The Gang of Four's book may be the most influential programming book, but it's probably done more harm than good. The patterns it describes fall into two categories: so painfully obvious that any programmer with more than a couple of months of experience should have invented them independently, or so obscure that you'll probably never find a real use for them. Unfortunately, once people have read it, they end up designing insanely overcomplicated systems that abuse patterns to death.
If you really want to understand design patterns, read some of Christopher Alexander's work, not the third-rate derivatives.
The statement of "goto is bad" isn't about the assembly language where there are no other choices, it's about using it in languages where there are other choices.
The original statement was 'Go To Statement Considered Harmful'. The essay - which really kick-started the structured programming revolution - was not about using goto, it was about providing it in the language at all. If you're using C, for example, then you don't have a go-to statement in the original sense. You have a very limited version that only allows flow control within a function scope. The go-to statements that Djikstra was complaining about were ones that made things like variable scoping impossible. They allowed jumping anywhere inside a program, without a return mechanism. This made reasoning about the code insanely hard. If you learned to program after the '70s, then you've probably never used a language (other than an assembly language) where this kind of thing was possible. The closest that we have now is setjmp()/longjmp().
Bullshit. I have a PhD in computer science, and I wouldn't recommend TAOCP to anyone. Sure, it covers the material, but it covers it amazingly badly. Knuth manages to take simple concepts and make them incomprehensible. Read pretty much any textbook on theoretical computer science other than TAOCP and you'll learn a lot more.
Couldn't agree more. TAOCP seems more like a book aimed at telling you how clever Knuth is than a book aimed at teaching. The idea of starting with a low-level programming language and building high-level constructs on top of it is great, but the execution of TAOCP is very poor. Part of the reason for this is that Knuth is a great theoretician and a terrible programmer. TeX should show you that TAOCP is not a book about programming - who, in the '70s, would invent a new programming language in which structured programming is impossible? Only Knuth. It combines the worst aspects of Lisp and an assembly language.
But you wouldn't know how to write, only how to type.
So, I'd only have the useful skill, not the useless one? I've now had 4 books published. Guess how many I wrote with a pen.
Word processors also fix your mistakes (automatically) which could lead to an overinflated grade
They didn't back then. I was using ClarisWorks 1.0. It did have a spell checker, but that's as far as it went. Now, autocorrect is the first thing I turn off with a computer. My spelling was atrocious when I was at school. Now, I'm the person people ask how to spell words. Why? Because I type a lot and my text editor red-underlines words that are not in the dictionary. If I type something wrongly, I have to go back and correct it. With a pen, if I misspelled a word then I didn't know until after I'd handed it in and lost marks for poor spelling. I never had the instant feedback that ingrained the correct spellings in my mind.
A piece of chalk and a board works just as well as a touchscreen and is more robust.
Really? My teachers at school had to book a TV and a VCR when they wanted to show us a video. With a touchscreen, they could just drag the relevant clip onto the board, run it, pause it in the middle, zoom in on relevant sections, and discuss it with the class a lot more easily.
We don't need more tech in a class room. We need more qualified teachers.
We don't need accurate textbooks in a classroom, we need more qualified teachers. See? It's just as nonsensical when phrased like that too. Computers, pens, calculators, books - these are all tools. They must be used correctly to provide a benefit.
Depends on where your fingers are. Not sure where it comes from, but when I was growing up (in England, home of English), crossing your fingers behind your back meant that you were lying.
Depends on how it's being used. For example, my average mark for English essays went from C to A when I started being able to use a word processor instead of a pen. I was able to focus on the ideas, rather than the mechanics of using a ridiculously archaic writing device. I spent over a decade at school having to use a pen. Now, I type more in an average day than I write with a pen in an average year. I'd probably have done better at school if I'd been allowed to type from the start.
The main use of both Twitter and Facebook is to communicate with other people
No, the main use for these systems is to build large databases of connections between people and between people and things for data mining by encouraging communications to flow via a single choke point. As a side effect, they also make it trivial for a single entity to restrict and censor communication. The fact that they provide some communication is a side effect of their main purpose.
