...I'd say the balance of power is most decidedly tipped in Congress's favor
You're referring to a balance of power between the parties, not between the branches of government. Our politicans think of themselves first as Democrat/Republican, then Legislative or Executive. It was supposed to be the other way around.
I'm taking it upon myself to get some solar panels
Perhaps you should also put these underground. Solar panels are expensive and could be damaged by flying debris in high winds!
Another poster suggested you should consider wind instead of solar. Here, you would want to bury the wind mills to protyect them, and also to protect flocks of birds. But the worm evangelists would come after you
To me, OpenLaszlo is not about the web. If you think about what it does, it allow syou to specify a complete user interface and logic in an XML file. The layout is done with XML, and the logic is done with ECMAScript (yes, that's what JavaScript became).
The first OpenLaszlo solution compiled this XML into Flash which can run in any browser. Then they made a new compiler which turns it into DHTML so youd on't need Flash any more. So now you can take the same application written once (as an LZX XML file) and compile it to Flash or DHTML and get the same behavior. Both of those are very ubiquitous mediums. If you read their roadmap, they also have plans for Java client.
My hope is that one day, there will just be clients that read the LZX XML directly. These clients could be written in Java,.NET, TclTk, C++, you name it. They would all read the same LZX XML and render it for the user. That's very much how various different browsers all read the same HTML file and render it. So you might be thinking that its no better than HTML, but:
It is a tighter specification than the original HTML which planted the seed for incompatibilities
It is designed from the beginning for user interface declaration, not text markup
It leverages two well-accepted standards: XML and ECMAScript
If I had to pick a solution for the world to use for rich internet applications, I'd choose OpenLaszlo over Java Applets, Java WebStart, Macromedia Flex, DHTML, etc.
I say to you that the VCR is to the American film producer and the American public as the Boston strangler is to the woman home alone.
Do you think maybe the MPAA hired someone to go strangle women -- later known as the Boston Strangler -- just so they could have a scary phantom to use as a simile when battling the VCR in court?
Actually, I think Sony was willing to license some of its technology, like Meory Stick, to others. But, I imagine their pricing and competition against cheaper or open standards kept other companies from even considering licensing Sony's proprietary standards.
As others have said, there's not much of a difference. And if this is the case, why aren't slot machine companies entering the voting machine business. I mean,s ure, they'd probbaly have to set up a separate ocmpany just to voting and gambling machines don't have the same company name attached raising public doubt. But tehcnologically, the machines could be quite similar.
Some things you just can't stamp out with brute force. Instead, you have to find out what is fueling it and try to reduce that source and acknowledge you'll never be able to fully extinguish it.
Don't be so quick to look at it through the geo-political polarizing sunglasses. YOu claim that people are upset because its the Chinese. YOu also say your dad "...firmly believes that people are getting paid by the US government to bash the Chinese government."
While Occam's razor might drive us towards the simplest solution (e.g. the bloggers just tricked us), it is always more fun to believe in a conspiracy theory because its more exiting, has more options, etc. So whether for the Chinese that conspiracy is the U.S. government pulling the strings, or if for the West that conspiracy is the Chinese government block bloggers, the true fact is that universally, conspiracy theories are more fun to believe!
Just today, I led a group of people to kill another group. Some fought back and both sides experienced casualties. But I had planned for this and made sure they were outnumbered. I didn't stop with killing those who fought back, I also killed the innocents. And I still didn't stop there. I also destroyed their houses and places of work, flattening every structure in sight.
Do you think this could be related to the Warcraft III map I played the night before?
I've heared once that the original hebrew text reads "Thou shalt not murder." If that's true, the contradiction is easy to resolve: Just define that killing ordered by god is no murder.
My understanding was similar but different. I had heard that the original ancient language of the bible did not have a rich engouh vocabulary to distingiuish between kill (e.g. an enemy) and murder (e.g. one in your own society), but the next most recent translation of the bible used the word "murder", not "kill".
The point is, when Moses was taking his tribe around the desert with their new commandments, they were to preserve their own society (which is what the 10 commandments promote), but if they had to kill competing tribes to survive, they could do so because it would be *killing*, not *murder*. Any society that condones unbridled murder within itself will quickly commit suicide.
