Is your application slow? This is because our civilization is dying.
With your permission, albeit a bit in the philosophical shell, I would like to continue the topic “Why are mobile web applications so slow?” , Which already appeared on Habré through another post: “HTML5 in mobile development - what to choose?” .
Translation of the post "Your App Is Slow Because Our World Is Ending" .
As usual, this week, exploring the Internet, I came across a post exploding the brain, which may become a precursor of fatal events. I mean the excellent Drew Crawford article “Why are mobile web apps so slow?” . It is voluminous and very informative, but you should definitely get acquainted with it if your level of knowledge is at a high enough level to follow the author’s idea.
')
In short, his idea goes something like this: HTML5 applications will significantly yield to native applications not only for the next year or two, but for the next 5-10 years, due to significant technical limitations inherent in mobile platforms. I had a couple of comments on some of his arguments, but he convinced me that two articles that I wrote two years ago, in which I urge iOS for garbage collection and argue that HTML5 will rule everything in a few years, should probably be revised in the light of new evidence.
Well, well, I will say this: I was wrong. But, in fairness, it should be noted that in my call for "garbage collection", I found a good company in the face of the very god of programming - John Carmack . “Good article Why mobile web ...? . I tend to think that garbage collection in some cases is justified even for games on large platforms, ” writes Carmack . But I'm also sure that HTML5 may not be the right solution for any truly non-trivial iOS application, either now or in the coming years; This position seems to me true for other mobile OSs.
When you start thinking about the root causes, this question becomes even more interesting or even in some way “apocalyptic”.
From early Crawford: “The solution is obvious! Just make ARM 10 times faster, and this will be enough to compete with x86, and then we will see instant JS execution! You want to agree, or not, it all depends on your belief in Moore's law against the background of the power capabilities of mobile chips with 3-ounce batteries. ”
This quote caught my eye, because not so long ago I noticed a surge of news that warned that "We have fallen behind Moore's Law since Intel rested on the“ Power Wall ” back in 2005”, - quoting Linley Gwenapp.
It would not be an exaggeration if I say that Moore’s law has been the single most important and powerful guide of our civilization for the past fifty years. For the reason that “technology, whether we like it or not , is the economic and cultural dominant of our time.” So there’s nothing wrong with it ...
But ask AMD , ask Broadcom ; A considerable amount of evidence seems to indicate that over the past few years, the dizzying exponential growth of computing power that created the high-tech industry as we know it, and has changed the whole world and the life of each individual person ... finally slows down.
Of course, people have repeatedly predicted its ending for decades, and somehow that capacity growth continued to grow, not paying attention to the forecasts. But now, it seems, we are trapped between the Scylla of heat and power dissipation on one side and the Charybdis of the fundamental quantum limits on the other. Although various innovations are being discussed at the moment, and some of them can actually succeed, let's not forget the original formulation of Moore's Law, which refers to the “number of components in an integrated circuit with minimal cost / cost”. It seems reasonable to be skeptical about new approaches / technologies that are unlikely to give us the similar price-quality ratio that we used to expect.
We’ve got used to thinking that every few years, technology becomes an order of magnitude faster / less / cheaper - but in fact, when you stop and think about it seriously, it may seem incredibly bizarre and crazy to you. Unlimited exponential growth must end sometime, by definition; and this, when this happens, will not be a big blow for us, but still it will make us sad. We will not hit a wall, we will simply ... begin ... slowly ... lose ... speed. And we note that this will happen primarily on most hardware devices with limited capabilities, that is, on our phones.
After we get into history, I think our era will be called the “era of Moore”. Now we may be witnessing the beginning of the end. This does not mean that progress will stop; even in the worst case, the shock wave from fifty years of rapid technical growth, like an echo, will affect all of humanity and all that we will do for several more decades. And it is also quite possible that Ray Kurzweil and the Singularitarianists are right, and that Moore's law will be extended or replaced by something even more insane. But it is also possible that we will finally come to the end of the most rapidly changing era in history and the future of humanity.
If so, then perhaps the new generations will look back on the historical period from 1963 to, well, let's say today, and sigh sadly: "Probably, this was the most exciting time in the whole history of mankind." Joke, of course. What can we know about the future? But still, just in case, if this is the end, let's do something like that, after which in the future we will be remembered only in a good tone.
