Sometimes it's not difficult to provide equivalent functionality in a standards-compliant way (for example, the MSN portal). In such cases, it is best to stick to standards.
Given the relative complexity of Google's Maps website, I think it's fair to give them the benefit of the doubt and accept the possibility that it is extremely hard to do this in a standards-compliant way. I think they may have chosen the option of giving each browser a slightly modified page and they just haven't gotten around to some of the other browsers yet.
Really? Since the whole point of.net is that you can call a Python method from C#, and subclass a C# object in Python, etc.,
You can still call any IL method from any language. It's just that the bytecode for the invocation (at the callsite) will be different because with Python you might have to look up the method at runtime.
I know going c#-IL-same C# program is very hard, but I'd have thought C#-IL-C# code that is obviously machine generated but nethertheless does the same thing and could be edited by a human was reasonably doable.
This thread was about being able to take a program and "edit it's source in any language I choose". The output of an IL disassembler is not good enough.
Besides, going C# -> IL -> C#, while hard, isn't as difficult as going Python -> IL -> C#.
Disassembly is not trivial. It's hard enough to go C# -> IL -> C#. It'll be much harder to go Python -> IL -> C#. For example, the bytecode for a Python method call is very different from the bytecode for a C# method call.
This isn't possible for languages that differ in any significant way. What happens to the variable declarations when you convert between Java and Python?
- Using XML requires that the complete grammar of the language can be captured accurately by an XML schema. Would this always be possible?
Just to be clear, you don't have to capture the language's grammar. You have to encode the AST, which isn't that difficult because ASTs are typically encoded with normal C-style structs. There's some additional formatting information you'd need to keep track of, but I don't think that'll be a major problem.
Also, compiling might be faster, because the parsing is less complex.
Bison-generated parsers are pretty fast. I think it's the lexical analyzer that takes the longest (since it has to look at every single character). I'm guessing that an XML formatted program, being more verbose, would be slower to compile. I'd rather have the program be stored in an efficient binary format, making parsing really fast.
If the customer wants a raw dump, it's easy to transform the XML into plain text. Docbook can do do it, you can too!
Ideally, your canonical data model should be losslessly representable in a relational DB schema or an XML schema. That way, you can convert between them whenever you need to.
The reason for not using XML is that the data model is fundamentally broken:
Attributes are useless (making it hard to represent records).
The necessary distinction between type names and field names isn't made.
Not to fuel the war, but when you code for the general case (templates) you aren't as fast as your hand optimized, coded for each data type algorithm. This is a simple fact.
This is true, but irrelevant.
First of all, you can do hand-optimized stuff in C++ as well. However, you usually don't want to waste manpower doing that. Normally, you'd like to program generically. Generic C++ algorithms and data structures are faster than generic C algorithms and data structures (and type-safe, as well).
But seriously, compiling a class with 9 inner classes is definately a strange thing to do
I'm not sure what he means here. If he means 9 classes nested within each other then yeah, that's kind of weird (though I don't see why code completion should be any harder if the algorithms were done properly).
But 9 inner classes all at the same level sounds fine. I don't understand why (without knowing the problem or looking at the code) that is "conceptually wrong."
Re:Cool! Just like form AutoComplete
on
Google Suggest
·
· Score: 1
The only next improvement I can think of is if it did related searches to the term you searched for, but I'm not sure how you would represent all of that data at once.
Gigablast does something like that. I think I've seen it in other places as well.
Absolute numbers don't help. Saying they save 115KB/s is less useful than saying what percentage increase this would be. I'm guessing that, using IEEE floats, the percentage increase would be zero. Besides, 115KB/s is less than the capacity of a residential broadband connection.
Sorry to spoil the MS-bashing party kids, but there's really nothing to see here.
There's some USPTO bashing that you've shown to be unjustified, but let the MS bashing continue. Whether or not they were awarded the patent doesn't excuse the application's idiocy and obfuscated wording. I can't understand how someone would be willing to put their name such a piece of trash.
