Agreed, we have massive sites serving millions of requests a day using Open Source relational databases and yet it seems everyone wants to use NoSQL because it's the hip new thing.
A million requests per day translates to 11.5 requests per second. That's a pretty trivial amount of traffic. A massive site like Facebook is probably serving about 4 orders-of-magnitude more requests that that.
Did you read your own link? Every poster who followed-up said that the problem was resolved when they realized that they had to slide (rather than simply touch) the answer button in order to answer calls.
What "lavish" benefits are you talking about? Lunches? Lunches pay for themselves because they all of a sudden take 25-30 minutes instead of an hour or more. At $100+ (sometimes way more than that) per hour it just makes sense for a company to pay for lunches. Buses to and from work? Umm. OK, I'll give you that (even though Microsoft also has buses). On-site gym that hardly anyone goes to? What else?
Tuition reimbursement up to $12,000 per year.
Back-Up Child Care
Charity gift matching
Adoption assistance
On-site doctor (though dental seems more useful to me), oil change, games rooms, car wash, laundry, dry cleaning, massage, barber, fitness classes, bike repair, tech talks (by Barrack Obama, Randall Munroe, etc.)
Annual ski trip and other random trips e.g. one
20% time (is that a benefit?)
Plus the usual as far as medical, dental, stock options, etc. And probably a bunch of other stuff that I don't know about.
Google is actually pretty bare bones on the inside.
If you search for tank man directly then you get thousands of results including 4 pictures of him in front of the tanks at the top of the results page.
The real scandal here is that they can do this at all. Why should corporations be allowed to negotiate taxes? Can individuals do that? "Hello, state! I am thinking of moving to your state and being productively employed! And some of the money I make will be spent in the local economy! Promise! I will do so if you cut me some slack on property taxes and state income taxes!"
You'd get laughed at. Scale? so what, could 200 independent single individuals do the same? Nope. But a corporation can.
You know, I'm not sure that a very wealthy individual could not negotiate in this fashion. What if Bill Gates said that he was thinking about moving his family to NC, wanted to build a $500,000,000 mansion there, and would permanently employ 200 people (gardeners, helicopter pilots, etc.) but the property taxes in NC are too high. You don't think that the government would consider cutting him a break?
So do I. And I don't do executive management for a living. It's an amusing situation, yes, but an unfair one. You don't expect an auto mechanic do also do the accounting and sales for the dealership; neither should you expect the comptroller to do a valve job.
I think that you are being overly generous. Steve Balmer was only called on to do something that an end user of his (consumer!) product would want to do. A better analogy would be that you shouldn't expect an executive at a car manufacturer to be able to drive the company's cars. But of course you would.
I think we should be allowed to mod the stories as well as the comments. This way we could get rid of both the dupes and the trolls like the current story.
I think that is a terrible idea - I view slashdot at +5 and I want to see more than just advertisements.
Actually, if we could comment on the advertisements, maybe it would be ok...
Java is a strongly typed language and this is not going to change ("dynamical typing" doesn't mean anything by the way).
It is interesting that you think that "dynamic typing is meaningless". It is widely used with the definition that typing is resolved at runtime (rather than at compile-time). How could you classify the following three cases?
Based on this (and other) behavior, many people would say that Perl is weakly-typed while Python and Java are strongly-typed. Python's typing is enforced at runtime while Java's typing is enforced at compile-time. Dynamic vs. static typing is a convenient set of terminology for expressing this difference.
How can you call Apple a hardware company? Because they put everything in a well designed box? All the components are 3rd party... Apple doesn't make processors, Apple doesn't make memory, Apple doesn't make harddrives or video cards or sound cards. They buy them from hardware companies, put them in a shiny box and then run *their software* on it.
Without the seven last words, this could also apply to Dell. I'm not disagreeing with your point, just observing that other companies seem to make money selling hardware using only 3rd party components.
NetFlix already has a healthy base of customers, and anyone interested in such a service already knows about them. What does Amazon have to bring to the table, other than not crushing them like a bug?
A stronger brand. Despite what you say, I doubt that most interested comsumers are aware of online DVD rental.
Empirical evidence has shown, time and time again, that low level languages like C and Forth produce more efficient, faster, and easier to maintain code than today's so called high level languages.
If that has been shown time and time again then providing references should be easy. I'm especially interested in convincing evidence that C and Forth are easier to maintain than "so called" high level languages.
There is a large difference. Gravity is accepted by probably about 99.9999999999% of the world. Evolution is accepted by a minority of the world.
