Re:Don't talk to me about Boost
on
Demise of C++?
·
· Score: 1
" How many people require a sort for every possible type? 99% of sorts will be integers or strings. And writing 2 functions is hardly a chore so its hardly theoretical."
WTF? You are kidding right?
No support for float or double?
C++ will be around for many a moon. Damn near every modern CADD program is writtenin C++. Oh and BOOST kicks ass. Obviously you don't write scientific software of system level software or you would see some of the obvious advantages of what BOOST has to offer.
Someone mod the paren AC up. Jesus google gives us tons for free...sure they make money off the sevice, but that isn' evil. Hey gmail rocks, google search rocks, google maps rock, common what do you mean give back to the community. OSS is not the end all be all people. I think they return more than enough, and there services are benifiting the world in a larger way than some OSS prjects.
Good point. I suspect that good C programmers do this. I am not down on C at all. I still program some things in pure C. I was moer directing my arguments towards the bashers of C/C++ programming in general.
For all your examples you use vectors and ierators. Who uses iterators with vectors?
vector aVec;
for (int i=0; i::iterator it = aVec.begin(); for (it; it != aVec.end(); it++) {..code }
As for the copy example...why would you say:
copy(vec1.begin(), vec2.begin(), vec2.end());
when you can say:
vec1 = vec2;
Your examples have a place with sets, maps, list, etc. Vectors and the STL eliminate so many common programming bugs that people who programm in C++ and don't use the STL are in my opinion not programming in C++. The STL reduces code time as well because it provides so many nice data structures, Red-black trees in the form of sets and maps, linked lists, deque's, queue's, stacks etc. If you code in C you have to write your own data structures for each of these cases. Sure it isn't that hard, but it provides more places for bugs.
Sure nothing is perfectly safe. All that fancy education we get in school has to be worth something. If all code could be written in Visual Basic by complete morons we would all be out of jobs. Show me a secure language that can do everythig for everyone? It doesn't exsist...never will.
Oh and for the people that complain how hard C++ is to learn. It really isn't. If you understand you algorithms, and data structures from school then the STL is pretty simple to pick up on. the rest of he langauge is almost just like Java, or C, or Python etc. It isn't that hard to learn the basics of any modern programming language. Sure Python is easier, but C++ is way more powerful.
Ok that is the end of my early morning rant. Of course we all know that programming lanuage debates are more emotional than religion or politics!!
"Because those kinds of systems are a maintenance nightmare and because training costs for them are far too high."
Actually mainentance costs are minimal on these kinds of systems. Training levels are at a minimum. Usually one person exports the various C++ classes to Python. Python coders/application devleopers code in python. C/C++ guy's code the backend. Once the basic integration system is implace coding is much faster than on pure compile language systems. You also get the benefit of being able to quickly move computationally expensive parts of your code to a compiled language while still maintaining the useability of Python/Ruby etc.
For more information on this approach you should check:
No. That is a big problem for people who are asked to develop large applications that are compute intensive and are to damned lazy to learn a Hard langauge.
Why must people resort to one or the other? Why not combine C++/Python. Or C++/Ruby. You get the quick dirty hacks form the scripting languages to test ideas and then you get the power of C++ to re-write the backend in a computationally efficient manner IF YOU NEED to.
Oh I no why? It is hard. Yeah but it is worth it. Most things in life that are worth it are hard. C++ is a touch language to master. Deal with it. Programming is a tough profession to master.
"On the other hand, smart pointers fall well short of what Java offers - since smart pointer based software typically suffer from reference count bugs, and don't handle references loops."
Yeah you need to kinds of smart points.
boost::shared_ptr boost::weak_ptr
They work great and they give you the good parts of garbage collection without the bad parts.
Piss on Makefiles. I have basically converted my entire project over to Jamfiles at this point. It was quite a pain, but wow is it ever easier to maintain than make files.
As for the IDE debate. I used Kdevelop for quite awhile. Then ported my app to windows using visual c++. Finally gave up on both. Vi/Emacs are both simply better for fast coding. yeah you have to take the time to learn them, but they really are better. Combine Vi with bjam and you are golden. At least that is my current opion.
