I have a real problem with using Python for large projects simply because of the lack of static typing. Dynamic typing is great for scripts, but well-documented interfaces become the most important factor when a project scales beyond a single team. Static typing forces you to put important interface information in the code, rather than hoping someone added it to comments.
Disagree, strongly. In practice, dynamic typing tends to greatly enhance interface reuse, and the benefits of dynamic typing become greater as the project scales. The minimal gain of having a type label in the function definition is far outweighed by the other benefits of dynamic systems.
Point of reference: my current project is just under 300,000 lines of Python code, though a bit under 1/3 of that is auto-generated code. It spans 15+ clients, each of whom uses a generic core, subclassing and extending it as needed.
Thay haven't used sodiun chloride on the roads for 5 years now in the northern states.
It's been a lot longer in some areas. I went through driver's ed in Maine in 1991, and at that time they said something along the lines of "back when they used salt, it was a good idea to wash your car off if it got salt on it, but since most communities have switched to more environmentally friendly compounds it's not a big issue anymore.
I don't read it your way; it seems to me that he's saying "all of the research shows small-to-moderate correlation between media violence and aggressive behavior and lesser correlation between media violence and violent behavior."
That's not inconclusive or indeterminate
Now you're misreading me and/or conflating media violence and video game violence.
I said that it reports that 1. Violent video games have indeterminate (not sufficiently studied to have any accurate knowledge) effect 2. All media violence has a measurable effect that is very small compared to other risk factors.
The Surgeon General's 2001 report on Youth Violence held that the effect of video game violence on youth behavior was indeterminate; this was after analyzing the Anderson report cited above and finding it insufficient to statistically demonstrate any link to aggressive behavior (and also pointing out as the earlier parent did that it does not study violent behavior at all). The same study held that aggregate media violence (not just video games) represented at most a very small risk factor for violent/aggressive behavior relative to other factors.
The JAMA in April of 2004 found that "Consensus is lacking on whether video games with violent content fuel aggressive behavior in children and adolescents."
that violent games and movies lead to measurable physiological responses and increased aggressive behavior in children.
Children are a whole other ball of wax; they show a remarkable ability to conflate reality and fantasy in all kinds of areas. I'm certainly in favor of policies prohibiting children from buying certain titles for themselves.
Regardless of that fact, these studies aren't very meaningful. There are a large number of studies indicating that playing competitive sports, or even simply excercising, leads to increased aggressive behavior (and obviously physical responses)--indeed, the research indicates that any exciting activity increases aggession.
Indeed, there have been many studies that show that playing action-oriented non-violent videogames causes the same aggression response as playing violent games--indeed, Craig Anderson himself (the guy in your second link) published a paper upholding that conclusion in the mid-1990s. In each of his earlier works where he would compare two action-oriented games, one violent and one not, he concluded that both had the same effect on aggression response. It wasn't until he switched to comparing a violent action game to a placid puzzle game (I think it was Wolfenstein 3d compared to Myst) that he found any difference.
In other words, the research seems to indicate that playing a fast-paced basketball video game would induce aggressive behavior, while playing a slow turn-based violent game like Nethack (where you can sacrifice other people to pagan gods, consort with succubi, practice cannibalism, commit genocide on entire species, etc) would not.
The evidence is compelling enough that the American Academy of Pediatrics has made a policy statement
And the American Medical Association and US Surgeon General have made statements to the contrary.
It's hilarious that you attribute all these social gains to a disgusting video game.
I did no such thing. I said that it's possible that there's a net gain, but certainly didn't imply that all the improvements we've seen over the last 20 years are because of games. But tearing down straw men is fun, I guess.
I'm simply saying that if you look at the numbers, these games haven't shown any real detrimental effects on society. Certainly not enough to justify attacking them out of concern for public health rather than attacking automobiles, skiing, drinking alcohol, or hundreds of other activities. And a few studies have shown that they may in fact be a beneficial force.
Now don't get me wrong, I'm fine with policys preventing children from purchasing certain kinds of titles. But cries to ban them are yet more invasive prohibitionism on the part of people who think their personal foibles should apply to everyone around them.
