“对do后面那串东西按monad那三条运算规则进行证明和优化基本上属于扯淡” 如果是说 IO Monad,我同意。 换做其它 Monad,就不是这样了。比如说 SQL 本质上就是 Monad,利用 Monad 性质优 化 query language 的工作二十年前就有人做了。Principles of programming with complex objects and collection types, P Buneman, S Naqvi, V Tannen, L Wong. 1995. 近 300 个引用不是白给的。
fp on avg programmer hand is just a free ride to maintenance disaster. 3 levels of functional reference will definitely kill any avg guy to improve the code base when original owner is gone. definitely not recommended for any laege scale system. might as well to rewrite it. receipt for job security though, if the executive is advantural enough to let it start with fp.
FP 最大的问题是method as parameter; 一个method的第一次被referenced 和其实际 使用地点往往相差甚远,其直接结果是只有概念上将被运行,却不知道实际运行的时间 和环境。这就要求reader有对整个系统的详细了解,知道该method在何时何地会被运行 ,然后才是debug/改进,而这恰恰是大系统下一般programmer所不具备的。 这恰恰是资深programmer的最爱, manager的梦魇,--- nobody can do anything unless I watch over it.
O*d
117 楼
用C++模板写一个简单的functional template int Sum() { return N + Sum(); } template <> int Sum<0>() { return 0; } int main (int argc, char** argv) { int a = Sum<9>(); printf("a = %d\n", a); // a == 9 + 8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 + 0 that is 45. return 0; }
【在 a*********a 的大作中提到】 : just using lambda is far from FP. : report back when the main stream languages ditch "for" and "while".
a*a
177 楼
raw C++ template metaprogramming dose not support iteration, nor side effects. just these two characters make it levels purer functional than most main stream programming language. When you write meta programs, most stuff you do are in the form of a serious superposed functions. MPL introduces some support for iteration but under the hood I believe, it is still a car/cdr recursion over an MPL container. Interestingly, Common Lisp later also introduced 'for' loop macro.
Overall q3 added more subscribers than last year. Domestic is gonna saturate at some point, it's all about international. If you want to be picky, you can always find flaw in earnings. I couldn't care less about day to day stock fluctuation.
the
【在 a*********a 的大作中提到】 : cheap or not, $1 increase in monthly fee on new subscribers only broker the : steady stream of increase in subscription number.
a*a
180 楼
a 25% drop within a few minutes and not recover in months to come is hardly day to day stock fluctuation. Incidentally when netflix last demanded a price increase the market value of the company was slaughtered too. At that time, netflix demanded a 100% price increase, i.e. (from $8 stream + mailer) to ($8 stream + $8 mailer), stock price dropped from 286 to 66 over the course of a few months and stayed there for over a year. this time it was an $1 increase to new subscribers only. that is enough to cause the subscriber growth to miss forecast. combined with news from HBO streaming, stock dropped 25% in a matter of few minutes after the news broke. note it was not that people don't expect the sub growth to saturate, it was that people was caught by surprise how much this $1 increase affected the growth. I am a long time netflix subscriber ever since it was a DVD mailer (before it even started to host blu ray discs :). And here is how cheap I am. The moment I learned about the spin off of mailer and that I must pay $8 for each of the mail and stream services, I suspended my account. it was only years after that I reopened them. Then earlier this year, when netflix became crawling slow on verizon. I chose to suspend my netflix stream account and stay with verizon because I m still in the 2-year promotion period with Verizon. That said, I d be very happy at this point to pay $16 a month and have netflix use the $8 extra to pay Verizon and other ISPs for bandwidth access.
saturate
【在 g*****g 的大作中提到】 : Overall q3 added more subscribers than last year. Domestic is gonna saturate : at some point, it's all about international. If you want to be picky, you : can always find flaw in earnings. I couldn't care less about day to day : stock fluctuation. : : the
g*g
181 楼
NFLX is volatile, it can go down 25% in a few minutes or it can go up 25%, it happens. It was a $50 stock just 2 years ago down from $300, so what? A stock no matter how good it is, doesn't go up straight, otherwise Wall street wouldn't make the most out of it. Feel free to short it if you like, I couldn't care less. I have the insider data and you don't. And history since NFLX IPO is on my side. So just save some crap.
hardly of broke.
【在 a*********a 的大作中提到】 : a 25% drop within a few minutes and not recover in months to come is hardly : day to day stock fluctuation. : Incidentally when netflix last demanded a price increase the market value of : the company was slaughtered too. At that time, netflix demanded a 100% : price increase, i.e. (from $8 stream + mailer) to ($8 stream + $8 mailer), : stock price dropped from 286 to 66 over the course of a few months and : stayed there for over a year. : this time it was an $1 increase to new subscribers only. that is enough to : cause the subscriber growth to miss forecast. combined with news from HBO : streaming, stock dropped 25% in a matter of few minutes after the news broke.
a*a
182 楼
Point taken. I m sure your CEO and CFO intimately share with you tons of insider information about how your company is managed and financed, just like mine with me.
【在 g*****g 的大作中提到】 : NFLX is volatile, it can go down 25% in a few minutes or it can go up 25%, : it happens. It was a $50 stock just 2 years ago down from $300, so what? A : stock no matter how good it is, doesn't go up straight, otherwise Wall : street wouldn't make the most out of it. Feel free to short it if you like, : I couldn't care less. I have the insider data and you don't. And history : since NFLX IPO is on my side. So just save : some crap. : : hardly : of
g*g
183 楼
I don't actually need a lot of data, I can see the daily subscriber, and the stock price pretty much depending on it.
【在 a*********a 的大作中提到】 : Point taken. : I m sure your CEO and CFO intimately share with you tons of insider : information about how your company is managed and financed, just like mine : with me.
