Actually, the iPhone has all of those except tactile keyboard, copy/paste, tethering, MMS, and Ogg support
AFAIK, the iPhone is getting some Apple-proprietary tethering with carrier restrictions. For 1/3 the price, you can get smart phones with no restrictions at all, and whose tethering consists of a standard 3G modem.
I think Sun has spent way too much money and effort on Java. The result has been a lot of bloated, badly designed libraries. If Oracle cuts that back, I think Java will actually improve technically.
If people can reasonably expect "wikipediaart.org" to be a site run by the same people as Wikipedia and concerned with art, then it's a trademark violation (it seems to me that it is).
Furthermore, Wikipedia has no choice in the matter: if it could conceivably be a trademark violation, they must get active against it.
Hollywood, in general, tends to support the left more than the right
Hollywood is socially liberal, but not when it comes to big business.
Consequently, my guess would be that the nominee would be someone who tends to favor Hollywood's interest
Democrats tend to take the public interest and public good into account more than Republicans. Hence, in picking a copyright czar, they are more likely to pick one that maintains shorter copyright terms and more public domain and fair use rights.
How so? I didn't realize copyright law enforcement was a particularly partisan issue in the United States
Democrats have a tendency to take the public good more into account than commercial rights. Therefore, a "copyright czar" appointed by a Democrat is more likely to balance public good and commercial interests better than a Republican.
That doesn't mean that there's a problem with Java or JIT compilation in general, it just means that someone implemented their JIT badly or that the JIT is coupled to a really bad interpreter that's slowing it down.
Neither party can afford to upset big, powerful industries, so don't expect any strong language from the president on copyright or media.
On balance, I suspect we're still better off with whoever Obama finds than with whatever a Republican president would have done and who he would have selected.
JIT-compiled Java is nearly as fast as C for inner loops. So, the sense in which JamVM "can be faster than JIT compiling" must be some sense of "faster" that I am not familiar with.
I'm glad Google trimmed some of the fat out of Java; bloat is one of the reasons Java has failed to go mainstream for desktop applications. The JVM and the JNI also were badly designed, and Dalvik improves on them.
If we're really lucky, Oracle will deprecate 90% of Java SE (since it's open source, you can still use it if you like).
Java is currently merely available under an open source license, as part of a dual licensing scheme. For example, it's still very difficult to fork the code base and Sun retains special rights (such as the right to license under non-open-source licenses). As a consequence, the project isn't run as an open source project.
(And if you want to argue that this is good for Java, no, it is not. The runtime and library designs and the codebase are both poor quality.)
site containing graphic imagery of aborted foetuses,
I don't understand why that would be subject to censorship. True, it's disturbing and it is motivated by a political agenda, but it doesn't seem to fall into any of the prohibited categories (violence, child pornography).
This makes a lot of sense: Sun is mostly about Java these days, but they haven't figured out hot to monetize Java. IBM, on the other hand, is making quite a bit of money with Java.
Sun has been running Java into the ground slowly. Hopefully, IBM can put Java on the right track again: fully open source it, fix its performance problems, provide better native interfaces, provide better integration with Linux, enable interoperability with Mono/.NET, etc.
Safari started from zero, so the first day, it had an infinite growth rate.
Comparing growth rates is meaningless for products that have such disparate market shares.
And 0.5%/day isn't even particularly impressive; that means it's going to grow six-fold in a year. That's singularly unimpressive for a newly released product.
On the other hand, IE and Firefox can't grow six fold in a year because they'd have more than 100% market share then.
Its still not as easy or elegant, and the compiler cannot optimize as aggressively
The compiler doesn't have to optimize as aggressively in OOLs because anything that needs to be high performance is written using traditional loops. Conversely, even the best functional programming language compilers aren't good enough to reliably compile functional control abstractions down to the same level of efficiency as loops.
and some stuff will take a whole lot more code.
Yes, it will. And it would be nice if someone actually produced a decent, high performance functional programming language. F# is not it. The only saving grace for F# is that you can drop down into C# or C++ if you need it to go fast.
