[Scooby_Doo]

From my own point of view, people shouldn't be naming variables and classes with the same name anyway, but in vast, multi-person projects, that is going to be extremely difficult to maintain. It's like the need to slap someone round the head when I see them using the Java 'This' command.

I actually like cased languages, it gives a lot more flexibility on variable names (and yet, for some reason, I have a habit of defining Booleans as all caps), and makes the language more 'readable' from a purely aesthetic position.

I've always defined variables thusly:

thisIsAVariable - therefore, the first letter is always a lower case, all further words lineated with Caps.

whereas:

ThisIsAClass - using this, it means I can tell whether the program is dealing with an array, an object or a field at a glance.

For example, you see a line of code the reads:

if CheckList(14) then XXXX

now, without casing, you might be looking at a value being passed to a test method, or the 15th object in a variable array, by using the naming convention, you know that what you are dealing with in this case is a Method, whereas checklist(14) is the 15th item in the array checklist.

Since C++ and java doesn't have properties, I'll agree with you. However, Delphi and C# do have properties and those can be used to "hide" the actual variable and allowing you to do design by contact on them. Also possibly giving them a more user friendly name.  For your boolean idea, I usually have a private variable named something like "dirty" and a public read only property "IsDirty".

Btw, for arrays isn't it common standard to use [] for arrays (at least those languages i can remember)


Oh wait... you said variables and classes nevermind...  I'll definetely agree with the "this" command, unless you absolutely need it, ala passing reference for self.

[phatosealpha]

text

Right.  So can we now say "Some people prefer case sensitivity for good reasons, some people prefer case insensitivity for good reasons" and call it a day?

Good.  Now can you please look over what I actually said and point to where I said "Managed code is absolutely the way to go?"

Cause you know, I didn't.  I didn't say Let's have a dot.net revolution, or lets start programming in java.  And I certainly didn't say I was going to program Crysis in VB, or a terabyte database backend in Java.


At which point, we got to where we started from.  And I still allege that C++ is unfriendly, and difficult to work in.  That's why those other languages even exist - because in most situations, C++ is trading ease of creation for efficiency which isn't actually needed.  Heck, you yourself go to C# at times.  Which tells me that you understand that it is easier to do some things in other languages.

Now, may I humbly suggest, that maybe, just maybe, we can do a lot better then C++ in terms of creation efficiency even without managed code?  And as programs grow ever more complicated, the amount of development time required for them is only going to increase, thus making development efficient is a pretty big priority, and maybe it's time to retire the old girl?  She's what, 30, 40 years old now, and she was a bunch of extra additional to a language already designed for procedural programming? 

Oh wait... you said variables and classes nevermind...  I'll definetely agree with the "this" command, unless you absolutely need it, ala passing reference for self.

I never really use java, but there didn't seem to be anything inherently evil about this.  What's so bad about it?  Performance hit for referencing, or what?

[Flipside]

I just hate people using things like, for example...

This House = New House();

As part of a method in a 'Street' class.

Meaning you now have a variable called House which is an instance of a class called House.

It's not 'wrong' it's just bloody irritating coding from my point of view, and can get very confusing on occasion.

[blackhole]

Now, may I humbly suggest, that maybe, just maybe, we can do a lot better then C++ in terms of creation efficiency even without managed code?  And as programs grow ever more complicated, the amount of development time required for them is only going to increase, thus making development efficient is a pretty big priority, and maybe it's time to retire the old girl?  She's what, 30, 40 years old now, and she was a bunch of extra additional to a language already designed for procedural programming? 

If D had a good IDE, I'd use it. Unfortunately, it doesn't. Its also not well-supported, which introduces heaps of problems. There's a really awesome game written in D using OpenGL somewhere called like... tube racer or something like that. TorusTrooper, thats what it was! The game is lightweight, unmanaged, and as totally kickass as C++, but sadly, D is young, and being 30 to 40 years old is more of an advantage to C++ then a disadvantage.

Also, C++ as a language works so well because it mimics how computers operate on a fundamental level. If we change how computers operate on a fundamental level, maybe then we'll need a new langauge for them. Until then, C++ still reigns supreme, and D is close behind, but D has issues of its own. Its these massive amounts of issues that keep new unmanaged languages from being developed. No huge company is going to put their money behind an unmanaged language, its all about managed stuff now, so its inevitable that C++ is probably going to still be here 30 to 40 years in the future. We still use roads, don't we?

COULD we do better than C++? Yeah, D is proof of that. Is that going to change anything? No, D is also proof of that. Everyone was like HOLY CRAP D IS AWESOME!

.... and then nothing happened.

I'm saying that next-gen games are always, always going to need the speed, power, and precise control that C++ offers, to such a degree that it will outweigh most advantages of managed code. Now, the higher-levels of games are moving to managed code for these precise reasons, things like scripting and whatnot, but at the end of the day, the game still needs to be coded primarily in C++, because thats the only way you can get peoples' jaws to drop at a conference room meeting.

I personally think this is just one of those things where no matter how much it sucks, its going to stay around until I'm so old I can't get out of my chair. However I also don't think that C++ is a bad language either, and I disagree that it is a horrifyingly mean and nasty language to program in. I think that it can be a b!tch to learn, but once you get the hang of it, its just another programming language. I transfer from C# to C++ every day with no problem.