The graviton is the force-carrying particle of gravity
This is slightly off topic, but I'm curious to know: What is the speed of gravity? For example, light from the sun takes 8 minutes to reach Earth. If the sun instantly disappeared form existence (e.g. the FSM took it), we'd still see ligth for 8 more minutes? But if the sun is gone, does its gravitation effect disappear at the moment it disappeared hurtling the Earth off into space, or would the planets continue to orbit the void for 8 more minutes?
The trend on this list is of (great) American liberals. Bush does not fit this mold imo, from various perspectives. Also importantly, the War on Terror is a much different type on conflict than the wars these Presidents faced. The enemy is borderless, uniformless, with unknown numbers, etc. This type of war is virtually endless, whether we are in Iraq or out of Iraq.
This is exactly why I am worried! Against my will, some of my liberty has been given up in the name of security. But its not even temporary!
Look, Bush might be an evil doer in disguise who just tricked us into givng up our rights. Or, he might be an idiot that someone else is controlling and tricked us out of our rights. He could even be an smart, honest, good man (who happens to seem like an idiot) who has our best interests at heart and is stopping dozens of terrorists attacks so we can sleep at night.
But even if it is the latter of those possibilities, Bush won't be in power forever. Someone else could eventually come along that fits the first two descriptions. Thanks to the situation that has arisen and shows no signs of being put in check, any future leader can swopp in and use these powers for whatever he wants. Especially, since we have an enemy that is borderless, uniformless, with unknown numbers, etc. Our leaders could have these temporary powers indefinitely, as long as it suits them.
So, even if Bush *is* the good little boy scout, he still needs judical oversight, evenif its for the sake of the future, not now.
I did some quick Googling, and couldn't find answers to an importantquestion about the wiretapping you seem to be holding up as justification for the current situation in the US: was the Italian wiretapping legal or illegal?. Maybe the Italian police got a warrant. Maybe Italian law doesn't require a warrant. Does anyone know?
...if it means a few calls to known terrorists are tapped.
And the number of people like me is growing, as witnessed by the 60% approval ratings for wiretapping actions that Bush enjoys.
Why is it that no poll can look like this: What do you think about the wiretapping?
Its OK no matter what
It would be ok if the admin got FISA to grant warrants
It is no tOK under any circumstances
Every time I debate this with people, they always talk about the fact that it's "known" terrorists on the other end so its excusable. I don't care if its your grandma on the other end. If an American at home is on the other end, why is it so imssposible for the administration to just get a warrant?! FISA grants almost every single request. FISA acts quickly, even in the middle of the night. FISA will even let you get the warrant after the fact! So...
Why won't the administration submit requests to FISA?
They're not wiretapping who they say they are so no court would actually grant such warrants
There are so many terrorist-connected calls coming out of America it would overwhelm the system
The administration is trying to save taxpayer dollars by cutting down on paper usage
This is a lot like the war on terra. They won't have successfully "won" until there isn't a single person out there making a copy of a song for someone else without paying. In short, it will never end...
So they don't care if married people pirate music? Rock on!
Actually, as I understand it, it goes something like this. The administration is only tapping those that have ties to known terrorists. How we're sure they are known terrorists is beyond me. If we've convicted them of being guilty, I assume it was in absentia since they're not in prison. Or, maybe we shoudl just trust the administrations opinion. After all, the administration did have their lawyers confirm that this wiretapping is legal. Of course, lawyers don't decide the law, judges do. But the amdinistration bypassed the judges by saying their lwyers felt they didn't need any judges...
So, now suppose that the administration thinks your mother is a terrorist. The administration feels its belief about your mother is enough to justify wiretapping your phone calls to your mom without having to convince a judge of probable cause. If they're SOOO sure, then the almost-never-deny-a-request FISA court wouuld certainly ok their requets -- even two weeks after they do the actual tapping! So, I can only assume the reason they need to bypass a FISA judge is that they're really doing something wrong.
