Sorry to sound a bit "curt", but honestly, why don't you just develop your application using something like Intel's Linux compiler?
I don't see how that would solve the problem, since we'd once again have to distribute the runtime. I didn't even know that ICC supplied its own runtime -- is that even true?
We weren't trying to weasel our way around license restrictions. We would have been happy to distribute a trimmed runtime and provide source for it, had it been easy enough. I TRIED.
We made open source contributions on multiple occasions. We found bugs in FreeType, fixed them, submitted patches. We found bugs in Leptonica, fixed them, submitted patches. We made enhancements to jbig2enc which were submitted back. During my last year there we took the plunge and paid $20000 for a commercial license to use xpdf in our product. We found that the error handling wasn't quite how we liked it, so we provided Derek with some suggestions and code, which he reworked in a way he liked a bit better, and put it into xpdf. You get the picture... We were trying to make profit from OUR technology, not by screwing over open source developers. Isn't this the way everyone wants it to work? Aren't we allowed to make profit somehow?
If you developed something for months, released a version labelled as "beta" for testing only, and found a/. review of it, do you really think you'd feel this was a good forum to provide feedback?
If it were me, I can't imagine why I'd deliberately ignore very specific comments regardless of their source. Obviously the issues should be tracked, so I'd spend a few hours entering the salient comments into my issue tracking system. The user is already doing significant work for me by even using my product. I'm not suggesting that I would scour the entire Internet, but Slashdot is a freebie as I read it on a daily basis.
Sounds like you learned several important lessons about development, none of which was "Indian programmers will screw you." It sounds harsh because it is, but that particular chain of events was completely foreseeable.
In every single one of those examples, the uniform serves a utilitarian FUNCTION. A doctor wears clothing which is easy to clean and makes it easy to see when it's contaminated with something (e.g. blood). A mechanic wears something that he doesn't mind getting covered in grease. A police officer wears something that makes him instantly recognizable as a legal authority figure, and it also gives him a place to hang or attach all of the equipment he carries. And do I really need to explain why a fireman needs a uniform?
When the car was first invented, the term "mechanic" called to mind an idiosyncratic genius who did magical things to make the cars work. Then they became commoditized. IT is heading the same way, in fact it's probably halfway through the transition already. You can't arrest that change. If you're really that smart (and egotistical) just switch to a field with higher social status.
I'm not IT, but if I was, I'd happily wear a uniform as long as it looked reasonably like an FBI windbreaker ("IT" in bold white sans-serif letters on the back). And I'd also carry a gun. An inventory gun. But in a holster. Yeah.
We briefly considered that, but decided it was unacceptable. The glibc binary is just too large. One of the things our customers consistently praised us for was that our.exe was under 1.5 megabytes, while the closest competing app was over 15 megabytes. glibc alone was equal to the size of the app. Slicing and dicing the code to the bare minimum wasn't acceptable either, because then it wasn't a stock library anymore and we would have had to put it through testing, and we were not interested in testing runtime libs. Not to mention that if we ever had to upgrade the library we'd have to do it all over again.
Actually, I briefly undertook a skunkworks effort to trim glibc down to the bare minimum. I gave up after just one evening when I discovered that simply calling printf() drags in almost the entire freaking library by reference. I was dumping linker dependency maps and it was clear that it would take MAJOR changes to make even MINOR effects on code size. The entire glibc codebase is so twisted and interdependent that I gave up in disgust. There's theory, then there's practice.
Anyway, somebody already pointed out that Chrome is BSD licensed, which I didn't know. In that case, your distro of choice should be building a compatible package for you. Patience!
What are those dependencies, and how are they licensed? Depending on the license, static linking could force them to open source the entire application.
This was a continuing source of irritation back when I worked on a closed source Linux app. The glibc people do not give a crap about binary-breaking changes. This resulted in us having to create multiple variants of our product to link against different versions of the runtime libs (in order to support older distros), multiplying our testing efforts by a factor of three. We desperately wanted to just link glibc statically, but that's a no-no because it's LGPL.
