Thread: Game Programming

Originally Posted by icuurd12b42

Steps I took for my first try.
Get animation strip of the web
Made sprite sets from strip for main character (Walk right, left, jump right, left and stop)
Created character and added code to move it right left and jump right and left and select right animation sprite for each case.
Added sprite for wall/floors
Added code from gravity
Added code for collision events when guy walks on wall or hits it.

I like it so far.
[/b]

well, all those event can be done simply in a "real" programming language. left/right is a simple if statement. gravity is 1 or two lines. The collision can be a simple if(hittest) and that would do what gamemaker does for collisions. But, heres the thing, you can program a custom hittest so that the character interacts with the environment like you want him to. Custom hittests are almost always more efficient too, which is important if you are going to have alot of hittests being proformed.

So, programming through a language is pretty much just as easy.

Plus, when you have a problem happening, you can actually understand better whats causing the problem because you know whats actually happening.

Are we fighting over programming languages now?

Oh boy.

sure wish I was better at drawing. got any tips?

Originally Posted by arby

well, all those event can be done simply in a "real" programming language. left/right is a simple if statement. gravity is 1 or two lines. The collision can be a simple if(hittest) and that would do what gamemaker does for collisions. But, heres the thing, you can program a custom hittest so that the character interacts with the environment like you want him to. Custom hittests are almost always more efficient too, which is important if you are going to have alot of hittests being proformed.

So, programming through a language is pretty much just as easy.

Plus, when you have a problem happening, you can actually understand better whats causing the problem because you know whats actually happening.
[/b]

Oh, absolutely. When you have access to the entire code set, you can apply a proper fix for a problem. This tool would most likely require major hack around when encountering a problem that I could easily fix if I had the code that causes the problem.

On the other hand, you wont catch me creating a game core engine ever again.

Anyway, it’s pretty early to tell what kinds of problems I’ll encounter. I have yet to try the scripting language, so I don’t know how powerful that is. Is all depends on the scope of access to the core features.

Originally Posted by arby

Are we fighting over programming languages now?

Oh boy.
[/b]

Not really. My argument is more about using tools that can make your job faster. In this case, the engine seems designed to make Mario Brother type games. So, I wont bash it because it cannot make any other thing.

I once worked with this guy who would totally refuse to buy components to finish a project. I can write this, I can write that, My code will be much more efficient than any of the components…. 20 minutes in, he had already listed 20 different components… Eventually, my boss summoned me to discuss the issue. We made a run down on how many components would have to be written. 240 components in 2 weeks (all readily existing commercial components written by experts in the field). We bought the components for 1500$, finished the job and the guy got canned. That guy would only write stuff in real programming language… Assembly. Crazy cook.

Moral: Giving up control is hard for the hard core programmer. Yet knowing when to do so can save your ass.

Originally Posted by arby

sure wish I was better at drawing. got any tips?
[/b]

Search the web for canned images or learn to draw or hire an artist. I suck at drawing too. If you pair up with someone who can, that would make all the difference.

Originally Posted by icuurd12b42

I once worked with this guy who would totally refuse to buy components to finish a project. I can write this, I can write that, My code will be much more efficient than any of the components…. 20 minutes in, he had already listed 20 different components… Eventually, my boss summoned me to discuss the issue. We made a run down on how many components would have to be written. 240 components in 2 weeks (all readily existing commercial components written by experts in the field). We bought the components for 1500$, finished the job and the guy got canned. That guy would only write stuff in real programming language… Assembly. Crazy cook.[/b]

So, I guess you are a programmer for a living then?

Originally Posted by icuurd12b42

I once worked with this guy who would totally refuse to buy components to finish a project. I can write this, I can write that, My code will be much more efficient than any of the components…. 20 minutes in, he had already listed 20 different components… Eventually, my boss summoned me to discuss the issue. We made a run down on how many components would have to be written. 240 components in 2 weeks (all readily existing commercial components written by experts in the field). We bought the components for 1500$, finished the job and the guy got canned. That guy would only write stuff in real programming language… Assembly. Crazy cook.

Moral: Giving up control is hard for the hard core programmer. Yet knowing when to do so can save your ass.
[/b]

haha, that guy is crazy. All assembly? XD

it must have taken him ages to do anything =P

How big were the components? what sort of components? Simple ones or huge ones?



as for graphics, you can try to draw them yourself, look at other games that are similar to what you want and see how the did it. If you keep drawing you'll get better.

