I have not thought of it too much either, though I know some issues would be easily checked (like preventing use of pointers outside low level classes... etc.), harder would be a declarative language for defining the acceptable paradigm, not leaving paradigms out, and probably a lot of difficult special cases of things that are part of people's current standards which would be hard to check with such a tool.
However: the advantage of not having a language built to the safe paradigm is that C++ and this tool would allow a choice between multiple paradigms (infinite? at any rate a large number). It would be meant to prevent mixed-paradigms on a module by module basis (e.g. if you have a pointerless paradigm you were using at the application level, you probably don't want that in the system classes used, etc.)
which makes me wonder... is what we want for C++ to catch up is a way to specify your coding standard to a tool that can check source to see... e.g. if a company uses a class system which supports a pointer free paradigm at the app level in C++, they specify their standards in a documents, and this would be a tool to check the code for compliance.
I understand this problem... C++ is multiparadigmed, but a program should adopt a single paradigm.
>In C, you can easily leak four bytes at a time, and good luck finding it.
you can easily use a debug version of the libraries that will tell you if there is a leak when you exit, it can tell you where the memory was allocated, and if the memory was overwritten.
Most of your answer is just "hey, you are just not good enough at Java"... while I cannot defend the parents skills with Java, I can that dumping the heap and writing scripts to analyse the output is not "not having to worry about memory". There is nothing wrong with that (quite the opposite, it shows your willingness to take the steps nec. to make a good product), but you also have to take into mind the time it took to get that expertise. You may have written those things in one hour (or was that one hour "programmer time"?), but you must also count how much time did you spend becoming familiar enough with that output to be able to write that in one hour.
I think Java has a lot going for it although I have only done limited work in it. However, I bristle at the idea (which you did not advocate but which is often heard) that somehow the answer to good programming isn't detailed, that memory issues are abstracted away, good programming does come down to things such as you have learned... how to make your VM dump memory and looking very closely at how you use memory.
OTOH: we have a java application which has been praised by a user that is still on OS/2 (don't ask)... try that with C/C++.
Could linux itself exist under a process that enforced focus? The code which is stable has become stable through an evolutionary process... tortoises have overtaken hares, and the competition of a software ecosystem is crucial.
As good as more focus would be for some things, I see no way open source can work that way. Since I do also agree with your point, maybe I'm saying there is no solution... but I still feel there is a ballance and room for the distributors to provide that kind of focus when it's important.
Still sensitive from an AC haunt that thinks I'm a zealot, no doubt inspired by my blowing off steam here and sometimes with semi-snarky humor, e.g. in the last Mono/C# vs. "C" thread, I want to be careful.
My pet peeves regarding these issues are not about what language to use in general. My pet peeves are somewhat more about semantics and about the arguments used themselves.
(1) C++ is high level, in fact, it's arbitrarilly high level. The distinction between C#/Java and C++ is better described with the term "managed code" as this really gets at what those languages offer that C++ cannot. Using the "higher level" language reinforces wrong ideas such as "you don't have garbage collection in C++" or "you are forced to use pointers in C++".
(2) It also peeves me when people crow about their language and insist it's the end all be all and your language is just lame. That goes for C++ too. It is not the end all be all of all languages or even compiled languages. These are all interesting technologies and we have to make decisions about them, to judge them, and which to use. If someone says "C/C++ is dead" or even "C/C++ will die" those are opinions they have a right to here on slashdot or anywhere else. However, if someone else, like me, says, "I don't buy it", and "if the idea is to not write the same low level code over and over, what do I get in managed code that I can't get with a class system or library in C++?" I know some answers to that question, and it's not "a garbage collector" and it's not "you don't have to think about memory", the answers is all about the managed code. Havoc points out some good ones, e.g. "A managed language solves some of the same problems COM was designed to solve, such as ABI encapsulation, versioning, and modularization." The answers to real questions tell you what circumstances a language is best suited for. And you will not find that one thing is taking over and everything else is dying.
I agree with your rough idea of programming languages as far as that goes, but C++ and C have had garbage collection for years... you use a garbage collecting library or class for your memory allocation.
Write once run anywhere is not possible as you say, write once compile everywhere is possible, but still problematic, but solvable and solved for many projects.
But I understand why people would want WORA from a single compile or from a runtime compile, I see where the C/C++ build environment posses problems. Nothing is perfect. I'm just of the opinion that this does not make C or C++ gone from the world, or soon to be gone, or purposeless. In C++'s case, it does not make C++ purposeless for high level work.
But you have made good points and each of these other languages no doubt has circumstances in which they are the adviseable tool to use.
saying that anonymously is not really "completely frank" in my book.
I'm not a zealot. Use whatever language you like. I'm entirely in favor of Java and can see many reasons to use it in many circumstances.
