The problem would arise if they start advocating a system that favors them particularly. Are they?
I dunno. Neither do you or anybody else, because nobody knows what kind of obscure little details they are burying in their arguments. And that's why their credibility and history matter.
Ham radio has become an obscure subculture not because people changed, but because it was obsoleted by technology. I see no indication that the same is happening to open source. Quite to the contrary.
I don't see why they shouldn't. It's obviously a big market and they obviously want to be in it.
And it's not like Apple's iPod, iTunes, or the iTunes music store are stunningly original either--Apple invented none of the key technologies or business models surrounding those products, they simply did a better job bundling, integrating, and marketing them. But we can expect that Apple, as usual, is going to try to play hardball by trying to entangle their competitors in patent and look-and-feel lawsuits.
If they were wildly successful in recent years geeks would complain.
With the kind of market power Microsoft wields comes responsibility, and they aren't living up to it yet. What kind of responsibility? Setting and documenting open standards, creating opportunities for competitors, making the entire market grow. Instead, Microsoft still operates like a small, aggressive start-up, out to kill anybody and everybody in their way.
Both IBM and Google are examples of successful, large companies that are behaving more responsibly. Microsoft is actually moving in that direction, but they still have a long way to go.
They'd rather anger the geeks than their investors.
That's a dangerous approach in the long term, because it's the geeks, not the investors, that ultimately support their business.
MS is always late the to party. Pioneers get the arrows. Settlers get the land.
You make it sound like some grand strategy; in fact, MS simply was lucky that being late to the party has worked for them in the past, mainly because they have been able to leverage their monopoly. That will stop working at some point.
It doesn't matter if Amazon is the worst patent abuser in the world...what Amazon has said should stand on its own merits.
Modifications to the patent system are highly complex, with difficult to predict implications. Therefore, it is impossible to assess its merits on its own or to determine whether Amazon has some hidden motives in making this proposal. That means that Amazon's history becomes an important part of evaluating their proposal. Since Amazon has proven by their past actions that they are attempting to abuse the patent and political process, anything coming from them needs to be viewed with suspicion.
The issue isn't Christian faith, it's the pope telling Hawking he shouldn't study the origins of the universe.
Don't go calling God petty and hateful just because you misunderstand Him.
I'm not calling God petty and hateful. I'm saying that the Catholic church's concept of God is one of a petty and hateful god. The implication of that statement is that the Catholic church does not represent God on earth and therefore is a fraud. Of course, that's a position protestants generally have subscribed to for centuries.
Catholics do not represent all Christians. You can't lump them in the same category as Fundamentalist Christians,
I didn't make a statement about all Christians, I made a statement about Catholicism.
There is nothing to stop a Catholic from investigating how God did it.
What a schizophrenic statement given the story that we are talking about; but, hey,
But, hey, we've made some progress: at least the pope doesn't have the power to torture and burn scientists at the stake anymore, and his armies don't go marauding across Europe anymore.
Geez, you talk like someone who programs by himself in his basement and fancies himself a good programmer. In the real world, memory management in C and C++ is a major issue on real-world software development projects: it has a profound effect on anything from API design to testing. Just the fact alone that errors in C/C++ are not isolated to the module they occurred in results in enormous increases in effort during testing and debugging. It's an even worse problem that people like you think that because they have somehow cobbled together a program with matching new/delete statements (most of the time), they have mastered memory management and are doing an efficient job; people like you don't even know how poor their memory management actually is in practice.
On top of that, being able to allocate and free memory when you want is part of the power of C/C++ meaning that it let's you work on the application's optimization process at a level that you simply can't control in Java.
That control is illusory; it means that C/C++ programmers often will choose inefficient storage management strategies and retain memory long after it has become garbage because doing things right is too hard for them.
Java is a hell of a language, just don't mistake the tools. C/C++ is for performance and scalability, not Java.
Actually, I think Java is a lousy language and I don't recommend you use it. But it got one thing right: it is a safe language with fault isolation that still has excellent performance for both memory management and compute-intensive tasks. Where Java fails is its inefficient I/O and inefficient standard libraries.
Regarding the typography, academic studies of both perceived readability (which font people prefer to look at) and actual readability (how fast and accurately people can read text in different fonts) have been carried out
Yes, and their results don't support your statements.
Likewise, there is a whole industry, albeit a relatively small one, dedicated to usability. Jakob Nielsen
Yes, and Nielsen couldn't back up his statements with actual facts if his life depended on it; usually, he takes a tiny factoid from some vaguely related experiment and completely overgeneralizes from it.
So no, this is not based on idle conjecture. My opinions have been formed after studying this field for several years, and the sources I'm most interested in are those that are not hypothetical but based on objective evidence.