Its just the mathematical model (not model as in a DSL model, but really, as in a math one) is extensive, so the formulas are very, very complex, even when done on paper.
That sounds like bullshitting to me. Complicated mathematical models generally do not require complicated programming constructs to implement. In fact, C#/F# has such a dearth of pre-existing mathematical libraries for it, that it is a poor choice for anything particularly mathematical.
One of primary reasons to use a functional language (aside that it goes from paper to code easier) is that its easier to prove the code afterward, which is required by law in some businesses.
The hard part in proving implementations of complicated mathematical models correct is not proving the control structures correct (which tend to be simple), it's proving the numerics correct.
Spectrum is a scarce resource, and to make the best use of it, it should go to companies willing to pay the largest amount of money for it. Otherwise, you're going to have companies hogging valuable spectrum with outdated technologies, and still charge whatever they can get away with.
As for price increases, T-Mobile has 32 million customers. $550 million/year in license fees ends up being less than $2/month/customer. The sky isn't falling.
And we need some way to pay the debts that GWB has racked up.
In my mind, the difficulty is ensuring that code marked "free" (to modify, redistribute, sell, and so on) isn't polluted by code involving patents
Free software, like proprietary software, inadvertently violates patents; that's just a fact and it's unavoidable. But so what? The worst that seems to happen in practice is that a judge orders the patented invention to be removed and people add a workaround.
If by "pollution" you mean "deliberate introduction", that's even less of a problem: source code is usually tracked openly. If the patent holder or one of their minions introduces a patented invention, they automatically donate them to the project under open source licenses (thank you). Third parties usually have no motivation to introduce someone else's patented inventions into an open source project. If they do, it's no worse than an accidental introduction.
And for situations like Tom Tom, it really doesn't matter either: Microsoft would have tried to screw Tom Tom whether they use Linux or whether they use some proprietary embedded operating system, and they don't need Linux-related patents to do it. The only way for Tom Tom not to get sued by Microsoft would have been to sell their soul to Microsoft from the start, pay hefty licensing fees, and build their systems on Windows CE.
"That said, I didn't use it much. What I can say, is we have an army of PhDs implementing -extremely- complex algorithms here (functions being passed around with several douzens levels of nested function types, to do very, very complex modeling in a couple lines of code...)"
If you do OOP, you're effectively doing the same thing when you're passing objects around: every method and every method of every class type that those methods work on is a "level of nested function type".
Also, it sounds like they're effectively using F# to build little domain specific languages. Functional programming languages are convenient for that purpose, but the end product tends to be worse than if you had actually designed a domain specific language from scratch.
Perhaps this is a sign that Redmond is finally starting to focus on being really good at a focused area instead of generally mediocre (or worse) at a huge number of things.
So, you are saying that Microsoft should become a computer vision company? I don't think so.
Besides, little of what they are showing is unique to Microsoft; lots of other companies and research groups have shown stitching and augmented reality.
I have always wondered whether Google Maps uses similar technology.
Image stitching has been around for a long time, and Photosynth is based on a lot of technology and research developed elsewhere.
Photosynth caught on because (1) there are lots of images to stitch now, (2) because they did a good engineering job stitching images that were taken under different conditions, and (3) they did a good engineering job on the UI.
In two patents, Microsoft basically tries to claim rights to running a general purpose OS on a computer designed for a car and having Internet access on such a machine. This is trying to patent a market niche.
In two other patents, they are trying to claim rights to the awful long/short filenames compatibility hack in FAT file systems. One patent is trying to claim allocating space from flash erasable memory in blocks. And the last patent is related to modes in user interfaces. All of these are trying to patent what any competent software developer would come up with when faced with such a programming task.
I hope Microsoft will be shredded to pieces in court.
Slashdot Mirror
Actually, the iPhone has all of those except tactile keyboard, copy/paste, tethering, MMS, and Ogg support
AFAIK, the iPhone is getting some Apple-proprietary tethering with carrier restrictions. For 1/3 the price, you can get smart phones with no restrictions at all, and whose tethering consists of a standard 3G modem.