Performing all activities as an unpriviliged user, with some method of securely and briefly authenticating to higher permissions when required
Services, scheduled tasks and mapped drives have long been able to been run as another user. Windows XP also has some newers tricks. It has the 'runas' command line tool (yeah, Windows has a command line!). Also, when I go to a shortcut, there is an 'Advanced' button where I can specify to have the program run as another user. In fact, a recent policy change at work took away my ability to use that last feature, and it ticks me off!
I was really only replying to the notion that java didn't have an equivalent mechanism for exposing the public interface of a class to a customer without revealing code.
To which I point out, every class has its public API embedded in it, in a well specified format. This information can easily be retrieved using a tool that comes wiht the JDK, javap. For example:
This will list the public, protected and package-protected classes, members and methods in java.lang.System, knowing only the information in the.class (which in this case comes from rt.jar in the default classpath for Java). It does not show implementation of those methods and classes, only their API.
What Java can't do, to my knowledge, is distribute a binary and prevent its API from being retrieved in this manner. Their obfuscators that can slow things down, but you can't prevent it.
What I started to hate in C++/Java/C# is that there's no easy and standard-conforming way to express complex data 'inline'.
I once worked on a Java programmer (written by a lot of recently C programmers) that had sometime like a Vector of HashMaps, and each HashMap had an element under a a known key that was an Array, and the first item in the array was a Vector of IDs, and the second item in the array was a Map of names to a Vectors where the 1st, 2nd and 3rd items were a crude struct. You couldn't debug crap with this, and even coding was hard unless you had a colorful diagram on your whiteboard explaining what everything meant.
This reminded me a lot of what I used to do with Perl before I started taking advantage of Perl's OO features.
I've often found places where I need to do somethign quick and dirty for a test program and wish I could just use Maps of Arrays rather than defining formal types. But for anything serious, it is much better to just make formal classes for each one. These classes are then types, and which gives not just your language some semantics, but the code you've written now has semantics!
Just a note: interface ==.h for java. Document all you want, and ship to your customer, along with a crypted jar with the actual binary.
I think I know what you're getting at, but I don't think its quite true.
Interfaces in Java are only used like a.h file once they're compiled. The compiled.class file is what you're actually importing... or rather, the API of that.class file. On the other hand, a header file in C++, IIRC, is inline at compile time as plain text before the compile does anything. Thus, you could include anything text you want.. imcomplete syntactical fragments, for example. In Java, this is not possible. This is essentially the difference between a preprocessor which treats source and includes as text and the Java compiler's treats imports as a way for the compiler to be aware of a pre-compiled API. That is, one happens in the text domain while the other happens in the domain of the compiler's understanding of the parsed source. This is nice in some ways, btu also means you can't do some cool tricks you could do with a preprocessor.
As for the crypted jar file, I'm not sure as I don't use them much. But what I am fairly certain of is that Java uses runtime linking, meaning the actual names of methods and classes is stored in the.class file itself so that they can be linked to the right code loaded from elsewhere by name. So in order for Java's runtime linking, the non-private API of a class must be known to the.class. This is why once you have a.class file, you can run javap against it and get its API. Its just dumping the same information that is used at runtime to match of what one class needs to what another provides.
So, I'd say Java's runtime linking info in a.class file is most similar to what C/C++ header files are used for a good deal of the time, but via a var different mechanism that those includes. Thus, it is much cleaner at doing that job, but weaker at doing everythign else you might do in a header file.
In this respect I would like C++ be more like Java (or TurboPascal for the matter) where interface declarations and compiled code are unified
One thing that Java does well is put the interface to something in its compiled code. This also means that anyone can look into your interfaces from your compiled code.:)
In C, you could choose not to give someone your headersand they could never (easily) use your code, but the other code you distributed which needs it could still use it. Consider a third party library, for example. In my C/C++ days, I often wanted to use a library that another library I was using called into, but I couldn't do so because I didn't have the headers for it.
In Java you can achieve some of this with signed jars, and good use of public/private/protected/etc., but you simply can't give expose a public method in your JAR to one fellow and not to another.
From a programmers point of view, a registry is just a one level higher abstraction above an INI file. The nice thing about an INI file is that its just text, so you can edit it with about any text editor. But a text editor editing an INI gives you no constraints on what goes in it. So you could but a non-numerical value for a line in an INI file where the system that uses it only supports a numerical value. So a programmer writing a program which uses such an INI file has to take extra precautions and use a bare bones text reading API, or find some existing library which handles it for him.