And anyone who has been around has seen this scenario play out at least once: you freeze the code, run your final tests, and the boss hands down a list of alterations from the client at the last minute.
... and then you tell the client that they have two options. Option A, they can wait while the changes are implemented and tested. Option B, they can accept a last-minute release and waive their contractual rights to sue your ass if the product doesn't work. There is no sane Option C. Any company that operates the way you described is full of idiots.
Maybe, but the contract is between two parties. Why does only one of them have copyrights? It's just as much your document as it is theirs. Let's assume that the contract, i.e. the physical paper it's written on, is a "work" under copyright law. When you apply your signature to that contract, you are modifying that work, and the fact that the company who drew up the contract doesn't immediately sue you for doing that, implies that your rights to that document are just as good as theirs. It is ludicrous to think that the company could sue you for publishing a work THAT YOU POSSESS RIGHTS TO.
What... the... hell? Two hours prior to RELEASE, you were BUILDING the product? Wow, you allocated what, an hour, for release testing? Or you just kinda skip that part? You don't freeze your code base months or weeks prior to release? You don't have a revision control system that allows you to roll back a bonehead change? Yes, there certainly is some incompetence at your company, but it's not the Indian.
If "relevance" is a requirement, then the government will have to produce a definition of "relevance." Wow, I love this idea. Instead of allowing the advancement of technology, we have to conform to a government definition, and if we rank our search results contrary to that definition, our search engine is ILLEGAL. And I'm sure the government won't abuse their ability to declare certain results orderings to be illegal.
Stay the hell away from my search engines. If I'm not happy with the one I'm using, I'll switch to another.
No matter how clever or paranoid you are, you're not clever and paranoid enough. Just use AES.
This sort of statement is equally dangerous by leading people to believe that just because they are using a strong cipher they are secure. Basically, unless a cryptography expert is designing your entire system, you're going to fuck SOMETHING up. There is no magic bullet.
As long as we are not restoring or replacing those habitats, what use is it to resurrect the animals that used to live in it?
That's obviously ridiculous, but it's not what I was thinking of. It isn't really that expensive or difficult to store the genomes indefinitely. Perhaps the remote possibility of being able to restore the extinct species will act as a motivator to restore the habitats they lived in. I don't think we should abandon hope just because we can't think further than a few years into the future.
But if they have skin samples, then they do have the mitochondrial DNA. We just don't have the ability to replace that part of the cell structure. Yet. Another problem is that the specimen is female, meaning there is no Y chromosome, so we could never create a male.
At this point we should probably be harvesting DNA from threatened species (from enough donors to form a not-completely-terrible breeding population) and storing it away somewhere.
Averaging 217 mph over a distance of 663 miles, supposedly connecting 20 cities... according to TFA, a trip of under three hours...
It just isn't possible. Assuming that, at each city, you have 3 minutes of deceleration, a stop time of 10 minutes, and 5 minutes of acceleration, that's 18*20 = 360 minutes, or 6 hours. That doesn't even include time at full speed. Okay, let's be insane and decelerate in only 1 minute and accelerate in 2 and stop for only 3 minutes, that's now two hours, leaving you one hour to travel 663 miles for an average speed of (duh) 663 miles per hour. That means top speed is somewhat more than that, but approximating the top speed as 663 MPH, then decelerating from 663 to 0 in one minute would give about 0.5 G's, which is going to be an uncomfortable experience for an entire minute. At the very least you would need to be facing backward so you'd be pressed into your seat instead of thrown out of it. It's just totally impossible.
That's because it's a different (more efficient) code. You could, in theory, use the same or similar code for 512-byte sectors. If you are changing the sector size, it's a good opportunity to switch codes.
That's understating it. Bit errors occur on every sector read from the disk. Let me repeat that: every single attempt to read a sector results in errors. Without ECC, hard disks simply would not work, period. The data densities are too great.
