As an example, the JDO implementation I use does take advantage of Oracle sequences if the target is an Oracle database. My application code doesn't know anything about that though, so when the target database is DB2, no code change is required.
This is a good thing, because our customers dictate what database our product runs on - not us.
CLR is a very good, thorough, book. (The book weighs about 100 lbs). For example, I've been working with a lot of graph theory lately, and CLR had some good information about relaxed heaps, fibonacci heaps, and even pointers to Ahuja's "Faster algorithms for the shortest path problem."
It looks like we're just about to hire an employee who has no college degree. He does, however, have decades of experience in the industry from microcontroller programming to webapp development. He also has a very strong drive to continously learn.
All that said, he will probably get a lesser salary than what he would have, had he obtained an applicable college degree.
On a side note, we're still looking to fill another position or two, so if you think you have what it takes, check out the job description.
I think the hardest part is trying to write code that is friendly to the host CPU's branch predictor.
This is where dynamic recompilation comes in. When you dynamically recompile the target binary, instead of interpreting it, you can remove all of the unpredictable branches you get when you interpret. Dynamic recompilation allows your code to be clean and fast.
I wrote a basic dynamic recompilation core for a Z80 emulator on the JVM.
I'd mod you up Insightful, but I just had to respond.
I have a road network pathing algorithm that runs approximately 3x faster on an AMD system than an equivalent-priced Intel system. The algorithm spends most of its time reading memory (the road network) and writing memory (data structures), so my gut feeling was that the AMD chip wasn't suffering as badly from stalls, and the memory subsystem was performing much better. I'm just blown away by the performance difference in identically priced machines.
Matching "Wolverine" is easy but how about: WeaponX, Wolvie, Woolvrine, Wepon Ex, etc?
Actually, as long as the words are phonetically similar (such as your Woolvrine/Wolverine and WeaponX/Wepon Ex), they can be trivially matched with the public domain SoundEx algorithm. It's what my team uses for geocoding, so we can find mistakes when people type "Mane St" instead of "Main St" for an address.
... to judge if a homosexual way of life is better or worst thant a Christian way of life?
A human being with free will. You're equally free to believe whatever you want.
In a free country, children would be presented both wolrd views
(You're obviously not a parent). You don't find your own statement hypocritical? You don't want Christians to legislate morality (neither do they), and yet you want to legislate morality yourself. Nice.
In the US gay bashing is a national sport
No. In the US, people have the right to exercise free speech, and Christians can freely say they believe the homosexual lifestyle is immoral. People who choose to live a homosexual lifestyle are free to go around calling Christians a bunch of moral prudes. Isn't that great?
We have a president that just got reelected primarily because he is overtly Christian and has made a huge effort to spread that religion using tax dollars and government influence in direct conflict with the constitution.
Proof? If you're talking about the faith-based initiatives, there are very strict rules about how government funds must be used, guaranteeing that the funds don't have any religious ties.
Here in the USA on planet Earth, Christians expend an enormous amount of time, energy and money forcing their religion on others through all sorts of avenues.
Really. And how exactly is this religion forced on you? Apparently, you must be quite powerful to resist the strong pressure they can exert.
All factions of Christianity spend hundreds of millions of tax-free dollars a year on missions -- domestic and abroad -- seeking to convert people to Christianity.
Christians demonstrate their love by caring for people in a material way (food, clothing, housing, etc...) and this is bad, because? Nobody is inducing anybody to do anything.
And, please show me some proof that public schools are forcing children to read the Quran.
Just do a Google search on "California student Koran Sekulow".
While you are at it, do you have proof that children are not being allowed to read the Bible during their free time at school? Are teachers confiscating Bibles?
Sure. Just look at all of the lawsuits which have been won by Jay Sekulow. They are public record.
The U.S. is not a theocracy like Iran, yet. But that's not for the lack of trying on the part of Christians.
Please. Christians don't want a theocracy anymore than any other group in the US wants a theocracy.
By the way, it's really difficult to listen to Christians whine as though they are the persecuted underdogs of our society when, in fact, they are majority and hold all the power.
And what power would that be? I know - we must have some sort of mystical mind control power we exercise over everyone.
I know many Christians like to pretend that the awful Hollywood liberals are some huge, menacing power, but it's just not the case.