Sorry, it doesn't work that way. Either you back up specific claims with specific evidence, or your statements are just hot air.
I note that you still have not provided references to back up your claims about typography.
Just think about what a pathetic concept of divinity that is: a supposedly almighty God who dislikes it when his creation looks at his works. That's in addition to all the smighting, shame, pain, and torture that Catholicism says God inflicts on the world.
I'm agnostic about whether there is some higher power. But a world created and ruled by the kind of schizophrenic and conflicted being that the Catholic church postulates makes no sense to me, and my faith tells me that they are wrong; no omnipotent being could sensibly be as petty and hateful towards mankind as the Catholic church claims God is.
No offense, but if you are getting faster Java code than C/C++ code then you are a terrible C/C++ programmer.
No, it's just that in the real world, there is a limited amount of time for developing a piece of software. Every minute I spend on worrying about memory allocation or tracking down a pointer bug is a minute I can't spend on putting in a more efficient data structure or profiling my code.
The best fonts are still hand-hinted by experienced typographers.
Is that merely your opinion, or do you have some concrete, experimental evidence to demonstrate that?
That's, in fact, a problem with lots of your statements and lots of the things people say about HCI: it's opinions, hunches, and anecdotal evidence, with little or no facts to back it up.
OS X wouldn't exist without open source software, and huge chunks other than the GUI in OS X are derived from open source software; yet, both Apple and NeXT have given back very little in return--even when the license forced them to open something, what they put up has often been completely useless to the original open source project.
Prior to OS X, Apple for years was shipping a clunky, single-tasking OS when other systems were already robust and multitasking, and at some point, Apple tried (and fortunately failed) at their attempts to shut out all other GUIs from the market.
I don't think Apple has ever been "in touch" with the tech community.
When companies impose weird intellectual property restrictions on their data sheets, then I'm all for making the process of getting the data sheets as cumbersome as possible--that way, FOSS developers will at least become aware that there is something funny going on.
Some other vendors hide a restrictive license ("if you look at this, we own stuff you do with it") somewhere in the documentation or behind a "Read This License" link, but people who look at the documentation never notice.
There is no obstacle to building bridges to Microsoft products on the open source side. In fact, a huge part of open source work is concerned with doing exactly that. The only obstacle is that Microsoft keeps protocols private, proprietary, and sometimes patented.
So, there really isn't any need for a "dialog". If Microsoft wants better open source support for Exchange, for SMB, for NTFS, for MS Office, for.NET, all they need to do is open up the specs, and the rest will happen automatically.
For each of the features you mention, there already exist similar implementations for Linux and I challenge you to demonstrate that there are unmet needs in any of these areas.
Note that just because Sun's marketing department says that ZFS or DTrace are superior to their Linux counterparts (or incorrectly states that they don't have any Linux counterparts) doesn't mean that they are right.
JITs amortize their overhead over other operations, so that, whatever you do, the JIT-related overhead is small. But that's just the amortized cost; you have to weigh against that the optimizations that JITs perform that no batch compiler can perform, and JITs often come out far ahead in that comparison.
The languages used may have changed, but the amount of (and use cases for) interpreted vs. native code hasn't changed that much over the decades. Shiny-new Java didn't change it, neither did.Net. Nor will Ruby on Rails. It's the same old song, covered by some fresh new 'hip' band.
If you lump together Java, C#, and interpreted languages, you really don't know what you're talking about. Java and C# are natively compiled languages, they just happen to perform compilation at runtime.
And for high performance computing, that's the future, because that's what's needed to adapt software to the specific hardware and data you have; batch compilers simply can't do that.
Java itself has lots of problems for high performance computing, but don't confuse a particular implementation with the approach. You'll likely see JIT-based Fortran implementations become widely used in the future and Fortran-specific JITs for high performance computing.
Ah, if it only were as simple as "Java is 5% slower than native". Unfortunately, performance can't be characterized that easily.
In real life, Java is as fast as C/C++ on microbenchmarks.
In real life, Java is also often slow as a pig, not because of the byte code or JIT, but because of the way class loading works, the way JNI works, library bloat, and lots and lots of other problems.
Java has given a bad name to JITs, but it has also demonstrated their potential.
Slashdot Mirror
The problem would arise if they start advocating a system that favors them particularly. Are they?
I dunno. Neither do you or anybody else, because nobody knows what kind of obscure little details they are burying in their arguments. And that's why their credibility and history matter.
Ham radio has become an obscure subculture not because people changed, but because it was obsoleted by technology. I see no indication that the same is happening to open source. Quite to the contrary.
I don't see why they shouldn't. It's obviously a big market and they obviously want to be in it.
