Better at 90's programming than modern programming?

Post Reply New Topic

Once I get to know them I love all IDE's and bug fixing is great. I think the OOP approach is a huge improvement over the procedural approach and now I just learnt services and love them too.

At the heart of any program, regardless of the programming paradigm lives a procedure or a set of procedures. OOP is a slick way of packaging programs that would leave the interfaces free from disruption and make the "guts" of the program easy to replace or enhance without doing violence to the external interfaces of the components. All programs are procedures. From Turing machine state specifications to OOP packages.

ruveyn

C++ is certainly a horrible language but it's not a product of academia. The nastiness of C++ is a more a product of design by committee, a bad habit of adding features which make code slightly more efficient in rare cases (often the default behavior), unnecessary redundancy, a complete unwillingness to correct braindead design decisions in future versions of the standard in order to maintain total backwards compatibility and a tendency to tack on new layers of abstraction to the language which clash badly with those that are already present.

Industry has had a lot of influence on the C++ standard. A better example of a language designed by academia would be Scheme or Haskell.



I don't know too much about Objective C but it looks fairly clean from what I know. What's so bad about it?

In any case it is true that there are languages which are even more horrible than C++ (e.g. COBOL).

How does C++ compare with Java?

I'm almost shocked to hear so many of you calling c++ horrible, considering how popular it is. Is it really so complicated as to interfere with writing programs on a regular basis? I haven't learned any compiled languages yet, and I was considering learning c++ first, but if it's that much worse than other languages like Java, then I won't even bother.

It's not too complicated for me, but it has its shortcomings and the learning curve is probably too high. As an experienced developer with 20 years of C++, the language doesn't really get in my way. I'm a sample of one though.

You have to explicitly delete your objects, or rely on scope based deletion. There is no garbage collector. I would say that is the biggest difference.

IMO, garbage collection is a brain damaged concept and I don't agree with it. This is one of the top reasons I am anti-Java. However, C++ is silent when memory is leaked, and I think that is also wrong.

I'm not a fan of Java either. In contrast to C++, Java tries to force you to do everything in one way, tends to require large amounts of boilerplate code for simple tasks and treats programmers as idiots who need to be projected from themselves for their own good.

In summary, C++ makes it easy to shoot yourself in the foot and even puls the trigger for you. Java does its best to stop you from shooting yourself in the foot even when that's exactly what you want to do.



Popularity is not a measure of quality. If you want to learn a compiled language, I'd recommend C.

As for C++, it's not that you can't write good code in C++; it's just that it's more painful and time consuming than in some other languages. Why should you fight with the language constantly when there are better ways? Anyway, you may want to learn C++ anyway at some point. Even though it's an ugly mess, there is a lot of C++ code out there so it's sometimes necessary to know a bit of C++ of you have to use a C++ library or other C++ code.

ObjC is basically C with infused Smalltalk OO. The syntax between the two languages is very different, and I don't think they play well together. I also happen to think the Smalltalk syntax is garbage.

C++ requires classes that share the same method to share a superclass. This is not a requirement in ObjC. However, all method calls in ObjC are slightly slower than virtual calls in C++, and there are times that virtual calls in C++ themselves are too slow for what I want. I don't find the requirement of sharing a superclass imposing, or even wrong in theory.

Hi,

I just noticed this interesting series of posts.
As one of the greatest programmers ever, I just have to clarify a few things.

Firstly non object orientated languages are not really 'ninetees' languages.

They ( C, cobol, basic, assembler ) are applicable to efficient embedded applications.

They are used because they are closely tied to what is happening in the processor.
There is not much change between what you write and what is actually happening.
Making them efficient in memory and speed.


C++
C++ is an object orientated experiment, a monstrosity that is abused. It has given rise to all manner of convoluted object orientated abstractions that aim to serve programmers rather than applications.
It is as complicated as the obscene abstractions dreamed up by the misguided fools who abuse it.
It is a stepping stone to better languages (see below) and should be abandoned to the waste basket of history.


C# in a proper IDE
This is the next step in object orientated programming, an example of where a good IDE has caught up with the power of the language.
Features such as, graphical GUI design, and Auto-complete make it aproximately 10 x fast than C++ programming in all circumstances.
It replaces C++.

Java
Refreshing improvement on C++, but not quite as easy as C#, c# copied alot of it's core advantages (a runtime-engine, garbage collection, large library of basic functionality ...strings, collections of data)
This also replaces C++
Java is prone to abuse, as are all object orientated languages (See Android development under Eclipse)


Labview.
Completely different from all of the above, check it out some time.



VB dot NET.
Almost exactly the same as C#
You may as well use C#.

The opposite argument could be made. The "proper" IDEs created by Microsoft, Apple, IBM, etc are attempts to compensate for the weaknesses of the languages they've created, by putting a layer on top of them to deal with their shortcomings. The IDE watches you try to program in the horribly broken language a company want you to use, and warns you when you're making mistakes with it. The result is a sort of arms race - languages get more screwball with every release, and IDEs get more bloated to compensate. Has anyone ever innovated with an IDE? Seems like that sort of thing starts showing up when a technology has reached maturity.

Not even close to being impossible. I guess it depends on the debugging capabilities of the IDE *AND* the coder. I once coded on Pocket PCs back in the early 2000's. Debugging was crap, so i set up a TCP socket on the device so i could dump variables from it over 802.11B to a telnet session on the PC

Also, testing methodologies have become better (automated Unit testing for example), some IDEs points out if you're trying to do something stupid (like initialise a varaible without a default value), you can still shoot yourself in the foot - if you want to. Intellisense has helped alot too, dont have to sit and browse MSDN all day.



The two have merged, you can have procedures (functions) on objects and procedures can contain objects, or adress external objects. Imaginagion is the sky. Pretty much every major language is object orientated and support classes. Dotnet and Java has been a good development, i.e. C# is basically the best of VB (the GUI/window properties) and the best of C++, .NET killed the horrible MFCs, and it does not matter what language you write your code in (C++/VB.NET/C#) because its the same compiler (JIT).

Speaking of that last part, i'm looking forward to the day when everything is 100% type safe and you can mix and match code from the major languages in the same project without having to have multiple projects open and have a minor hell maintaining all that code.

Still, i look back and smile at the days when i were coding on the C64...

C# has too many ties to Microsoft.

Guess C was designed by professionals, and C++ was designed by bureaucrats

Saying C++ was "designed" is going too far C++ sort of just happened randomly. The new 2011 standard has warts to fix other warts!

The return value after the function arguments in C++ 2011 has got to be when C++ jumped the shark. Surely a compiler would not need the coder to move the function return value? Surely in 2011, compiler technology is advanced enough that the compiler could figure this stuff out?

That's what I'm saying about IDEs - they cloak the real problems - the new IDEs will probably automatically move the function return value to after the arguments for you by next year, but a compiler should be smart enough to figure out the return value - I mean, the days of one-pass C compilers are long over, right? If a compiler can figure out template specifications and functors, surely it could wait for a few tokens before it resolved the return value of a function?

The further away from the command line or the machine instruction we get, the more modern we are.

Can you give a more detailed example of this?