Let alone the couple of cases I witnessed where Word was unable to open.docs created by itself, on the same machine by the same user. Opening it in openoffice and saving it again in.doc format saved the day.
1) Bad intel? From what planet? Where I live everyone knew that Saddam was leading one of the most secular governments in the region. In particular the US has funded him to fight the ayatollah's of yesteryear. Saddam being in any way involved with 9/11 is and was a red herring. No sense in discussing that.
2) ID. Conceded. Bush is just generally stupid, but that's not the point.
4) Partly conceded. When the WMD lies (or was it again just bad intel? Boy, these people believe anything they already believe in!) didn't pan out, it was used as a justification in retrospect to such an extent that a general scepsism against the war in Iraq is met by the question if you would have liked Saddam in place. My answer these days would be that I can answer that in 10 years time. But indeed it's just expected political weaseling, nothing really untoward.
4&5) conceded
6) Not conceded. The illegal combatant status, dispension of rights (and subsequent gitmo issues) has led to the absurd situation that democratic rights are suspended to save democracy.
Side effect:
Some allies (the Netherlands two days ago) no longer extradite people suspected by the US for terrorism due to the suspension of rights for the suspects. This is an expected side-effect of disregarding human rights.
7) Wrongly phrased by me, the phrase I ment was: "Global warming is hotly debated in science". It isn't. That's a lie. However, I did do some checking and indeed, you are right. It's only airheads in blogs that ventilate this crap, the US gov does believe in global warmint, it just uses 8 to not do 9. 7,8 and 9 are generally the results of some discussions I witnessed in blogs by self-proclaimed Reps. I expected that at least they would know their government's pov. Apparently not.
10) Saying the UN wants to take over (power in) the US as something to fear does remind me of that elephant that was afraid to be strangled by an ant. The discrepancy in power is so big that even suggesting it is ridicolous.
So, where are we. lots of points conceded. Some of them (the important ones I think) not. I would like to add a few more irrationalities in for good measure. I'm trying to get together a good list.
More irrationalities:
War against drugs
War against terrorism
War against another concept
(And as a side-note, has anyone ever noticed that the US has been at war with something or another since 1941? Will this go into the history books as the 100 years war?)
The parallel was the war against rational thought. The Nazis (as well as the commies) were masters at obscuring rational discourse. Some examples:
Saddam Hussein was involved in 9/11
Intelligent Design is a science
Who is not for us is against us
The war in Iraq was started to liberate the Iraqi people.
We are winning the war in Iraq
God speaks to me
There is something called an illegal combatant
Global warming doesn't occur
In the case global warming does occur humans do not contribute to it
In the case humans contribute to global warming, any attempt to remedy the effects will instantly melt down the whole economy
United Nations are out to take over US autonomy
... more, much more...
Not all of these things can be attributed only to the current administration, although they take the war on rationality to new heights. As far as lies and creeping fascism go, the Dems are just as dangerous as the Reps. The fact that you only hear those 'two' sides of the story is part of the problem.
Ah, those lovely Java references: combining the strengths of pointers (being NULL) with the strengths of references (no pointer arithmetic). In any case, good luck with your hunt. At times I also wished that Java had a way of specifying that an object is not supposed to be null in a function signature, that would help a lot. But until that happens we're stuck at specifying it in the Javadoc...
In 1946, the Nazis where the most terrible thing that had happened to the world. In 1935 however, the Nazis were not deemed particularly dangerous, also not by the US. Looking at US now and Nazi Germany 1935, some parallels can be drawn. What is in particular frightening me is the ongoing war on all forms of rational thought that is fought from the trenches in the White House. That's a well-known totalitarian tactic, and scary as hell.
I do have some hope nowadays as it seems that the US is finally waking up, and maybe this nightmare will not be coming through.
I personally like not to worry about pointers when I'm not doing low level work.