You would be hard pressed to convince me that if a learned person replaced their intellectual persuits with television their IQ would go up.
It doesn't have to be all one or the other.
It could be that different learning tools all have diminishing returns, and mixing them is a good strategy. So maybe spending 100% of your learning time reading books is better than spending it watching TV, but mixing them 50/50 is even more effective than reading alone. Or that, say, spending some time going to plays, reading, watching movies, watching tv, playing video games, playing games, going to concerts, etc is better for your intellectual development than just doing a couple of those.
And it could easily be the case that new media have contributed to this intellectual diversity--they may not be the best single way to learn, but adding them to the mix helps out the learned as well as the average.
You may think it's just a game but the longer you entertain such thoughts the more likely you are to program yourself into believing that crap.
As violent games have become commonplace, violent crime rates have declined dramatically. Since GTA3 came out, women have made more strides toward pay equity, more positions of power, and there's been a decline in both domestic violence and rape.
All the evidence I've seen indicates that most people are, in fact, able to distinguish games from reality and there may even be a net benefit to society from "dangerous, immoral" games (acting out impulses in fantasies/games may make one less likely to act them out in reality).
And the parent's statement that "Beating up bitches and killing cops definitely helps on the IQ" was clearly intentionally using such language to show a counterintuitive contrast.
I switched in 1994, so my original reasons no longer apply:
1. I couldn't type my name in Microsoft Word. Every time I wrote "Sumner", it changed it to "Summer". 2. All the Windows network stacks crashed hard if I opened more than one connection at a time.
Yes. That's why I specified that they were an illegal monopoly.
Furthermore, AT&T broke up voulentarily. It wasn't court-mandiated, it was an AT&T inspired solution for settlement.
True. Though they came to this decision after a number of decisions in federal court began eroding their monopoly (Execunet, and it was apparent that Judge Greene was prepared to break them up--it wasn't until he ruled that "the testimony and the documentary evidence adduced by the government demonstrate that the Bell System has violated the antitrust laws in a number of ways over a lengthy period of time" that they broke down and offered a settlement.
You're entirely wrong. Groovy is another language that targets the JVM. Groovy is either compiled to Java bytecode (and from there, JIT compiled to machine code by the JVM), or it's compiled at runtime in a JIT fashion. There is no interpreter written in Java that runs Groovy code. Groovy is just another language that targets the JVM.
This is indeed a good way to integrate seamlessly with Java.
Jython takes the same approach with Python (Python source is compiled to Java bytecodes) to allow tight Java integration. It has a nice intereactive shell (dynamically compiles to Java bytecodes) so you can do dynamic experimentation/debugging of large Java programs. Pretty cool. http://www.jython.org/
Re:Ma Bell? Yo no entiendo
on
Ma Bell is Back
·
· Score: 2, Informative
AT&T's Bell System (aka Ma Bell) was the large telephone system that served the US until the 1980s.
It was ruled an illegal monopoly and broken up into many smaller regional companies (the so-called "Baby Bells"). SBC was one of the baby bells.
Unfortunately I know that there exists no language (it might be possible in some scripting languages) that has such a feature. Perhaps when I get around to finishing MODL2 (my own damn language, the second) I'll add in not just operator overloading, but operation definition
Most languages with real (not preprocessor) macro support (notably LISP, Scheme, Dylan) support this.
The functionality you mention has nothing to do with an IDE (which is about integrating the development chain) and has been readily available in most decent programming editors for over a decade.
* If I see a class name I don't recognize, I can control-click to go to the class definition. If the source is not attached, I at least get a view of all the method signatures.
If you hit ctrl-t in vim, you go to the class definition (or function definition, variable definition, etc). I can't remember the keybinding in emacs, but it's in the context menu or M-x follow-tag.
* If my code invokes a method I don't recognize, I can hover the mouse over the method and the Javadoc description of the method will pop up, telling me what the method does, what the arguments are for, and what the return value is.
K in vim brings up the man pages or info pages for the defined function. I use it with Python docs; Java programmers use it with Javadoc. Emacs has similar capability.