The registry provides constraints. When you have a DWORD value, you can't put letters into it (unless, of course, you're using the hex mode of the field your favorite registry editor). It's hierarchial in nature meaining you can infer a lot of context, too. And as a programmer, you have a very simple API to go against for 90% of what you need to do.
That said, there are some things I which the registry did better. For one, I wish it were "registries" instead of "the registry". Why not let each program store its data in its own sort of registry file. You could very surely know you've removed a program by not letting it write to the system or user registry, but only to its own globval or user-specific entrues -- and delete those files when you want. I also wish every key and value had a description with it. I also wish there were richer data types, such as "existing file path", "file path", etc. so you could build even better basic tools for working with them. And I wish the registry were an open format that worked on any platform. Then I could take the registry file for application X I configured on my Windows machine and copy it to my Linux machine (granted, there's other cross-platform issues when you start intrereting the data on disparate platforms, such as file paths).
There is a difference between testable and ease of testing. Evolution can certainly be tested. Take a population of some organism, place them in to equal environments, vary the environments differently over time.
The difficulty if in the magnitude of some of the variables. How big of an environment do you need? How many organisms in each population? How much time to let it transpire? Each of these are probably answered with large numbers. However, just because it is difficult to test, doesn't mean its not testable.
Consider some competing theories from the past, like the stars being points of light fixed to a class sphere encasing the earth. To test this, you can go up high enough and see if indeed there is a glass sphere.
Now when you look an intelligent design, "everything you see was put in its current state by God.", that's a lot harder to test. That's not a theory of an on going process, but an event from the past whose conditions and variables you cannot know.
Intelligent design is not testable; however, evolution is, difficult though it may be.
Slashdot Mirror
You're referring to a balance of power between the parties, not between the branches of government. Our politicans think of themselves first as Democrat/Republican, then Legislative or Executive. It was supposed to be the other way around.
I'm taking it upon myself to get some solar panels
Perhaps you should also put these underground. Solar panels are expensive and could be damaged by flying debris in high winds!
Another poster suggested you should consider wind instead of solar. Here, you would want to bury the wind mills to protyect them, and also to protect flocks of birds. But the worm evangelists would come after you
To me, OpenLaszlo is not about the web. If you think about what it does, it allow syou to specify a complete user interface and logic in an XML file. The layout is done with XML, and the logic is done with ECMAScript (yes, that's what JavaScript became).
The first OpenLaszlo solution compiled this XML into Flash which can run in any browser. Then they made a new compiler which turns it into DHTML so youd on't need Flash any more. So now you can take the same application written once (as an LZX XML file) and compile it to Flash or DHTML and get the same behavior. Both of those are very ubiquitous mediums. If you read their roadmap, they also have plans for Java client.
My hope is that one day, there will just be clients that read the LZX XML directly. These clients could be written in Java, .NET, TclTk, C++, you name it. They would all read the same LZX XML and render it for the user. That's very much how various different browsers all read the same HTML file and render it. So you might be thinking that its no better than HTML, but:
If I had to pick a solution for the world to use for rich internet applications, I'd choose OpenLaszlo over Java Applets, Java WebStart, Macromedia Flex, DHTML, etc.
I say to you that the VCR is to the American film producer and the American public as the Boston strangler is to the woman home alone.
Do you think maybe the MPAA hired someone to go strangle women -- later known as the Boston Strangler -- just so they could have a scary phantom to use as a simile when battling the VCR in court?
Nah, they wouldn't stoop that low... would they?
nd of course it only worked with sony stuff.
Actually, I think Sony was willing to license some of its technology, like Meory Stick, to others. But, I imagine their pricing and competition against cheaper or open standards kept other companies from even considering licensing Sony's proprietary standards.
what's the difference? - Really?
As others have said, there's not much of a difference. And if this is the case, why aren't slot machine companies entering the voting machine business. I mean,s ure, they'd probbaly have to set up a separate ocmpany just to voting and gambling machines don't have the same company name attached raising public doubt. But tehcnologically, the machines could be quite similar.