Otherwise, you can make the game using placeholders. A placeholder is something simple like a circle or a stickfigure that is used when you dont have finalized graphics. If you make a game using placeholders then when you're finished then engine or a decent demo, you can show it to artists and you have a good chance for some to want to work with you since they see what you are capable of. Thats what i'm doing right now with that game i showed you because I can't draw for beans either =P

ok. well i think i've got movement and borders and stuff down, but now i need to learn how to say change scenes when in collision with an object and stuff like that to make a nice side scroller. i gotta figure out how to make the background scroll and stuff too.

here is what i've got so far:

Float!

Yeah, I program games, but I do it with Game Maker. Its not great, but it gets the job done. I was able to get the basic engine of the original LOZ, some megaman, where one plays as the bosses, and I worked on a Dragoon platform game. The dragoon game stopped during the process of making enemy ninja AI, and I haven't fully developed any game to completion. I get so fed up with one and move on to another, to later return to the original game.

Originally Posted by TweaK

So, I guess you are a programmer for a living then?
[/b]

20 years in now. Long enough to have had many “What language is the best” debates. I used most of them in my career.

Originally Posted by TweaK

haha, that guy is crazy. All assembly? XD

it must have taken him ages to do anything =P

How big were the components? what sort of components? Simple ones or huge ones?
[/b]

At that time, ASM was still popular with hard core programmers. C++ Interdev was released along with Power Builder and VB 3. ActiveX was still called COM.… To give you a time line.

Most of the components were image manipulating functions. Some communication (network and modem).

They wanted to port simple Unix controlled video surveillance that used standard TV and cameras to CCD captured video transmitted via any computer communication device and display security details (alarm detectors) and video directly on the computer monitor (on Win95) as well as control real world objects at remote locations through same communication system.

We got Lead Tools (Straight DLL, Non COM enabled) for all the imaging stuff and other various communication controls.

We could now concentrate on the control code that glued the whole system together.

There was a lot of complaining at the time about how DLL must be slower and shit. The complainers did not realize that all Windows API call were actually a DLL rpc. And that coding in VB using mostly rpc was as fast as C or C++.

The debate crumbled when the application was written in 2 weeks and was faster than the first C and ASM based DOS prototype (Who‘s code was totally not portable to Windows). Oh yeah, that felt good.

Originally Posted by Casualtie

ok. well i think i've got movement and borders and stuff down, but now i need to learn how to say change scenes when in collision with an object and stuff like that to make a nice side scroller. i gotta figure out how to make the background scroll and stuff too.

here is what i've got so far:

Float!
[/b]

and you said you can't draw? =P

thats nice.

making the backround scrool is as easy as making it move instead of the character =P

Just tell me we're not fighting over program languages. Just tell it to my admittedly ugly face.

Originally Posted by Kaniaz

Just tell me we're not fighting over program languages. Just tell it to my admittedly ugly face.
[/b]

We're not fighting over programming languages.

We're debating programming languages. =P

I didn't draw it. I got it off google. Its a new pokemon I think.

..It's from fuggin Mario.

Originally Posted by TweaK

..It's from fuggin Mario.
[/b]

haha... I THOUGHT it looked a lil familiar....

Well, I’m done playing with Game Maker for now. Here is My quick evaluation.

For learning Game Programming
9
For learning what is required in a game, required recourses and how they are used
10
For learning objectized programming
9
For learning event driven programming
9
For learning map driven games
9

Really liked the sprite management tool
Really liked the multiple view on games
Really liked the ability to stretch the view on the game
Really liked the view following the player object
Really liked the built in room persistance (can I save and load room? I don't know)
Really liked the one file distribution (only need to copy it)
Really liked that you can set a ZOrder to the objects so you can move behind them

Liked the development environment
Liked the way the objects are structured and work
Liked the way to convert the game features into 3D mapping (Looks good on paper, could not try sample with my unregistered version)

Disliked the collision system and supplied methods to correct errors
Disliked that deviating for the proposed implementation method causes grief and weirdness
Disliked the code writing system (the objectized language is OK)
Disliked that I could not debug or print the value of a variable[edit you can debug and add watch but cannot set initial break point]. Not even in a message box (Not using code)
Disliked that there is no hint in the development environment about what variables and objects are available. Though there are object picker systems and variable viewer, there is no auto complete system in the code editor or a variable picker in the visual programming..