* If I'm navigating through someone else's class hierarchy, by selecting the class name and pressing Ctrl+T I can see all interfaces this class implements, and its superclasses from which it inherits methods. If I hit Ctrl+T again, I can see all classes that inherit from this class, and what classes inherit from them, etc.
Sounds like emacs' oobr or vim's cscope interface (which is used for more than just C in vim).
If I hot Ctrl+O, I can see all the methods callable from my current cursor position. I can also see all variables within scope.
This is the only point that vim doesn't do by default yet (there are 3rd party packages to give similar completions, and vim 7 that encompasses this and a more powerful Intellisense(tm)-type thing). Emacs has this feature, though.
The problem with that test is that MySQL uses functions that tell the kernel to flush writes to disk. These functions are supposed to block so that the program can't continue until the writes are done and the data is stored on the platter. On OS X, this is exactly what happens, and every time MySQL requests data be flushed, the thread doing the flushing has to wait until the data is on the platter (or at the very least, in the drive's write cache). On Linux, this function returns instantly, as Linux (apparently) assumes that hard drives and power supplies are infallible, and obviously if you're that concerned about your data, get a UPS.
This is not true. The only time Linux returns after an fsync (or fdatasync) without the data being physically on disk is when the device lies about the write being complete. MacOS also returns immediately in that case. The problem is not an OS problem, it's that a lot of IDE drives for the PC come shipped with write caching enabled.
then they should fsking static link in ALL dependencies then... don't rely on the library being there... make sure you've compiled your code so it has everything it needs.
Then you have to include all your object files with your binaries (or otherwise fulfill LGPL requirements), as well as anything else needed to re-link. That is, assuming that you use any LGPL'd libraries (like glibc).
Which is similar to why Google more or less wiped out Yahoo in searching. At the time, Yahoo was seen as the searching site that couldn't been beaten
Basic point is right, but this seems revisionist; Yahoo was mainly a categorized set of links (like dmoz.org) with a halfassed search engine tacked on as an afterthought. Indeed, at the outset it didn't even have a search engine--and even after adding one, it was never really a search engine back then. Altavista was the "preferred" search engine at the time (having replaced Lycos when Lycos decided to go in the portal direction).
But yeah, Google came on the scene and did just searching, and did it better than Altavista and Lycos at a time when they had lost the search focus and were adding relatively crappy portal features instead of improving search.
just hitting Esc if you're not in Normal mode followed by ZZ to save/quit is quite the useful sort of thing you'd expect from vim (or equivalent in emacs)
Of course with a modern vim (last 5 years) you could: 1. Select save-exit from the menu. 2. Select save from the menu, then select exit from the menu. 3. Select "exit" from the menu, then respond yes to the "do you want to save your work" prompt 4. Any of the above, using Meta-F to access the file menu and hitting S for save, x for exit, or v for save-exit.
No need to hit escape before any of those options, either.
Betamax was never better than VHS. It was theoretically superior on picture quality, though most side-by-side reviews at the time of real equipment gave the edge to VHS except for a few short periods of a couple of months when Beta would implement a picture advance before VHS. But Beta was vastly inferior on sound quality and on tape length.
Of course, whether or not Dvorak is any better than Qwerty is open to debate; I've seen at least one study showing that Qwerty's design (alternate hands to avoid typewriter jams) also turns out to be quite helpful to typing speed (alternating hands is faster than typing successive characters on the same hand), and that Dvorak typists with repetitive stress injury benefit by switching to Qwerty in exactly the same way that Qwerty RSI sufferers benefit by switching to Dvorak. The only studies I know of show major Dvorak benefits were done by the inventor and salesman of the Dvorak layout.
If you think Java's syntax is radically different from C's syntax
Syntax is one of the least important features of a language, and OP never said anything about them having different syntax.
OP wrote: "Try C, LISP, SmallTalk or Ruby and you start to feel that the language is helping you. Cobol & Java fell like they are getting in the way."
One can infer that he's in favor of strongly, dynamically typed languages with the exception of C (weakly, statically typed), and against strongly, statically typed languages (or at least those without type inference).
I can venture a guess with reasonable confidence that he'd like python and scheme and would strongly dislike Pascal and Ada.