Why not use a language like C++ were you have the option? In non compute intensive section you use smart pointers and in other sections you use raw pointers. For what it's worth I pretty much always use smart points and never ever notice performance issues. I write scientific code. Performance is everything to me.
As a comparison I wrote a similar Molecular Dynamics simulator in Java and the performance was significantly worse...no numbers as this was a couple of years ago.
My point is that you can use basic garbage collection facilities in a language like C++ and still retain the down-to-the-metal part of the language.
Actually I am not familiar...I have spent a total of perhaps 1.5 hours developing on a Mac. What you are saying totally rocks and adds support to my wanting to get a Mac to develop on!!!
That is interesting indeed. However the issue still remains that I can't justify buying two Mac's. As it stands I don't support them at all. Ulitmately my primary users will be Windows and Linux any way. I am just looking forward to developing on the Mac. If however the majority of my users ran Macs I would certainly support as both old and new.
"Once you've built support for Carbon or Cocoa (or whatever it's called), the change in CPU architecture is just a matter of telling GCC which target to use."
Agreed. But why would a cross platform developer use Carbon or Cocoa? My shop uses Qt and I know quite a few others who seem to be jumping on the Qt bandwagon. However when it comes to code optimization the processor that you are coding for becomes very important. Should I use float or double? What vectorization support does the processor have...will they have it in the future? My main point is that with OXS being BSD based and running Intel it will surely be easier to support Linux/Mac that Linux/Windows as I now do. Although I have just about sreamlined the processes. I have being toying with the idea of putting together a good cross platform developer web site...just don't have the time.
Got a place were my business can purchase that? Sure when I was younger I could justify pirating a copy, but with a business to run and the liability involved...nah. I will just purchase the Intel Mac when it comes out.
"If your CPU handles SSE2 instructions, you should do it, if you want so much to get into it"
It does and i have been into it for quite a while. I am just not releasing any products with SSE inlined at this time....probably soon tho.
The little porting I have done for my Qt based OpenGL app has been as simple as setting the QTDIR varialbe and typing make. The key is to use a cross platform window system. Screw Macs special API's, Windows APIs, and X's apis. That is what I pay the boy's/girls at troll tech to take care of. I have to much else on my plate to worry about this. So if you are programming with Qt as your windowing system it is pretty much a recompile. I am sure there are a few kinks I will run into. I have only done limited testing. Others using Qt seem to be having a pretty simple time with it.
"'all' developers coding on Unix platforms? WTF? The majority have never even used it."
Ok let me refraphse that: Most independant developers...and any developer who has never developed on a Unix platform is bound to be at least a little incompetent.. I have never met a single developer who has NEVER user Unix. I know some that get paid to work in windows shops and they don't tinker much at home. By they all at least have done development in Unix. Most people who truly love to code will argue that Unix is the superior platform. I am arguing that once they are given the chance to work on a platform as polished as OSX with the same tools they love from Linux many will switch..I am one of them.
"OSX just isn't standard enough for cross platform work, btw. the kernel is Unix but the filesystem layout is nonstandard (not to mention the case insensitive filesystem). I also doesn't run X by default so GUI work is out."
Not standard to Linux. Very similar to Irix. I like th eLInux layout better myself, but then I have used Linux almost exclusivly for about 6 years now.
Right and Windows is? Come one. You can set your development environment on OSX however you like...I have a friend who runs his company with all development on OSX and ports to Windows, Linux, Irix, and Sun. Most users are windows users and he would flat out say you have never given it a chance...in fact I argued for quite awhile myself until he walked me throught some of his setups. OSX blows away other Unices for out-of-the-box configuration. A file system hierarchy is irrelvevant. In production code you do your best to get static compiles and filesystem dependancies out anyway. I mean the difference between Windows and Linux is night and Day yet we still manage to do it. OSX moving to Intel platform will open up a lot of doors to developers who need to code very close to the hardward. Me and the guy I mentioned are both scientific coders. It is important to be able to optimize to one instruction set. With all major platforms moving to Intel many of us will start inlining SSE/SSE2 code to gain maximum performance. Yeah most application won't. by scientific code will. Games will. Graphics intesive application will.