This is a discussion.
I differ with you on certain arguments ABOUT languages, and about HOW GENERAL some tools are or are not, too bad you can't handle that, but there is it.
But what you totally miss, and why I'm responding to an AC yet again is: THERE ARE CLASS AND FUNCTION LIBRARIES TO TAKE AWAY THE GRUNT WORK FOR GOBSHITE'S LIKE ME YOU TWIT! This is the problem, not with Java or Python or C#, but fools like you that think using those languages is the only way in the world to get rid of grunt work.
first, thank you for the conversation. Our difference comes to this.
>>C# and Java ARE higher level than C++.
because I agree with what else you say. I have not seen that. You base many arguments on this assumption, that the direction of higher level is to use Java or C# or indeed, Python, etc.
I understand the perspective, but I think it's taking "high" level for granted. You draw comparisons between C++ v. C# as if I could have said the same thing about ASM v. C or C v. C++. But no, ASM does not have the syntax to support some abstractions. C++ can be used with no pointers. A C++ class can be just as high level an class as a C#'s! Seriously what is C++ missing as far as high level goes? It's not missing something... it's got too much... it still allows you to go low level. And this is a different issue. An important one, but...
One other thing, probably semantic. Memory handling does not get abstracted away. It gets abstracted higher. You are, as you know, still handling memory, still having to plan efficient data structures and layout.
I value these other language. I praise Java's wide multi-platform libraries. I have nothing against any tool or language, but there are certain arguments about them which seem wrong to me.
I used to work with MSVC... people said how they could code things faster with VB? This was the environment, I think... MSVC provides a very contained C++ environment, you can use Wizards to make MFC based applications. It was just as fast to create those forms and write the C++ code behind them as the VB... but you give up much less freedom. C++ is just as high a level.
The big pain about C++ is the antiquated development environment... the build process which can have a lot of loose ends and is still platform specific. When the build environment is coherent, the language itself performs as a high level tool.
I know the path goes on and ever higher, we agree on that.
The difference, you see, is that the higher level lanaguage already provides the features whereas YOU have to program them yourself in lower level languages.
why am I not allowed to use any of the thousands of libraries and class systems?
And you need to understand memory because the machine does. You cannot abstract it away. However, I'm not saying don't use higher level memory abstractions, just understand what's being abstracted. Not wanting to understand what a pointer is is a terrible reason to use a VM based language. But if you don't want to mess with pointers, that's a different matter, and it's also, by the way, a paradigm available within C++.
The thing that bothers me is the argument that high-level == VM based languages. That holds no water whatever. C++ is plodding along with it's "high level" + "no paying for what you don't use" philosophy and doing quite well with it.
I have not python only a little so far (although I use Zope), and yes, it's a very nice tool.
Keep in mind that my comment is in general... learn how memory works... not that you can't use a language (or C++ class system) that manages memory at a higher level, but learn how it works anyway. It's important. I suspect you in particular probably have.
But I don't like the VM excuse from people that think the concept of pointer is too complex. I feel it's important to understand how the machine thinks of what you tell it. I love abstraction and see no limit to valuable high level abstractions and languages whatever. But in the end it is a machine, and how it wants to work is the most important thing for an engineer.
I've figured out this simple question. And I've also heard it repeated a thousand times. Don't like pointers? Don't write in assempbly, use a class system that allows you to not use pointers.
You can tell how well hidden my agenda is.
You are a trusting fool to think the bugs you can do nothing about in a VM are always better to have than bugs in you own systems that you could fix or avoid with engineering practices.
But certainly, use whatever language you like. And don't forget... with opinions like your it's best to keep posting anonymously! But when speaking about bugs... remember to speak for yourself.
I believe that is actually converting primatives into objects... boxing them in an object... for wierd languages where everything is an object, except the things that aren't. Also known as languages where operator overloading is stupid and useless... except for string. Also known as a language where the containers take only objects.
I wouldn't be surpised if this, like polymorphism, has an array of official definitions.
Slashdot Mirror
you bastards.
-1 Flamebait. I don't think so. That is totally inflamable. But you don't know that because you are paranoid bastards.
but before you misunderstand any longer... I loathe our Big Brother overlords... but not on the internet, because they're watching you idiots!
this is for your own good!
btw: this is +1 Funny!
you've played right into their hands... things posted to the net are now considered "Microsoft Internal".
:)
!!!!
I turn the TV off multiple times EVERY day already.
Slashdot at it's best.
/. at IT'S best!
Not that it's VERY good... but still,
No matter how well they appear to cover their tracks, stompers get a reputation and no one trusts what they say. Including bosses.
that sounds really naive to me.
