The Big Programming Thread - Page 35

On February 11 2011 23:13 Morfildur wrote:


Other advantages of C# compared to java:

Functional extensions, first class functions, common namings of methods and functions in the .NET library (depending on which java array class you use, it's either .size(), .getLength(), .length, .getCount(), .count(), .getLength(), ... i guess thats the drawback on being open source with no early concept), extension methods, ...

Functional extensions:
For example if i want the ID properties of all customer objects in an array where the customer registered before 2011 (which reminds me, C# properties are really cool, Python has similar, Java doesn't):

Oh, did i mention that the last time i had to use Javas date handling libraries i nearly had to scream because of how complex and unlogical it was? calendar... uhm... yes.... hope for you java folks they improved on that.

Oh, and if we want to make that multi-threaded because the sourceArray is very big... we have to use third party libraries in Java or write a lot of code.

C# multi-threaded (.NET 4.0):
IEnumberable<Int32> results = sourceArray.AsParallel().Where(c => c.RegisteredDate < (new DateTime(2010, 1, 1)).Select(c => c.ID);

Extension methods:

I recently had the case that i wanted to have the position of units on a map, i stored them as instances of the Point class. I later had the case where i wanted to calculate the distances between units, so instead of creating a new class and using that everywhere, i simply created a "PointExtensions" class and added a GetDistanceTo() method to the original Point class.

Oh, did i mention that visual studio 2010 is simply the best IDE i've ever used? Neither eclipse, netbeans, kdevelop, etc even comes close.

I had to write Java for a year, but my knowledge is still limited and dated, if anyone wants to correct me i always welcome to learn something new, but imho C# is just a more modern language, though you still pay for it by knowing less about what happens in the background.

EDIT: If you would dabble in java, to be effective/productive you should really learn proper OO design to utilize the language to its max.

On February 12 2011 00:38 heishe wrote:As such they are nice for mediocre software engineers and software architects who need some codemonkeys to code their carefully designed code-structure, and of course in general for GUI designers, because of Swing and Windows Forms (even though I personally think that the swing GUI looks really ugly), but even after working with all 3 languages (Java, C#, C++) extensively, I have always and will always prefer C++.

On February 12 2011 01:32 Kambing wrote:


Personal preference is one thing, but declaring that C# and Java are for medicore or unskilled developers and C++ is for experts is naive. Each language has its own pros and cons that make them suitable in different situations. Just like choosing the proper algorithm or architecture for your project, picking a programming language is very much a design decision with trade-offs that you need to weigh appropriately.

On February 12 2011 01:32 Kambing wrote:
I would venture to say that many people think C++ is the "best" language for their project when in reality C# or Java are a better choice since it will save them significantly in productivity and maintainability (over any possible performance gains with C++).

On February 12 2011 03:46 Eventh wrote:
To argue which programming language is the best is meaningless, as they depend on the problem you wish to solve.

On February 12 2011 05:00 heishe wrote:
Name one thing that's objectively better about the Java or C# language than in C++.

On February 12 2011 05:00 heishe wrote:


In general yes, but for this discussion it's simply not true, as we're not talking about languages in general, but about Java, C# and C++, with Java and C# being highly similar to C++, just with less features.

You guys are producing the same meaningless statements over and over again. Try to reason please, and don't just put general statements like the one I quoted out there. A basic argument should consist of thesis and reasoning/proof (and an example doesn't hurt). All you are doing is putting thesises out there with no backup at all.

Come on. Name one thing that's objectively better about the Java or C# language than in C++.

try {
  // ...
} finally {
  s1.dispose();
  s2.dispose();
}

using (Socket s1 = /* ... */, Socket s2 = /* ... */)
{
  // ...
}
/**
  * dispose is automatically called once we leave the block similarly
  * to how the destructor of a stack-allocated C++ object is always
  * called once the object leaves scope.
  */

On February 12 2011 05:36 Kambing wrote:


In general, there's no objective measure we can apply to programming languages because most language are able to express any computation (that is actually computable). So the right question to ask is not "can you do X" in a programming language but instead "how well can you implement X" which is strictly subjective (with answers being a mix of verbiage, lines of code, expertise, understandability, etc.).

But I'll bite anyways (with the above caveat) with the obvious example. Complete garbage collection is the major benefit that Java and C# provides over C++ within the family of C-like languages. Good programming practices in C++ such as smart pointers or approximate GC libraries alleviate much of the burden of manual memory management, but not all of it.

The caveat is that, e.g., in Java, with garbage collection you gain automatic memory management at the cost of easy deterministic cleanup (which you obtain in C++ through raii). Since Java only provides non-deterministic clean-up via finalizers, you are forced into the try-finally pattern which is itself brittle, e.g., the following code always cleans up after s1 and s2 unless dispose() also throws an exception.

try {
  // ...
} finally {
  s1.dispose();
  s2.dispose();
}

C# gets around this by providing built-in support for the dispose pattern via the using construct:

using (Socket s1 = /* ... */, Socket s2 = /* ... */)
{
  // ...
}
/**
  * dispose is automatically called once we leave the block similarly
  * to how the destructor of a stack-allocated C++ object is always
  * called once the object leaves scope.
  */

On February 12 2011 05:36 Eventh wrote:

Java RMI, and all you get for free with built in reflection.

On February 12 2011 06:07 heishe wrote:
Automatic garbage collection is only an advantage if you're a bad code designer. It has no upside except for making the lives of bad programmers easier, but it has the downside (as I already said a few posts up) of being a MAJOR slowdown in classes and algorithms which rely on the constant creation (and eventually destruction) of new objects.

It saves the need for typing a few extra lines of code (and I really mean just a few. If you constantly have the need to delete tons of pointers in the destructor of a class, you're doing something wrong), but that is almost entirely irrelevant in practical circumstances. The average software engineer probably creates something like 10 lines of code per hour, since the vast majority of time is spending thinking (unless you're in a structure where you're a codemonkey under someone who designs the broad picture of the code, in which case, as I already pointed out, Java and C# are good for you and your boss, the actual software engineer).

On February 12 2011 06:07 heishe wrote:
And in the very near future, Java and C# will fall even farther behind in that department, as C++ is introducing Rvalue references, and together with that, move semantics and perfect forwarding.

On February 12 2011 06:33 Willes wrote:
Modern software becomes bigger, development of new software raises because markets/ppl/industries/whatever need new software and here you have your biggest advantage over c++: bug-resistance

the chance of producing bugs is alot higher with c++ compared to c#, java, because of the nature of the languages o.c., especially memory organisation produces bugs...
besides technical differences this is a huge factor in a time where speed isnt that much of a factor because of faster cpu's and all that stuff that belongs to it...

programming languages have huge success because they give ppl what they need at a given time, in this case easy oop-able languages are more effectively in large software-projects than c++, different requirements - different languages , thats simply all.

time = money, garbage collectors safe time, also for good programmers, especially in large projects...

mfg

On February 12 2011 07:01 Ace wrote:


That isn't necessarily true. C++ is an older language and as such, components you'd use (aka other people's code and well known solutions to problems) will be extensively documented.

No matter what language you use you will always have bugs. Garbage collectors, like time spent coding, aren't major issues if you are a Software Engineer. It's nice to have but isn't going to make someone that isn't experienced in programming better.

On February 12 2011 05:36 Eventh wrote:

Java RMI, and all you get for free with built in reflection.

// Typechecks fine but will throw an exception at
// runtime if d does not have a f() member.
int foo(dynamic d) { d.f(); }