Really frustrated when moving from C# to C++
-
And you haven't even scratched the surface... The frustration will last for another five years aftetr which, if you make it through, you might get addicted.
FILETIME to time_t
| FoldWithUs! | sighist | WhoIncludes - Analyzing C++ include file hierarchyWait until the OP finds out you can't put two greater than signs next to each other when using templates without the compiler barfing. :)
-
My observation is that people who discount the importance of knowing how things really work (i.e. how things work in the absence of an unseen, out-of-thread garbage collector) tend to be people who look at everything from a very PC-centered perspective. On a PC, or anything resembling a PC, it probably does not make sense for application code to need to do anything special just to effect proper garbage collection. But a real computer scientist, or even a reasonably good (i.e. versatile) programmer, should realize that not everything is a PC, and that there is value to being able to run things on devices that are not PCs (e.g. microcontrollers retailing for $2). This is just one of a whole family of spurious, PC-centric arguments I see on the Internet, which has, after all, become a network largely comprised of PCs. Statements like "every application should have logging" or "every application should be OO" fall into the same category. Such assertions really seem nonsensical if one considers the code that runs inside of digital wristwatches, DVD drive controllers, and such.
Can you really blame them? That's what everybody keeps teaching them. And they are used to having a comfortable IDE, a big fat framework, strong mulitcore processors with a strong GPU to help out, a few gigabytes RAM and a few terabytes mass storage at their disposal. This also means that they are used to clubbing down many problems simply with brute force and getting away with it. And when they don't get away with it, they usually are at a total loss to explain why, since they have done everything the right way.
"I just exchanged opinions with my boss. I went in with mine and came out with his." - me, 2011 ---
I am endeavoring, Madam, to construct a mnemonic memory circuit using stone knives and bearskins - Mr. Spock 1935 and me 2011 -
Wait until the OP finds out you can't put two greater than signs next to each other when using templates without the compiler barfing. :)
But of course you can[^]! I've once stumbled over some guru discussion about the different ways to solve that, make it compatible with the existing C++ standard, figuring out side effects and code that might break silently etc. It's downright scary-amazing to see how complicated it is to fgit such a fix in.
FILETIME to time_t
| FoldWithUs! | sighist | WhoIncludes - Analyzing C++ include file hierarchy -
You're really gonna miss the String class. :)
- S 50 cups of coffee and you know it's on! Code, follow, or get out of the way.
What's so bad about CString? ducks
FILETIME to time_t
| FoldWithUs! | sighist | WhoIncludes - Analyzing C++ include file hierarchy -
Interesting, I hadn't considered the problem it that way, and I think for your points are valid. That said, I've only really developed for PC and I've only really been a .net developer professionally, I enjoy tinkering with other languages and I miss c++, which I use to write my dissertation. I think the deepest separation is between those who like programming and computing for its own sake, and those who do it solely as their day job. I think a large proportion of the latter group fall into your PC-centric category as this is what pays and what is easiest to get work in. Got my 5 anyway!
Sort of a cross between Lawrence of Arabia and Dilbert.[^]
-Or-
A Dead ringer for Kate Winslett[^]Just immagine you had a C64 from 1983. A small 8 bit CPU at less than one MHz and about 50k RAM which you can use. None of the brute force methods you are used to would get you anywhere. Multitasking / multithreading? From the hardware side that would have been no problem. All you need is a hardware timer and an interrupt. But then the interrupt routine, which would have to handle the scheduling of the threads and also the bookkeeping, would nibble away at your little memory. And the effectiveness of having a CPU, which with luck executes 100000 - 200000 instructions per second, do this is more than questionable. Just-in-time-compiler? Or any compiler at all? An IDE which goes beyond a simple text editor? How? CPU, memory and available mass storage (two flopy drives with about 160k disk space each, if you are lucky) are simply not enough. A framework? How many libraries do you think you can load into your memory? And, with only very slow floppies available, some kind of swapping mechanism is also out of the question. Good code? That means wrapping everything up in functions and/or objects. Now and back then, calling a function or method was the same procedure: Push parameters onto the stack, call the function (which then cleans up the stack by removing the parameters), the function then does its job, pushes the return value onto the stack, returns and finally leaves it to the calling code to remove the return value from the stack again. Very quickly the little CPU is working more on the stack than on the actual program. Obviously, good code back then was defined a little differently than today. Those were only a few examples where the brute force approach has become the normal way to go. The more limited your system gets, the less useful the 'normal' approaches become. And I have enough things which were considered obsolete come back.