Honestly, in my C++ work, I don't like to worry about pointers either. To wit, two weeks ago I was doing some quick and dirty work and suddenly had to call delete on a simple array. I really had to think about what the syntax for that was again, as I hadn't used that for years. And for Java not having pointers, seen any NullPointerExceptions lately? At least my C++ references cannot be NULL (they can be invalid though;)
That's all true, and well known, but only points to the fact that you are not supposed to use reference counting everywhere. You use it for simple shared memory, not for non-shared stuff. Unlike Java, you are not forced to use one particular mode of memory management and stick with it. One uses refcounting for 'shared memory' only, and local memory (i.e., simple scoped memory) for other purposes. When that shared memory involves cycles that cannot be factored out, you might throw in a garbage collector for that piece of the code. This doesn't happen often however, as many cycles can be resolved by using a combination of shared_ptr and weak_ptr. The latter points to the same mem as a shared_ptr, but doesn't do reference counting. Who owns what can be simply found by looking at the type.
But just a question for GC lovers (in Java). Have you ever worked with some library/other person's code, gotten and object (say an int[]) back and asked yourselves: can I safely change these values or would the whole system fall over? GC does not solve memory ownership, it only gets rid of garbage. Memory sharing is difficult (as it breaks encapsulation), and need to be treated as such. In my experience, most Java programs have horrible encapsulation, simply because most code give references to their private parts out freely;)
So, tell me, how is C++ implementing virtual method calss, without virtual method tables? Oh, and yes, fuck this synchronisation stuff, one thread will be enough for all applications!
If a C++ class has virtual functions, there will be a virtual method table. However, if it doesn't, there will be none. A simple class for data storage will have zero byte overhead. Add synchronization (I think you mean semaphores?) and this will cost you that extra few bytes. But only when you need it.
A Java object always comes with overhead, there is no way to get it out. The big difference between the philosophies of C++ and Java are that in C++ you don't pay for what you don't use. In Java you simply pay for everything, even if you don't need it. Don't need inheritance, but only data encapsulation? you pay for inheritance. Don't need synchronization? You pay.
Just a question: how much memory does an ArrayList<Integer> of size 1000 take? In C++, a vector<int>(1000) takes 1000*4 bytes + 12 (3 pointers for vector). Last time I checked, a single Integer in Java takes 16 (!!!) bytes, 4 times more than strictly necessary. Arraylist itself comes at a nice 80 bytes (versus 12 for C++). Further, how much would a non-virtual class in C++ take that has a single integer data member? Yes, again 4 bytes. Add a virtual function table, and you're at 8 bytes, still half of that of Java. All these unavoidable bytes quickly add up, hence Java being a memory hog. One million of objects floating around? (can easily happen as you hardly have a choice avoiding objects) Eight megabyte gone for Object overhead alone.
What the article basically says is that java memory allocations/deallocations are absolutely not slow. Anyone who knows a bit about garbage collection knows that this should be the case. However, what it forgets to mention is that java simply uses such stupendous amounts of memory that it is bound to slow down even the fastest allocation technique. Traversing 512 MB of allocated memory/garbage takes time.
... then you add a method which releases that resource - there's no need to deallocate the object.
And the issue with this is that the Resource Acquisition Is Initialization idiom is broken, and you are suddenly forced to do resource management. In a language like Perl/Python or even C++ using shared pointers, you don't do explicit deallocations, and also no explicit resource management, and all works even in the presence of exceptions. Resource management is not as ubiquitous as memory management, but errors in it can be even more dangerous. I prefer to open a file and know that it gets closed when it goes out of scope.
For this kind of stuff, shared memory, a sane C++ developer uses a shared_ptr. A simple reference counted pointer scheme that will take care of deallocation. It's trivial to write your own, or just use boost.
Modern C++ simply doesn't use raw pointers. One uses the memory management scheme that is most suitable for that part of the code, but hardly ever use 'delete' anymore.
Ok, so you want to distribute your binaries, but you can't use the GPL, because you can't distribute (a part of) your source? And this due to export restrictions? Are you sure you can distribute your binaries?
Or are you simply doing a research project where you don't distribute the binaries/source outside of your institute? In that case, the GPL will work fine for you, as there's nothing in the GPL that says that you MUST distribute. IF you distribute, certain rules apply. If you don't, nothing prevents you from using it inside your institute.
Actually, I'm doing that right now. I'm working together on a project with one other developer (I'm kind of the architect/designer/co-developer of the product), and I'm continuously pushing to him small sketches of UI using only pencil and paper. Much faster than actually creating a design document or a mockup (5 mins. versus 1 hour). I'm planning on putting all these sketches into a design 'book' at some point, just to prove a point. I've got about 5 A4s now. Information worth: about 40 pages of painstacking work electronically.