Or rather, we see most of Hollywood deriding Christian values. (The reason why Gibson's Passion of The Christ received so much media attention). You don't see Christian's going out and attempting to get laws enacted against movies which glorify premarital sex, violence, the homosexual lifestyle, etc... We do what every other special interest group does - avoid the material, and possibly boycott the producers.
The theme here, is that we respect the rights of others to choose to do as they will. If I don't like what you're doing, I can show my displeasure by choosing not to patronize/support you. Your rights extend all the way up to the point to where they start harming others. Contrast this to other groups. For example, gay organizations which are lobbying to create laws which would force schools to teach children that the homosexual lifestyle is good. That, unfortunately, is a very scary reality that we've almost come to.
You don't see banning abortion/gay marriage etc. as a problem because it fits right into your beliefs.
I don't see banning abortion as a problem, because it's a simple human rights issue. The right to live trumps another person's right to convenience. It's as simple as that.
I don't see banning "gay marriage" as a problem, because marriage has a very specific definition and meaning. Call "gay marriage" what you want - civil union - whatever, but it sure isn't marriage. These people are free to do whatever they want - cohabitate, whatever. They are grownups after all. I'm just not going to personally support their behavior with my tax dollars.
If the last thing a Christian wants to do is force his religion on someone else, then who or what is in the White House?
Right, because this nation has never had president's that acknowledge the Christian God.
George Washington:
By 1778, George Washington had so often witnessed God's intervention that on August 20, he wrote Thomas Nelson that:
The Hand of providence has been so conspicuous in all this, that he must be worse than an infidel that lacks faith, and more than wicked, that has not gratitude enough to acknowledge his obligations.
Abraham Lincoln:
"When any church will inscribe over its altar, as its sole qualification for membership, the Savior's condensed statement of the substance of both law and Gospel, 'Thou shalt love the lord thy God with all thy heart, and with all thy soul and thy neighbor as thyself' that church will I join with all my heart and all my soul."
Theodore Roosevelt: "The true Christian is the true citizen."
Those of course only reflect the first few presidents I did a Google search on. I think you'll find that contemporaries like Reagan and Bush are in good company. No president has ever forced their faith on anyone, though they did publicly announce their own personal faith. Presidents have been driven by their faith, to uphold basic human rights. For example, Abraham Lincoln waged war over slavery. George Bush has worked to pass legislature protecting the rights of the unborn.
Re:Which is really surprising to me
on
Decompiling Java
·
· Score: 1
who, as a compiler hacker, would have expected an optimization pass to transform the first form into the second form before generating the bytecode
Almost all of the Java compilers out there do little to no optimizations while generating bytecode. Rather, the focus of the compilers is to generate bytecode that is easily understood and optimized by the virtual machine.
Since the bytecode for Java is a well-defined standard, there are many more bytecode processors for Java compared to say object-code processors for C or C++. So, a non-optimizing compiler also has the side-effect of making it easier for all of the people writing bytecode processors.
I keep hearing this sentiment expressed on Slashdot, yet the last thing a Christian wants to do is force his religion on someone else. Our forefathers fled Britain because of religious persecution, and our constitution grants freedom of religion.
In this day and age, students in public schools are forced to read from the Quran and pray Muslim prayers, but are not allowed to read from the Bible. It strikes me strange that people can use the term "Christian theocracy" with any sort of straight face.
I find this particularly funny, because the exit polls show that Bush received more votes from people who had a high-school education or college degree, while Kerry received more votes from people without a high-school degree.
My project, Jace, is one of the tier 1 projects. It'll probably be one of the simpler ones to port, because it was already written with portability in mind. I was all "WOO HOO" when I found out I was getting an iPod Mini. Then I was all "DARN" when I realized I couldn't port my own project and win a G5.;)
On a more serious note, almost all C++ applications prefer std::vector to std::list, and almost all Java applications prefer ArrayList over LinkedList - neither of which require any pointers.
These data structures tend to perform much better, because they are naturally cache friendly. It can take over 100,000 elements in a list to make it more expensive to do an insert or a remove in an array-based list over a link-based list.
In a language like C++, you can try to mitigate these factors by using a special-purpose allocator that greedily reserves large pools of contiguous memory to improve spatial locality.
Re:varargs is *not* an enhancement
on
Java 1.5 vs C#
·
· Score: 1
varargs is *not* an enhancement
Although I'm not incredibly fond of varargs, you mischaracterize them greatly. Varargs in Java are entirely type safe. For example take,
void foo( String... values ) {
}
The compiler will only allow me to call that function with a list of Strings, which then gets boxed as a String[] on its way into foo.