Computer science if it's the only option, but if your university offers something like software engineering or electronic engineering that would be more help. Modern cars contain a huge number of microprocessors, and understanding this side of things as well as you understand the mechanical parts of the engine will give you a huge advantage.
The "founders" lived in a land of farmers and small, independent businesses, where getting from town to town took a day, and from state to state took a week.
Judging by their policies, I think getting back to that is pretty high on the Tea Party's agenda...
The problem with handwriting recognition is twofold. First, it's actually a really hard problem (hell, humans have problem recognising my handwriting, what chance do machines have?). The second is that a pen is actually a pretty poor input device for text. I can type a lot faster than I can write with a pen, and the result is always consistent, while with a pen it can vary in quality depending on a whole variety of factors.
That said, the Newton had some amazing technology. The UI for copy and paste was beautiful. The drawing application, which let you sketch shapes and then turned them into vector primitives, was amazing to use. Newton Soup was an great idea, as was the agent.
There's one important thing about having the spammers win: you can appeal if you lose a case, you can't appeal if you win. If the judge had ruled against the spammers, then they could have appealed at the next court up and wasted more time and money. This judgement means that Spamhaus can appeal if they want (which, I presume, they won't), but the spammers have to accept the judgement. The point of an award like this is to say 'you are technically in the right as a point of law, but you shouldn't actually win, no go home.'
That's only if the court doesn't have jurisdiction over the offence, not if it doesn't have jurisdiction over one of the parties. If I stand in Germany and shoot someone across the border in France, then a French court can still prosecute me in absentia, but they will need to extradite me to carry out the sentence. In contrast, a British court would throw the case out, because things that happen in France are outside its jurisdiction.
Ah, no, this is not just normal WiFi, this is two-way WiFi. With old one-way WiFi, you had to send a packet, then eject the card, turn it around, and insert it the other way around to receive the reply. With this new two-way wireless magic your card can both send and receive! It's exciting and new!
In English, the term school is never used to mean university, for the reasons that I outlined. This is a peculiarity of American English, and of the US educational system (US universities have a lot more supervision than in the UK and are closer to schools than to [UK] universities in their approach to education).
Part of that is the lack of good metrics for measuring developer quality. We still use things like lines of code, which are absolutely appalling measures of productivity. I've had a couple of days where I've made things faster, less buggy, and added features just by simplifying some existing code and then slightly tweaking the simplified version. The original author wrote more code though, so he'd win the lines-of-code count metric.
This is why I prefer to work for small companies, or on small teams. In that sort of environment, the person in charge can more easily track what real contributions someone has made, and see who is the dead weight on a team.
If you write something that is grammatically incorrect, then it makes it harder for people to read it. The same is true of poor spelling, especially for non-native speakers for whom homophones may be difficult to follow. If your writing is bad, then it tells me one of the following:
You made a typo and didn't notice it. Everyone does this from time to time, and if it's just an occasional mistake then I'll usually ignore it.
You're writing about grammar and are therefore forced by Eris to make the most embarrassing mistake that you've made for a long time.
You are not a native English speaker.
You are too stupid to know how to write properly.
You do know how to write properly, but you think that saving you a few seconds by writing badly is worth more than saving your readers a few seconds each by writing well.
In most cases, it's one of the last two options. In short, it means that you're an asshat or an idiot. Either way, it's not worth my time to work out which.
Nope. Universities do not teach students, and they do not employ teachers. Universities employ lecturers, who present information to students and place them in an environment where they can learn. The students are supposed to teach themselves. This is the difference between a university and a school. You go to a school to be taught, you go to a university to assisted in learning. Failure to understand this difference is one of the biggest reasons why people drop out of university.