As other will certainly point out, science is not a democracy. So I'll just dispute your claim that the theory of gravitation is universily accepted. I'm sure that most people would accept the observation that they tend not to float away from the Earth but do you really believe that everyone knows (and believes) that the phenomena is caused by attraction between masses?
Yes, Java's GC would notice that nothing is referring to them and remove the objects. This unlike a simple reference counting gc such as python's which would not notice this. Java's GC can even relocate memory on the fly to minimize page misses and avoid memory fragmentation.
Python actually uses a generational garbage collection system that can break cycles to reclaim unused objects. It also performs certain optimizations to avoid unnecessary memory allocations and deallocations.
In Python, reference counting is a combination of a historical artifact and a performance optimization.
To me, it's more important to measure the potential of a programmer, more than the experience of the programmer. Creating new ideas and systems does not require experience alone -- it requires a great deal of potential. Simply breaking and fixing code in a realtime setting is good for the day to day stuff (and that is valuable for Google), but when a customer or boss asks to you build something new and difficult, many experienced programmers answer with, "that can't be done" when less experienced programmers will try to do it, with limited success.
I did the Google Code Jam last year and found that they required a lot more than basic coding skills.
One question involved analysing sequences of resource lock and unlocks, represented by upper- and lowercase characters in strings e.g. "ABba", "BAba" and "CADadBbc". Given such strings, you had to write a program to identify potential deadlock conditions e.g. the first two strings can deadlock at position 1 because the first string will hold a lock for A while waiting for B and the second string will hold a lock for B while waiting for A. Of course the real tests were more difficult than this. I think that less than 10 people out of 250 got this one.
Another question involved graph searching but with the cost of node traversal changing with time according to a pseudorandom algorithm. My solution was only twenty or thirty lines of code but it took a while to think how to do a time-based graph search.
The challenge phase is fairly challenging (excuse the pun) as well because you only have thirty minutes to make all of your challenges and you lose points if you are incorrect i.e. you claim that a certain valid input combination will cause the program to fail and it doesn't. Thirty minutes is not a lot of time to attack the 30 potentially flawed solutions that you are looking at. You have to pick a set of likely flaws (based on your own design) and quickly scan each set of code for it. Then, when you find a likely candidate, you have to read it in detail to make sure that you aren't wrong.
My point is that the Google Code Jam is for more than experienced programmers who can quickly bash out simple algorithms - you have to be able to devise novel algorithms under time pressure and devise strategies based on the psycology of your fellow coders
You should say "easier to read then perl" because I don't think it's easier to read then ruby or php or even java
I think that Python is a lot easier to read than Ruby or Java. Ruby allows a lot of the same punctuation-based idioms that make Perl so difficult to read and Java is too verbose to be easy to read. Consider the Java version of hello world:
class HelloWorldApp { public static void main(String[] args) { System.out.println("Hello World!"); } }
To be fair to Nero, the complete name listed in the Start menu is "Nero - Burning Rom".
Of course you could argue that the name should be something like "Create CD". Unfortunately for Nero, that is exactly what the Start menu extry for Adaptec Easy CD Creator is.
In any case I think that it is a huge mistake to not include the word "CD" in the menu entry.
Re:A (hopefully) unbiased opinion on Perl v. Pytho
on
Python in a Nutshell
·
· Score: 1
>Overall, as a learner I found Perl's documentation to be the best, and as an advanced developer I find PHP's to be supreme, bar none. Python's is a disgrace, useful to neither beginning nor advanced users.
Yep, I've come to those exact same conclusions. Take a look at the "win32*" API docs hosted at ActiveState, for example win32net.DriveAdd()
I tried to find win32 bindings for Perl and couldn't. Could someone point them out to me please? I did notice that the only Perl COM solution seems to involve paying money. The Python solution is free.
Which might be why the Python win32 documentation isn't very strong: it is not done by the core Python team, nor by a company that charges you money, but by one really smart guy (Mark Hammond) who, in addition to doing work that pays the bills, develops and documents the Python win32 bindings pretty much by himself.
I think that the core Python library reference is excellent.
Cheers,
Brian
Re:A (hopefully) unbiased opinion on Perl v. Pytho
on
Python in a Nutshell
·
· Score: 1
Python 1.5.1 doesn't even come with any documentation by default; you have to install that separately
You are basing your experience on a Python release that is over five years old!
But Python's missing many of the goodies that make slow languages great. Macros.
Not having macros was a deliberate design decision. One of Python's key stengths is easy readability and adding the ability to add/modify language contructs might compromise this.