For languages that fall outside of the spectrum he described (say, Tcl and VBscript, or Dylan and ML) I have no feel for what he'd like.
Those guesses have nothing to do with syntax.
Re:Why are we hiding from the police, daddy?
on
Vim 6.4 Released
·
· Score: 1
Is vi the most effective editor? I doubt it. By so much that you can actually win back the time you spent learning? I strongly doubt it. Remember, we're talking tenths of a second here.
I've often found it saved me a half hour or more over the readily available alternatives. It's just about the easiest install on a Windows box if you need to do a single easily scripted edit (munging field order and formatting, stuff like that).
It's also one of the easiest code browsers to install if you get dropped in somewhere where you need to navigate an unfamiliar source tree quickly (just ctags all the files and fire up vim), especially if there's more than just C/C++/Java/HTML/VB/ASP in the tree.
Re:gaim works for me, but loses ground from here
on
Linux Instant Messengers
·
· Score: 2, Interesting
I'm just curious, while I know most young kids out there in school and college use IM quite a lot, what about those of you out there in the 'earning a living' category? Do you IM very much?
Absolutely. I almost never use it outside of work. But it's very convenient at work, and every office I've been in has used some form of IM extensively. It's a hell of a lot easier to throw URLs and commands back and forth in IM than in email (let alone by voice).
Examples: Developer1: Do you know the syntax off the top of your head to make an existing column autoincrement in mysql? Developer2: alter table modify column type autoincrement
or
Developer: What's next most important for CLIENT? Project Manager: Finish up the issue you're working on first, then take a look at this: http://bugtracker-bugurl/?bugid or: Secretary: Reminder to all: company meeting is starting in 1 minute in the conference room
It also has the advantage of letting you keep in touch with peers who work elsewhere--I have friends who I know are familiar with certain commands or whatever, and for some packages it's a lot quicker for me to ask them (and vice-versa) "what's the flag to do X" than to dig through the docs.
Also, how are the young people of today going to react when they find they can't use a tool they are kinda of hooked on daily in the real work world?
Badly, as they should if the company they work for doesn't allow them a valuable communications tool.
I'm curious as to the particulars of your project. How many people are working on it? How long has it been around? Does your company employ short-term contractors?
The current project isn't really a C project, it's about 300,000 lines of Python (with an equivalent amount of template code) and maybe 500 lines of C; we have 5 full-time developers and 1 part-time (maybe 10 or so people have worked on it at some point). About a third of the code is machine-generated via tools we've written; half the remainder is generic library code used by many client implementations, and the other half is specific to client needs. I've been on it for a year, the project has been around for 4 years.
The previous project was about 50,000 lines of C, 15,000 lines of C++, and 30,000 lines of Java. Because of institutional inefficiency, it had as many as 15 people work on it at peak (maybe 20-25 total over the life of the project) for about 6 months. Normal staff over the project lifespan was 3 C developers, 4 Java developers, and 2 C++ developers.
It's still in production (large network server servicing millions of requests daily--it's over the 1 billion served mark in its lifetime). No short-term contractors were used. It's in static production now, with a few weeks of development each year if new features are needed--it's been in production for about 6 years (active development for the first couple), with a year of development lead-in before that.
It's essentially 7-8 seperate back-end processes (C, one C++ component) interfaced by a desktop UI for data entry and system management (Java), a reporting front-end (server-side Java web interface) and a number of logging and summarizing processes (C mainly, one small Perl component).
I myself was originally a big advocate of C, but I've come to the conclusion the average programmer is simply too stupid to avoid 1)memory leaks, 2) buffer overruns, and 3) my personal favorite - passing references to stack variables back to the calling frame
In my experience, such programmers are also bad at avoiding business and programming logic errors. You're far better off going with fewer skilled programmers. Not only will the bug rate decrease, but the productivity will be higher--and although the cost per programmer will be higher, the total amount spent on programmers will be lower for more work done.
Essentially, if memory leaks and buffer overruns represent a significant amount of debugging time for your staff then you're going to be woefully underproductive no matter what language you choose.
That said, shaving a few percent off is worth it, and higher level languages can give you much more dramatic productivity gains in other arenas (depending on the task).