There is no equivalent to varargs in C++. When a function uses an ellipsis, it has absolutely no idea what data is actually being passed to it. The compiler does no enforcing of types upon invocation, and you have no type information at runtime. If the programmer goofed, you are fubar.
No, he's wrong. The Java 1.5 compiler stores the generic type attributes in the class file in what are called "Signature attributes". Generic types are perfectly safe across compilation boundaries. Where they aren't safe is when they are combined with raw types. That behavior is by design, and it is the tradeoff paid for having mostly seamless compatibility with an existing codebase. Whether that was an appropriate decision is an entirely different discussion.
Heh... I worked with images of Lenna in my DSP classes while working on my CmpE at Georgia Tech many moons ago. This article seems to disagree with you about the image being a good benchmark.
If you dislike Sun for just owning software patents, you should dislike just about every tech company that exists. Practically every company nowadays owns some bogus set of software patents, and companies like IBM (who regularly receives the most patent grants worldwide every year), own huge software patent portfolios. Fortunately, most of these companies (including Sun) don't go around abusing these bogus patents left and right in billion dollar lawsuits.
Companies are never going to stop patenting software as long as it can be done, and there will always be rotten apples abusing the system, so we need to stop this problem at its roots. New case law needs to be arrived at, or new legislation needs to be passed, to kill software patents.
No, for example, I listed the change that the old synthetic attributes have now changed to synthetic access specifers, so anything that checks for conforming 1.4 class files would barf on the goofed access specifier (gcj did this up until recently). Same goes for the additional varargs specifier. There are also new JVM changes that will happen soon that will increase things like the limits on the constant pool size, the max amount of bytecode in a method, etc... If you generate a class that goes over the original limits, you're also SOL.
Do you have a reference to a specification of the changes involved?
No, autoboxing suffers an entirely new set of problems. For example, let's take a container, which is the most commom usage case of generics.
You will be forced to box and unbox upon every single access to that container. Boxing can be quite expensive, compared to simple operations, because it will usually involve a memory allocation - which can suffer even more under multiprocessor boxes due to heap contention.
Your container will use a much larger amount of memory compared to an array of primitives. Compare an int[] to a List<Integer>. The int[] will cost you 4*N bytes. The List<Integer> will cost you 16*N bytes - 4*N for the Object[], 8*N for the two-word object header for each Integer, and 4*N for the actual int value stored in each Integer.
For systems like mine, that store 100 million objects in memory, a 4X memory increase is hugely unacceptable, and the access penalty is also unacceptable. My situation is not uncommon, and it is the reason why libraries like fastutil exist and are so popular.
I think.NET is mostly a rip-off of the JVM with very little innovation, but they seem to have a much better approach to primitives, with JIT type-specialization
I have my own fair share of criticisms to level against Java Generics (no support for primitives, no type reification which leads to bastardisms like Class<T> and limitations like no new T[], etc...), but the cost of a cast has to be the worst criticism I've heard yet to date. A cast is implemented as a single VM instruction (checkcast) which generally takes nanoseconds to execute.
It's most definitely not identical. There are several changes that went into the classfile/bytecode format for 1.5. Just a few examples off the top of my head:
new Signature attributes (for generics)
new Varargs access specifier
the JVM LDC instruction now supports class literals
new Annotation attributes
New Synthetic access specifiers replace old Synthetic attributes
Above and beyond the bytecode/classfile format changes, using -target 1.5 will cause the compiler to automatically take advantage of new code that is only available in the 1.5 runtime. For example, enums depend upon java.lang.Enum, the new for loop depends upon java.lang.Iterable, the compiler supports string concatenation using the new StringBuilder class instead of StringBuffer, etc...
Retroweaver takes care of all of that, and was even given a thumbs up by the 1.5 compiler writer, Neal Gafter.
Slashdot Mirror
As an example, the JDO implementation I use does take advantage of Oracle sequences if the target is an Oracle database. My application code doesn't know anything about that though, so when the target database is DB2, no code change is required.
This is a good thing, because our customers dictate what database our product runs on - not us.
CLR is a very good, thorough, book. (The book weighs about 100 lbs). For example, I've been working with a lot of graph theory lately, and CLR had some good information about relaxed heaps, fibonacci heaps, and even pointers to Ahuja's "Faster algorithms for the shortest path problem."
