. Even if we take the 4x speedup at face value, there aren't yet killer OS features or applications that make it a piece de resistance. So currently there is not much of a use case. Your second paragraph is on spot w.r.t. the conservative development and playing it safe. The iPhone became a hit and a success because they did not play safe then, and the improvements seem to taper off, while Google might steal the show with new types of user interactions (Google Glasses) and other phones put more emphasis on higher bandwidth between man and machine (which is why it would have been cool to put a wider screen in the same physical dimensions).
How is it a big jump from the iPhone 4? Everything is slightly better, but I may stick to 4 or get an android. I don't care about Siri. A big jump was 4, quadrupling resolution, laminating screen glasses, multitasking and a cool new design. When it came, my 2-year old iPhone felt very slow esp. with the OS of the time.
The iPhone 5 should have introduced a wider, side to side screen and informative lock/home screen widgets at least, or using the back of the phone as a touchpad, or a google glasses like peripheral or something deeply AR, or all of this for a standing ovation.
> I work 50-60 hours a week trying to improve what I already have.
How do you find the time to teach, between 50-60 hours a week for improving what you have, and the occasional lunchbreak Slashdot presence since... seeing your uid... last millennium? I delivered some guest lectures and don't do that anymore due to how much time is needed for preparation.
> Although you can't replicate facial-expression-based judgement calls
Why not? We are just scratching the surface as far as on-line education is concerned. With essentially all devices possessing a camera, and having developed somewhat reliable facial expression detection, it's not a stretch of the imagination to see how, tens of thousands of students provide feedback through their facial expression. It is possible to set up the teacher, who - in this example - delivers the lecture or consultation in real time, and the several types of facial expressions are mimicked for him on his screen, a couple of dozen virtual students conveying the facial expressions. So it could even show the distribution of expressions, e.g. a polarized classroom or clusters of student types.
Just because Udacity, or even worse, the posterchild of non-innovative UI that is Coursera, does not emulate or supersede something that's a traditional property of the classroom does not mean it's impossible or even avoidable.
Slashdot Mirror
I made this mistake, and its contents (i.e., Windows 8 on portables might consolidate mobile and PC use cases) could have been summed up in half the length of the slashdot summary, which in turn was just astroturfing. Then, as usual with Microsoft propaganda trying to outdumb readers, the slashdot summary does NOT say: "Windows mobile version xxx is pretty good / growing / will be the good for whatever", but, adapting the message to the slashdot crowd, first it tactically discredits MS just to buy our agreement, then slip on the real message in parentheses, almost subliminally, which is the assumed 5% market share. It's not even near that except in certain Seattle and Espoo cafeterias.
The article itself is full of cheap dramatization like "When it was pointed out he forgot Microsoft, Schmidt said this wasn’t a mistake. (then in a standalone paragraph) I believe it is." and patronizing passages like "Seriously stop and think about this for a minute. [...] give it 12 months and try again."
> why not just install gnome or mate if you're unhappy with unity?
because XFCE looks better and is at least as good?
It's almost as if someone asked what equipment he needs for performing bypass surgery. First requirement would be a bookshelf (to be filled with books for one's study), and second is a cabinet for keeping degrees, diplomas and continuing education. I'd love to know what motivated the question to begin with, it sounds almost eerie, esp. the professional part. Maybe a PHB who wants to monitor his engineers' purchase requests or a lottery winner with his dreams.
Just put in a Fresnel lens; will not help with the sound, but fun to clean it.
Well, it's a glaring error coming from someone who teaches kids, period. You're right, we're not in class, but perhaps said teacher is eager to learn about and correct this basic issue. In addition, naming himself "VVrath" can be seen as solicitation for pointing out improvement opportunities, and who are you to interfere with that?
I realize that many slashdotters don't give a flying rat's ass about grammar (mixing their with there, it's with its), so feel free to mod this to oblivion while modding up the equally uninteresting grammar denialist ("shit like this", "we're not in class", wow this IS novel and insightful). I like the fact that some people have the spine to express what is worth expressing despite the wrath of the plebs.