Either way, I'm going to go play TorusTrooper and then use my handy dandy C++ skills to help improve Freespace 2.

[Stormkeeper]

@Flipside
I name my variables that way too; I got influenced by my lecturer, so I put either a small m or small g: mTexture/gTexture

[IPAndrews]

Assembly language. Self modifying code for the win.

[Mongoose]

I find the "hard to learn" complaints about C++ kind of amusing, since my school's introductory programming classes taught C/++ as a matter of course.  (And unless you're doing it in emacs at a Unix prompt, you're doing it wrong.)  I never had any real trouble wrapping my mind around the mechanics of the language, and I'd never programmed a day in my life before that.  Granted, I've never really used C++ for anything substantial (other than a bigass read-from/write-to/sort type of program), but at least I feel that I'd be able to do so were I to go paging back through my textbooks.

(Actually, the only other language I ever touched besides C/++ was Scheme.  Dear lord, Scheme...)

[phatosealpha]

You know, a little bit of looking and D really looks kind of interesting.  Without a major backer, I can see why it has trouble, and it obviously needs a good dev environment.   Is there anything for D even in the realm of visual studio?

[blackhole]

http://dsource.org/projects/poseidon

Still needs quite a bit of configuration, but I think its pretty much as good as your going to get.

[Xelion]

I've used

• C# which I found easy to pick up
• Visual Basic I thought was a joke of a program to use
• JAVA had way too many elements to reference
• JavaScript is fun and easy to learn
• PHP was ok but Perls got a lot more to offer
• ActionScript was a great language until it got updated to version 2 and 3 were it turned JAVA-like, which is why I dropped it
• Perl, a language I finally can appreciate because of its regex setup and that it can be used for a myriad of purposes other than web orientated.

I used to like Delphi, then switched to C++, now I'm liking C#.  Not a big fan of perl, php, python, don't like non-compiled weakly typed languages.

Non-compiled languages are just as good as compiled languages, it just depends on what your using it for. C++ is a weakly typed language as well just like Perl.

http://www.ods.com.ua/win/eng/program/Perl5Unleashed/ch1.phtml#InterpretedLanguage1
Perl is interpreted. This can be either an advantage or disadvantage, depending on your needs. For example, Perl has a short development cycle compared to compiled languages, but it will never execute as fast as a compiled language. I discuss the disadvantages in the section called, "What Are the Negatives of Using Perl?," but there are some definite advantages.

One advantage of an interpreted language for tool or application development is that you can perform incremental, iterative development and testing without having to go through a create/compile/test/debug/fix cycle. By eliminating the compile portion of the cycle, interpreted languages can speed the development cycle drastically. It can also be helpful if you are evolving your application by implementing it with minimal capabilities and adding advanced capabilities later.

Because it is interpreted and relatively C-like, you can also use Perl as a prototyping language. This can be especially useful with complex or technically difficult projects such as network communication. You can use Perl's shortened development cycle to evaluate your design and then, once it is proven, rewrite the code in the language of your choice. By the way, C and C++ are good choices because Perl is a lot like C and supports much the same functionality.

I use VB.net primarily - though, that's less about VB.net itself, and more about how much better VS is when using VB then C#.  That and it's case insensitive, which frankly, all languages should be.

Case sensitive exists so you can use variables names that don't conflict with existing names that may be used for methods, functions, etc

We do not need more languages to argue about.

Agreed . There are already way too many...

[Colonol Dekker]

Freakin basic rues,,,,, no sexp i see... but sexust dib;t you thinkk?....

[Flipside]

Step away from the pint-glass with your hands up....

[Colonol Dekker]

Garghhhhh, you were wattchinn mee.

[blackhole]

C++ is a weakly typed language as well just like Perl.

Only if you consider weakly typed to mean implicit conversions, otherwise C++ is strongly typed. A good example being this:

Code: [Select]

float a = 1/4; //Evaluates to 0 since 1 and 4 are integers, and the implicit conversion only occurs after they have been resolved

[Mika]

Me, QBASIC, Visual Basic by some extend (this was years ago), Assembly, C++, Scheme, and lastly MATLAB. Since my stuff is basically scientific computing, MATLAB is the choice of language here. In these cases you cannot beat massive amounts of libraries that are tested by lots of people, and the graphical output libraries stuff is way above what I could code with C++. The faster I get to test some hypothesis, the better.

If I absolutely wanted something to happen fast, (instead of slow and sure), then it is C++ time, but since I have never actually needed object oriented programming, I'm probably losing half of the benefits. I simply want a faster dot or cross product than MATLAB provides... or that there is no way to get rid of those four nested for-loops...

For graphics people, it is Assembly and C++. For electronical engineering, ditto. But I do understand the software level abstraction to get rid of POPs and PUSHs. And remove the chances of code becoming self-modifying and thus reduce risk of program gaining self-awareness (and thus starting nuclear war against all human race).

As an end chuckles, here is a list of undocumented assembly language mnemonics:
http://www.geocities.com/capecanaveral/lab/3550/assembly.htm

I personally like ARG, RPM, VDP, SCE, QVC, IDD and of course LNM (Launch Nuclear Missiles)

You will probably not find this stuff funny if you haven't written stuff with Assembler...

Mika