It looks like we're just about to hire an employee who has no college degree. He does, however, have decades of experience in the industry from microcontroller programming to webapp development. He also has a very strong drive to continously learn.
All that said, he will probably get a lesser salary than what he would have, had he obtained an applicable college degree.
On a side note, we're still looking to fill another position or two, so if you think you have what it takes, check out the job description.
This is where dynamic recompilation comes in. When you dynamically recompile the target binary, instead of interpreting it, you can remove all of the unpredictable branches you get when you interpret. Dynamic recompilation allows your code to be clean and fast.
I wrote a basic dynamic recompilation core for a Z80 emulator on the JVM.
I'd mod you up Insightful, but I just had to respond.
I have a road network pathing algorithm that runs approximately 3x faster on an AMD system than an equivalent-priced Intel system. The algorithm spends most of its time reading memory (the road network) and writing memory (data structures), so my gut feeling was that the AMD chip wasn't suffering as badly from stalls, and the memory subsystem was performing much better. I'm just blown away by the performance difference in identically priced machines.
Actually, as long as the words are phonetically similar (such as your Woolvrine/Wolverine and WeaponX/Wepon Ex), they can be trivially matched with the public domain SoundEx algorithm. It's what my team uses for geocoding, so we can find mistakes when people type "Mane St" instead of "Main St" for an address.
A human being with free will. You're equally free to believe whatever you want.
(You're obviously not a parent). You don't find your own statement hypocritical? You don't want Christians to legislate morality (neither do they), and yet you want to legislate morality yourself. Nice.
No. In the US, people have the right to exercise free speech, and Christians can freely say they believe the homosexual lifestyle is immoral. People who choose to live a homosexual lifestyle are free to go around calling Christians a bunch of moral prudes. Isn't that great?
Christians demonstrate their love by caring for people in a material way (food, clothing, housing, etc...) and this is bad, because? Nobody is inducing anybody to do anything.
Just do a Google search on "California student Koran Sekulow".Sure. Just look at all of the lawsuits which have been won by Jay Sekulow. They are public record.
Please. Christians don't want a theocracy anymore than any other group in the US wants a theocracy.
And what power would that be? I know - we must have some sort of mystical mind control power we exercise over everyone.
Or rather, we see most of Hollywood deriding Christian values. (The reason why Gibson's Passion of The Christ received so much media attention). You don't see Christian's going out and attempting to get laws enacted against movies which glorify premarital sex, violence, the homosexual lifestyle, etc... We do what every other special interest group does - avoid the material, and possibly boycott the producers.
The theme here, is that we respect the rights of others to choose to do as they will. If I don't like what you're doing, I can show my displeasure by choosing not to patronize/support you. Your rights extend all the way up to the point to where they start harming others. Contrast this to other groups. For example, gay organizations which are lobbying to create laws which would force schools to teach children that the homosexual lifestyle is good. That, unfortunately, is a very scary reality that we've almost come to.
I don't see banning abortion as a problem, because it's a simple human rights issue. The right to live trumps another person's right to convenience. It's as simple as that.
I don't see banning "gay marriage" as a problem, because marriage has a very specific definition and meaning. Call "gay marriage" what you want - civil union - whatever, but it sure isn't marriage. These people are free to do whatever they want - cohabitate, whatever. They are grownups after all. I'm just not going to personally support their behavior with my tax dollars.
Right, because this nation has never had president's that acknowledge the Christian God.
George Washington:
By 1778, George Washington had so often witnessed God's intervention that on August 20, he wrote Thomas Nelson that:
The Hand of providence has been so conspicuous in all this, that he must be worse than an infidel that lacks faith, and more than wicked, that has not gratitude enough to acknowledge his obligations.
Abraham Lincoln:
"When any church will inscribe over its altar, as its sole qualification for membership, the Savior's condensed statement of the substance of both law and Gospel, 'Thou shalt love the lord thy God with all thy heart, and with all thy soul and thy neighbor as thyself' that church will I join with all my heart and all my soul."
Theodore Roosevelt: "The true Christian is the true citizen."
Those of course only reflect the first few presidents I did a Google search on. I think you'll find that contemporaries like Reagan and Bush are in good company. No president has ever forced their faith on anyone, though they did publicly announce their own personal faith. Presidents have been driven by their faith, to uphold basic human rights. For example, Abraham Lincoln waged war over slavery. George Bush has worked to pass legislature protecting the rights of the unborn.