I don't know where you get your numbers from and of what unit of measure. It's clear that thermal insulation is night and day between parallel panes with a distance in the teens (mm) filled with gas (i.e. standard fare double-pane window) vs. non-parallel pane windows which are filled with air (other gas would escape and would be too expensive at >10x the volume to begin with). Have you ever held your hand near a single-pane window in wintertime, as a more accessible experiment? The air falls down quite fast. Maybe it's not 5-6mph but between non-parallel panes air would rotate at several turns a minute.
If economical window sound insulation is needed, stick to parallel panes as the main window, filled with gas, and use glasses of unequal thickness, the heavier the better. An ex-relative of mine lives in a very busy city road and their three-pane window (getting more prevalent in Europe) does wonders without exotic things like non-parallel panes that insulate poorly and you can't open them. On a visit, it was quite an eerie feeling when they opened their window at once and all that noise came in.
Correct, since the libraries are either-or and 3.whatever was spottily supported the last time around (12 months ago), there are effectively two languages, Python2 (you need for those libraries) and Python3 (you'd rather use). Python's attractiveness diminished so much so that neither of these two languages were chosen.
| I rent. If I want, a simple phone call to one person ends my relationship and I can move.
In the US it was customary to lease for a minimum of one year, and upon renewal, it typically needs to be renewed for one year. Now the problem is, if you rent, it is usually because you have a relatively mobile lifestyle in that you always move where there is a good job opportunity. Now it is statistically unlikely that your wish to move for a job will coincide with the lease cycle, unless you swim in job offers or have children and need to adhere to a school year. Though I'd be OK with a minimum lease term, it is a rip-off to have to renew for another year as it either adds a constraint to your life or adds cost when you violate that constraint.
Or did the rental market become more reasonable since the mortgage crisis?
Don't consult acoustics with a constructor. Generally, constructors know how to sell what they can do, and they will happily construct your idea or in the absence of it (or despite of it) their own. But they don't need to live with the consequences that their work is ineffective or suboptimal. Rather, obtain knowledge from an acoustician, from building acoustics books/education or Slashdot, with the obvious tradeoff between accuracy and cost. Use constructors to construct.
Yes, read it again, that's the point; if the air gap is too large, then the air has an easier way of circulating between the panes, leading to worse insulation.
Between you and me, I'd pick the 12" concrete wall, or better yet, a 5" concrete wall and a 4" concrete wall with some air in between, or rockwool. The jeans solution didn't say what surrounds the cavity - chickenwire? drywall? And also, the old jeans are uneven; will rot, compact, get wet, catch fire, house mice or smell bad. While a mass-spring-mass system is better than the same mass with no spring between them, I'd argue for the merits of the 12" concrete wall; one could always attach another layer if needed (e.g. some rockwool and drywall that is not fixed to the concrete wall). There are all kinds of weird effects going on; it is possible to worsen the acoustic situation, for example a new mass layer highlights some frequency, or due to the way it is suspended, it conducts vibrations, turning them into giang speakers. Where's an acoustician when we need one?
Never heard of this one and while it sounds plausible, I find its effectiveness limited, due to the typically very small angles achievable (if the window should be openable). What I read in a building acoustics book is that that glass panes of different thickness are helpful, because they act as mass-spring-mass systems and their pass-through filter frequency will be different. Ideally there is an air gap between the panes, but multiple panes are useful too. For example, 4mm glass + 10mm air + 2x3mm glass can be a good window choice.
Also, insulation will be worse with planes that are not parallel; by necessity it means that on average you have a significantly larger air gap between the two planes, and the internal air (or noble gas mixture) is better able to circulate, reducing insulation. There is a reason why windows are the way they are, at least in Europe (I found the prevalence of single-glazed windows in the UK and the US appalling from an acoustic and environmental viewpoint, their insulation capability is a small fraction of that of double-glazed ones, often worsened by gaping holes).
It's a miracle to me how easy it is to forget about some key benefits of car ownership.
1. A lot of people own a car not because they have to but because they need to.
2. I would not be as comfortable in a car if it was used by the commons the same way public transportation is, and it would actually be worse: people are better at refraining from vandalism and disgusting things in the public, while they would be left to their own devices in a privately used car. Probably hardcore vandalism like knifed seats would be rare, due to renter identification, but there are a lot of bodily excretions that are even worse but harder to detect or penalize.