Sun an old company?! gee... I'm like 1000 years old then. :)
I have not thought of it too much either, though I know some issues would be easily checked (like preventing use of pointers outside low level classes... etc.), harder would be a declarative language for defining the acceptable paradigm, not leaving paradigms out, and probably a lot of difficult special cases of things that are part of people's current standards which would be hard to check with such a tool.
However: the advantage of not having a language built to the safe paradigm is that C++ and this tool would allow a choice between multiple paradigms (infinite? at any rate a large number). It would be meant to prevent mixed-paradigms on a module by module basis (e.g. if you have a pointerless paradigm you were using at the application level, you probably don't want that in the system classes used, etc.)
lint++?
... superlint? lint++?
A tool that loads a coding standard schema to check code (on it's way into revision control and/or at build time?).
which makes me wonder... is what we want for C++ to catch up is a way to specify your coding standard to a tool that can check source to see... e.g. if a company uses a class system which supports a pointer free paradigm at the app level in C++, they specify their standards in a documents, and this would be a tool to check the code for compliance.
I understand this problem... C++ is multiparadigmed, but a program should adopt a single paradigm.
>In C, you can easily leak four bytes at a time, and good luck finding it.
you can easily use a debug version of the libraries that will tell you if there is a leak when you exit, it can tell you where the memory was allocated, and if the memory was overwritten.
Most of your answer is just "hey, you are just not good enough at Java"... while I cannot defend the parents skills with Java, I can that dumping the heap and writing scripts to analyse the output is not "not having to worry about memory". There is nothing wrong with that (quite the opposite, it shows your willingness to take the steps nec. to make a good product), but you also have to take into mind the time it took to get that expertise. You may have written those things in one hour (or was that one hour "programmer time"?), but you must also count how much time did you spend becoming familiar enough with that output to be able to write that in one hour.
I think Java has a lot going for it although I have only done limited work in it. However, I bristle at the idea (which you did not advocate but which is often heard) that somehow the answer to good programming isn't detailed, that memory issues are abstracted away, good programming does come down to things such as you have learned... how to make your VM dump memory and looking very closely at how you use memory.
OTOH: we have a java application which has been praised by a user that is still on OS/2 (don't ask)... try that with C/C++.
but slightly.
Could linux itself exist under a process that enforced focus? The code which is stable has become stable through an evolutionary process... tortoises have overtaken hares, and the competition of a software ecosystem is crucial.
As good as more focus would be for some things, I see no way open source can work that way. Since I do also agree with your point, maybe I'm saying there is no solution... but I still feel there is a ballance and room for the distributors to provide that kind of focus when it's important.
Still sensitive from an AC haunt that thinks I'm a zealot, no doubt inspired by my blowing off steam here and sometimes with semi-snarky humor, e.g. in the last Mono/C# vs. "C" thread, I want to be careful.
My pet peeves regarding these issues are not about what language to use in general. My pet peeves are somewhat more about semantics and about the arguments used themselves.
(1) C++ is high level, in fact, it's arbitrarilly high level. The distinction between C#/Java and C++ is better described with the term "managed code" as this really gets at what those languages offer that C++ cannot. Using the "higher level" language reinforces wrong ideas such as "you don't have garbage collection in C++" or "you are forced to use pointers in C++".
(2) It also peeves me when people crow about their language and insist it's the end all be all and your language is just lame. That goes for C++ too. It is not the end all be all of all languages or even compiled languages. These are all interesting technologies and we have to make decisions about them, to judge them, and which to use. If someone says "C/C++ is dead" or even "C/C++ will die" those are opinions they have a right to here on slashdot or anywhere else. However, if someone else, like me, says, "I don't buy it", and "if the idea is to not write the same low level code over and over, what do I get in managed code that I can't get with a class system or library in C++?" I know some answers to that question, and it's not "a garbage collector" and it's not "you don't have to think about memory", the answers is all about the managed code. Havoc points out some good ones, e.g. "A managed language solves some of the same problems COM was designed to solve, such as ABI encapsulation, versioning, and modularization." The answers to real questions tell you what circumstances a language is best suited for. And you will not find that one thing is taking over and everything else is dying.
I agree with your rough idea of programming languages as far as that goes, but C++ and C have had garbage collection for years... you use a garbage collecting library or class for your memory allocation.
Write once run anywhere is not possible as you say, write once compile everywhere is possible, but still problematic, but solvable and solved for many projects.
But I understand why people would want WORA from a single compile or from a runtime compile, I see where the C/C++ build environment posses problems. Nothing is perfect. I'm just of the opinion that this does not make C or C++ gone from the world, or soon to be gone, or purposeless. In C++'s case, it does not make C++ purposeless for high level work.
But you have made good points and each of these other languages no doubt has circumstances in which they are the adviseable tool to use.
btw, YOU know perfectly well I'm not a zealot and would be happy to work in Java but YOU wouldn't let me!