[from translator] Stumbled upon an interesting article on this subject Moore 's Law against nanometers . It can be very useful to those who want to penetrate into this topic and argue with the author of the topic.
Translation of the post "Your App Is Slow Because Our World Is Ending" .
As usual, this week, exploring the Internet, I came across a post exploding the brain, which may become a precursor of fatal events. I mean the excellent Drew Crawford article “Why are mobile web apps so slow?” . It is voluminous and very informative, but you should definitely get acquainted with it if your level of knowledge is at a high enough level to follow the author’s idea.
')
In short, his idea goes something like this: HTML5 applications will significantly yield to native applications not only for the next year or two, but for the next 5-10 years, due to significant technical limitations inherent in mobile platforms. I had a couple of comments on some of his arguments, but he convinced me that two articles that I wrote two years ago, in which I urge iOS for garbage collection and argue that HTML5 will rule everything in a few years, should probably be revised in the light of new evidence.
Well, well, I will say this: I was wrong. But, in fairness, it should be noted that in my call for "garbage collection", I found a good company in the face of the very god of programming - John Carmack . “Good article Why mobile web ...? . I tend to think that garbage collection in some cases is justified even for games on large platforms, ” writes Carmack . But I'm also sure that HTML5 may not be the right solution for any truly non-trivial iOS application, either now or in the coming years; This position seems to me true for other mobile OSs.
When you start thinking about the root causes, this question becomes even more interesting or even in some way “apocalyptic”.
From early Crawford: “The solution is obvious! Just make ARM 10 times faster, and this will be enough to compete with x86, and then we will see instant JS execution! You want to agree, or not, it all depends on your belief in Moore's law against the background of the power capabilities of mobile chips with 3-ounce batteries. ”
This quote caught my eye, because not so long ago I noticed a surge of news that warned that "We have fallen behind Moore's Law since Intel rested on the“ Power Wall ” back in 2005”, - quoting Linley Gwenapp.
It would not be an exaggeration if I say that Moore’s law has been the single most important and powerful guide of our civilization for the past fifty years. For the reason that “technology, whether we like it or not , is the economic and cultural dominant of our time.” So there’s nothing wrong with it ...
But ask AMD , ask Broadcom ; A considerable amount of evidence seems to indicate that over the past few years, the dizzying exponential growth of computing power that created the high-tech industry as we know it, and has changed the whole world and the life of each individual person ... finally slows down.
Of course, people have repeatedly predicted its ending for decades, and somehow that capacity growth continued to grow, not paying attention to the forecasts. But now, it seems, we are trapped between the Scylla of heat and power dissipation on one side and the Charybdis of the fundamental quantum limits on the other. Although various innovations are being discussed at the moment, and some of them can actually succeed, let's not forget the original formulation of Moore's Law, which refers to the “number of components in an integrated circuit with minimal cost / cost”. It seems reasonable to be skeptical about new approaches / technologies that are unlikely to give us the similar price-quality ratio that we used to expect.
We’ve got used to thinking that every few years, technology becomes an order of magnitude faster / less / cheaper - but in fact, when you stop and think about it seriously, it may seem incredibly bizarre and crazy to you. Unlimited exponential growth must end sometime, by definition; and this, when this happens, will not be a big blow for us, but still it will make us sad. We will not hit a wall, we will simply ... begin ... slowly ... lose ... speed. And we note that this will happen primarily on most hardware devices with limited capabilities, that is, on our phones.
After we get into history, I think our era will be called the “era of Moore”. Now we may be witnessing the beginning of the end. This does not mean that progress will stop; even in the worst case, the shock wave from fifty years of rapid technical growth, like an echo, will affect all of humanity and all that we will do for several more decades. And it is also quite possible that Ray Kurzweil and the Singularitarianists are right, and that Moore's law will be extended or replaced by something even more insane. But it is also possible that we will finally come to the end of the most rapidly changing era in history and the future of humanity.
If so, then perhaps the new generations will look back on the historical period from 1963 to, well, let's say today, and sigh sadly: "Probably, this was the most exciting time in the whole history of mankind." Joke, of course. What can we know about the future? But still, just in case, if this is the end, let's do something like that, after which in the future we will be remembered only in a good tone.
[from translator] Stumbled upon an interesting article on this subject Moore 's Law against nanometers . It can be very useful to those who want to penetrate into this topic and argue with the author of the topic.
Source: https://habr.com/ru/post/186660/
All Articles