And it's not like Apple's iPod, iTunes, or the iTunes music store are stunningly original either--Apple invented none of the key technologies or business models surrounding those products, they simply did a better job bundling, integrating, and marketing them. But we can expect that Apple, as usual, is going to try to play hardball by trying to entangle their competitors in patent and look-and-feel lawsuits.
If they were wildly successful in recent years geeks would complain.
With the kind of market power Microsoft wields comes responsibility, and they aren't living up to it yet. What kind of responsibility? Setting and documenting open standards, creating opportunities for competitors, making the entire market grow. Instead, Microsoft still operates like a small, aggressive start-up, out to kill anybody and everybody in their way.
Both IBM and Google are examples of successful, large companies that are behaving more responsibly. Microsoft is actually moving in that direction, but they still have a long way to go.
They'd rather anger the geeks than their investors.
That's a dangerous approach in the long term, because it's the geeks, not the investors, that ultimately support their business.
MS is always late the to party. Pioneers get the arrows. Settlers get the land.
You make it sound like some grand strategy; in fact, MS simply was lucky that being late to the party has worked for them in the past, mainly because they have been able to leverage their monopoly. That will stop working at some point.
Discoloration doesn't happen on the aluminum cases. Metal seems to be a good material for the wrist rests.
It doesn't matter if Amazon is the worst patent abuser in the world...what Amazon has said should stand on its own merits.
Modifications to the patent system are highly complex, with difficult to predict implications. Therefore, it is impossible to assess its merits on its own or to determine whether Amazon has some hidden motives in making this proposal. That means that Amazon's history becomes an important part of evaluating their proposal. Since Amazon has proven by their past actions that they are attempting to abuse the patent and political process, anything coming from them needs to be viewed with suspicion.
The statement that I quoted:
For someone who's so keen on justifying "specific claims with specific evidence", your criticisms of my argument here are remarkably troll-like.
It's quite clear what you sound like: a weasel.
Microsoft has held back the development of personal computing by at least a decade; only in recent years have they come close to the state of the art.
You grossly misunderstand the Christian faith.
The issue isn't Christian faith, it's the pope telling Hawking he shouldn't study the origins of the universe.
Don't go calling God petty and hateful just because you misunderstand Him.
I'm not calling God petty and hateful. I'm saying that the Catholic church's concept of God is one of a petty and hateful god. The implication of that statement is that the Catholic church does not represent God on earth and therefore is a fraud. Of course, that's a position protestants generally have subscribed to for centuries.
Catholics do not represent all Christians. You can't lump them in the same category as Fundamentalist Christians,
I didn't make a statement about all Christians, I made a statement about Catholicism.
There is nothing to stop a Catholic from investigating how God did it.
What a schizophrenic statement given the story that we are talking about; but, hey,
But, hey, we've made some progress: at least the pope doesn't have the power to torture and burn scientists at the stake anymore, and his armies don't go marauding across Europe anymore.
Geez, you talk like someone who programs by himself in his basement and fancies himself a good programmer. In the real world, memory management in C and C++ is a major issue on real-world software development projects: it has a profound effect on anything from API design to testing. Just the fact alone that errors in C/C++ are not isolated to the module they occurred in results in enormous increases in effort during testing and debugging. It's an even worse problem that people like you think that because they have somehow cobbled together a program with matching new/delete statements (most of the time), they have mastered memory management and are doing an efficient job; people like you don't even know how poor their memory management actually is in practice.
On top of that, being able to allocate and free memory when you want is part of the power of C/C++ meaning that it let's you work on the application's optimization process at a level that you simply can't control in Java.
That control is illusory; it means that C/C++ programmers often will choose inefficient storage management strategies and retain memory long after it has become garbage because doing things right is too hard for them.
Java is a hell of a language, just don't mistake the tools. C/C++ is for performance and scalability, not Java.
Actually, I think Java is a lousy language and I don't recommend you use it. But it got one thing right: it is a safe language with fault isolation that still has excellent performance for both memory management and compute-intensive tasks. Where Java fails is its inefficient I/O and inefficient standard libraries.
Regarding the typography, academic studies of both perceived readability (which font people prefer to look at) and actual readability (how fast and accurately people can read text in different fonts) have been carried out
Yes, and their results don't support your statements.
Likewise, there is a whole industry, albeit a relatively small one, dedicated to usability. Jakob Nielsen
Yes, and Nielsen couldn't back up his statements with actual facts if his life depended on it; usually, he takes a tiny factoid from some vaguely related experiment and completely overgeneralizes from it.
So no, this is not based on idle conjecture. My opinions have been formed after studying this field for several years, and the sources I'm most interested in are those that are not hypothetical but based on objective evidence.