Slashdot Mirror
You're very clever, young man, very clever. But it's turtles all the way down.
http://en.wikipedia.org/wiki/Turtles_all_the_way_
I have a real problem with using Python for large projects simply because of the lack of static typing. Dynamic typing is great for scripts, but well-documented interfaces become the most important factor when a project scales beyond a single team. Static typing forces you to put important interface information in the code, rather than hoping someone added it to comments.
Disagree, strongly. In practice, dynamic typing tends to greatly enhance interface reuse, and the benefits of dynamic typing become greater as the project scales. The minimal gain of having a type label in the function definition is far outweighed by the other benefits of dynamic systems.
Point of reference: my current project is just under 300,000 lines of Python code, though a bit under 1/3 of that is auto-generated code. It spans 15+ clients, each of whom uses a generic core, subclassing and extending it as needed.
Thay haven't used sodiun chloride on the roads for 5 years now in the northern states.
It's been a lot longer in some areas. I went through driver's ed in Maine in 1991, and at that time they said something along the lines of "back when they used salt, it was a good idea to wash your car off if it got salt on it, but since most communities have switched to more environmentally friendly compounds it's not a big issue anymore.
I don't read it your way; it seems to me that he's saying "all of the research shows small-to-moderate correlation between media violence and aggressive behavior and lesser correlation between media violence and violent behavior."
That's not inconclusive or indeterminate
Now you're misreading me and/or conflating media violence and video game violence.
I said that it reports that
1. Violent video games have indeterminate (not sufficiently studied to have any accurate knowledge) effect
2. All media violence has a measurable effect that is very small compared to other risk factors.
Which seems to agree with your reading here.
The Surgeon General's 2001 report on Youth Violence held that the effect of video game violence on youth behavior was indeterminate; this was after analyzing the Anderson report cited above and finding it insufficient to statistically demonstrate any link to aggressive behavior (and also pointing out as the earlier parent did that it does not study violent behavior at all). The same study held that aggregate media violence (not just video games) represented at most a very small risk factor for violent/aggressive behavior relative to other factors.
The JAMA in April of 2004 found that "Consensus is lacking on whether video games with violent content fuel aggressive behavior in children and adolescents."
Yes, you're right. Please disregard my parent post.
that violent games and movies lead to measurable physiological responses and increased aggressive behavior in children.
Children are a whole other ball of wax; they show a remarkable ability to conflate reality and fantasy in all kinds of areas. I'm certainly in favor of policies prohibiting children from buying certain titles for themselves.
Regardless of that fact, these studies aren't very meaningful. There are a large number of studies indicating that playing competitive sports, or even simply excercising, leads to increased aggressive behavior (and obviously physical responses)--indeed, the research indicates that any exciting activity increases aggession.
Indeed, there have been many studies that show that playing action-oriented non-violent videogames causes the same aggression response as playing violent games--indeed, Craig Anderson himself (the guy in your second link) published a paper upholding that conclusion in the mid-1990s. In each of his earlier works where he would compare two action-oriented games, one violent and one not, he concluded that both had the same effect on aggression response. It wasn't until he switched to comparing a violent action game to a placid puzzle game (I think it was Wolfenstein 3d compared to Myst) that he found any difference.
In other words, the research seems to indicate that playing a fast-paced basketball video game would induce aggressive behavior, while playing a slow turn-based violent game like Nethack (where you can sacrifice other people to pagan gods, consort with succubi, practice cannibalism, commit genocide on entire species, etc) would not.
The evidence is compelling enough that the American Academy of Pediatrics has made a policy statement
And the American Medical Association and US Surgeon General have made statements to the contrary.
It's hilarious that you attribute all these social gains to a disgusting video game.
I did no such thing. I said that it's possible that there's a net gain, but certainly didn't imply that all the improvements we've seen over the last 20 years are because of games. But tearing down straw men is fun, I guess.
I'm simply saying that if you look at the numbers, these games haven't shown any real detrimental effects on society. Certainly not enough to justify attacking them out of concern for public health rather than attacking automobiles, skiing, drinking alcohol, or hundreds of other activities. And a few studies have shown that they may in fact be a beneficial force.