A good, scientifically-oriented numerical facility.
What facilities are you looking for, in particular? Python has a double-precision floating point type, unbounded integers and complex numbers as part of the core language. Are you looking for something else, like rationals?
Or is it libraries that you are looking for? If so, here are the docs for Numeric
Python is odd: it has both models. Underneath the model is prototype-based. Kludged on top of this is a class model. What the heck?
I don't understand your complaint. Can you discuss it in terms of semantics e.g. I would expect method to get resolved like this but they get resolved like that....
I get the feeling that what happened was Python started as a procedural Perl-style language, then hacked on a prototype model, then due to, I dunno, misinformed "customer complaints", hacked on another class-based model on top of that. What a mess.
Python is fundamentally object-oriented. Even integers are treated as objects.
Cheers, Brian
Re:A (hopefully) unbiased opinion on Perl v. Pytho
on
Python in a Nutshell
·
· Score: 1
The biggest problem with Python, IMHO, is the online documentation.
I partially disagree: I think that the library reference is excellent but that the language reference is terrible for beginners. I would suggest that you find a tutorial to learn the language and use the official documentation to learn the libraries.
% gunzip Python-2.2.2.tgz % tar -xf Python-2.2.2.tar % cd Python-2.2.2 %./configure % make install
Cheers, Brian
Re:A (hopefully) unbiased opinion on Perl v. Pytho
on
Python in a Nutshell
·
· Score: 1
you can't implement it [Python] in itself
Actually, there are people who are in the process of implementing Python in Python right now. The project is called "Minimal Python".
it has no strong typing (even Perl's use strict is ridiculously better)
Actually, Python is strongly typed.
an OO system with no support for data hiding
Most Python developers consider data hiding a misfeature. Unless you are operating in a secure environment or can enumerate every possible usage case for your code and prove that no one could ever need access to a private method, you should allow access. Just document the fact that the implementation could change in the future.
If anyone is interested I'm willing to share my stories about being burned by C++ library developers who were too fond of data hiding:-)
and it's too dog-slow to be used for real CS applications
I'm not sure what "real CS applications are" but Python is frequently used for scientific computing.
Slashdot Mirror
Agreed, we have massive sites serving millions of requests a day using Open Source relational databases and yet it seems everyone wants to use NoSQL because it's the hip new thing.
Naturally I start thinking of this: http://xtranormal.com/watch/6995033
A million requests per day translates to 11.5 requests per second. That's a pretty trivial amount of traffic. A massive site like Facebook is probably serving about 4 orders-of-magnitude more requests that that.
Is that a bad thing not to want to worry about the infrastructure?
Yes, it's a VERY VERY bad thing if your business and it's reputation relies on said infrastructure.
I agree. Which is why I would assume that your company manages the following infrastructure internally:
Compare that with the hottest new Samsung Galaxy S which sometimes fails at receiving a basic phone call..
Did you read your own link? Every poster who followed-up said that the problem was resolved when they realized that they had to slide (rather than simply touch) the answer button in order to answer calls.
What "lavish" benefits are you talking about? Lunches? Lunches pay for themselves because they all of a sudden take 25-30 minutes instead of an hour or more. At $100+ (sometimes way more than that) per hour it just makes sense for a company to pay for lunches. Buses to and from work? Umm. OK, I'll give you that (even though Microsoft also has buses). On-site gym that hardly anyone goes to? What else?
Plus the usual as far as medical, dental, stock options, etc. And probably a bunch of other stuff that I don't know about.
Google is actually pretty bare bones on the inside.
Compared to?
If you search for tank man directly then you get thousands of results including 4 pictures of him in front of the tanks at the top of the results page.
I'm not sure what your point is here.
The N900 has 45% more volume than the N1 but also has a few more features (but mostly they seems pretty similar). Is that really surprising?
In other news my desktop has a bigger HD than my laptop, my apartment has better rain protection than my tent, ...
Cheers,
Brian
The real scandal here is that they can do this at all. Why should corporations be allowed to negotiate taxes? Can individuals do that? "Hello, state! I am thinking of moving to your state and being productively employed! And some of the money I make will be spent in the local economy! Promise! I will do so if you cut me some slack on property taxes and state income taxes!"
You'd get laughed at. Scale? so what, could 200 independent single individuals do the same? Nope. But a corporation can.