Some things you just can't stamp out with brute force. Instead, you have to find out what is fueling it and try to reduce that source and acknowledge you'll never be able to fully extinguish it.
- War on Piracy
- War on Poverty
- War on Drugs
- War on Terror
- ...
Did I forget any?Don't be so quick to look at it through the geo-political polarizing sunglasses. YOu claim that people are upset because its the Chinese. YOu also say your dad "...firmly believes that people are getting paid by the US government to bash the Chinese government."
While Occam's razor might drive us towards the simplest solution (e.g. the bloggers just tricked us), it is always more fun to believe in a conspiracy theory because its more exiting, has more options, etc. So whether for the Chinese that conspiracy is the U.S. government pulling the strings, or if for the West that conspiracy is the Chinese government block bloggers, the true fact is that universally, conspiracy theories are more fun to believe!
Just today, I led a group of people to kill another group. Some fought back and both sides experienced casualties. But I had planned for this and made sure they were outnumbered. I didn't stop with killing those who fought back, I also killed the innocents. And I still didn't stop there. I also destroyed their houses and places of work, flattening every structure in sight.
Do you think this could be related to the Warcraft III map I played the night before?
I've heared once that the original hebrew text reads "Thou shalt not murder." If that's true, the contradiction is easy to resolve: Just define that killing ordered by god is no murder.
My understanding was similar but different. I had heard that the original ancient language of the bible did not have a rich engouh vocabulary to distingiuish between kill (e.g. an enemy) and murder (e.g. one in your own society), but the next most recent translation of the bible used the word "murder", not "kill".
The point is, when Moses was taking his tribe around the desert with their new commandments, they were to preserve their own society (which is what the 10 commandments promote), but if they had to kill competing tribes to survive, they could do so because it would be *killing*, not *murder*. Any society that condones unbridled murder within itself will quickly commit suicide.
The graviton is the force-carrying particle of gravity
This is slightly off topic, but I'm curious to know: What is the speed of gravity? For example, light from the sun takes 8 minutes to reach Earth. If the sun instantly disappeared form existence (e.g. the FSM took it), we'd still see ligth for 8 more minutes? But if the sun is gone, does its gravitation effect disappear at the moment it disappeared hurtling the Earth off into space, or would the planets continue to orbit the void for 8 more minutes?
The trend on this list is of (great) American liberals. Bush does not fit this mold imo, from various perspectives. Also importantly, the War on Terror is a much different type on conflict than the wars these Presidents faced. The enemy is borderless, uniformless, with unknown numbers, etc. This type of war is virtually endless, whether we are in Iraq or out of Iraq.
This is exactly why I am worried! Against my will, some of my liberty has been given up in the name of security. But its not even temporary!
Look, Bush might be an evil doer in disguise who just tricked us into givng up our rights. Or, he might be an idiot that someone else is controlling and tricked us out of our rights. He could even be an smart, honest, good man (who happens to seem like an idiot) who has our best interests at heart and is stopping dozens of terrorists attacks so we can sleep at night.
But even if it is the latter of those possibilities, Bush won't be in power forever. Someone else could eventually come along that fits the first two descriptions. Thanks to the situation that has arisen and shows no signs of being put in check, any future leader can swopp in and use these powers for whatever he wants. Especially, since we have an enemy that is borderless, uniformless, with unknown numbers, etc. Our leaders could have these temporary powers indefinitely, as long as it suits them.
So, even if Bush *is* the good little boy scout, he still needs judical oversight, evenif its for the sake of the future, not now.
Wiretapping also works: the Al Qaeda cell in Italy that was planning to outdo 9-11 was caught by wiretapping.
I did some quick Googling, and couldn't find answers to an importantquestion about the wiretapping you seem to be holding up as justification for the current situation in the US: was the Italian wiretapping legal or illegal?. Maybe the Italian police got a warrant. Maybe Italian law doesn't require a warrant. Does anyone know?
if he started coveting his neighbour's ass you better believe he'd get smacked down.
You mean, like Clinton?
And the number of people like me is growing, as witnessed by the 60% approval ratings for wiretapping actions that Bush enjoys.