ECC means "error correcting code." It isn't just a way of detecting errors, it's a way of operating even in the face of them. As such, the length of the ECC chunk is proportional to the sector payload size. So going to a larger sector does not reduce ECC overhead, unless you switch to a more efficient ECC code (and you can do that without changing sector size). The other stuff, like sync, is fixed size and your argument does apply to that. The real reason for using a 4-kilobyte sector is because it exactly matches the size of a VM page on most consumer architectures, therefore guaranteeing at a hardware level that a single page worth of information cannot be fragmented.
It's a brother, not a father. I'm just picking up on the very salient lack of any comment about whether the kid WANTS to learn any of this stuff. When I was younger I tried several times to get my own brother to be interested in things I was interested in, and vice versa. Some of those discussions ended up in fist fights. Eventually I realized that my brother is who he is, and I am who I am.
"I'd love to teach him how to program"... "Lead him down a path"... What the hell is your problem? It's a human being, not a parakeet. He's not a plaything to amuse yourself with or a piece of clay to mold however is most entertaining for you.
Curious what you think is wrong with MythBusters. As far as I've ever seen, there's no interpersonal drama, although some of the characters are a bit goofy. Are you objecting to the fact that it's not dry and boring, or something else? Seriously, I want to know. As long as you don't delude yourself into thinking they are conducting scientific experiments, I've always thought it's interesting and quite worth watching.
Yes, what the hell happened to Discovery anyway? Actually, I'm just being rhetorical, it's fairly obvious what happened, and it took several sad years. Watching Discovery turn to shit was like watching a relative die of cancer. The only people left worth watching are Jamie, Adam, and Mike Rowe (Mike Rowe, by the way, should be given a fucking medal and a gigantic bronze statue for the comments he makes about safety fascism in modern America). If these guys had the guts they should start their own channel and give Discovery the big fat finger.
Slashdot Mirror
Sorry to sound a bit "curt", but honestly, why don't you just develop your application using something like Intel's Linux compiler?
I don't see how that would solve the problem, since we'd once again have to distribute the runtime. I didn't even know that ICC supplied its own runtime -- is that even true?
We weren't trying to weasel our way around license restrictions. We would have been happy to distribute a trimmed runtime and provide source for it, had it been easy enough. I TRIED.
We made open source contributions on multiple occasions. We found bugs in FreeType, fixed them, submitted patches. We found bugs in Leptonica, fixed them, submitted patches. We made enhancements to jbig2enc which were submitted back. During my last year there we took the plunge and paid $20000 for a commercial license to use xpdf in our product. We found that the error handling wasn't quite how we liked it, so we provided Derek with some suggestions and code, which he reworked in a way he liked a bit better, and put it into xpdf. You get the picture... We were trying to make profit from OUR technology, not by screwing over open source developers. Isn't this the way everyone wants it to work? Aren't we allowed to make profit somehow?
If you developed something for months, released a version labelled as "beta" for testing only, and found a /. review of it, do you really think you'd feel this was a good forum to provide feedback?
If it were me, I can't imagine why I'd deliberately ignore very specific comments regardless of their source. Obviously the issues should be tracked, so I'd spend a few hours entering the salient comments into my issue tracking system. The user is already doing significant work for me by even using my product. I'm not suggesting that I would scour the entire Internet, but Slashdot is a freebie as I read it on a daily basis.
Sounds like you learned several important lessons about development, none of which was "Indian programmers will screw you." It sounds harsh because it is, but that particular chain of events was completely foreseeable.
In every single one of those examples, the uniform serves a utilitarian FUNCTION. A doctor wears clothing which is easy to clean and makes it easy to see when it's contaminated with something (e.g. blood). A mechanic wears something that he doesn't mind getting covered in grease. A police officer wears something that makes him instantly recognizable as a legal authority figure, and it also gives him a place to hang or attach all of the equipment he carries. And do I really need to explain why a fireman needs a uniform?