You know, I'm not sure that a very wealthy individual could not negotiate in this fashion. What if Bill Gates said that he was thinking about moving his family to NC, wanted to build a $500,000,000 mansion there, and would permanently employ 200 people (gardeners, helicopter pilots, etc.) but the property taxes in NC are too high. You don't think that the government would consider cutting him a break?
So do I. And I don't do executive management for a living. It's an amusing situation, yes, but an unfair one. You don't expect an auto mechanic do also do the accounting and sales for the dealership; neither should you expect the comptroller to do a valve job.
I think that you are being overly generous. Steve Balmer was only called on to do something that an end user of his (consumer!) product would want to do. A better analogy would be that you shouldn't expect an executive at a car manufacturer to be able to drive the company's cars. But of course you would.
I think we should be allowed to mod the stories as well as the comments. This way we could get rid of both the dupes and the trolls like the current story.
I think that is a terrible idea - I view slashdot at +5 and I want to see more than just advertisements.
Actually, if we could comment on the advertisements, maybe it would be ok...
Java is a strongly typed language and this is not going to change ("dynamical typing" doesn't mean anything by the way).
It is interesting that you think that "dynamic typing is meaningless". It is widely used with the definition that typing is resolved at runtime (rather than at compile-time). How could you classify the following three cases?
Perl: 5 + "5" (result: 10)
Python: 5 + "5" (type exception at runtime)
Java: 5 + "5" (type exeption at compile-time)
Based on this (and other) behavior, many people would say that Perl is weakly-typed while Python and Java are strongly-typed. Python's typing is enforced at runtime while Java's typing is enforced at compile-time. Dynamic vs. static typing is a convenient set of terminology for expressing this difference.
...gravity continues it's matter-attraction strategy.
How can you call Apple a hardware company? Because they put everything in a well designed box? All the components are 3rd party... Apple doesn't make processors, Apple doesn't make memory, Apple doesn't make harddrives or video cards or sound cards. They buy them from hardware companies, put them in a shiny box and then run *their software* on it.
Without the seven last words, this could also apply to Dell. I'm not disagreeing with your point, just observing that other companies seem to make money selling hardware using only 3rd party components.
NetFlix already has a healthy base of customers, and anyone interested in such a service already knows about them. What does Amazon have to bring to the table, other than not crushing them like a bug?
A stronger brand. Despite what you say, I doubt that most interested comsumers are aware of online DVD rental.
Empirical evidence has shown, time and time again, that low level languages like C and Forth produce more efficient, faster, and easier to maintain code than today's so called high level languages.
If that has been shown time and time again then providing references should be easy. I'm especially interested in convincing evidence that C and Forth are easier to maintain than "so called" high level languages.
There is a large difference. Gravity is accepted by probably about 99.9999999999% of the world. Evolution is accepted by a minority of the world.
As other will certainly point out, science is not a democracy. So I'll just dispute your claim that the theory of gravitation is universily accepted. I'm sure that most people would accept the observation that they tend not to float away from the Earth but do you really believe that everyone knows (and believes) that the phenomena is caused by attraction between masses?
So, you're proud of a system that requires somebody to have a large amount of money in order that they can defend themselves?
Hey, if I were a lawyer, I'd be proud of such a system too.
Yes, Java's GC would notice that nothing is referring to them and remove the objects. This unlike a simple reference counting gc such as python's which would not notice this. Java's GC can even relocate memory on the fly to minimize page misses and avoid memory fragmentation.
Python actually uses a generational garbage collection system that can break cycles to reclaim unused objects. It also performs certain optimizations to avoid unnecessary memory allocations and deallocations.
In Python, reference counting is a combination of a historical artifact and a performance optimization.
To me, it's more important to measure the potential of a programmer, more than the experience of the programmer. Creating new ideas and systems does not require experience alone -- it requires a great deal of potential. Simply breaking and fixing code in a realtime setting is good for the day to day stuff (and that is valuable for Google), but when a customer or boss asks to you build something new and difficult, many experienced programmers answer with, "that can't be done" when less experienced programmers will try to do it, with limited success.
I did the Google Code Jam last year and found that they required a lot more than basic coding skills.
One question involved analysing sequences of resource lock and unlocks, represented by upper- and lowercase characters in strings e.g. "ABba", "BAba" and "CADadBbc". Given such strings, you had to write a program to identify potential deadlock conditions e.g. the first two strings can deadlock at position 1 because the first string will hold a lock for A while waiting for B and the second string will hold a lock for B while waiting for A. Of course the real tests were more difficult than this. I think that less than 10 people out of 250 got this one.