Just to be on the safe side, I do enter them into an electronic document at some point, but only after the fact: when the implementation is finished.
Common, get of it. Viruses are mostly social engineering attacks (worms are the ones that attack engineering flaws). Microsoft providing a free anti-virus toolkit is simply common sense. Any OS-vendor should do this as this should be part of the service. When linux gets targeted due to being successfull on the desktop (any year now) you would expect it to come with basic virus-detection tools, right?
Symantec is just going the way of Trumpet Winsock. Remember? Building a TCP/IP stack for windows didn't turn out to be a viable long-term strategy. Bye Symantec, and good riddens.
The alternative is to work old-style, in a room full of desks. Would you prefer that?
A most definite YES from here. Cubicles are horrible, give me a room full of desks anytime. Make the rooms contain 8 people max and it's a perfect work environment.
Not sure about staroffice, but more than once I had to help someone rescue a Word document by loading it into OO and save it again as Word doc. Word couldn't do it. Mind you, this was a document that never was touched by any other version of Word apart from the one that was used to create it. That one couldn't load it in at some point. In essence, Word is not even compatible with itself.
there was never a proletarian revolution in China, since there wasn't an industrial proletariat
What I gather from the news these days, there sure is an industrial proletariat in China right now! It's the real thing, probably worse than the one that prompted Marx to write Das Kapital. Maybe they'll actually go communist at some stage.
Are you that naive? Do you honestly believe that the Prez of the US of A has any more important mission than to progress the wealth of the corporations of that same US of A? Do they actually teach history where you come from? For the last century or so, the USA foreign policy has been primarily focussed on making sure that US corporations can make money. Why do you think that has changed?
To spell it out: democracy in prez-speak means freedom for corporations to make money. Democratic values is the value of being able to make money without being hampered too much. You see, those damn Arabs don't buy much of our stuff, so we need to bring democracy to them.
Okay, a bit over the top this, but the point is that from the US point of view, business comes first, democratic values come second. This has always been the case, and I don't see a change under the current administration. You don't think that the American benevolence towards Iraq has nothing to do with the fact that Iraq is important for business? Unlike, say Sudan?
tsk, foulmoothing on so little pretext. Yes the JVM is written in an unsafe language. This simply means that the JVM is a single point of failure. However, if the JVM is safe, all java apps are safe. Now try to argue the same thing with every C-app, and envision the amount of effort that goes into (a) ensuring that the JVM is safe and (b) ensuring that every c-application on the face of the earth is safe. Then estimate the chances of success for (a) and (b). Furthermore try to envision the amount of effort that has gone into ensuring that the Java sandbox is foolproof, compared with the effort in avoiding buffer overruns in your random c-app. Only when carefully thinking this through, start calling people dumbasses, dumbass.
Ok, just read the rest of the comments and noted that BT is exactly this. Never bothered to read up on the actual protocol, hence the confusion. Ah well, another comment destined to the bitbucket.
Not quite true. Indeed in a simple prisoner's dilemma, defecting (cheating) is optimal, but in an iterated prisoner's dilemma it shifts to tit-for-tat being optimal. BT is iterated, if cheating for one block of the download can possibly ban you for a next block. In this scenario tit-for-tat makes sense. The only problem is the final block of code, for defecting to get that one is again optimal. To actually get full downloads it does seem that you need to have a few suckers around. The seeder is such a sucker, but all in all, I doubt that people will bother with writing cheats to *only* get the last block for free.
I think the GP is dead-on and a swarm-based tit-for-tat system would be a real stable p2p network. There's been twenty years of research after Axelrod's initial experiments and as far as I know, tit-for-tat has not been convincingly beaten.
Slashdot Mirror
Let alone the couple of cases I witnessed where Word was unable to open .docs created by itself, on the same machine by the same user. Opening it in openoffice and saving it again in .doc format saved the day.
The answer is 0.918181818181... Do I get a bit of candy now?
2) ID. Conceded. Bush is just generally stupid, but that's not the point.
3)The exact quote is: you are either for us or agains us in the fight against terror . Note that critique against the US approach towards the fight, puts you straight in the against camp.