>>you know damned well what is driving this ignorant tirade of yours.
blowing off steam on slashdot?
you know what? The market will decide. The end.
saying that anonymously is not really "completely frank" in my book.
I'm not a zealot. Use whatever language you like. I'm entirely in favor of Java and can see many reasons to use it in many circumstances.
This is a discussion.
I differ with you on certain arguments ABOUT languages, and about HOW GENERAL some tools are or are not, too bad you can't handle that, but there is it.
But what you totally miss, and why I'm responding to an AC yet again is: THERE ARE CLASS AND FUNCTION LIBRARIES TO TAKE AWAY THE GRUNT WORK FOR GOBSHITE'S LIKE ME YOU TWIT! This is the problem, not with Java or Python or C#, but fools like you that think using those languages is the only way in the world to get rid of grunt work.
first, thank you for the conversation. Our difference comes to this.
>>C# and Java ARE higher level than C++.
because I agree with what else you say. I have not seen that. You base many arguments on this assumption, that the direction of higher level is to use Java or C# or indeed, Python, etc.
I understand the perspective, but I think it's taking "high" level for granted. You draw comparisons between C++ v. C# as if I could have said the same thing about ASM v. C or C v. C++. But no, ASM does not have the syntax to support some abstractions. C++ can be used with no pointers. A C++ class can be just as high level an class as a C#'s! Seriously what is C++ missing as far as high level goes? It's not missing something... it's got too much... it still allows you to go low level. And this is a different issue. An important one, but...
One other thing, probably semantic. Memory handling does not get abstracted away. It gets abstracted higher. You are, as you know, still handling memory, still having to plan efficient data structures and layout.
I value these other language. I praise Java's wide multi-platform libraries. I have nothing against any tool or language, but there are certain arguments about them which seem wrong to me.
I used to work with MSVC... people said how they could code things faster with VB? This was the environment, I think... MSVC provides a very contained C++ environment, you can use Wizards to make MFC based applications. It was just as fast to create those forms and write the C++ code behind them as the VB... but you give up much less freedom. C++ is just as high a level.
The big pain about C++ is the antiquated development environment... the build process which can have a lot of loose ends and is still platform specific. When the build environment is coherent, the language itself performs as a high level tool.
I know the path goes on and ever higher, we agree on that.
cheers.
C++ is a high level language.
The difference, you see, is that the higher level lanaguage already provides the features whereas YOU have to program them yourself in lower level languages.
why am I not allowed to use any of the thousands of libraries and class systems?
And you need to understand memory because the machine does. You cannot abstract it away. However, I'm not saying don't use higher level memory abstractions, just understand what's being abstracted. Not wanting to understand what a pointer is is a terrible reason to use a VM based language. But if you don't want to mess with pointers, that's a different matter, and it's also, by the way, a paradigm available within C++.
The thing that bothers me is the argument that high-level == VM based languages. That holds no water whatever. C++ is plodding along with it's "high level" + "no paying for what you don't use" philosophy and doing quite well with it.
and there is also no open bracket.
I have not python only a little so far (although I use Zope), and yes, it's a very nice tool.
Keep in mind that my comment is in general... learn how memory works... not that you can't use a language (or C++ class system) that manages memory at a higher level, but learn how it works anyway. It's important. I suspect you in particular probably have.
But I don't like the VM excuse from people that think the concept of pointer is too complex. I feel it's important to understand how the machine thinks of what you tell it. I love abstraction and see no limit to valuable high level abstractions and languages whatever. But in the end it is a machine, and how it wants to work is the most important thing for an engineer.
javagator, the comment you reply to was self referential... not about you!
or at least, it seems that way to me.
I've figured out this simple question. And I've also heard it repeated a thousand times. Don't like pointers? Don't write in assempbly, use a class system that allows you to not use pointers.
You can tell how well hidden my agenda is.
You are a trusting fool to think the bugs you can do nothing about in a VM are always better to have than bugs in you own systems that you could fix or avoid with engineering practices.
But certainly, use whatever language you like. And don't forget... with opinions like your it's best to keep posting anonymously! But when speaking about bugs... remember to speak for yourself.
"Microsoft - Duplicates anything they don't own and control"
and that is only reason #13!
I believe that is actually converting primatives into objects... boxing them in an object... for wierd languages where everything is an object, except the things that aren't. Also known as languages where operator overloading is stupid and useless... except for string. Also known as a language where the containers take only objects.
I wouldn't be surpised if this, like polymorphism, has an array of official definitions.
I can write a web server in bash in one line! AND it installs itself.
apt-get install apache
verily... bash has gazillionized my productivity... it used to take me over six months to write a decent web server.