Now don't get me wrong, I'm fine with policys preventing children from purchasing certain kinds of titles. But cries to ban them are yet more invasive prohibitionism on the part of people who think their personal foibles should apply to everyone around them.
You would be hard pressed to convince me that if a learned person replaced their intellectual persuits with television their IQ would go up.
It doesn't have to be all one or the other.
It could be that different learning tools all have diminishing returns, and mixing them is a good strategy. So maybe spending 100% of your learning time reading books is better than spending it watching TV, but mixing them 50/50 is even more effective than reading alone. Or that, say, spending some time going to plays, reading, watching movies, watching tv, playing video games, playing games, going to concerts, etc is better for your intellectual development than just doing a couple of those.
And it could easily be the case that new media have contributed to this intellectual diversity--they may not be the best single way to learn, but adding them to the mix helps out the learned as well as the average.
You may think it's just a game but the longer you entertain such thoughts the more likely you are to program yourself into believing that crap.
As violent games have become commonplace, violent crime rates have declined dramatically. Since GTA3 came out, women have made more strides toward pay equity, more positions of power, and there's been a decline in both domestic violence and rape.
All the evidence I've seen indicates that most people are, in fact, able to distinguish games from reality and there may even be a net benefit to society from "dangerous, immoral" games (acting out impulses in fantasies/games may make one less likely to act them out in reality).
And the parent's statement that "Beating up bitches and killing cops definitely helps on the IQ" was clearly intentionally using such language to show a counterintuitive contrast.
I switched in 1994, so my original reasons no longer apply:
1. I couldn't type my name in Microsoft Word. Every time I wrote "Sumner", it changed it to "Summer".
2. All the Windows network stacks crashed hard if I opened more than one connection at a time.
Monopolies aren't illegal in the USA.
Yes. That's why I specified that they were an illegal monopoly.
Furthermore, AT&T broke up voulentarily. It wasn't court-mandiated, it was an AT&T inspired solution for settlement.
True. Though they came to this decision after a number of decisions in federal court began eroding their monopoly (Execunet, and it was apparent that Judge Greene was prepared to break them up--it wasn't until he ruled that "the testimony and the documentary evidence adduced by the government demonstrate that the Bell System has violated the antitrust laws in a number of ways over a lengthy period of time" that they broke down and offered a settlement.
You're entirely wrong. Groovy is another language that targets the JVM. Groovy is either compiled to Java bytecode (and from there, JIT compiled to machine code by the JVM), or it's compiled at runtime in a JIT fashion. There is no interpreter written in Java that runs Groovy code. Groovy is just another language that targets the JVM.
This is indeed a good way to integrate seamlessly with Java.
Jython takes the same approach with Python (Python source is compiled to Java bytecodes) to allow tight Java integration. It has a nice intereactive shell (dynamically compiles to Java bytecodes) so you can do dynamic experimentation/debugging of large Java programs. Pretty cool.
http://www.jython.org/
AT&T's Bell System (aka Ma Bell) was the large telephone system that served the US until the 1980s.
It was ruled an illegal monopoly and broken up into many smaller regional companies (the so-called "Baby Bells"). SBC was one of the baby bells.
Unfortunately I know that there exists no language (it might be possible in some scripting languages) that has such a feature. Perhaps when I get around to finishing MODL2 (my own damn language, the second) I'll add in not just operator overloading, but operation definition
Most languages with real (not preprocessor) macro support (notably LISP, Scheme, Dylan) support this.
The functionality you mention has nothing to do with an IDE (which is about integrating the development chain) and has been readily available in most decent programming editors for over a decade.
* If I see a class name I don't recognize, I can control-click to go to the class definition. If the source is not attached, I at least get a view of all the method signatures.
If you hit ctrl-t in vim, you go to the class definition (or function definition, variable definition, etc). I can't remember the keybinding in emacs, but it's in the context menu or M-x follow-tag.
* If my code invokes a method I don't recognize, I can hover the mouse over the method and the Javadoc description of the method will pop up, telling me what the method does, what the arguments are for, and what the return value is.
K in vim brings up the man pages or info pages for the defined function. I use it with Python docs; Java programmers use it with Javadoc. Emacs has similar capability.