So yeah OSX isn't standard. But it works. It has all my developer tools and many that I wish I had. I will pay for that. Porting back to Linux is as easy as setting the PATH, LIBRARY, and a few other application specific environment variables. Windows will alway's be the biggest pain in the ass.
And just to clarify I am not really a Mac fanboy...yet. I don't own one and never have. But I will.
Oh and as far as the cool aid...nah. It is christmas time and the local pub as about 150 new Christmas beers in stock. So yeah I am on the "sauce" but it isn't cool aid!!!
"Apple, which over the years has been losing support from software and peripheral vendors, and may some day soon wake up and find itself in third place on the desktop, behind Linux"
I would agree except for the fact the Apple is switching to Intel processors which means we will have a BSD unix OS with enough company dollars to add all the polish to the final product. Remember many of us don't mind paying for a good product. The reason I use Linux is because Windows is a horrible platform for developers and Macs are to expensive with obscure hardware (good, but obscure).
Having just used (ok played around with) a friends developer edition Intel Mac I can say that I am sold. We are both Unix developers who port applications to Windows. He does all his development on the Mac side and let me just say I am envious. For the most part he has a solid Unix platform that just plain works. That and combined with some of the developer tools on the Mac like the shark profiler...shit I am sold. I love Linux like a little brother. A naggin little brother that alway's needs tinkering with. When it comes time to for development I don't want to dink around with getting video cards working so my OpenGL app renders correctly on every card (this is a bigger problem than you might expect). I also don't want to test the application on every damned Linux distribution out there. Nope I want to sit down and code to a standard machine (Intel-OSX) and then port it to the other platforms. basically I am a wanna be Mac developer. I can't justify the cost of a Mac right now since they are switching platforms, but soon...ah yes. soon.
However I am a developer and not Joe six pack (Joe doesn't like the same beers as me!!). However I do believe that many developers will be sold on the Mac platform and this will create a large spike in applications running on both Mac and Linux. I personally think the OSX/Linux combo will be the 1-2 punch for Microsoft. Steve just has to get the computers to a reasonable price...still charge a premium but make it affordable for Joe Belgian style six pack.
disclaimer: For the most part I am an idiot...I tried to audition for an idiot role in a film, but did not get the part. I have never slept in a Holiday in Express so you should probably take what I say with a small grain of salt.
"The Mac is a niche market and always will be and Firefox isn't a real threat to IE "
I think you will eat those words in 2 years time.
Damn near all developers are coding on Unix platforms and porting to wintel these day's. With the Mac move to intel processors OSX will become the defacto standard platform that all cross platform applications will work on. Microsoft see's the real threat. Many developers (myself included) only support Unix/Linux and Windows because 3 is just to hard. But with OSX moving to intel it will be a total cake walk...in fact it will probably hurt Linux some. I know my primary development machine will become an Apple with OSX on intel procs.
Perhaps, but there are quite a few banks that have support for Firefox. Two Banks I use (Key Bank and Citizens..formally charter one) both work fine with Firefox.
Like a previous poster said, vote with your dollars. Yeah it is a pain in the ass to switch banks but it is for a good cause...and everyone needs at least one good cause worth fighting for.
Like the previous poster said. Once you learn vi or emacs the speed with which you can operate is far far faster than a point and click WYSIWYG editor. Don't get me wrong I do my fair share of coding in kdevelop, but for quick fixes etc, nothing beats vi or emacs...oh and on the vi/emacs topic learn both. The both excel in different areas.
Slashdot Mirror
" How many people require a sort for every possible type? 99% of sorts will
be integers or strings. And writing 2 functions is hardly a chore so its
hardly theoretical."
WTF? You are kidding right?
No support for float or double?
C++ will be around for many a moon. Damn near every modern CADD program is writtenin C++. Oh and BOOST kicks ass. Obviously you don't write scientific software of system level software or you would see some of the obvious advantages of what BOOST has to offer.