Were those meant to be hard C++ questions? Because C++ is one of the languages that I despise and avoid where possible, yet they all looked pretty trivial to me. Oh, and:
The real answer is in the event you are concerned for space you don't want the overhead of the vptr
The real answer is a lot more complex than that. Your vptr is going to be the first word and allocations are always aligned, meaning that if you access any fields in the structure, you're likely pulling the vptr into cache. For an array of small structures, the vptr can burn a lot of dcache, which is still a scarce resource on most systems. Additionally, most C++ compilers suck at code flow analysis across virtual calls (I've done some work on this to add speculative inlining, and others have done some devirtualisation stuff, but by and large the results are nowhere near what you get from just deleting the 'virtual'), meaning that not only do you lose inlining, you also lose all of the interprocedural analysis, so you miss out on things like function specialisation, and also on more subtle things, like code placement (i.e. placing functions that will be called together nearby so that they're in the same cache line).
Oh, and it gets even more complicated if you're writing a shared library boundary for x86, because you actually have less indirection with a virtual call than you do with a non-virtual call, so the virtual call will be faster. So, for large classes with large methods, adding virtual can actually speed things up. Although, given that they're large methods, not by very much because the cost of the call won't be significant.
In fact, the real answer is 'because you don't need any'. Unless you work in HPC or embedded systems development, if you're thinking about this level of micro-optimisation, then you're doing it wrong.
Hahahaha! C++ added a whole category of bad pointer errors. You see, in C++, a pointer cast can actually be a pointer arithmetic expression (i.e. the numerical value of the pointer changes as a result of the cast). In some cases, these casts can be implicit. Worse, sometimes the C++ type system is so paranoid that it forces you to add explicit casts in the safe case, so it hides when it's an unsafe cast. And then you spend a day wondering why the code passes all of the tests on one architecture but fails them on another.
A youngster would have a hard time finding a C programming job, too.
I'm going to be 30 next year, so I'm not sure if I qualify as a youngster anymore, but I have no problem at all finding C programming jobs. I periodically get unsolicited offers of jobs (which I'm not interested in) and contracting work (which I am). When I turn down the job offers, I'm frequently asked if I know anyone else who can write good C code. I do, but they're also all happily employed.
Slashdot Mirror
The Gang of Four's book may be the most influential programming book, but it's probably done more harm than good. The patterns it describes fall into two categories: so painfully obvious that any programmer with more than a couple of months of experience should have invented them independently, or so obscure that you'll probably never find a real use for them. Unfortunately, once people have read it, they end up designing insanely overcomplicated systems that abuse patterns to death.
If you really want to understand design patterns, read some of Christopher Alexander's work, not the third-rate derivatives.
The statement of "goto is bad" isn't about the assembly language where there are no other choices, it's about using it in languages where there are other choices.
The original statement was 'Go To Statement Considered Harmful'. The essay - which really kick-started the structured programming revolution - was not about using goto, it was about providing it in the language at all. If you're using C, for example, then you don't have a go-to statement in the original sense. You have a very limited version that only allows flow control within a function scope. The go-to statements that Djikstra was complaining about were ones that made things like variable scoping impossible. They allowed jumping anywhere inside a program, without a return mechanism. This made reasoning about the code insanely hard. If you learned to program after the '70s, then you've probably never used a language (other than an assembly language) where this kind of thing was possible. The closest that we have now is setjmp()/longjmp().
Bullshit. I have a PhD in computer science, and I wouldn't recommend TAOCP to anyone. Sure, it covers the material, but it covers it amazingly badly. Knuth manages to take simple concepts and make them incomprehensible. Read pretty much any textbook on theoretical computer science other than TAOCP and you'll learn a lot more.
Couldn't agree more. TAOCP seems more like a book aimed at telling you how clever Knuth is than a book aimed at teaching. The idea of starting with a low-level programming language and building high-level constructs on top of it is great, but the execution of TAOCP is very poor. Part of the reason for this is that Knuth is a great theoretician and a terrible programmer. TeX should show you that TAOCP is not a book about programming - who, in the '70s, would invent a new programming language in which structured programming is impossible? Only Knuth. It combines the worst aspects of Lisp and an assembly language.