Sorry, it doesn't work that way. Either you back up specific claims with specific evidence, or your statements are just hot air.
I note that you still have not provided references to back up your claims about typography.
Just think about what a pathetic concept of divinity that is: a supposedly almighty God who dislikes it when his creation looks at his works. That's in addition to all the smighting, shame, pain, and torture that Catholicism says God inflicts on the world.
I'm agnostic about whether there is some higher power. But a world created and ruled by the kind of schizophrenic and conflicted being that the Catholic church postulates makes no sense to me, and my faith tells me that they are wrong; no omnipotent being could sensibly be as petty and hateful towards mankind as the Catholic church claims God is.
No offense, but if you are getting faster Java code than C/C++ code then you are a terrible C/C++ programmer.
No, it's just that in the real world, there is a limited amount of time for developing a piece of software. Every minute I spend on worrying about memory allocation or tracking down a pointer bug is a minute I can't spend on putting in a more efficient data structure or profiling my code.
He or she was talking about games. Does anyone really write games in anything but C or C++?
Most games actually are not "written in C or C++", they are written in a game engine or scripting language.
The best fonts are still hand-hinted by experienced typographers.
Is that merely your opinion, or do you have some concrete, experimental evidence to demonstrate that?
That's, in fact, a problem with lots of your statements and lots of the things people say about HCI: it's opinions, hunches, and anecdotal evidence, with little or no facts to back it up.
OS X wouldn't exist without open source software, and huge chunks other than the GUI in OS X are derived from open source software; yet, both Apple and NeXT have given back very little in return--even when the license forced them to open something, what they put up has often been completely useless to the original open source project.
Prior to OS X, Apple for years was shipping a clunky, single-tasking OS when other systems were already robust and multitasking, and at some point, Apple tried (and fortunately failed) at their attempts to shut out all other GUIs from the market.
I don't think Apple has ever been "in touch" with the tech community.
When companies impose weird intellectual property restrictions on their data sheets, then I'm all for making the process of getting the data sheets as cumbersome as possible--that way, FOSS developers will at least become aware that there is something funny going on.
Some other vendors hide a restrictive license ("if you look at this, we own stuff you do with it") somewhere in the documentation or behind a "Read This License" link, but people who look at the documentation never notice.
There is no obstacle to building bridges to Microsoft products on the open source side. In fact, a huge part of open source work is concerned with doing exactly that. The only obstacle is that Microsoft keeps protocols private, proprietary, and sometimes patented.
.NET, all they need to do is open up the specs, and the rest will happen automatically.
So, there really isn't any need for a "dialog". If Microsoft wants better open source support for Exchange, for SMB, for NTFS, for MS Office, for
with Sun's Kernel and everything else that Solaris10 brings to the table, why would you use any of the Linux Variants?
Well, obviously, millions of Linux users disagree.
Solaris *IS* better than Linux.
"Better" in what sense?
As for me, running AMD64 using Solaris AND Sun Hardware is the way to go.
How nice for you. But you are in a tiny minority.
For each of the features you mention, there already exist similar implementations for Linux and I challenge you to demonstrate that there are unmet needs in any of these areas.
Note that just because Sun's marketing department says that ZFS or DTrace are superior to their Linux counterparts (or incorrectly states that they don't have any Linux counterparts) doesn't mean that they are right.
JITs amortize their overhead over other operations, so that, whatever you do, the JIT-related overhead is small. But that's just the amortized cost; you have to weigh against that the optimizations that JITs perform that no batch compiler can perform, and JITs often come out far ahead in that comparison.
The languages used may have changed, but the amount of (and use cases for) interpreted vs. native code hasn't changed that much over the decades. Shiny-new Java didn't change it, neither did .Net. Nor will Ruby on Rails. It's the same old song, covered by some fresh new 'hip' band.
If you lump together Java, C#, and interpreted languages, you really don't know what you're talking about. Java and C# are natively compiled languages, they just happen to perform compilation at runtime.
And for high performance computing, that's the future, because that's what's needed to adapt software to the specific hardware and data you have; batch compilers simply can't do that.
Java itself has lots of problems for high performance computing, but don't confuse a particular implementation with the approach. You'll likely see JIT-based Fortran implementations become widely used in the future and Fortran-specific JITs for high performance computing.
Ah, if it only were as simple as "Java is 5% slower than native". Unfortunately, performance can't be characterized that easily.
In real life, Java is as fast as C/C++ on microbenchmarks.
In real life, Java is also often slow as a pig, not because of the byte code or JIT, but because of the way class loading works, the way JNI works, library bloat, and lots and lots of other problems.
Java has given a bad name to JITs, but it has also demonstrated their potential.