When the car was first invented, the term "mechanic" called to mind an idiosyncratic genius who did magical things to make the cars work. Then they became commoditized. IT is heading the same way, in fact it's probably halfway through the transition already. You can't arrest that change. If you're really that smart (and egotistical) just switch to a field with higher social status.
I'm not IT, but if I was, I'd happily wear a uniform as long as it looked reasonably like an FBI windbreaker ("IT" in bold white sans-serif letters on the back). And I'd also carry a gun. An inventory gun. But in a holster. Yeah.
If you think the Chrome developers aren't reading this very thread, you're crazy. This is a good a forum as any to provide feedback.
We briefly considered that, but decided it was unacceptable. The glibc binary is just too large. One of the things our customers consistently praised us for was that our .exe was under 1.5 megabytes, while the closest competing app was over 15 megabytes. glibc alone was equal to the size of the app. Slicing and dicing the code to the bare minimum wasn't acceptable either, because then it wasn't a stock library anymore and we would have had to put it through testing, and we were not interested in testing runtime libs. Not to mention that if we ever had to upgrade the library we'd have to do it all over again.
Actually, I briefly undertook a skunkworks effort to trim glibc down to the bare minimum. I gave up after just one evening when I discovered that simply calling printf() drags in almost the entire freaking library by reference. I was dumping linker dependency maps and it was clear that it would take MAJOR changes to make even MINOR effects on code size. The entire glibc codebase is so twisted and interdependent that I gave up in disgust. There's theory, then there's practice.
Anyway, somebody already pointed out that Chrome is BSD licensed, which I didn't know. In that case, your distro of choice should be building a compatible package for you. Patience!
What are those dependencies, and how are they licensed? Depending on the license, static linking could force them to open source the entire application.
This was a continuing source of irritation back when I worked on a closed source Linux app. The glibc people do not give a crap about binary-breaking changes. This resulted in us having to create multiple variants of our product to link against different versions of the runtime libs (in order to support older distros), multiplying our testing efforts by a factor of three. We desperately wanted to just link glibc statically, but that's a no-no because it's LGPL.
"Don't criticize it, it's a beta." That's nonsense. The whole reason you release a beta is to get feedback.
As far as the KDE thing, though, I agree. Exactly what sort of "integration" with KDE was expected?
And anyone who has been around has seen this scenario play out at least once: you freeze the code, run your final tests, and the boss hands down a list of alterations from the client at the last minute.
... and then you tell the client that they have two options. Option A, they can wait while the changes are implemented and tested. Option B, they can accept a last-minute release and waive their contractual rights to sue your ass if the product doesn't work. There is no sane Option C. Any company that operates the way you described is full of idiots.
Maybe, but the contract is between two parties. Why does only one of them have copyrights? It's just as much your document as it is theirs. Let's assume that the contract, i.e. the physical paper it's written on, is a "work" under copyright law. When you apply your signature to that contract, you are modifying that work, and the fact that the company who drew up the contract doesn't immediately sue you for doing that, implies that your rights to that document are just as good as theirs. It is ludicrous to think that the company could sue you for publishing a work THAT YOU POSSESS RIGHTS TO.
What... the... hell? Two hours prior to RELEASE, you were BUILDING the product? Wow, you allocated what, an hour, for release testing? Or you just kinda skip that part? You don't freeze your code base months or weeks prior to release? You don't have a revision control system that allows you to roll back a bonehead change? Yes, there certainly is some incompetence at your company, but it's not the Indian.
If "relevance" is a requirement, then the government will have to produce a definition of "relevance." Wow, I love this idea. Instead of allowing the advancement of technology, we have to conform to a government definition, and if we rank our search results contrary to that definition, our search engine is ILLEGAL. And I'm sure the government won't abuse their ability to declare certain results orderings to be illegal.
Stay the hell away from my search engines. If I'm not happy with the one I'm using, I'll switch to another.
No matter how clever or paranoid you are, you're not clever and paranoid enough. Just use AES.