I think Sun has spent way too much money and effort on Java. The result has been a lot of bloated, badly designed libraries. If Oracle cuts that back, I think Java will actually improve technically.
If people can reasonably expect "wikipediaart.org" to be a site run by the same people as Wikipedia and concerned with art, then it's a trademark violation (it seems to me that it is).
Furthermore, Wikipedia has no choice in the matter: if it could conceivably be a trademark violation, they must get active against it.
Hollywood, in general, tends to support the left more than the right
Hollywood is socially liberal, but not when it comes to big business.
Consequently, my guess would be that the nominee would be someone who tends to favor Hollywood's interest
Democrats tend to take the public interest and public good into account more than Republicans. Hence, in picking a copyright czar, they are more likely to pick one that maintains shorter copyright terms and more public domain and fair use rights.
How so? I didn't realize copyright law enforcement was a particularly partisan issue in the United States
Democrats have a tendency to take the public good more into account than commercial rights. Therefore, a "copyright czar" appointed by a Democrat is more likely to balance public good and commercial interests better than a Republican.
That doesn't mean that there's a problem with Java or JIT compilation in general, it just means that someone implemented their JIT badly or that the JIT is coupled to a really bad interpreter that's slowing it down.
Neither party can afford to upset big, powerful industries, so don't expect any strong language from the president on copyright or media.
On balance, I suspect we're still better off with whoever Obama finds than with whatever a Republican president would have done and who he would have selected.
JIT-compiled Java is nearly as fast as C for inner loops. So, the sense in which JamVM "can be faster than JIT compiling" must be some sense of "faster" that I am not familiar with.
I'm glad Google trimmed some of the fat out of Java; bloat is one of the reasons Java has failed to go mainstream for desktop applications. The JVM and the JNI also were badly designed, and Dalvik improves on them.
If we're really lucky, Oracle will deprecate 90% of Java SE (since it's open source, you can still use it if you like).
Don't let the door hit you in the behind on your way out.
What do you mean "fully open source" it?
Java is currently merely available under an open source license, as part of a dual licensing scheme. For example, it's still very difficult to fork the code base and Sun retains special rights (such as the right to license under non-open-source licenses). As a consequence, the project isn't run as an open source project.
(And if you want to argue that this is good for Java, no, it is not. The runtime and library designs and the codebase are both poor quality.)
Shell's "investment" in non-oil energy sources seems to have been miniscule anyway; this is merely bringing marketing in line with reality.
Does this no.make business sense? If they operate with a view on quarterly or annual stock performance, probably yes. Long term? Probably not.
site containing graphic imagery of aborted foetuses,
I don't understand why that would be subject to censorship. True, it's disturbing and it is motivated by a political agenda, but it doesn't seem to fall into any of the prohibited categories (violence, child pornography).
Perhaps, but I think that IBM would be getting one hell of a sweet deal
I don't think so. Sun's core server and OS business is in deep trouble, and Java is under threat as well.
This makes a lot of sense: Sun is mostly about Java these days, but they haven't figured out hot to monetize Java. IBM, on the other hand, is making quite a bit of money with Java.
Sun has been running Java into the ground slowly. Hopefully, IBM can put Java on the right track again: fully open source it, fix its performance problems, provide better native interfaces, provide better integration with Linux, enable interoperability with Mono/.NET, etc.
Safari started from zero, so the first day, it had an infinite growth rate.
Comparing growth rates is meaningless for products that have such disparate market shares.
And 0.5%/day isn't even particularly impressive; that means it's going to grow six-fold in a year. That's singularly unimpressive for a newly released product.
On the other hand, IE and Firefox can't grow six fold in a year because they'd have more than 100% market share then.
Its still not as easy or elegant, and the compiler cannot optimize as aggressively
The compiler doesn't have to optimize as aggressively in OOLs because anything that needs to be high performance is written using traditional loops. Conversely, even the best functional programming language compilers aren't good enough to reliably compile functional control abstractions down to the same level of efficiency as loops.
and some stuff will take a whole lot more code.