But you wouldn't know how to write, only how to type.
So, I'd only have the useful skill, not the useless one? I've now had 4 books published. Guess how many I wrote with a pen.
Word processors also fix your mistakes (automatically) which could lead to an overinflated grade
They didn't back then. I was using ClarisWorks 1.0. It did have a spell checker, but that's as far as it went. Now, autocorrect is the first thing I turn off with a computer. My spelling was atrocious when I was at school. Now, I'm the person people ask how to spell words. Why? Because I type a lot and my text editor red-underlines words that are not in the dictionary. If I type something wrongly, I have to go back and correct it. With a pen, if I misspelled a word then I didn't know until after I'd handed it in and lost marks for poor spelling. I never had the instant feedback that ingrained the correct spellings in my mind.
A piece of chalk and a board works just as well as a touchscreen and is more robust.
Really? My teachers at school had to book a TV and a VCR when they wanted to show us a video. With a touchscreen, they could just drag the relevant clip onto the board, run it, pause it in the middle, zoom in on relevant sections, and discuss it with the class a lot more easily.
We don't need more tech in a class room. We need more qualified teachers.
We don't need accurate textbooks in a classroom, we need more qualified teachers. See? It's just as nonsensical when phrased like that too. Computers, pens, calculators, books - these are all tools. They must be used correctly to provide a benefit.
Depends on where your fingers are. Not sure where it comes from, but when I was growing up (in England, home of English), crossing your fingers behind your back meant that you were lying.
Depends on how it's being used. For example, my average mark for English essays went from C to A when I started being able to use a word processor instead of a pen. I was able to focus on the ideas, rather than the mechanics of using a ridiculously archaic writing device. I spent over a decade at school having to use a pen. Now, I type more in an average day than I write with a pen in an average year. I'd probably have done better at school if I'd been allowed to type from the start.
The main use of both Twitter and Facebook is to communicate with other people
No, the main use for these systems is to build large databases of connections between people and between people and things for data mining by encouraging communications to flow via a single choke point. As a side effect, they also make it trivial for a single entity to restrict and censor communication. The fact that they provide some communication is a side effect of their main purpose.
Computer science if it's the only option, but if your university offers something like software engineering or electronic engineering that would be more help. Modern cars contain a huge number of microprocessors, and understanding this side of things as well as you understand the mechanical parts of the engine will give you a huge advantage.
The "founders" lived in a land of farmers and small, independent businesses, where getting from town to town took a day, and from state to state took a week.
Judging by their policies, I think getting back to that is pretty high on the Tea Party's agenda...
The problem with handwriting recognition is twofold. First, it's actually a really hard problem (hell, humans have problem recognising my handwriting, what chance do machines have?). The second is that a pen is actually a pretty poor input device for text. I can type a lot faster than I can write with a pen, and the result is always consistent, while with a pen it can vary in quality depending on a whole variety of factors.
That said, the Newton had some amazing technology. The UI for copy and paste was beautiful. The drawing application, which let you sketch shapes and then turned them into vector primitives, was amazing to use. Newton Soup was an great idea, as was the agent.
There's one important thing about having the spammers win: you can appeal if you lose a case, you can't appeal if you win. If the judge had ruled against the spammers, then they could have appealed at the next court up and wasted more time and money. This judgement means that Spamhaus can appeal if they want (which, I presume, they won't), but the spammers have to accept the judgement. The point of an award like this is to say 'you are technically in the right as a point of law, but you shouldn't actually win, no go home.'
That's only if the court doesn't have jurisdiction over the offence, not if it doesn't have jurisdiction over one of the parties. If I stand in Germany and shoot someone across the border in France, then a French court can still prosecute me in absentia, but they will need to extradite me to carry out the sentence. In contrast, a British court would throw the case out, because things that happen in France are outside its jurisdiction.