Another question involved graph searching but with the cost of node traversal changing with time according to a pseudorandom algorithm. My solution was only twenty or thirty lines of code but it took a while to think how to do a time-based graph search.
The challenge phase is fairly challenging (excuse the pun) as well because you only have thirty minutes to make all of your challenges and you lose points if you are incorrect i.e. you claim that a certain valid input combination will cause the program to fail and it doesn't. Thirty minutes is not a lot of time to attack the 30 potentially flawed solutions that you are looking at. You have to pick a set of likely flaws (based on your own design) and quickly scan each set of code for it. Then, when you find a likely candidate, you have to read it in detail to make sure that you aren't wrong.
My point is that the Google Code Jam is for more than experienced programmers who can quickly bash out simple algorithms - you have to be able to devise novel algorithms under time pressure and devise strategies based on the psycology of your fellow coders
I think that Python is a lot easier to read than Ruby or Java. Ruby allows a lot of the same punctuation-based idioms that make Perl so difficult to read and Java is too verbose to be easy to read. Consider the Java version of hello world:
Cheers,
Brian
To be fair to Nero, the complete name listed in the Start menu is "Nero - Burning Rom".
Of course you could argue that the name should be something like "Create CD". Unfortunately for Nero, that is exactly what the Start menu extry for Adaptec Easy CD Creator is.
In any case I think that it is a huge mistake to not include the word "CD" in the menu entry.
>Overall, as a learner I found Perl's documentation to be the best, and as an advanced developer I find PHP's to be supreme, bar none. Python's is a disgrace, useful to neither beginning nor advanced users.
Yep, I've come to those exact same conclusions. Take a look at the "win32*" API docs hosted at ActiveState, for example win32net.DriveAdd()
I tried to find win32 bindings for Perl and couldn't. Could someone point them out to me please? I did notice that the only Perl COM solution seems to involve paying money. The Python solution is free.
Which might be why the Python win32 documentation isn't very strong: it is not done by the core Python team, nor by a company that charges you money, but by one really smart guy (Mark Hammond) who, in addition to doing work that pays the bills, develops and documents the Python win32 bindings pretty much by himself.
I think that the core Python library reference is excellent.
Cheers, Brian
Python 1.5.1 doesn't even come with any documentation by default; you have to install that separately
You are basing your experience on a Python release that is over five years old!
Cheers,
Brian
But Python's missing many of the goodies that make slow languages great. Macros.
Not having macros was a deliberate design decision. One of Python's key stengths is easy readability and adding the ability to add/modify language contructs might compromise this.
A good, scientifically-oriented numerical facility.
What facilities are you looking for, in particular? Python has a double-precision floating point type, unbounded integers and complex numbers as part of the core language. Are you looking for something else, like rationals?
Or is it libraries that you are looking for? If so, here are the docs for
Numeric
Python is odd: it has both models. Underneath the model is prototype-based. Kludged on top of this is a class model. What the heck?
I don't understand your complaint. Can you discuss it in terms of semantics e.g. I would expect method to get resolved like this but they get resolved like that....
I get the feeling that what happened was Python started as a procedural Perl-style language, then hacked on a prototype model, then due to, I dunno, misinformed "customer complaints", hacked on another class-based model on top of that. What a mess.
Python is fundamentally object-oriented. Even integers are treated as objects.
Cheers,
Brian
I partially disagree: I think that the library reference is excellent but that the language reference is terrible for beginners. I would suggest that you find a tutorial to learn the language and use the official documentation to learn the libraries.
Take a look at:
http://www.python.org/doc/Intros.html
abandoning it quickly because the documentation (and installation process) were so opaque
What part of the installation process is opaque?
The Windows installer is a GUI installer that only asks you the install location.
The UNIX install process uses the usual autoconf process. On Linux, for example, all you have to do is:Cheers,
Brian
you can't implement it [Python] in itself
:-)
Actually, there are people who are in the process of implementing Python in Python right now. The project is called "Minimal Python".
it has no strong typing (even Perl's use strict is ridiculously better)
Actually, Python is strongly typed.
an OO system with no support for data hiding
Most Python developers consider data hiding a misfeature. Unless you are operating in a secure environment or can enumerate every possible usage case for your code and prove that no one could ever need access to a private method, you should allow access. Just document the fact that the implementation could change in the future.
If anyone is interested I'm willing to share my stories about being burned by C++ library developers who were too fond of data hiding
and it's too dog-slow to be used for real CS applications
I'm not sure what "real CS applications are" but Python is frequently used for scientific computing.
Cheers,
Brian