3. I can hop into my car at a moment's notice.
4. One can choose a car based on brand or performance preferences.
So there are lots of reasons for not burying the concept of car ownership just because of an upcoming invention. If you look at economy alone, why don't you put people in the rooms that you don't occupy? A large percent of rooms spend a large percent of time empty. Why don't you invite others to build on your lot, in the corner that is unused 99% of the time? Because man is not a pure economical animal.
And I don't personalize the interior of the car, don't clean it too often and don't drive a lot.
In fact there are arguments for increased car ownership with self-driving cars, too. Someone without a driver's license (for whatever reason) might own one. Someone who can't park near the house or the office might want one (or someone who can't park safely, or who is afraid to drive). Or anyone who uses public transportation so he can read along (or drink before heading back home) may become a buyer. I'm not suggesting that all these individual gains are good for the society, but some benefits of ownership will remain present, and there will be totally new benefits as well.
Harvest the mechanical energy of the keystroke. Either by using a manual typewriter, or using magnets and coils in the keys to generate tiny currents, which in turn charge a capacitor, which drive the mW level processor and LCD. If that does not work, make the kids exercise by taking turns on a stationary bicycle with a dynamo.
> You found 10ms to shave off a single calculation? 10 ms is practically infinity, even 10 ns would be important on most modern machines.
That's nothing, kiddos! Where I work, even sub-nanosecond time differences matter a lot on single calculations! And we use the fastest electrons we can get!
> 80% of programmers can program, 20% can do it efficiently.
And beyond these, there is another 20% who can't program or even add up to 100%.
What's your interview technique? I'm just asking because an interview setting is very different from a work setting. To give an example, once I was interviewed for a position and questions about C/C++ came out, and I said, I haven't programmed in it, so I probably scored -2 on a scale of 1 to 5 (and didn't get hired). Then at work, soon a problem came up that required C programming, and guess what, with no experience in it whatsoever I was able to create working code (not a big task but still about a thousand lines, incl. GPGPU, OpenGL and use of template libraries). The development was kind of slow (maybe twice as long as of what I use daily) and obviously a good learning experience.
However the code worked nicely and reliably; I know where it was awesome and where I cut some corners (being an alpha version); what the tradeoffs were (performance, memory, verbosity, portability etc.); and most of all, I had a full grasp of the functional model and the inner workings of the solution.
Whereas I have seen programmers who had used language X for many years and would have aced typical interview questions such as write a heap sort or what's the best data representation for something (e.g. doubly linked list) and would still make a worse programmer than me. On multiple occasions I inherited code in a language new to me and rewrote much of it of higher coherence.
A closing thought is that it turned out, on an interview I can even fail questions where I'm at home; in an interactive discussion maybe the best approach does not come to mind, and I can't program on paper (i.e. without a PC with quick lookups as needed). So depending on how you interview people, the question is how much bias you introduce for people who fit the interview style and against those who get the job done in good quality in a work setting.
With some current languages (khm.. Java ... khm... C++) there is so much broilerplate and there are so many frameworks and libraries that unless you spend a lot of time with it, the most productive approach is to go by example. Also, consider that macros, library functions and template (generic) programming is programming by example at its core. It is no accident that frameworks and "solutions" come with an SDK, typically with examples.
On the other hand, if you have the time, you can start with more of a RISC-like language, and/or do programming expressly for the fun of it. Such languages: Lisp, Scheme, Logo, Forth or Python. If you do something long enough for the fun of it in an interactive, relaxed setting that allows you to get into the groove (i.e. give you the flow experience when hours just fly by) then you can build up significantly. For the flow experience, there are preconditions, such as don't set the bar too high or too low and make it real interactive with a series of quick modifications and tests rather than writing 20 lines of code then test. If you can't find fun in it, that's the first thing to reprogram inside. Think about whatever application interests you - control a robot, draw something interesting like a fractal (generative art), or program the simplest game you can think of. If you are interested.
Is this an attempt to prefer or glorify the job title "developer" over "programmer"? Let's not pretend. We should not create a false dichotomy which is programmer vs. developer. The layering continuum, if we go into that, is vast. If he writes code to solve financial problems, he might be a quant. Or he may be a researcher. Or relative to an organization, a team lead, a software architect or project manager.