"I just exchanged opinions with my boss. I went in with mine and came out with his." - me, 2011 ---
I am endeavoring, Madam, to construct a mnemonic memory circuit using stone knives and bearskins - Mr. Spock 1935 and me 2011 -
Mostly because of how pointers and references work, and also function definition seems very different. :((
Once you formally understand data types, memory addresses and l-values, getting pointers should not be that difficult. A pointer is a data type whose admissible values are memory addresses, and whose operations are referencing (input: an l-value, output: the l-value's memory address) and dereferencing (input: a memory address, output: whatever data is in that memory address as an l-value, exception thrown if the original memory address is null). Once you formally understand pointers, getting references is very easy: A reference is a pseudo-data type that is implemented a hidden non-null pointer. When you declare a reference, it must be passed an l-value, which will be referenced. As we have already seen, the referencing operation yields a memory address, which will be used to initialize the hidden pointer (hence, the pointer will be non-null). When you use an already declared reference, the hidden pointer is automatically dereferenced, so you do not have to do it explicitly. As a result, no pointer stuff is explicitly coded. So the message is: If you want to understand pointers in C++, learn what an l-value is. --- Myself, I migrated the other way (C++ to C# 2.0), and had a hard time understanding delegates, because some idiot said that delegates were, deep inside, nothing but function pointers, and I even more stupidly believed him. (Or her. I still do not know the gender of whoever wrote that article.) But that is not true: a delegate contains two pointers: a (member) function pointer, and a this pointer, which may be null (if you pass a static method).
Eduardo León
-
And you haven't even scratched the surface... The frustration will last for another five years aftetr which, if you make it through, you might get addicted.
FILETIME to time_t
| FoldWithUs! | sighist | WhoIncludes - Analyzing C++ include file hierarchy -
Mostly because of how pointers and references work, and also function definition seems very different. :((
It always took a lot of hard work to move from C to A
A hidden needle is way more effective than an unsheathed sword. That is, in the hand of professionals. What about you? Just pray your enemies are blind
-
I feel the same way about C++ to C#. I've found you still need to check for null just as often and C#, without the .NET library (libraries are really a different topic...IMHO), is hardly the C++ 2.0 most C/C++ programmers had been longing for. I would like a reinvented C/C++ without all the historical baggage but I'd like to keep pointers and references the way they are. I had high hopes for C# but I think they missed a great opportunity to "fix" what was wrong with C/C++ and instead ended up trying to "fix" what was wrong with Java.
bob16972 wrote:
instead ended up trying to "fix" what was wrong with Java.
That might have been true for C# 1.0, but C# now has closures, delegates, anonymous types, lazy evaluation (yield) which AFAIK, neither C++ nor Java have. Of course, you could write code yourself to get the same effect, but I hope that's not what we are talking here.
Regards Senthil _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
-
bob16972 wrote:
instead ended up trying to "fix" what was wrong with Java.
That might have been true for C# 1.0, but C# now has closures, delegates, anonymous types, lazy evaluation (yield) which AFAIK, neither C++ nor Java have. Of course, you could write code yourself to get the same effect, but I hope that's not what we are talking here.