So, if you stick with the implied method for implementing a game then you are OK. Divert from it and there could be trouble (read on).


After playing around with my first game, I started to realise that most of the code that handled things was associated with the main character object. So, if you were to want to add new world objects to the map, you have to go into the character object and add code to handle it.

Also, to deal with weird interaction with objects, odd things were being done like turning off the gravity when hitting and repositioning objects and turning it back on when not touching an object.

This style did not agree with me. I wanted to create generic objects and add them to the map without the need to reprogram the main actor or any other actor for that matter. I want a bullet to kill what it hits, I want a wall to block any actor, I want a floor to support any actor. And so on.

So, I started to transfer code from my old Sonic actor to a newer more generic one. Created a floor object and added the collision event (in ObjFloor) to set the colliding object’s vertical velocity to 0. Not touching the gravity witch I planed to leave to 1g.

This is where I started to be disappointed. The collision event has no generic version so I had to add a Sonic specific collision event. Time for sub classing! I created a generic actor object, derived my Sonic object from it and added a Collide With ActorGeneric event in ObjFloor. Still, I lost flexibility as the floor object is now dependant on the ActorGeneric class. That’s OK, as long as the users of the object do derive their objects from it.

I derived ActorPlayer1 from ActorSonic, set the room with a floor (make with ObjFloor instances), set the rooms viewer to show 30% of the room, following the ActorPlayer1 object (cool)

This is where I saw how bad the collision and the way to handle it was.

The method to fix collision that were detected where not good. I could not reposition the colliding object properly. The collision event seems to be triggered only once so, If I misplaced the character and placed him still within the floor, the event would not trigger again. The vertical velocity would increase until my guy disappeared (It seems).

Plus, the code to move right of left was somehow nullified while the game was stuck in that particular conundrum.

Noticed that the world is drawn before you handle collision events, making my guy bouncy when collision was working as written. If I moved the guy one pixel to high, the gravity would bring him down, the world would be drawn and the collision code would trigger. If I moved the guy one pixel less high (in collision) the code would freak and my guy would freeze for a few seconds (not moving left or right or jump) and would disappear. No, I could not set the guy at the right place, there was no right place, either bouncy guy or code freeze.

I got fed up trying many kinds of way to fix this bug.

I’ll probably get back to it soon. I’ll keep you posted.

let's ramp this arguement up a notch.....

Cobol Vs. Fortran

ready....Go!

Originally Posted by Ynot

let's ramp this arguement up a notch.....

Cobol Vs. Fortran

ready....Go!
[/b]

Cobol obviously wins because if you take the b out, you're left with cool.

That pwns Fortran.

And what you're describing there is exactly why I dislike gamemaker =P

I like my flexibility =P

But ya... as a very basic learning tool it may prove semi-useful. But otherwise I reccomend to stay away.

Wanna try Actionscript? I'm interested in your opinion on it =P

Originally Posted by Ynot

let's ramp this arguement up a notch.....

Cobol Vs. Fortran

ready....Go!
[/b]

A long time ago, in a galaxy far far away, Cobol was invented for huge business management software.

Meanwhile, the more academic mathematicians needed a language tailored to their skills. Fortran was the gift form the Programming Language gods.

Cobol will never go away. It is implanted in the business psyche. Still today, billions of lines of Cobol code written in the 70 and 80 are executed everyday by banks and governments. To add features and make changes, your are condemned to do it in this language.

Fortran code is probably still used by scientific purists, running scientific experiments..

Fortran has a compiler, I can’t remember if Cobol is compiled.

But, today, linking to DLLs and COM is supported by most languages so, really, programming becomes a multi language endeavour.

Usually, you select a language that does most of what you want well and tack on the rest with DLLs and COM.

The code you write now is likely to be 5% of the entire code executed by your application. That 5% (if written in interpreted code compiled to byte code) runs at 95% of the speed of a fully compiled code. So, using this information, a mixed language code, who’s glue code is byte code, will run at 97.5% the speed it would if fully compiled.