As you say, you end up with a slippery slope that leads to AOT compilation, which would be fine, except that it makes you ask the question, why are we jitting in the first place?
Although "Just In Time" does imply that compilation is only done immediately prior to execution, that is only accurate of the early implementations. Though newer VMs still say JIT, I think JIT has come to mean "dealing with a canonical representation that is not directly executable." Caching is a transparent optimization in that the semantics of what's going on doesn't change. The VM's "CompileMethod()" function is still called and still has the same function signature. The only difference is that it might run faster (when a cached copy is available).
JIT is a generalization of the AOT workflow. In a JIT framework, you can still compile everything ahead of time and only once but you have the option of trying other things.
But you're overcomplicating the matter. A signed contract can add ethical issues to anything.
True... What I *meant* to convey was that the intent of "anti-decompilation" clause might be to hide proprietary company secrets -- to protect an R&D investment. Maybe the seller doesn't really mind if you decompile to interoperate, but can't figure out a safe way to say that in the contract. In such a situation, there are ethical issues related to how much of the decompiled source code can you look at without "stealing" the company's ideas.
I made no agreement with the seller. If the seller wants an agreement, he can give me the forms to sign and I will put my John Hancock on the line if I am willing to make that agreement. Just because I clicked past some text to get the program to install doesn't mean I agreed to it.
There's some confusion here... I (and some other people) were saying that decompilation is immoral if the agreement you had doesn't allow it. You (and others) are arguing that a post-purchase EULA doesn't constitute a mutual agreement. I'm not exactly how those two issues got mixed up in this thread, but it was probably when jmorris42 implied that there are never any ethical issues involved in decompilation:
What ethical problems? Decompiling is perfectly moral and ethical.
He isn't saying that laws and morality are unrelated, only that the relationship ought to be associated with a clear, unidirectional causality, and therefore that moral decisions are not dictated by the law.
If it's a unidirectional causality, then a moral assessment should never depend on any particular laws. The original poster claimed that unapproved decompilation was immoral. Then jmorris42 came around and started citing specific laws to prove that he had the legal right to do so.
What's worse is he assumed that hypothetical agreement wasn't in writing and proceeded to draw further conclusions from that assumption, sounding very much like someone who is used to exploiting another's lack of knowledge about specific laws.
Everyone understands the side effects of electoral math. Your wisdom is not needed here, neither is your hypothetical since the situation you describe is exactly what happened in 2000.
If you had correctly read my post, you would see that my comment argued against the idea that raw vote counting is stupid. It implied that the electoral college system is not the most accurate. That said, it's not completely clear which is better. Alan Keys makes a decent argument in support of the electoral college system.
It's tricky because goal of an election is not necessarily to follow the will of the majority. I don't know exactly what the goal is, but it's something like "make things better." After all, allowing everybody to vote is probably not a good way to select the best person for the job (allowing everybody in a corporation to vote on every issue would be disasterous). But the sense of individual empowerment you get from voting could be good for society and could ultimately offset the fact that the outcome of an election may not be the best one.
If you don't like the conditions of the offer, don't buy the product.
That is exactly what the Uniform Commercial Code is about, ensuring everyone knows and can agree in a meaningful way to the terms of the transaction. Without a signed contract specifying different terms though, the UCC says that a sale of goods transfers a clear title to those goods, meaning there can't be any conditions attached.
You're talking about specific laws here. The grandparent post was merely saying that violating an agreement is immoral. Why you you assume that the agreement wasn't in writing? Even if it isn't, the point is that violating an agreement is immoral (since you, at the time, were talking about morals and not the specifics of certain laws).
I agree. Also, in the World Series, they should just add up the number of total runs for all the games, and declair the team that got the most runs in the series the winner.