4) Partly conceded. When the WMD lies (or was it again just bad intel? Boy, these people believe anything they already believe in!) didn't pan out, it was used as a justification in retrospect to such an extent that a general scepsism against the war in Iraq is met by the question if you would have liked Saddam in place. My answer these days would be that I can answer that in 10 years time. But indeed it's just expected political weaseling, nothing really untoward.
4&5) conceded
6) Not conceded. The illegal combatant status, dispension of rights (and subsequent gitmo issues) has led to the absurd situation that democratic rights are suspended to save democracy.
Side effect:
Some allies (the Netherlands two days ago) no longer extradite people suspected by the US for terrorism due to the suspension of rights for the suspects. This is an expected side-effect of disregarding human rights.
7) Wrongly phrased by me, the phrase I ment was: "Global warming is hotly debated in science". It isn't. That's a lie. However, I did do some checking and indeed, you are right. It's only airheads in blogs that ventilate this crap, the US gov does believe in global warmint, it just uses 8 to not do 9. 7,8 and 9 are generally the results of some discussions I witnessed in blogs by self-proclaimed Reps. I expected that at least they would know their government's pov. Apparently not.
10) Saying the UN wants to take over (power in) the US as something to fear does remind me of that elephant that was afraid to be strangled by an ant. The discrepancy in power is so big that even suggesting it is ridicolous.
So, where are we. lots of points conceded. Some of them (the important ones I think) not. I would like to add a few more irrationalities in for good measure. I'm trying to get together a good list.
More irrationalities:
War against drugs
War against terrorism
War against another concept
(And as a side-note, has anyone ever noticed that the US has been at war with something or another since 1941? Will this go into the history books as the 100 years war?)
Saddam Hussein was involved in 9/11
Intelligent Design is a science
Who is not for us is against us
The war in Iraq was started to liberate the Iraqi people.
We are winning the war in Iraq
God speaks to me
There is something called an illegal combatant
Global warming doesn't occur
In the case global warming does occur humans do not contribute to it
In the case humans contribute to global warming, any attempt to remedy the effects will instantly melt down the whole economy
United Nations are out to take over US autonomy
Not all of these things can be attributed only to the current administration, although they take the war on rationality to new heights. As far as lies and creeping fascism go, the Dems are just as dangerous as the Reps. The fact that you only hear those 'two' sides of the story is part of the problem.
Ah, those lovely Java references: combining the strengths of pointers (being NULL) with the strengths of references (no pointer arithmetic). In any case, good luck with your hunt. At times I also wished that Java had a way of specifying that an object is not supposed to be null in a function signature, that would help a lot. But until that happens we're stuck at specifying it in the Javadoc...
I do have some hope nowadays as it seems that the US is finally waking up, and maybe this nightmare will not be coming through.
Honestly, in my C++ work, I don't like to worry about pointers either. To wit, two weeks ago I was doing some quick and dirty work and suddenly had to call delete on a simple array. I really had to think about what the syntax for that was again, as I hadn't used that for years. And for Java not having pointers, seen any NullPointerExceptions lately? At least my C++ references cannot be NULL (they can be invalid though ;)
But just a question for GC lovers (in Java). Have you ever worked with some library/other person's code, gotten and object (say an int[]) back and asked yourselves: can I safely change these values or would the whole system fall over? GC does not solve memory ownership, it only gets rid of garbage. Memory sharing is difficult (as it breaks encapsulation), and need to be treated as such. In my experience, most Java programs have horrible encapsulation, simply because most code give references to their private parts out freely ;)
So, tell me, how is C++ implementing virtual method calss, without virtual method tables? Oh, and yes, fuck this synchronisation stuff, one thread will be enough for all applications!
If a C++ class has virtual functions, there will be a virtual method table. However, if it doesn't, there will be none. A simple class for data storage will have zero byte overhead. Add synchronization (I think you mean semaphores?) and this will cost you that extra few bytes. But only when you need it.
A Java object always comes with overhead, there is no way to get it out. The big difference between the philosophies of C++ and Java are that in C++ you don't pay for what you don't use. In Java you simply pay for everything, even if you don't need it. Don't need inheritance, but only data encapsulation? you pay for inheritance. Don't need synchronization? You pay.
