De-coding BZ Source code, its possible?

[Ded10c]

Developers are typically more responsible for making that call than publishers. It's simply common sense; why re-invent the wheel? We still see games today that can be traced back to the Quake and Doom engines.

[Red Devil]

A mountain does not become lower just because someone climbed it before you.

too true; i have enough trouble understanding someone else's working code without coming up with some of my own.

[Roscoe]

Everybody else's code is CRAP!

[General BlackDragon]

What a shotty thing to say...

[Roscoe]

It was a joke based on years of seeing people react to other people's code.

[Red Devil]

i decoded the BZ2 code awhile back. i have it on a CD...somewhere.

[Apollo]

i decoded the BZ2 code awhile back. i have it on a CD...somewhere.

You too?

[arcanise]

you could start by looking at the game engine BZ uses which is i76 from their decoding it shouldnt be "that" hard

[Ded10c]

First, decompiling BZ2 would be illegal. People have been sued for this, and the only cases I've seen where they were acquitted were where they had to decompile the program to get it to work at all (that's not the case here).

Second, decompiling BZ2 would not produce a perfect result. Even on a program as simple as Hello World, you're unlikely to get a perfect result from the decompiler. In a game like BZ2, which would be hundreds of thousands of lines of code at the minimum, you're not even going to get close.

Frankly, I think Apollo and Red Devil's claims are nothing but hyperbole. Unless they can back their claims up, of course. Anything that deviates from the norm should be dismissed until evidence is provided.

[Sporkinator]

I don't know what this "De-coding" thing is, but I'm assuming we're referring to "decompiling", attempting to reverse the process of compiling a programming language down to byte code or machine code with a decompiler. Converting binary machine code back to its original source code is virtually impossible. There are many different programming languages. Let's suppose I write a program or a game in Pascal and compile it. Let's also suppose you have no idea what language the program was written in, and you just assume C++. Suppose you run it through some kind of C++ decompiler and it (somehow) generates some code. Even if the original program had been written in C++, the identifiers (variable names, object names, constant names, etc) aren't actually stored anywhere in the compiled code, they only exist for the understanding of the human programmer. So now you've got all kinds of generic identifiers such as var1, var2, etc. The entire structure of this code will most likely be different as well, nothing like the original. Comments aren't kept either, as they aren't even essential for the program to run. At this point you might be better off with obfuscated code. In this theoretical example, since the original program wasn't even written in C++, you're even further from having the original code, you've got something new.

I don't actually have any experience with decompiling machine code, so this post's accuracy is not guaranteed.

[Ded10c]

No Spork, you're perfectly correct. I haven't decompiled anything myself because I've never had the need, but that's what happens when you try to reach high-level languages from a low-level starting point. It just doesn't work.

[arcanise]

well if you do some research and fine that the i76 engine or the mechwarrior2 engine used C/C++ which means decompiling is a bit easier, both of these engines are the best candidates for BZ

[Ded10c]

Over thousands, even millions, of lines of code? A perfect - hell, even a *working* replication - is, as Spork said, virtually impossible.

[DuoRanger]

isn't like MW4/Mercs that Jeho managed to decode that in a year and few weeks to build the mekpacks(manly because he works for nasa, and MS gave the code for him and mektek)

[Eddy]

Decompiling typically produces assembly language code. Even if you find a C++ decompiler, the code it produces would not be the same as the original code. It's a version that should produce the same results but will be significantly different. The moment you make a change to the decompiled program, it will likely cause side effects that will drive more changes. Sounds like a rat hole to me...

A good test is to use the Hello World example - write it in GCC, compile it, run it through a decompiler and compare the original code to the machine code. Then try something more sophisticated and compare the code again. I think you will end up with two completely different programs that produce the same results.

Eddy