The rules of baseball can be arbitrary. The goal is for it to be entertaining. The same thing goes for "Final Jeopardy." The goal of an election is different. I can't think of a good way to describe that goal, but it's closer to "accuracy" than it is to "entertainment."
First you imply that the actual laws are irrelevant to your views on morality:
What ethical problems? Decompiling is perfectly moral and ethical. Whether it is illegal is a seperate and, for me, almost irelevant issue.
Then when someone argues on moral grounds:
If I write software and say I will sell it to you for $x on condition that you do Y (perhaps Y is not decompiling the source), and you agree to these terms, I think it is morally repugnant of you to break our agreement and decompile.
You go back to the law.
While you are welcome to your delusions, but out here in the real world we have some things called laws.
If you don't like the conditions of the offer, don't buy the product. But if you do buy the product, don't you think it is immoral to break the agreement you made with the seller (whether it is a signed commercial agreement or not)?
Ford can't forbid me from taking apart a Caddy and not only making, but SELLING plans;
So if you buy a copy of MS Office and figure out how to break the copy protection, you think it's OK to sell copies of the cracked version?
Slashdot Mirror
CLI is fine too. They have a lot of acronyms. If it starts with a "C", it probably means something in .Net.
Gambas. I really wish it wasn't based off of Basic, though.
Sometimes it's not difficult to provide equivalent functionality in a standards-compliant way (for example, the MSN portal). In such cases, it is best to stick to standards.
Given the relative complexity of Google's Maps website, I think it's fair to give them the benefit of the doubt and accept the possibility that it is extremely hard to do this in a standards-compliant way. I think they may have chosen the option of giving each browser a slightly modified page and they just haven't gotten around to some of the other browsers yet.
You can still call any IL method from any language. It's just that the bytecode for the invocation (at the callsite) will be different because with Python you might have to look up the method at runtime.
This thread was about being able to take a program and "edit it's source in any language I choose". The output of an IL disassembler is not good enough. Besides, going C# -> IL -> C#, while hard, isn't as difficult as going Python -> IL -> C#.
Disassembly is not trivial. It's hard enough to go C# -> IL -> C#. It'll be much harder to go Python -> IL -> C#. For example, the bytecode for a Python method call is very different from the bytecode for a C# method call.
The .Net compilers don't handle this at all. They convert source code to CLR bytecode. They don't convert between source languages.
This isn't possible for languages that differ in any significant way. What happens to the variable declarations when you convert between Java and Python?
Just to be clear, you don't have to capture the language's grammar. You have to encode the AST, which isn't that difficult because ASTs are typically encoded with normal C-style structs. There's some additional formatting information you'd need to keep track of, but I don't think that'll be a major problem.
Bison-generated parsers are pretty fast. I think it's the lexical analyzer that takes the longest (since it has to look at every single character). I'm guessing that an XML formatted program, being more verbose, would be slower to compile. I'd rather have the program be stored in an efficient binary format, making parsing really fast.
They're quite different at the core, even though they use similar syntax and are used to do similar things.
If the customer wants a raw dump, it's easy to transform the XML into plain text. Docbook can do do it, you can too!
Ideally, your canonical data model should be losslessly representable in a relational DB schema or an XML schema. That way, you can convert between them whenever you need to.
The reason for not using XML is that the data model is fundamentally broken:
This is true, but irrelevant.
First of all, you can do hand-optimized stuff in C++ as well. However, you usually don't want to waste manpower doing that. Normally, you'd like to program generically. Generic C++ algorithms and data structures are faster than generic C algorithms and data structures (and type-safe, as well).
I'm not sure what he means here. If he means 9 classes nested within each other then yeah, that's kind of weird (though I don't see why code completion should be any harder if the algorithms were done properly).
But 9 inner classes all at the same level sounds fine. I don't understand why (without knowing the problem or looking at the code) that is "conceptually wrong."
Gigablast does something like that. I think I've seen it in other places as well.
BTW, anybody else tired of the Google fanboys?
A JIT for a statically typed language can do even better (theoretically).