Search for area 51. Pretty darned cool!
Someone mod the paren AC up. Jesus google gives us tons for free...sure they make money off the sevice, but that isn' evil. Hey gmail rocks, google search rocks, google maps rock, common what do you mean give back to the community. OSS is not the end all be all people. I think they return more than enough, and there services are benifiting the world in a larger way than some OSS prjects.
Thanks google.
Good point. I suspect that good C programmers do this. I am not down on C at all. I still program some things in pure C. I was moer directing my arguments towards the bashers of C/C++ programming in general.
For all your examples you use vectors and ierators. Who uses iterators with vectors?
..code
vector aVec;
for (int i=0; i::iterator it = aVec.begin();
for (it; it != aVec.end(); it++) {
}
As for the copy example...why would you say:
copy(vec1.begin(), vec2.begin(), vec2.end());
when you can say:
vec1 = vec2;
Your examples have a place with sets, maps, list, etc. Vectors and the STL eliminate so many common programming bugs that people who programm in C++ and don't use the STL are in my opinion not programming in C++. The STL reduces code time as well because it provides so many nice data structures, Red-black trees in the form of sets and maps, linked lists, deque's, queue's, stacks etc. If you code in C you have to write your own data structures for each of these cases. Sure it isn't that hard, but it provides more places for bugs.
Sure nothing is perfectly safe. All that fancy education we get in school has to be worth something. If all code could be written in Visual Basic by complete morons we would all be out of jobs. Show me a secure language that can do everythig for everyone? It doesn't exsist...never will.
Oh and for the people that complain how hard C++ is to learn. It really isn't. If you understand you algorithms, and data structures from school then the STL is pretty simple to pick up on. the rest of he langauge is almost just like Java, or C, or Python etc. It isn't that hard to learn the basics of any modern programming language. Sure Python is easier, but C++ is way more powerful.
Ok that is the end of my early morning rant. Of course we all know that programming lanuage debates are more emotional than religion or politics!!
Cheers
"Because those kinds of systems are a maintenance nightmare and because training costs for them are far too high."
Actually mainentance costs are minimal on these kinds of systems. Training levels are at a minimum. Usually one person exports the various C++ classes to Python. Python coders/application devleopers code in python. C/C++ guy's code the backend. Once the basic integration system is implace coding is much faster than on pure compile language systems. You also get the benefit of being able to quickly move computationally expensive parts of your code to a compiled language while still maintaining the useability of Python/Ruby etc.
For more information on this approach you should check:
http://www.boost-consulting.com/writing/bpl.html
As an aside many computataional chemistry companies are taking this appoach. Mine included (Dynamol).
No. That is a big problem for people who are asked to develop large applications that are compute intensive and are to damned lazy to learn a Hard langauge.
Why must people resort to one or the other? Why not combine C++/Python. Or C++/Ruby. You get the quick dirty hacks form the scripting languages to test ideas and then you get the power of C++ to re-write the backend in a computationally efficient manner IF YOU NEED to.
Oh I no why? It is hard. Yeah but it is worth it. Most things in life that are worth it are hard. C++ is a touch language to master. Deal with it. Programming is a tough profession to master.
"On the other hand, smart pointers fall well short of what Java offers - since smart pointer based software typically suffer from reference count bugs, and don't handle references loops."
Yeah you need to kinds of smart points.
boost::shared_ptr
boost::weak_ptr
They work great and they give you the good parts of garbage collection without the bad parts.
Piss on Makefiles. I have basically converted my entire project over to Jamfiles at this point. It was quite a pain, but wow is it ever easier to maintain than make files.
As for the IDE debate. I used Kdevelop for quite awhile. Then ported my app to windows using visual c++. Finally gave up on both. Vi/Emacs are both simply better for fast coding. yeah you have to take the time to learn them, but they really are better. Combine Vi with bjam and you are golden. At least that is my current opion.
