Sign in to follow this  

How does the programmer merge systems together in the industry

This topic is 1641 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

To be more specific, I was looking at the credits on the game manual and it listed the programmers behind the menu system, battle system, field system, etc a bunch system.

 

While I never been a part of a programming team, I have developed a small game that involved features on a much smaller scale. But I cannot imagine how does the team merge the systems together.

 

Does it always work the first time? Is the main programmer the person that does the merging?

 

I would think it would require a lot of communication and planning before even writing the code. How do they even keep track of this massive project and the flow of the program? 

 

When they start the project, is it usually a prototype just to see if it works before merging more systems?

 

Seems pretty daunting to me. I need perspective and insight so if I ever face a situation in the industry, I would have something. To bear in mind about. The idea is exciting yet scary.

Edited by warnexus

Share this post


Link to post
Share on other sites

You usually try to keep the merging to a minimum.

You define an API that borders the code you are working on now with others.

This API-border is very well defined.

 

The API is a set of functions like: Enemy.move(x,y).

 

One programer will create the function "move" which moves the enemy. The other programer will create the AI function which tells the enemy when to move.

 

The only thing they have to decide in advance is that enemy has a move function that takes x & y. This is the API.

 

The AI programer does not care about how the enemy moves (collison detection etc...).

The "move" programer does not care about why the enemy moves ( the AI ).

And the only thing they decided in advance, is that move takes two parameters: x & y

 

You can define such API borders in advance any time you want to split the job.

 

So basically there is a well defined border, and usually a lot of discussion goes in to that. The better you are at the design stage, the less you work you will have merging it later. 

Edited by SillyCow

Share this post


Link to post
Share on other sites

Does it always work the first time?

More like does it ever work first time. In general there's always a bit of back and forth with it. The more planning that went in the more it pays off and good communication between the developers themselves is crucial, in fact you can gauge how communicative they were just by looking at the amount/ugliness of the glue-code between systems.
 

Is the main programmer the person that does the merging?

Sometimes, but not always. As SillyCow mentioned the important thing is to have an API boundary that both sides can agree on and work with. I find these APIs work best when created by one person who has taken the time to understand what the other guy needs.
 

I would think it would require a lot of communication and planning before even writing the code.

It can and probably should. A lot of the time though these things arrive organically.
 

How do they even keep track of this massive project and the flow of the program?

By having well-compartmentalised systems with clearly defined API boundaries. The overall flow is oft defined by a sort of backbone, like the main game loop, or a messaging system with a reflex-arc triggered by something tangible (like a user clicking a button).

This organisation doesn't always arrive though - I've inherited projects where the communication between systems is blurry with complex, hard to tease apart interactions and dependencies. Keeping track of things then becomes something that only the original developers can do, so access to pick their brain is pretty darn useful.
 

When they start the project, is it usually a prototype just to see if it works before merging more systems?

I like prototyping, I think it's seriously under-valued as an approach. A lot of the time a prototype isn't made or, if it is, then it'll just be a single "prototype" and regardless of it's successes/failures proceeds to become the full-fledged product anyway.

Edited by dmatter

Share this post


Link to post
Share on other sites

As SillyCow mentioned the important thing is to have an API boundary that both sides can agree on and work with. I find these APIs work best when created by one person who has taken the time to understand what the other guy's needs.
 

By having well-compartmentalised systems with clearly defined API boundaries. The overall flow is oft defined by a sort of backbone, like the main game loop, or a messaging system with a reflex-arc triggered by something tangible (like a user clicking a button).


These are the key bits really; there is no great 'merge' just systems which happen to work together.

For example I work in the core rendering group which means it's my job to define systems which interface with the 3D API to get stuff on the screen but what we produce is dictated by the game rendering group who want to instruct the hardware in a certain way so they provide the requirements and we produce the solution; they don't care how it is done under the hood just that it is efficient to work with.

At the same time we need to get resources into the renderer's memory this generally means grabbing data from disk in some manner at which point we interface with and work with another group who are responsible for loading the data and telling the correct system that the data has arrived. This is, again, a back and forth - we don't care how the data gets to us we just know that our resource factories have to have a certain interface and work in a certain way.

In both cases the API is what defines the interaction - both sides of the API have input but the implementation comes down to groups and there is no 'merge' - a group lead might well sign off on the design of something but it goes directly into the code base everyone else is working with (well, via a few branches of source control but the point is that it is part of a whole).

Share this post


Link to post
Share on other sites


how communicative they were just by looking at the amount/ugliness of the glue-code between systems.

 

can you post a pseudo code of ugliness of the glue-code? I cannot picture it.

Share this post


Link to post
Share on other sites


I like prototyping, I think it's seriously under-valued as an approach. A lot of the time a prototype isn't made or, if it is, then it'll just be a single "prototype" and regardless of it's successes/failures proceeds to become the full-fledged product anyway.

 

i like prototyping and play-testing too. Check to see if the game is fun and I found 4 game-breaking bugs through the course of development.

Share this post


Link to post
Share on other sites

 

 


As SillyCow mentioned the important thing is to have an API boundary that both sides can agree on and work with. I find these APIs work best when created by one person who has taken the time to understand what the other guy's needs.
 

 

By having well-compartmentalised systems with clearly defined API boundaries. The overall flow is oft defined by a sort of backbone, like the main game loop, or a messaging system with a reflex-arc triggered by something tangible (like a user clicking a button).


These are the key bits really; there is no great 'merge' just systems which happen to work together.

For example I work in the core rendering group which means it's my job to define systems which interface with the 3D API to get stuff on the screen but what we produce is dictated by the game rendering group who want to instruct the hardware in a certain way so they provide the requirements and we produce the solution; they don't care how it is done under the hood just that it is efficient to work with.

At the same time we need to get resources into the renderer's memory this generally means grabbing data from disk in some manner at which point we interface with and work with another group who are responsible for loading the data and telling the correct system that the data has arrived. This is, again, a back and forth - we don't care how the data gets to us we just know that our resource factories have to have a certain interface and work in a certain way.

In both cases the API is what defines the interaction - both sides of the API have input but the implementation comes down to groups and there is no 'merge' - a group lead might well sign off on the design of something but it goes directly into the code base everyone else is working with (well, via a few branches of source control but the point is that it is part of a whole).

you mean like in your case, you can just feed code in and it doe


o basically there is a well defined border, and usually a lot of discussion goes in to that. The better you are at the design stage, the less you work you will have merging it later.

 

interesting thanks!

Share this post


Link to post
Share on other sites

This topic is 1641 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this