* If I'm navigating through someone else's class hierarchy, by selecting the class name and pressing Ctrl+T I can see all interfaces this class implements, and its superclasses from which it inherits methods. If I hit Ctrl+T again, I can see all classes that inherit from this class, and what classes inherit from them, etc.
Sounds like emacs' oobr or vim's cscope interface (which is used for more than just C in vim).
If I hot Ctrl+O, I can see all the methods callable from my current cursor position. I can also see all variables within scope.
This is the only point that vim doesn't do by default yet (there are 3rd party packages to give similar completions, and vim 7 that encompasses this and a more powerful Intellisense(tm)-type thing). Emacs has this feature, though.
The problem with that test is that MySQL uses functions that tell the kernel to flush writes to disk. These functions are supposed to block so that the program can't continue until the writes are done and the data is stored on the platter. On OS X, this is exactly what happens, and every time MySQL requests data be flushed, the thread doing the flushing has to wait until the data is on the platter (or at the very least, in the drive's write cache). On Linux, this function returns instantly, as Linux (apparently) assumes that hard drives and power supplies are infallible, and obviously if you're that concerned about your data, get a UPS.
This is not true. The only time Linux returns after an fsync (or fdatasync) without the data being physically on disk is when the device lies about the write being complete. MacOS also returns immediately in that case. The problem is not an OS problem, it's that a lot of IDE drives for the PC come shipped with write caching enabled.
then they should fsking static link in ALL dependencies then... don't rely on the library being there... make sure you've compiled your code so it has everything it needs.
Then you have to include all your object files with your binaries (or otherwise fulfill LGPL requirements), as well as anything else needed to re-link. That is, assuming that you use any LGPL'd libraries (like glibc).
Which is similar to why Google more or less wiped out Yahoo in searching. At the time, Yahoo was seen as the searching site that couldn't been beaten
Basic point is right, but this seems revisionist; Yahoo was mainly a categorized set of links (like dmoz.org) with a halfassed search engine tacked on as an afterthought. Indeed, at the outset it didn't even have a search engine--and even after adding one, it was never really a search engine back then. Altavista was the "preferred" search engine at the time (having replaced Lycos when Lycos decided to go in the portal direction).
But yeah, Google came on the scene and did just searching, and did it better than Altavista and Lycos at a time when they had lost the search focus and were adding relatively crappy portal features instead of improving search.
just hitting Esc if you're not in Normal mode followed by ZZ to save/quit is quite the useful sort of thing you'd expect from vim (or equivalent in emacs)
Of course with a modern vim (last 5 years) you could:
1. Select save-exit from the menu.
2. Select save from the menu, then select exit from the menu.
3. Select "exit" from the menu, then respond yes to the "do you want to save your work" prompt
4. Any of the above, using Meta-F to access the file menu and hitting S for save, x for exit, or v for save-exit.
No need to hit escape before any of those options, either.
Betamax was never better than VHS. It was theoretically superior on picture quality, though most side-by-side reviews at the time of real equipment gave the edge to VHS except for a few short periods of a couple of months when Beta would implement a picture advance before VHS. But Beta was vastly inferior on sound quality and on tape length.
t ory/0,12449,881780,00.html
See, for instance, http://technology.guardian.co.uk/online/comment/s
Of course, whether or not Dvorak is any better than Qwerty is open to debate; I've seen at least one study showing that Qwerty's design (alternate hands to avoid typewriter jams) also turns out to be quite helpful to typing speed (alternating hands is faster than typing successive characters on the same hand), and that Dvorak typists with repetitive stress injury benefit by switching to Qwerty in exactly the same way that Qwerty RSI sufferers benefit by switching to Dvorak. The only studies I know of show major Dvorak benefits were done by the inventor and salesman of the Dvorak layout.
See, for instance, http://wwwpub.utdallas.edu/~liebowit/keys1.html
If you think Java's syntax is radically different from C's syntax
Syntax is one of the least important features of a language, and OP never said anything about them having different syntax.
OP wrote:
"Try C, LISP, SmallTalk or Ruby and you start to feel that the language is helping you. Cobol & Java fell like they are getting in the way."