Why not use a language like C++ were you have the option? In non compute intensive section you use smart pointers and in other sections you use raw pointers. For what it's worth I pretty much always use smart points and never ever notice performance issues. I write scientific code. Performance is everything to me.
As a comparison I wrote a similar Molecular Dynamics simulator in Java and the performance was significantly worse...no numbers as this was a couple of years ago.
My point is that you can use basic garbage collection facilities in a language like C++ and still retain the down-to-the-metal part of the language.
Yeah probably because Yahoo bought his company for who knows how much money. He is probably set for life.
Are you having a Homer Simpson DUH!!! moment yet?
Actually I am not familiar...I have spent a total of perhaps 1.5 hours developing on a Mac. What you are saying totally rocks and adds support to my wanting to get a Mac to develop on!!!
Thanks for the cool links and the info!
Cheers.
Depends. Are you using grants from tax payer money for this interesting research or are you self funding it?
Not all research is worht funding just for the sake of pure research....most probably is tho.
That is interesting indeed. However the issue still remains that I can't justify buying two Mac's. As it stands I don't support them at all. Ulitmately my primary users will be Windows and Linux any way. I am just looking forward to developing on the Mac. If however the majority of my users ran Macs I would certainly support as both old and new.
No it means that I will start supporting the newer Macs. As it stands now I don't support them.
"Once you've built support for Carbon or Cocoa (or whatever it's called), the change in CPU architecture is just a matter of telling GCC which target to use."
Agreed. But why would a cross platform developer use Carbon or Cocoa? My shop uses Qt and I know quite a few others who seem to be jumping on the Qt bandwagon. However when it comes to code optimization the processor that you are coding for becomes very important. Should I use float or double? What vectorization support does the processor have...will they have it in the future? My main point is that with OXS being BSD based and running Intel it will surely be easier to support Linux/Mac that Linux/Windows as I now do. Although I have just about sreamlined the processes. I have being toying with the idea of putting together a good cross platform developer web site...just don't have the time.
"why don't you just put Mac OS x86 on your PC?"
Got a place were my business can purchase that? Sure when I was younger I could justify pirating a copy, but with a business to run and the liability involved...nah. I will just purchase the Intel Mac when it comes out.
"If your CPU handles SSE2 instructions, you should do it, if you want so much to get into it"
It does and i have been into it for quite a while. I am just not releasing any products with SSE inlined at this time....probably soon tho.
The little porting I have done for my Qt based OpenGL app has been as simple as setting the QTDIR varialbe and typing make. The key is to use a cross platform window system. Screw Macs special API's, Windows APIs, and X's apis. That is what I pay the boy's/girls at troll tech to take care of. I have to much else on my plate to worry about this. So if you are programming with Qt as your windowing system it is pretty much a recompile. I am sure there are a few kinks I will run into. I have only done limited testing. Others using Qt seem to be having a pretty simple time with it.
"'all' developers coding on Unix platforms? WTF? The majority have never even used it."
Ok let me refraphse that: Most independant developers...and any developer who has never developed on a Unix platform is bound to be at least a little incompetent.. I have never met a single developer who has NEVER user Unix. I know some that get paid to work in windows shops and they don't tinker much at home. By they all at least have done development in Unix. Most people who truly love to code will argue that Unix is the superior platform. I am arguing that once they are given the chance to work on a platform as polished as OSX with the same tools they love from Linux many will switch..I am one of them.
"OSX just isn't standard enough for cross platform work, btw. the kernel is Unix but the filesystem layout is nonstandard (not to mention the case insensitive filesystem). I also doesn't run X by default so GUI work is out."
Not standard to Linux. Very similar to Irix. I like th eLInux layout better myself, but then I have used Linux almost exclusivly for about 6 years now.