Ah, no, this is not just normal WiFi, this is two-way WiFi. With old one-way WiFi, you had to send a packet, then eject the card, turn it around, and insert it the other way around to receive the reply. With this new two-way wireless magic your card can both send and receive! It's exciting and new!
In English, the term school is never used to mean university, for the reasons that I outlined. This is a peculiarity of American English, and of the US educational system (US universities have a lot more supervision than in the UK and are closer to schools than to [UK] universities in their approach to education).
Part of that is the lack of good metrics for measuring developer quality. We still use things like lines of code, which are absolutely appalling measures of productivity. I've had a couple of days where I've made things faster, less buggy, and added features just by simplifying some existing code and then slightly tweaking the simplified version. The original author wrote more code though, so he'd win the lines-of-code count metric.
This is why I prefer to work for small companies, or on small teams. In that sort of environment, the person in charge can more easily track what real contributions someone has made, and see who is the dead weight on a team.
In most cases, it's one of the last two options. In short, it means that you're an asshat or an idiot. Either way, it's not worth my time to work out which.
Both professions could benefit from a reliable way of differentiating the good from the bad.
Nope. Universities do not teach students, and they do not employ teachers. Universities employ lecturers, who present information to students and place them in an environment where they can learn. The students are supposed to teach themselves. This is the difference between a university and a school. You go to a school to be taught, you go to a university to assisted in learning. Failure to understand this difference is one of the biggest reasons why people drop out of university.
The real answer is in the event you are concerned for space you don't want the overhead of the vptr
The real answer is a lot more complex than that. Your vptr is going to be the first word and allocations are always aligned, meaning that if you access any fields in the structure, you're likely pulling the vptr into cache. For an array of small structures, the vptr can burn a lot of dcache, which is still a scarce resource on most systems. Additionally, most C++ compilers suck at code flow analysis across virtual calls (I've done some work on this to add speculative inlining, and others have done some devirtualisation stuff, but by and large the results are nowhere near what you get from just deleting the 'virtual'), meaning that not only do you lose inlining, you also lose all of the interprocedural analysis, so you miss out on things like function specialisation, and also on more subtle things, like code placement (i.e. placing functions that will be called together nearby so that they're in the same cache line).
Oh, and it gets even more complicated if you're writing a shared library boundary for x86, because you actually have less indirection with a virtual call than you do with a non-virtual call, so the virtual call will be faster. So, for large classes with large methods, adding virtual can actually speed things up. Although, given that they're large methods, not by very much because the cost of the call won't be significant.
In fact, the real answer is 'because you don't need any'. Unless you work in HPC or embedded systems development, if you're thinking about this level of micro-optimisation, then you're doing it wrong.
No, Muslims are not allowed to worship the Lords of COBOL.
I could code circles around the 30+ crowd when I was 15, and the 15 year olds I know now can outcode me on any language that came out after 2005
If that's really true, then it means that you suck as a programmer, or you know some exceptionally talented 15 year olds.
bad pointers. ...which C++ eliminated
Hahahaha! C++ added a whole category of bad pointer errors. You see, in C++, a pointer cast can actually be a pointer arithmetic expression (i.e. the numerical value of the pointer changes as a result of the cast). In some cases, these casts can be implicit. Worse, sometimes the C++ type system is so paranoid that it forces you to add explicit casts in the safe case, so it hides when it's an unsafe cast. And then you spend a day wondering why the code passes all of the tests on one architecture but fails them on another.
A youngster would have a hard time finding a C programming job, too.
I'm going to be 30 next year, so I'm not sure if I qualify as a youngster anymore, but I have no problem at all finding C programming jobs. I periodically get unsolicited offers of jobs (which I'm not interested in) and contracting work (which I am). When I turn down the job offers, I'm frequently asked if I know anyone else who can write good C code. I do, but they're also all happily employed.
Only in the USA. Note that this is a topic about a UK-based project...