“对do后面那串东西按monad那三条运算规则进行证明和优化基本上属于扯淡” 如果是说 IO Monad,我同意。 换做其它 Monad,就不是这样了。比如说 SQL 本质上就是 Monad,利用 Monad 性质优 化 query language 的工作二十年前就有人做了。Principles of programming with complex objects and collection types, P Buneman, S Naqvi, V Tannen, L Wong. 1995. 近 300 个引用不是白给的。
fp on avg programmer hand is just a free ride to maintenance disaster. 3 levels of functional reference will definitely kill any avg guy to improve the code base when original owner is gone. definitely not recommended for any laege scale system. might as well to rewrite it. receipt for job security though, if the executive is advantural enough to let it start with fp.
FP 最大的问题是method as parameter; 一个method的第一次被referenced 和其实际 使用地点往往相差甚远,其直接结果是只有概念上将被运行,却不知道实际运行的时间 和环境。这就要求reader有对整个系统的详细了解,知道该method在何时何地会被运行 ,然后才是debug/改进,而这恰恰是大系统下一般programmer所不具备的。 这恰恰是资深programmer的最爱, manager的梦魇,--- nobody can do anything unless I watch over it.
O*d
264 楼
用C++模板写一个简单的functional template int Sum() { return N + Sum(); } template <> int Sum<0>() { return 0; } int main (int argc, char** argv) { int a = Sum<9>(); printf("a = %d\n", a); // a == 9 + 8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 + 0 that is 45. return 0; }
【在 a*********a 的大作中提到】 : just using lambda is far from FP. : report back when the main stream languages ditch "for" and "while".
a*a
324 楼
raw C++ template metaprogramming dose not support iteration, nor side effects. just these two characters make it levels purer functional than most main stream programming language. When you write meta programs, most stuff you do are in the form of a serious superposed functions. MPL introduces some support for iteration but under the hood I believe, it is still a car/cdr recursion over an MPL container. Interestingly, Common Lisp later also introduced 'for' loop macro.
Overall q3 added more subscribers than last year. Domestic is gonna saturate at some point, it's all about international. If you want to be picky, you can always find flaw in earnings. I couldn't care less about day to day stock fluctuation.
the
【在 a*********a 的大作中提到】 : cheap or not, $1 increase in monthly fee on new subscribers only broker the : steady stream of increase in subscription number.
a*a
327 楼
a 25% drop within a few minutes and not recover in months to come is hardly day to day stock fluctuation. Incidentally when netflix last demanded a price increase the market value of the company was slaughtered too. At that time, netflix demanded a 100% price increase, i.e. (from $8 stream + mailer) to ($8 stream + $8 mailer), stock price dropped from 286 to 66 over the course of a few months and stayed there for over a year. this time it was an $1 increase to new subscribers only. that is enough to cause the subscriber growth to miss forecast. combined with news from HBO streaming, stock dropped 25% in a matter of few minutes after the news broke. note it was not that people don't expect the sub growth to saturate, it was that people was caught by surprise how much this $1 increase affected the growth. I am a long time netflix subscriber ever since it was a DVD mailer (before it even started to host blu ray discs :). And here is how cheap I am. The moment I learned about the spin off of mailer and that I must pay $8 for each of the mail and stream services, I suspended my account. it was only years after that I reopened them. Then earlier this year, when netflix became crawling slow on verizon. I chose to suspend my netflix stream account and stay with verizon because I m still in the 2-year promotion period with Verizon. That said, I d be very happy at this point to pay $16 a month and have netflix use the $8 extra to pay Verizon and other ISPs for bandwidth access.
saturate
【在 g*****g 的大作中提到】 : Overall q3 added more subscribers than last year. Domestic is gonna saturate : at some point, it's all about international. If you want to be picky, you : can always find flaw in earnings. I couldn't care less about day to day : stock fluctuation. : : the
g*g
328 楼
NFLX is volatile, it can go down 25% in a few minutes or it can go up 25%, it happens. It was a $50 stock just 2 years ago down from $300, so what? A stock no matter how good it is, doesn't go up straight, otherwise Wall street wouldn't make the most out of it. Feel free to short it if you like, I couldn't care less. I have the insider data and you don't. And history since NFLX IPO is on my side. So just save some crap.
hardly of broke.
【在 a*********a 的大作中提到】 : a 25% drop within a few minutes and not recover in months to come is hardly : day to day stock fluctuation. : Incidentally when netflix last demanded a price increase the market value of : the company was slaughtered too. At that time, netflix demanded a 100% : price increase, i.e. (from $8 stream + mailer) to ($8 stream + $8 mailer), : stock price dropped from 286 to 66 over the course of a few months and : stayed there for over a year. : this time it was an $1 increase to new subscribers only. that is enough to : cause the subscriber growth to miss forecast. combined with news from HBO : streaming, stock dropped 25% in a matter of few minutes after the news broke.
a*a
329 楼
Point taken. I m sure your CEO and CFO intimately share with you tons of insider information about how your company is managed and financed, just like mine with me.
【在 g*****g 的大作中提到】 : NFLX is volatile, it can go down 25% in a few minutes or it can go up 25%, : it happens. It was a $50 stock just 2 years ago down from $300, so what? A : stock no matter how good it is, doesn't go up straight, otherwise Wall : street wouldn't make the most out of it. Feel free to short it if you like, : I couldn't care less. I have the insider data and you don't. And history : since NFLX IPO is on my side. So just save : some crap. : : hardly : of
g*g
330 楼
I don't actually need a lot of data, I can see the daily subscriber, and the stock price pretty much depending on it.
【在 a*********a 的大作中提到】 : Point taken. : I m sure your CEO and CFO intimately share with you tons of insider : information about how your company is managed and financed, just like mine : with me.