Right and Windows is? Come one. You can set your development environment on OSX however you like...I have a friend who runs his company with all development on OSX and ports to Windows, Linux, Irix, and Sun. Most users are windows users and he would flat out say you have never given it a chance...in fact I argued for quite awhile myself until he walked me throught some of his setups. OSX blows away other Unices for out-of-the-box configuration. A file system hierarchy is irrelvevant. In production code you do your best to get static compiles and filesystem dependancies out anyway. I mean the difference between Windows and Linux is night and Day yet we still manage to do it. OSX moving to Intel platform will open up a lot of doors to developers who need to code very close to the hardward. Me and the guy I mentioned are both scientific coders. It is important to be able to optimize to one instruction set. With all major platforms moving to Intel many of us will start inlining SSE/SSE2 code to gain maximum performance. Yeah most application won't. by scientific code will. Games will. Graphics intesive application will.
So yeah OSX isn't standard. But it works. It has all my developer tools and many that I wish I had. I will pay for that. Porting back to Linux is as easy as setting the PATH, LIBRARY, and a few other application specific environment variables. Windows will alway's be the biggest pain in the ass.
And just to clarify I am not really a Mac fanboy...yet. I don't own one and never have. But I will.
Oh and as far as the cool aid...nah. It is christmas time and the local pub as about 150 new Christmas beers in stock. So yeah I am on the "sauce" but it isn't cool aid!!!
Cheers.
"Apple, which over the years has been losing support from software and peripheral vendors, and may some day soon wake up and find itself in third place on the desktop, behind Linux"
I would agree except for the fact the Apple is switching to Intel processors which means we will have a BSD unix OS with enough company dollars to add all the polish to the final product. Remember many of us don't mind paying for a good product. The reason I use Linux is because Windows is a horrible platform for developers and Macs are to expensive with obscure hardware (good, but obscure).
Having just used (ok played around with) a friends developer edition Intel Mac I can say that I am sold. We are both Unix developers who port applications to Windows. He does all his development on the Mac side and let me just say I am envious. For the most part he has a solid Unix platform that just plain works. That and combined with some of the developer tools on the Mac like the shark profiler...shit I am sold. I love Linux like a little brother. A naggin little brother that alway's needs tinkering with. When it comes time to for development I don't want to dink around with getting video cards working so my OpenGL app renders correctly on every card (this is a bigger problem than you might expect). I also don't want to test the application on every damned Linux distribution out there. Nope I want to sit down and code to a standard machine (Intel-OSX) and then port it to the other platforms. basically I am a wanna be Mac developer. I can't justify the cost of a Mac right now since they are switching platforms, but soon...ah yes. soon.
However I am a developer and not Joe six pack (Joe doesn't like the same beers as me!!). However I do believe that many developers will be sold on the Mac platform and this will create a large spike in applications running on both Mac and Linux. I personally think the OSX/Linux combo will be the 1-2 punch for Microsoft. Steve just has to get the computers to a reasonable price...still charge a premium but make it affordable for Joe Belgian style six pack.
disclaimer: For the most part I am an idiot...I tried to audition for an idiot role in a film, but did not get the part. I have never slept in a Holiday in Express so you should probably take what I say with a small grain of salt.
Cheers
"The Mac is a niche market and always will be and Firefox isn't a real threat to IE "
I think you will eat those words in 2 years time.
Damn near all developers are coding on Unix platforms and porting to wintel these day's. With the Mac move to intel processors OSX will become the defacto standard platform that all cross platform applications will work on. Microsoft see's the real threat. Many developers (myself included) only support Unix/Linux and Windows because 3 is just to hard. But with OSX moving to intel it will be a total cake walk...in fact it will probably hurt Linux some. I know my primary development machine will become an Apple with OSX on intel procs.
Perhaps, but there are quite a few banks that have support for Firefox. Two Banks I use (Key Bank and Citizens..formally charter one) both work fine with Firefox.
Like a previous poster said, vote with your dollars. Yeah it is a pain in the ass to switch banks but it is for a good cause...and everyone needs at least one good cause worth fighting for.
Well said.
Like the previous poster said. Once you learn vi or emacs the speed with which you can operate is far far faster than a point and click WYSIWYG editor. Don't get me wrong I do my fair share of coding in kdevelop, but for quick fixes etc, nothing beats vi or emacs...oh and on the vi/emacs topic learn both. The both excel in different areas.