One can infer that he's in favor of strongly, dynamically typed languages with the exception of C (weakly, statically typed), and against strongly, statically typed languages (or at least those without type inference).
I can venture a guess with reasonable confidence that he'd like python and scheme and would strongly dislike Pascal and Ada.
For languages that fall outside of the spectrum he described (say, Tcl and VBscript, or Dylan and ML) I have no feel for what he'd like.
Those guesses have nothing to do with syntax.
Is vi the most effective editor? I doubt it. By so much that you can actually win back the time you spent learning? I strongly doubt it. Remember, we're talking tenths of a second here.
I've often found it saved me a half hour or more over the readily available alternatives. It's just about the easiest install on a Windows box if you need to do a single easily scripted edit (munging field order and formatting, stuff like that).
It's also one of the easiest code browsers to install if you get dropped in somewhere where you need to navigate an unfamiliar source tree quickly (just ctags all the files and fire up vim), especially if there's more than just C/C++/Java/HTML/VB/ASP in the tree.
Absolutely. I almost never use it outside of work. But it's very convenient at work, and every office I've been in has used some form of IM extensively. It's a hell of a lot easier to throw URLs and commands back and forth in IM than in email (let alone by voice).
Examples:
Developer1: Do you know the syntax off the top of your head to make an existing column autoincrement in mysql?
Developer2: alter table modify column type autoincrement
or
Developer: What's next most important for CLIENT?
Project Manager: Finish up the issue you're working on first, then take a look at this: http://bugtracker-bugurl/?bugid
or:
Secretary: Reminder to all: company meeting is starting in 1 minute in the conference room
It also has the advantage of letting you keep in touch with peers who work elsewhere--I have friends who I know are familiar with certain commands or whatever, and for some packages it's a lot quicker for me to ask them (and vice-versa) "what's the flag to do X" than to dig through the docs.
Also, how are the young people of today going to react when they find they can't use a tool they are kinda of hooked on daily in the real work world?
Badly, as they should if the company they work for doesn't allow them a valuable communications tool.
I'm curious as to the particulars of your project. How many people are working on it? How long has it been around? Does your company employ short-term contractors?
The current project isn't really a C project, it's about 300,000 lines of Python (with an equivalent amount of template code) and maybe 500 lines of C; we have 5 full-time developers and 1 part-time (maybe 10 or so people have worked on it at some point). About a third of the code is machine-generated via tools we've written; half the remainder is generic library code used by many client implementations, and the other half is specific to client needs. I've been on it for a year, the project has been around for 4 years.
The previous project was about 50,000 lines of C, 15,000 lines of C++, and 30,000 lines of Java. Because of institutional inefficiency, it had as many as 15 people work on it at peak (maybe 20-25 total over the life of the project) for about 6 months. Normal staff over the project lifespan was 3 C developers, 4 Java developers, and 2 C++ developers.
It's still in production (large network server servicing millions of requests daily--it's over the 1 billion served mark in its lifetime). No short-term contractors were used. It's in static production now, with a few weeks of development each year if new features are needed--it's been in production for about 6 years (active development for the first couple), with a year of development lead-in before that.
It's essentially 7-8 seperate back-end processes (C, one C++ component) interfaced by a desktop UI for data entry and system management (Java), a reporting front-end (server-side Java web interface) and a number of logging and summarizing processes (C mainly, one small Perl component).
I myself was originally a big advocate of C, but I've come to the conclusion the average programmer is simply too stupid to avoid 1)memory leaks, 2) buffer overruns, and 3) my personal favorite - passing references to stack variables back to the calling frame
In my experience, such programmers are also bad at avoiding business and programming logic errors. You're far better off going with fewer skilled programmers. Not only will the bug rate decrease, but the productivity will be higher--and although the cost per programmer will be higher, the total amount spent on programmers will be lower for more work done.
Essentially, if memory leaks and buffer overruns represent a significant amount of debugging time for your staff then you're going to be woefully underproductive no matter what language you choose.
That said, shaving a few percent off is worth it, and higher level languages can give you much more dramatic productivity gains in other arenas (depending on the task).