Why is it that no poll can look like this: What do you think about the wiretapping?
Every time I debate this with people, they always talk about the fact that it's "known" terrorists on the other end so its excusable. I don't care if its your grandma on the other end. If an American at home is on the other end, why is it so imssposible for the administration to just get a warrant?! FISA grants almost every single request. FISA acts quickly, even in the middle of the night. FISA will even let you get the warrant after the fact! So...
Why won't the administration submit requests to FISA?
This is a lot like the war on terra. They won't have successfully "won" until there isn't a single person out there making a copy of a song for someone else without paying. In short, it will never end...
So they don't care if married people pirate music? Rock on!
500 individuals that had some tie to evildoers
Actually, as I understand it, it goes something like this. The administration is only tapping those that have ties to known terrorists. How we're sure they are known terrorists is beyond me. If we've convicted them of being guilty, I assume it was in absentia since they're not in prison. Or, maybe we shoudl just trust the administrations opinion. After all, the administration did have their lawyers confirm that this wiretapping is legal. Of course, lawyers don't decide the law, judges do. But the amdinistration bypassed the judges by saying their lwyers felt they didn't need any judges...
So, now suppose that the administration thinks your mother is a terrorist. The administration feels its belief about your mother is enough to justify wiretapping your phone calls to your mom without having to convince a judge of probable cause. If they're SOOO sure, then the almost-never-deny-a-request FISA court wouuld certainly ok their requets -- even two weeks after they do the actual tapping! So, I can only assume the reason they need to bypass a FISA judge is that they're really doing something wrong.
It's hard for some of us not to look at grammatical or spelling errors and wince
You would get along well with Lynne Truss, the author of Eats, Shoots & Leaves.
Performing all activities as an unpriviliged user, with some method of securely and briefly authenticating to higher permissions when required
Services, scheduled tasks and mapped drives have long been able to been run as another user. Windows XP also has some newers tricks. It has the 'runas' command line tool (yeah, Windows has a command line!). Also, when I go to a shortcut, there is an 'Advanced' button where I can specify to have the program run as another user. In fact, a recent policy change at work took away my ability to use that last feature, and it ticks me off!
I was really only replying to the notion that java didn't have an equivalent mechanism for exposing the public interface of a class to a customer without revealing code.
To which I point out, every class has its public API embedded in it, in a well specified format. This information can easily be retrieved using a tool that comes wiht the JDK, javap. For example:
"%JAVA_HOME%\bin\javap.exe" -package java.lang.System
This will list the public, protected and package-protected classes, members and methods in java.lang.System, knowing only the information in the .class (which in this case comes from rt.jar in the default classpath for Java). It does not show implementation of those methods and classes, only their API.
What Java can't do, to my knowledge, is distribute a binary and prevent its API from being retrieved in this manner. Their obfuscators that can slow things down, but you can't prevent it.
What I started to hate in C++/Java/C# is that there's no easy and standard-conforming way to express complex data 'inline'.
I once worked on a Java programmer (written by a lot of recently C programmers) that had sometime like a Vector of HashMaps, and each HashMap had an element under a a known key that was an Array, and the first item in the array was a Vector of IDs, and the second item in the array was a Map of names to a Vectors where the 1st, 2nd and 3rd items were a crude struct. You couldn't debug crap with this, and even coding was hard unless you had a colorful diagram on your whiteboard explaining what everything meant.
This reminded me a lot of what I used to do with Perl before I started taking advantage of Perl's OO features.
I've often found places where I need to do somethign quick and dirty for a test program and wish I could just use Maps of Arrays rather than defining formal types. But for anything serious, it is much better to just make formal classes for each one. These classes are then types, and which gives not just your language some semantics, but the code you've written now has semantics!
Consider:
(String)((Vector)((Map)((Object[])((Map)vec.get( 0)).get("foo"))[1]).get("email")).get(0)
vs.
customers.getCustomer(0).getContact("foo").getEm ailInfo().getEmailAddresses().getEmailAddress(0)
Just a note: interface == .h for java. Document all you want, and ship to your customer, along with a crypted jar with the actual binary.
I think I know what you're getting at, but I don't think its quite true.