[edit the 97.5 calc may not have been calculate right. will come back to confirm]

Of course, if your code consists mainly of embedded language specific calls then the code runs at 95% of the speed it could run if fully compiled. That, of course (the 95%), is if the compiled code mirrors exactly the original byte code. I’m not taking into account that you can delete a huge amount of useless (but still executed) code from your application when you have full control over the code to fully optimize things.

I program game, in C++ i don't use microsoft visual C++ i use Dev-cpp, i think it's great, I've made this game It uses Allegro, it's a very good graphics library, it also has support for the mouse, keyboard, timing functions and MIDI support. So i would recommend it for game programming. I not exactly sure what you're asking, but start with the basics ie. pong. You don't even have to make the game use two human players the A.I. in pong is really simple
example: (C++)
if(ball_y>computer_y) computer_y++;
else
if(ball_y>computer_y) computer_y--;

p.s. sorry if this post seems to jump to different subjects.

I remember using Allegro with C++. Fun times, I had never compiled anything before and was following tutorials to the letter. Happy days.

I made a game in it called "revenge of the blue dot", in which a blue pixel would chase you directly and there was absolutely no gameplay strategy except 'keep your lead', but, well, shut up.

Another graphics engine for game programming is the Irrlicht engine. You'll need to add network support and sound yourself though. Strategist and Dawn of Daria both use a modified version of this engine. (Now who said 2D-games are dead?)

CasualTie, game programming might sound interesting, but it is very complex and not a good introduction to programming. To be a good game programmer you will have to be a good programmer, and the latter itself takes a good five years. I don't want to discourage you if you are really serious about it, but you'll have to determine if game development is really worth the time and effort needed to put into it. Do you see this as a temporary hobby or as a lifelong pursuit?

Now that all that is out of the way, let me give you my advise. First, learn a high-level language that doesn't lock you into a specific platform (i.e. don't start with .NET languages). Python is an excellent choice for many reasons. While Eric Raymond lays out myriad reasons why Python is a great first programming language, I will list just the highlights:

1. Compact - easy to learn, remember, and use; fewer bugs
2. Concise - thanks to dynamic typing and builtin high-level data types, Python code is about 7x shorter than Java and 10x shorter than C++
3. Productive - shorter development cycle because it's interpreted and concise
4. Clear syntax - fewer bugs in large projects
5. Cross-platform - runs on post architectures and OSes; wxPython for cross-platform GUIs
6. Huge problem domain - useful for: quick scripts, command-line apps, GUI apps, web apps, CGI scripts, etc.

Python also has a great tutorial to get beginners up to speed. Once you learn the basics (expect 2 weeks to a month), stick with it for at least a year and study other people's code. Then learn the PyGame API. Once you're comfortable with that, then you might consider learning a lower-level language like C++ (still higher-level than C or assembly). C++ will give you even more power than Python and certainly higher performance, but you have to be willing to spend years to become a good programming, especially with a language like C++. Don't think programming, game programming no less, is a simple task!

Originally Posted by Pyrofan1

I program game, in C++ i don't use microsoft visual C++ i use Dev-cpp, i think it's great, I've made this game It uses Allegro, it's a very good graphics library, it also has support for the mouse, keyboard, timing functions and MIDI support. So i would recommend it for game programming. I not exactly sure what you're asking, but start with the basics ie. pong. You don't even have to make the game use two human players the A.I. in pong is really simple
example: (C++)
if(ball_y>computer_y) computer_y++;
else
if(ball_y>computer_y) computer_y--;

p.s. sorry if this post seems to jump to different subjects.
[/b]

Dev-C++ is what I use. I thought you criticized my compiler

Anyway, for the absolute beginner to game programming, you can use something like game maker found at gamemaker.nl. It's an excellent introduction to sprite manipulation, variables, conditionals, etc. without any coding at all, as its all drag-and-drop. If you have a TI graphing calculator, you can make programs in the built-in BASIC language, which is very easy to use. You can make a good snake game in an hour with it. You also can program in the TI calculator's assembly language, though I wouldn't recommend that

That should give you experience and knowledge of what it takes to make a game; after that, learning the actual code for a language like C++ (which is actually incredibly easy) is the last step before you can make your own games.

Anyway those are just suggestions. I find making games more fun that playing them. Have fun!