Yes, it will. And it would be nice if someone actually produced a decent, high performance functional programming language. F# is not it. The only saving grace for F# is that you can drop down into C# or C++ if you need it to go fast.
Its just the mathematical model (not model as in a DSL model, but really, as in a math one) is extensive, so the formulas are very, very complex, even when done on paper.
That sounds like bullshitting to me. Complicated mathematical models generally do not require complicated programming constructs to implement. In fact, C#/F# has such a dearth of pre-existing mathematical libraries for it, that it is a poor choice for anything particularly mathematical.
One of primary reasons to use a functional language (aside that it goes from paper to code easier) is that its easier to prove the code afterward, which is required by law in some businesses.
The hard part in proving implementations of complicated mathematical models correct is not proving the control structures correct (which tend to be simple), it's proving the numerics correct.
Spectrum is a scarce resource, and to make the best use of it, it should go to companies willing to pay the largest amount of money for it. Otherwise, you're going to have companies hogging valuable spectrum with outdated technologies, and still charge whatever they can get away with.
As for price increases, T-Mobile has 32 million customers. $550 million/year in license fees ends up being less than $2/month/customer. The sky isn't falling.
And we need some way to pay the debts that GWB has racked up.
In my mind, the difficulty is ensuring that code marked "free" (to modify, redistribute, sell, and so on) isn't polluted by code involving patents
Free software, like proprietary software, inadvertently violates patents; that's just a fact and it's unavoidable. But so what? The worst that seems to happen in practice is that a judge orders the patented invention to be removed and people add a workaround.
If by "pollution" you mean "deliberate introduction", that's even less of a problem: source code is usually tracked openly. If the patent holder or one of their minions introduces a patented invention, they automatically donate them to the project under open source licenses (thank you). Third parties usually have no motivation to introduce someone else's patented inventions into an open source project. If they do, it's no worse than an accidental introduction.
And for situations like Tom Tom, it really doesn't matter either: Microsoft would have tried to screw Tom Tom whether they use Linux or whether they use some proprietary embedded operating system, and they don't need Linux-related patents to do it. The only way for Tom Tom not to get sued by Microsoft would have been to sell their soul to Microsoft from the start, pay hefty licensing fees, and build their systems on Windows CE.
"That said, I didn't use it much. What I can say, is we have an army of PhDs implementing -extremely- complex algorithms here (functions being passed around with several douzens levels of nested function types, to do very, very complex modeling in a couple lines of code...)"
If you do OOP, you're effectively doing the same thing when you're passing objects around: every method and every method of every class type that those methods work on is a "level of nested function type".
Also, it sounds like they're effectively using F# to build little domain specific languages. Functional programming languages are convenient for that purpose, but the end product tends to be worse than if you had actually designed a domain specific language from scratch.
Perhaps this is a sign that Redmond is finally starting to focus on being really good at a focused area instead of generally mediocre (or worse) at a huge number of things.
So, you are saying that Microsoft should become a computer vision company? I don't think so.
Besides, little of what they are showing is unique to Microsoft; lots of other companies and research groups have shown stitching and augmented reality.
I have always wondered whether Google Maps uses similar technology.
Image stitching has been around for a long time, and Photosynth is based on a lot of technology and research developed elsewhere.
Photosynth caught on because (1) there are lots of images to stitch now, (2) because they did a good engineering job stitching images that were taken under different conditions, and (3) they did a good engineering job on the UI.
The patents in question are idiotic.
In two patents, Microsoft basically tries to claim rights to running a general purpose OS on a computer designed for a car and having Internet access on such a machine. This is trying to patent a market niche.
In two other patents, they are trying to claim rights to the awful long/short filenames compatibility hack in FAT file systems. One patent is trying to claim allocating space from flash erasable memory in blocks. And the last patent is related to modes in user interfaces. All of these are trying to patent what any competent software developer would come up with when faced with such a programming task.
I hope Microsoft will be shredded to pieces in court.
Get them approved in writing by senior management. If you don't, it really does look suspicious.