Interfaces in Java are only used like a .h file once they're compiled. The compiled .class file is what you're actually importing... or rather, the API of that .class file. On the other hand, a header file in C++, IIRC, is inline at compile time as plain text before the compile does anything. Thus, you could include anything text you want.. imcomplete syntactical fragments, for example. In Java, this is not possible. This is essentially the difference between a preprocessor which treats source and includes as text and the Java compiler's treats imports as a way for the compiler to be aware of a pre-compiled API. That is, one happens in the text domain while the other happens in the domain of the compiler's understanding of the parsed source. This is nice in some ways, btu also means you can't do some cool tricks you could do with a preprocessor.
As for the crypted jar file, I'm not sure as I don't use them much. But what I am fairly certain of is that Java uses runtime linking, meaning the actual names of methods and classes is stored in the .class file itself so that they can be linked to the right code loaded from elsewhere by name. So in order for Java's runtime linking, the non-private API of a class must be known to the .class. This is why once you have a .class file, you can run javap against it and get its API. Its just dumping the same information that is used at runtime to match of what one class needs to what another provides.
So, I'd say Java's runtime linking info in a .class file is most similar to what C/C++ header files are used for a good deal of the time, but via a var different mechanism that those includes. Thus, it is much cleaner at doing that job, but weaker at doing everythign else you might do in a header file.
In this respect I would like C++ be more like Java (or TurboPascal for the matter) where interface declarations and compiled code are unified
One thing that Java does well is put the interface to something in its compiled code. This also means that anyone can look into your interfaces from your compiled code. :)
In C, you could choose not to give someone your headersand they could never (easily) use your code, but the other code you distributed which needs it could still use it. Consider a third party library, for example. In my C/C++ days, I often wanted to use a library that another library I was using called into, but I couldn't do so because I didn't have the headers for it.
In Java you can achieve some of this with signed jars, and good use of public/private/protected/etc., but you simply can't give expose a public method in your JAR to one fellow and not to another.
Still, I manage to just fine without this :)
From a programmers point of view, a registry is just a one level higher abstraction above an INI file. The nice thing about an INI file is that its just text, so you can edit it with about any text editor. But a text editor editing an INI gives you no constraints on what goes in it. So you could but a non-numerical value for a line in an INI file where the system that uses it only supports a numerical value. So a programmer writing a program which uses such an INI file has to take extra precautions and use a bare bones text reading API, or find some existing library which handles it for him.
The registry provides constraints. When you have a DWORD value, you can't put letters into it (unless, of course, you're using the hex mode of the field your favorite registry editor). It's hierarchial in nature meaining you can infer a lot of context, too. And as a programmer, you have a very simple API to go against for 90% of what you need to do.
That said, there are some things I which the registry did better. For one, I wish it were "registries" instead of "the registry". Why not let each program store its data in its own sort of registry file. You could very surely know you've removed a program by not letting it write to the system or user registry, but only to its own globval or user-specific entrues -- and delete those files when you want. I also wish every key and value had a description with it. I also wish there were richer data types, such as "existing file path", "file path", etc. so you could build even better basic tools for working with them. And I wish the registry were an open format that worked on any platform. Then I could take the registry file for application X I configured on my Windows machine and copy it to my Linux machine (granted, there's other cross-platform issues when you start intrereting the data on disparate platforms, such as file paths).
It is not possible to test evolution
I disagree.
There is a difference between testable and ease of testing. Evolution can certainly be tested. Take a population of some organism, place them in to equal environments, vary the environments differently over time.
The difficulty if in the magnitude of some of the variables. How big of an environment do you need? How many organisms in each population? How much time to let it transpire? Each of these are probably answered with large numbers. However, just because it is difficult to test, doesn't mean its not testable.
Consider some competing theories from the past, like the stars being points of light fixed to a class sphere encasing the earth. To test this, you can go up high enough and see if indeed there is a glass sphere.
Now when you look an intelligent design, "everything you see was put in its current state by God.", that's a lot harder to test. That's not a theory of an on going process, but an event from the past whose conditions and variables you cannot know.
Intelligent design is not testable; however, evolution is, difficult though it may be.