This is a nice way of saying that most the remaining low-skill jobs got outsourced to other countries (the rest being in the process of transitioning to the "gig economy"), and all the remaining inconvenience and cost was pushed onto customers in the form of increased waiting times, hidden fees, self-service, etc.
Well that doesn't mean they're worthless to many. Regardless, a less misleading headline would be: "Apple is the 7000th vendor to open source their key value store after discovering the space is too crowded".
It probably depends what you mean by "project". I doubt that 90% of "failed" projects cost anywhere near the same as a "successful" project becaus most fail early.
I also suspect that most new pharmaceutical projects don't originate within a pharmaceutical company, but rather within academia.
Ummm, North Korea isn't a communist state. There aren't any communist states, never were in modern history.
For the record, there aren't any capitalist states or libertarian states, either. Fundamentalism in any form is an inherently impractical form of government, especially in the modern era.
I remember when Agile was called XP. That was before the seminar and certification industry got hold of it.
XP made sense, but only by comparison with what came before it. If there's one thing I've learned after 25 years in this business, it's that fundamentalists make shitty products.
That depends on if you include the file system in the kernel.
NTFS is a driver that happens to be compiled in (because it makes it easier to boot the damn thing if nothing else), so you could go either way on that one
FWIW, I mostly agree. NTOSKRNL is very well-designed kernel by 1990 standards, and it still mostly holds up today. The lack of a unified event system, for example (e.g. wait on a socket and a condvar), is one of the outstanding issues in Linux. (For completeness, FreeBSD's kqueue almost does the job, and libevent makes things a bit less painful.)
The main problem with Windows NT is not the kernel design, it's almost everything else.
As I understand it, Apple was compelled to violate the basics of microkernels at many points, [...]
You understand wrong, but you're not alone. Apple broke the rules of 1990s microkernel fundamentalism, which are not the same as "basics of microkernels".
It's probably not a common scenario, but it's certainly true for me some of the time. It may be possible to rearchitect a Big Simulation (where, say, touching every part of a 20G data structure in a quasi-random order every simulation step is normal) to have better locality for debugging purposes, but is it worth it just to avoid debugging over an SSH connection?
No one is 'fixing' a Java, C++ program installed on a production system with SSH and VI.
Correct. However, "fixing" a program installed on a test environment is common enough. This is especially true in the scientific computing space (which I currently work in) where the best cut-down test case that you can find in a hurry won't run on your laptop with only 16G RAM.
You are going for a computer science degree. You must be able to express your ideas on paper, a white board, napkin, back of your hand,....anywhere.
I agree, but I'm going to make one proviso. Yes, programming exams on paper. But marks should not be taken off for trivial syntax errors or anything else that anyone with half a brain could easily fix when the compiler refuses to compile the code.
Your job, as a programmer or computer scientist, to express your ideas, and as such, it is the intent of the program that must be checked before feeding it to the compiler. It is not your job to do the work of a machine. It is the compiler's job to free you from checking syntax, types, and typos.
If the exam is in a language whose semantics are so dynamic that it will accept type-incorrect code, then you need to get it right under exam conditions. Or write unit tests with your pencil. Or something. I don't want to think about it, to be honest.
Slashdot Mirror
What did happen is businesses got more efficient.
This is a nice way of saying that most the remaining low-skill jobs got outsourced to other countries (the rest being in the process of transitioning to the "gig economy"), and all the remaining inconvenience and cost was pushed onto customers in the form of increased waiting times, hidden fees, self-service, etc.
FoundationDBKit?
Well that doesn't mean they're worthless to many. Regardless, a less misleading headline would be: "Apple is the 7000th vendor to open source their key value store after discovering the space is too crowded".
Burt Bacharach was wrong. What the world needs now is another distributed key value store.
Europe doesn't need to elect a next Mussolini because a Berlusconi is a more effective kleptocratic mismanager in the modern economy.
It probably depends what you mean by "project". I doubt that 90% of "failed" projects cost anywhere near the same as a "successful" project becaus most fail early.
I also suspect that most new pharmaceutical projects don't originate within a pharmaceutical company, but rather within academia.
Yes it is. Shame that few listened.
Ummm, North Korea isn't a communist state. There aren't any communist states, never were in modern history.
For the record, there aren't any capitalist states or libertarian states, either. Fundamentalism in any form is an inherently impractical form of government, especially in the modern era.
I remember when Agile was called XP. That was before the seminar and certification industry got hold of it.
XP made sense, but only by comparison with what came before it. If there's one thing I've learned after 25 years in this business, it's that fundamentalists make shitty products.
That depends on if you include the file system in the kernel.
NTFS is a driver that happens to be compiled in (because it makes it easier to boot the damn thing if nothing else), so you could go either way on that one
Besides, the good is the enemy of the perfect. If we can't do everything, we should do nothing.
FWIW, I mostly agree. NTOSKRNL is very well-designed kernel by 1990 standards, and it still mostly holds up today. The lack of a unified event system, for example (e.g. wait on a socket and a condvar), is one of the outstanding issues in Linux. (For completeness, FreeBSD's kqueue almost does the job, and libevent makes things a bit less painful.)
The main problem with Windows NT is not the kernel design, it's almost everything else.
IIRC, this issue was pretty much fixed since the advent of PCID. Also, it was never as much of an issue on ARM.
As I understand it, Apple was compelled to violate the basics of microkernels at many points, [...]
You understand wrong, but you're not alone. Apple broke the rules of 1990s microkernel fundamentalism, which are not the same as "basics of microkernels".
Has it occurred to you that Trump may be The Mule?
AOL failed because ...
Will Rogers never met Donald Trump.
That makes an odd kind of sense when you think about it.
It's probably not a common scenario, but it's certainly true for me some of the time. It may be possible to rearchitect a Big Simulation (where, say, touching every part of a 20G data structure in a quasi-random order every simulation step is normal) to have better locality for debugging purposes, but is it worth it just to avoid debugging over an SSH connection?
There is as yet insufficient data for a meaningful answer.
No one is 'fixing' a Java, C++ program installed on a production system with SSH and VI.
Correct. However, "fixing" a program installed on a test environment is common enough. This is especially true in the scientific computing space (which I currently work in) where the best cut-down test case that you can find in a hurry won't run on your laptop with only 16G RAM.
I can tell you that the IDE is not why we have so many Date classes.
If anything, it's the other way around. The main job of a modern IDE is to make working with badly-designed high-surface-area APIs less painful.
You are going for a computer science degree. You must be able to express your ideas on paper, a white board, napkin, back of your hand, ....anywhere.
I agree, but I'm going to make one proviso. Yes, programming exams on paper. But marks should not be taken off for trivial syntax errors or anything else that anyone with half a brain could easily fix when the compiler refuses to compile the code.
Your job, as a programmer or computer scientist, to express your ideas, and as such, it is the intent of the program that must be checked before feeding it to the compiler. It is not your job to do the work of a machine. It is the compiler's job to free you from checking syntax, types, and typos.
If the exam is in a language whose semantics are so dynamic that it will accept type-incorrect code, then you need to get it right under exam conditions. Or write unit tests with your pencil. Or something. I don't want to think about it, to be honest.
What "evangelical" seems to mean in the United States is that Calvin, Wesley, Spurgeon, etc were liberal wusses.
There are lots of things that look good to eat that aren't, and lots of things that are delicious that don't look good.
You take my Tide pods out of my cold, poisoned hands.
Yes. There is clear evidence that research causes cancer in rats.
Blinn's Law is never wrong.