24 minutes ago, SillyCow said:

The number of times I have to open a browser and google something when I work with Unreal is much larger, because C++ is harder for the IDE to parse. 

Well C++ does work with intellisense, It may not work as well as with C#. I can't speak to that because I haven't programmed in C# much. I find the main trick to using intellisense with C++ is to keep your code in a pretty good state. That means compile often to at least get rid of syntax errors, especially where templates are involved.

I don't really have a problem with C# or Java, but they do have some major performance hits, and put some big constraints on how you can do things.

1 hour ago, gaxio said:

A few things in your code are needlessly verbose.

Picking up my programming habits, sorry about that; at this point it is ingrained into me.

Because I make a lot of games on mobile I have run into some problems when multiplying vectors with other data types. So because vector * float is slower than float * float I avoid it.

The way I would have done this if used in my own game is like this:

void MoveObject(float InXAxis, float InYAxis){
	transform.position +=
      new vector3(InXAxis * Speed * Time.DeltaTime, InYAxis * Speed * Time.DeltaTime,0);
}

2 hours ago, gaxio said:

And you're really all over the place with your naming

As you can see my real naming methods are even worse. I add "In" to everything that goes into a function. All functions must have two descriptive properties in the name, so I get less naming problems like damage.damage(). Lastly all my own code should start with a capital letter, that way I can see what functions are my own; because it is more likely to have bugs.

All these tricks are things I developed while coding my own games. I am not a programmer, everything I know I learned the hard way. They are redundant but has helped me makeover sixteen games from start to end now.

The other problems is probably from me not testing the code.

As a non programmer I prefer the whole over naming thing if it helps make things clear.

I only avoid words like "manager" and "system" that don't actually mean much, even so I still use them from time to time. Object here tells me the function moves a Unity object and won't work on moving a pixel or something else.

22 minutes ago, Scouting Ninja said:

Because I make a lot of games on mobile I have run into some problems when multiplying vectors with other data types. So because vector * float is slower than float * float I avoid it.

What ? You mean you have run into performance problems because of something like that ? That doesn't sound likely at all. All vector * float really is, is:
 

    public static Vector3 operator *(Vector3 lhs, float rhs)
    {
        return new Vector3(lhs.x * rhs, lhs.y * rhs, lhs.z * rhs);
    }

Apart from that, that code snippet isn't too bad. If you're working with the separate components instead of directly using the constructor, that's really just style at that point. ("this" is pretty redundant though, yeah.)

Btw.

Quote

I want to point out that Java is no longer supported by Unity. C# is Unity's main language for users now.

That's "Javascript" or rather UnityScript right there. Doesn't have too much in common with Java.

3 minutes ago, SillyCow said:

The best code completion and refactoring tools for the other popular languages. (Javascript[webstorm] ,Python[pycharm] ,C++[visual-assist] ,etc...)

do not come close to what you get for C# and Java.

In C# and Java I sometimes create whole classes and interfaces just by using automatic refactoring tools. In C++ for example refactoring doesn't yield the same trusted result, and usually fails for all but the most trivial refactoring tasks. Same goes for auto completion. I think that the preprocessor is to blame for this. And many engines make heavy use of preprocessor directives.

I don't think I want whole classes being created for me, but I suppose that's a matter of preference.  Auto completion with C++ however works fine for me in Visual Studio. If it ever stops working it means I've messed something up big time which means I immediately go and fix it.

For me most of this is kind of irrelevant however, because again, I don't want to take the performance hit, I don't want to use a GC, I  sometimes use multiple inheritance which C# doesn't support, and I often use arrays of classes and/or structs (not references to them) which Java doesn't support. (correct me if I wrong on any of this because it's been a while since I programmed in Java or C#, so things may have changed)

1 minute ago, SillyCow said:

If it's only the GC that's bothering you, you can circumvent this problem by not using the "new" keyword too often and recycling your classes. Using "new" in C++ isn't such a great idea either because it will lead to alot memory fragmentation. I tend to try and avoid heap memory allocation in real time games altogether. In unity's C# for example "new Vector3" does not create a reference.

 

I don't know how anyone can avoid heap allocation in programs of any size. That goes for Java, C# or C++. Objects on the stack go away when you leave a routine. The only other option is static/global objects, so yeah I guess if you want to go back to old Fortran style you can do it, but you can end up with a lot of wasted memory.

New in C++ is only as bad as the implementation and I would guess in most performance sensitive cases it's generally overloaded. I almost never use stock new.  I only use it for the simplest things. My current heap implementation uses VirtualAlloc and has slabs, free-lists,8 byte aligned 32 bit pointers (on 64 bit machines) and other such things to make allocation as fast as possible and use as little memory as I can get away with.

Quote

You are completely right. But what kind of game are you making (and on what hardware?) where this is a real performance issue? Today's smartphones are powerful enough to overcome most of this. And the really core performance code in engines is written in C++ anyway.

 

Well I'm working on an MMO planet engine. It uses voxels, run time fractal functions and stuff like that. It has to be capable of generating terrain for a whole planet without actually storing it on disk.  It has to generate two versions of the planet, one for viewing and a very localized version for physics, although admittedly the physics isn't so complex. It's just basically pill to mesh collision and response, and gravity, but since the planet is spherical I can't use any cheats that require one axis to always be "up". 

Quote

With CPUs, unless you are writing your own physics, or some really fancy AI I doubt any of this would matter. Unless you are note optimising your code (making too many real time memory allocations and such). But you can write bad code in any language...

 While it's true that some applications can afford to give away performance, that's far from universally true especially with cutting edge games.  With efficient code you can always do more with less. If nothing else, the game will run decently on older computers. I've bench-marked C++ vs C# vs Java many times, and C++ always comes out much faster, sometimes several times faster.

One more comment, memory allocations can be nearly free.  Most of time time when I allocate something it just pops a bit of memory off of a free list. The code for it is even in the header file so it's inlined. Only on occasion do I have to allocate a new slab. Freeing memory is the same story. I would much rather have a good heap system then trying to reuse memory from old objects in the code of the application.  That sounds like a nightmare to me.