Almost all of the Java compilers out there do little to no optimizations while generating bytecode. Rather, the focus of the compilers is to generate bytecode that is easily understood and optimized by the virtual machine.
Since the bytecode for Java is a well-defined standard, there are many more bytecode processors for Java compared to say object-code processors for C or C++. So, a non-optimizing compiler also has the side-effect of making it easier for all of the people writing bytecode processors.
I keep hearing this sentiment expressed on Slashdot, yet the last thing a Christian wants to do is force his religion on someone else. Our forefathers fled Britain because of religious persecution, and our constitution grants freedom of religion.
In this day and age, students in public schools are forced to read from the Quran and pray Muslim prayers, but are not allowed to read from the Bible. It strikes me strange that people can use the term "Christian theocracy" with any sort of straight face.
I find this particularly funny, because the exit polls show that Bush received more votes from people who had a high-school education or college degree, while Kerry received more votes from people without a high-school degree.
Sorry, it's a tier 2 app. You do win a PowerMac G5.
Java rocks.
Read the rules. They give you free access to POWER machines.
They are dual boot. It's posted somewhere on the rules page.
My project, Jace, is one of the tier 1 projects. It'll probably be one of the simpler ones to port, because it was already written with portability in mind. I was all "WOO HOO" when I found out I was getting an iPod Mini. Then I was all "DARN" when I realized I couldn't port my own project and win a G5. ;)
On a more serious note, almost all C++ applications prefer std::vector to std::list, and almost all Java applications prefer ArrayList over LinkedList - neither of which require any pointers.
These data structures tend to perform much better, because they are naturally cache friendly. It can take over 100,000 elements in a list to make it more expensive to do an insert or a remove in an array-based list over a link-based list.
In a language like C++, you can try to mitigate these factors by using a special-purpose allocator that greedily reserves large pools of contiguous memory to improve spatial locality.
Although I'm not incredibly fond of varargs, you mischaracterize them greatly. Varargs in Java are entirely type safe. For example take,
void foo( String... values ) { }The compiler will only allow me to call that function with a list of Strings, which then gets boxed as a String[] on its way into foo.
There is no equivalent to varargs in C++. When a function uses an ellipsis, it has absolutely no idea what data is actually being passed to it. The compiler does no enforcing of types upon invocation, and you have no type information at runtime. If the programmer goofed, you are fubar.
No, he's wrong. The Java 1.5 compiler stores the generic type attributes in the class file in what are called "Signature attributes". Generic types are perfectly safe across compilation boundaries. Where they aren't safe is when they are combined with raw types. That behavior is by design, and it is the tradeoff paid for having mostly seamless compatibility with an existing codebase. Whether that was an appropriate decision is an entirely different discussion.
Heh... I worked with images of Lenna in my DSP classes while working on my CmpE at Georgia Tech many moons ago. This article seems to disagree with you about the image being a good benchmark.
If you dislike Sun for just owning software patents, you should dislike just about every tech company that exists. Practically every company nowadays owns some bogus set of software patents, and companies like IBM (who regularly receives the most patent grants worldwide every year), own huge software patent portfolios. Fortunately, most of these companies (including Sun) don't go around abusing these bogus patents left and right in billion dollar lawsuits.
Companies are never going to stop patenting software as long as it can be done, and there will always be rotten apples abusing the system, so we need to stop this problem at its roots. New case law needs to be arrived at, or new legislation needs to be passed, to kill software patents.
For systems like mine, that store 100 million objects in memory, a 4X memory increase is hugely unacceptable, and the access penalty is also unacceptable. My situation is not uncommon, and it is the reason why libraries like fastutil exist and are so popular.
I think .NET is mostly a rip-off of the JVM with very little innovation, but they seem to have a much better approach to primitives, with JIT type-specialization
- new Signature attributes (for generics)
- new Varargs access specifier
- the JVM LDC instruction now supports class literals
- new Annotation attributes
- New Synthetic access specifiers replace old Synthetic attributes
Above and beyond the bytecode/classfile format changes, using -target 1.5 will cause the compiler to automatically take advantage of new code that is only available in the 1.5 runtime. For example, enums depend upon java.lang.Enum, the new for loop depends upon java.lang.Iterable, the compiler supports string concatenation using the new StringBuilder class instead of StringBuffer, etc...Retroweaver takes care of all of that, and was even given a thumbs up by the 1.5 compiler writer, Neal Gafter.