Aside from the labeling, we can say for certainty that there are experienced programmers who should have picked another career (provable via the existence proof). Good programming definitely requires aptitude. But it's not true that there is a separate animal called "developer" which embodies these qualities; it just looks better on a resume. Programmer, developer, software engineer, code jedi, play with the words.
Maybe it's more to do with finding programming fun to begin with, or at least the ability to find fun in whatever he chooses to do, e.g. programming.
Shedding unnecessary layers is a step in the right direction. But to be really disruptive, it's not enough of a jump, because by the time it hits the markets, the current, expensive CPUs will be the low end. Instead, you have to work on eliminating components altogether. Currently, smartphones have a separate chip for baseband and CPU, although stronger and stronger CPUs, these days A9 cores, are integrated with baseband. The writing is on the wall (or in this PDF: http://tinyurl.com/9bjsgv7). To create a truly impressive performance gap, additional performance gains should be obtained. To benefit both the user and the cellular network, convert verbose HTML, CSS and JS directly into ARM Thumb instructions on the server side - it is a compressed instruction set, and the machine code could call natively coded, expressive functions on the phone so it could still result in denser code than e.g. Java or LLVM byte code. You need to look at what mobile technology will be real cheap, and target the solution to that. Probably it's going to be a SoC that integrates baseband, CPU, memory, storage or SD card controller, LCD driver and any sensors.
In the farther future, all these components will be integrated on the back of LCD panels, so the ultimate phone will have two components: an LCD panel with all necessary electronic components including touch, backlight, CPU and baseband, and the battery, which will be a polymer shaped to become the housing of the phone.There will always be low-end CPUs, because at every transition point, there is an economic incentive for smaller, cooler, less power hungry processors. To target that, you need to go native code, but in smarter ways than NaCl / PNaCl.
Yet another benefit with native code is that for things like games or immersive apps, Javascript JIT creates an unpredictable performance characteristic, and lags and stutters may result.
> The CPU and GPU will be about 4x the performance of the iPhone 4. Best guess is that the "5" has dual ARM Cortex A15 processors running at around 1200MHz
A Cortex A15 is an optimistic guess, there is some speculation on the web that the memory subsystem was improved. Even if we take the 4x speedup at face value, there aren't yet killer OS features or applications that make it a piece de resistance. So currently there is not much of a use case. Your second paragraph is on spot w.r.t. the conservative development and playing it safe. The iPhone became a hit and a success because they did not play safe then, and the improvements seem to taper off, while Google might steal the show with new types of user interactions (Google Glasses) and other phones put more emphasis on higher bandwidth between man and machine (which is why it would have been cool to put a wider screen in the same physical dimensions).
How is it a big jump from the iPhone 4? Everything is slightly better, but I may stick to 4 or get an android.
I don't care about Siri. A big jump was 4, quadrupling resolution, laminating screen glasses, multitasking and a cool new design. When it came, my 2-year old iPhone felt very slow esp. with the OS of the time.
The iPhone 5 should have introduced a wider, side to side screen and informative lock/home screen widgets at least, or using the back of the phone as a touchpad, or a google glasses like peripheral or something deeply AR, or all of this for a standing ovation.
> I work 50-60 hours a week trying to improve what I already have.
How do you find the time to teach, between 50-60 hours a week for improving what you have, and the occasional lunchbreak Slashdot presence since ... seeing your uid ... last millennium? I delivered some guest lectures and don't do that anymore due to how much time is needed for preparation.
> Although you can't replicate facial-expression-based judgement calls
Why not? We are just scratching the surface as far as on-line education is concerned. With essentially all devices possessing a camera, and having developed somewhat reliable facial expression detection, it's not a stretch of the imagination to see how, tens of thousands of students provide feedback through their facial expression. It is possible to set up the teacher, who - in this example - delivers the lecture or consultation in real time, and the several types of facial expressions are mimicked for him on his screen, a couple of dozen virtual students conveying the facial expressions. So it could even show the distribution of expressions, e.g. a polarized classroom or clusters of student types.
Just because Udacity, or even worse, the posterchild of non-innovative UI that is Coursera, does not emulate or supersede something that's a traditional property of the classroom does not mean it's impossible or even avoidable.