Regards Senthil _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
Delegates can be useful in C#, but I'd have to say that I could only be glad if the other 3 never got fully implemented in any language. I'm sure there will be those who differ in opinion with my view and all I can say is, have fun coding/debugging/tracing and reading code that utilizes those features. You have my sympathies. :sigh:
-
Delegates can be useful in C#, but I'd have to say that I could only be glad if the other 3 never got fully implemented in any language. I'm sure there will be those who differ in opinion with my view and all I can say is, have fun coding/debugging/tracing and reading code that utilizes those features. You have my sympathies. :sigh:
Well, I personally love any language feature that helps express my intent concisely, and I must say I use closures and lazy evaluation a lot. Besides, it's not like these features were invented by the C# language designers - they've been around for a long time in functional programming languages. I'm curious though - do you dislike these features because they increase the level of abstraction from the machine?
Regards Senthil _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
-
What's so bad about CString? ducks
FILETIME to time_t
| FoldWithUs! | sighist | WhoIncludes - Analyzing C++ include file hierarchy -
I've just moved to Python - you can't even tell a thread to terminate or pass variables by reference!
Join the cool kids - Come fold with us[^] "Program as if the technical support department is full of serial killers and they know your home address" - Ray Cassick Jr., RIP
-
Mostly because of how pointers and references work, and also function definition seems very different. :((
I've found three courses very useful: these are the Programming Methodology (Java), Programming Abstractions (C++) and Programming Paradigms (Advanced C++) available for free at the "Stanford Engineering Everywhere" site: http://see.stanford.edu/see/courses.aspx[^] You'll be surprised to note that they teach pointers in the Java course. This is what makes you a real programmer. After that, learning about C++ going through the second and third courses is easier. In fact, if you search for "CS106B", you'll reach the course site, where you'll find a C++ Course Reader, which is very well written. Keep in mind that all of this is for free.
-
Mostly because of how pointers and references work, and also function definition seems very different. :((
Bite on it big boy, you are in a mans world now! :cool:
Dr D Evans "The whole idea that carbon dioxide is the main cause of the recent global warming is based on a guess that was proved false by empirical evidence during the 1990s" financialpost
-
Mostly because of how pointers and references work, and also function definition seems very different. :((
Hmm. I went the opposite direction. I spent years programming in C++, and only a couple of years ago started working in C# as well. Like a lot of other C++ -> C# folks, I'm mainly frustrated by the lack of a deterministic destructor (and don't bring up
IDisposable; just don't). I believe you have the harder way to go. I simply had to learn a certain amount of ".NET-fu". I also had to learn to recognize when I was working too hard at the problem at hand. A lot of times I tried to re-invent the wheel, when .NET had a perfectly good set of wheels already available. While the C++ does include a substantial body of 'plumbing' code available in the form of standard and open source libraries, most of it is less systematized and organized than the .NET libraries. This makes it much more likely that you will be tempted to write your own 'plumbing' code. Best quote I've heard regarding C++ came from Bjarne Stroustrup, creator of the language: "C makes it easy to shoot yourself in the foot. With C++, you blow your entire leg off." My addendum, regarding C#: "With C#, the good news is you're shooting rubber bullets."Software Zen:
delete this; -
And even worse, pointers do exist in C# and work very much the same way.
"I just exchanged opinions with my boss. I went in with mine and came out with his." - me, 2011 ---
I am endeavoring, Madam, to construct a mnemonic memory circuit using stone knives and bearskins - Mr. Spock 1935 and me 2011 -
Well, I personally love any language feature that helps express my intent concisely, and I must say I use closures and lazy evaluation a lot. Besides, it's not like these features were invented by the C# language designers - they've been around for a long time in functional programming languages. I'm curious though - do you dislike these features because they increase the level of abstraction from the machine?
Regards Senthil _____________________________ My Home Page |My Blog | My Articles | My Flickr | WinMacro
-
I feel the same way about C++ to C#. I've found you still need to check for null just as often and C#, without the .NET library (libraries are really a different topic...IMHO), is hardly the C++ 2.0 most C/C++ programmers had been longing for. I would like a reinvented C/C++ without all the historical baggage but I'd like to keep pointers and references the way they are. I had high hopes for C# but I think they missed a great opportunity to "fix" what was wrong with C/C++ and instead ended up trying to "fix" what was wrong with Java.