Yeah...1000 req/s * 118 bytes ~ 115 KB/s. That's why I used the capital 'B'. It's within the capacity of a residential broadband line.
Absolute numbers don't help. Saying they save 115KB/s is less useful than saying what percentage increase this would be. I'm guessing that, using IEEE floats, the percentage increase would be zero. Besides, 115KB/s is less than the capacity of a residential broadband connection.
There's some USPTO bashing that you've shown to be unjustified, but let the MS bashing continue. Whether or not they were awarded the patent doesn't excuse the application's idiocy and obfuscated wording. I can't understand how someone would be willing to put their name such a piece of trash.
Although "Just In Time" does imply that compilation is only done immediately prior to execution, that is only accurate of the early implementations. Though newer VMs still say JIT, I think JIT has come to mean "dealing with a canonical representation that is not directly executable." Caching is a transparent optimization in that the semantics of what's going on doesn't change. The VM's "CompileMethod()" function is still called and still has the same function signature. The only difference is that it might run faster (when a cached copy is available).
JIT is a generalization of the AOT workflow. In a JIT framework, you can still compile everything ahead of time and only once but you have the option of trying other things.
True... What I *meant* to convey was that the intent of "anti-decompilation" clause might be to hide proprietary company secrets -- to protect an R&D investment. Maybe the seller doesn't really mind if you decompile to interoperate, but can't figure out a safe way to say that in the contract. In such a situation, there are ethical issues related to how much of the decompiled source code can you look at without "stealing" the company's ideas.
There's some confusion here... I (and some other people) were saying that decompilation is immoral if the agreement you had doesn't allow it. You (and others) are arguing that a post-purchase EULA doesn't constitute a mutual agreement. I'm not exactly how those two issues got mixed up in this thread, but it was probably when jmorris42 implied that there are never any ethical issues involved in decompilation:
If it's a unidirectional causality, then a moral assessment should never depend on any particular laws. The original poster claimed that unapproved decompilation was immoral. Then jmorris42 came around and started citing specific laws to prove that he had the legal right to do so.
What's worse is he assumed that hypothetical agreement wasn't in writing and proceeded to draw further conclusions from that assumption, sounding very much like someone who is used to exploiting another's lack of knowledge about specific laws.
Everyone understands the side effects of electoral math. Your wisdom is not needed here, neither is your hypothetical since the situation you describe is exactly what happened in 2000.
If you had correctly read my post, you would see that my comment argued against the idea that raw vote counting is stupid. It implied that the electoral college system is not the most accurate. That said, it's not completely clear which is better. Alan Keys makes a decent argument in support of the electoral college system.
It's tricky because goal of an election is not necessarily to follow the will of the majority. I don't know exactly what the goal is, but it's something like "make things better." After all, allowing everybody to vote is probably not a good way to select the best person for the job (allowing everybody in a corporation to vote on every issue would be disasterous). But the sense of individual empowerment you get from voting could be good for society and could ultimately offset the fact that the outcome of an election may not be the best one.
You're talking about specific laws here. The grandparent post was merely saying that violating an agreement is immoral. Why you you assume that the agreement wasn't in writing? Even if it isn't, the point is that violating an agreement is immoral (since you, at the time, were talking about morals and not the specifics of certain laws).
The rules of baseball can be arbitrary. The goal is for it to be entertaining. The same thing goes for "Final Jeopardy." The goal of an election is different. I can't think of a good way to describe that goal, but it's closer to "accuracy" than it is to "entertainment."
First you imply that the actual laws are irrelevant to your views on morality:
Then when someone argues on moral grounds:
You go back to the law.
If you don't like the conditions of the offer, don't buy the product. But if you do buy the product, don't you think it is immoral to break the agreement you made with the seller (whether it is a signed commercial agreement or not)?
So if you buy a copy of MS Office and figure out how to break the copy protection, you think it's OK to sell copies of the cracked version?