This sort of statement is equally dangerous by leading people to believe that just because they are using a strong cipher they are secure. Basically, unless a cryptography expert is designing your entire system, you're going to fuck SOMETHING up. There is no magic bullet.
As long as we are not restoring or replacing those habitats, what use is it to resurrect the animals that used to live in it?
That's obviously ridiculous, but it's not what I was thinking of. It isn't really that expensive or difficult to store the genomes indefinitely. Perhaps the remote possibility of being able to restore the extinct species will act as a motivator to restore the habitats they lived in. I don't think we should abandon hope just because we can't think further than a few years into the future.
But if they have skin samples, then they do have the mitochondrial DNA. We just don't have the ability to replace that part of the cell structure. Yet. Another problem is that the specimen is female, meaning there is no Y chromosome, so we could never create a male.
At this point we should probably be harvesting DNA from threatened species (from enough donors to form a not-completely-terrible breeding population) and storing it away somewhere.
Averaging 217 mph over a distance of 663 miles, supposedly connecting 20 cities... according to TFA, a trip of under three hours...
It just isn't possible. Assuming that, at each city, you have 3 minutes of deceleration, a stop time of 10 minutes, and 5 minutes of acceleration, that's 18*20 = 360 minutes, or 6 hours. That doesn't even include time at full speed. Okay, let's be insane and decelerate in only 1 minute and accelerate in 2 and stop for only 3 minutes, that's now two hours, leaving you one hour to travel 663 miles for an average speed of (duh) 663 miles per hour. That means top speed is somewhat more than that, but approximating the top speed as 663 MPH, then decelerating from 663 to 0 in one minute would give about 0.5 G's, which is going to be an uncomfortable experience for an entire minute. At the very least you would need to be facing backward so you'd be pressed into your seat instead of thrown out of it. It's just totally impossible.
That's because it's a different (more efficient) code. You could, in theory, use the same or similar code for 512-byte sectors. If you are changing the sector size, it's a good opportunity to switch codes.
That's understating it. Bit errors occur on every sector read from the disk. Let me repeat that: every single attempt to read a sector results in errors. Without ECC, hard disks simply would not work, period. The data densities are too great.
ECC means "error correcting code." It isn't just a way of detecting errors, it's a way of operating even in the face of them. As such, the length of the ECC chunk is proportional to the sector payload size. So going to a larger sector does not reduce ECC overhead, unless you switch to a more efficient ECC code (and you can do that without changing sector size). The other stuff, like sync, is fixed size and your argument does apply to that. The real reason for using a 4-kilobyte sector is because it exactly matches the size of a VM page on most consumer architectures, therefore guaranteeing at a hardware level that a single page worth of information cannot be fragmented.
It's a brother, not a father. I'm just picking up on the very salient lack of any comment about whether the kid WANTS to learn any of this stuff. When I was younger I tried several times to get my own brother to be interested in things I was interested in, and vice versa. Some of those discussions ended up in fist fights. Eventually I realized that my brother is who he is, and I am who I am.
"I'd love to teach him how to program" ... "Lead him down a path" ... What the hell is your problem? It's a human being, not a parakeet. He's not a plaything to amuse yourself with or a piece of clay to mold however is most entertaining for you.
Curious what you think is wrong with MythBusters. As far as I've ever seen, there's no interpersonal drama, although some of the characters are a bit goofy. Are you objecting to the fact that it's not dry and boring, or something else? Seriously, I want to know. As long as you don't delude yourself into thinking they are conducting scientific experiments, I've always thought it's interesting and quite worth watching.
Yes, what the hell happened to Discovery anyway? Actually, I'm just being rhetorical, it's fairly obvious what happened, and it took several sad years. Watching Discovery turn to shit was like watching a relative die of cancer. The only people left worth watching are Jamie, Adam, and Mike Rowe (Mike Rowe, by the way, should be given a fucking medal and a gigantic bronze statue for the comments he makes about safety fascism in modern America). If these guys had the guts they should start their own channel and give Discovery the big fat finger.