Just a question: how much memory does an ArrayList<Integer> of size 1000 take? In C++, a vector<int>(1000) takes 1000*4 bytes + 12 (3 pointers for vector). Last time I checked, a single Integer in Java takes 16 (!!!) bytes, 4 times more than strictly necessary. Arraylist itself comes at a nice 80 bytes (versus 12 for C++). Further, how much would a non-virtual class in C++ take that has a single integer data member? Yes, again 4 bytes. Add a virtual function table, and you're at 8 bytes, still half of that of Java. All these unavoidable bytes quickly add up, hence Java being a memory hog. One million of objects floating around? (can easily happen as you hardly have a choice avoiding objects) Eight megabyte gone for Object overhead alone.
What the article basically says is that java memory allocations/deallocations are absolutely not slow. Anyone who knows a bit about garbage collection knows that this should be the case. However, what it forgets to mention is that java simply uses such stupendous amounts of memory that it is bound to slow down even the fastest allocation technique. Traversing 512 MB of allocated memory/garbage takes time.
And the issue with this is that the Resource Acquisition Is Initialization idiom is broken, and you are suddenly forced to do resource management. In a language like Perl/Python or even C++ using shared pointers, you don't do explicit deallocations, and also no explicit resource management, and all works even in the presence of exceptions. Resource management is not as ubiquitous as memory management, but errors in it can be even more dangerous. I prefer to open a file and know that it gets closed when it goes out of scope.
I think he prefers jikes, written in C for speed.
Modern C++ simply doesn't use raw pointers. One uses the memory management scheme that is most suitable for that part of the code, but hardly ever use 'delete' anymore.
Or are you simply doing a research project where you don't distribute the binaries/source outside of your institute? In that case, the GPL will work fine for you, as there's nothing in the GPL that says that you MUST distribute. IF you distribute, certain rules apply. If you don't, nothing prevents you from using it inside your institute.
Just to be on the safe side, I do enter them into an electronic document at some point, but only after the fact: when the implementation is finished.
Symantec is just going the way of Trumpet Winsock. Remember? Building a TCP/IP stack for windows didn't turn out to be a viable long-term strategy. Bye Symantec, and good riddens.
Great! Digg! That's where the cool kids hang out. That will leave slashdots for the ... erm ... nerds?
A most definite YES from here. Cubicles are horrible, give me a room full of desks anytime. Make the rooms contain 8 people max and it's a perfect work environment.
Not sure about staroffice, but more than once I had to help someone rescue a Word document by loading it into OO and save it again as Word doc. Word couldn't do it. Mind you, this was a document that never was touched by any other version of Word apart from the one that was used to create it. That one couldn't load it in at some point. In essence, Word is not even compatible with itself.
What I gather from the news these days, there sure is an industrial proletariat in China right now! It's the real thing, probably worse than the one that prompted Marx to write Das Kapital. Maybe they'll actually go communist at some stage.
To spell it out: democracy in prez-speak means freedom for corporations to make money. Democratic values is the value of being able to make money without being hampered too much. You see, those damn Arabs don't buy much of our stuff, so we need to bring democracy to them.
Okay, a bit over the top this, but the point is that from the US point of view, business comes first, democratic values come second. This has always been the case, and I don't see a change under the current administration. You don't think that the American benevolence towards Iraq has nothing to do with the fact that Iraq is important for business? Unlike, say Sudan?
And a fourth.
tsk, foulmoothing on so little pretext. Yes the JVM is written in an unsafe language. This simply means that the JVM is a single point of failure. However, if the JVM is safe, all java apps are safe. Now try to argue the same thing with every C-app, and envision the amount of effort that goes into (a) ensuring that the JVM is safe and (b) ensuring that every c-application on the face of the earth is safe. Then estimate the chances of success for (a) and (b). Furthermore try to envision the amount of effort that has gone into ensuring that the Java sandbox is foolproof, compared with the effort in avoiding buffer overruns in your random c-app. Only when carefully thinking this through, start calling people dumbasses, dumbass.
Ok, just read the rest of the comments and noted that BT is exactly this. Never bothered to read up on the actual protocol, hence the confusion. Ah well, another comment destined to the bitbucket.
I think the GP is dead-on and a swarm-based tit-for-tat system would be a real stable p2p network. There's been twenty years of research after Axelrod's initial experiments and as far as I know, tit-for-tat has not been convincingly beaten.