In the early days of video games (from the early 1970s to mid-1980s), a game programmer also took on the job of a designer and artist. This was generally because the abilities of early computers were so limited that having specialized personnel for each function was unnecessary. Game concepts were generally light and games were only meant to be played for a few minutes at a time, but more importantly, art content and variations in gameplay were constrained by computers' limited power.
Later, as specialized arcade hardware and home systems became more powerful, game developers could develop deeper storylines and could include such features as high-resolution and full color graphics, physics, advanced artificial intelligence and digital sound. Technology has advanced to such a great degree that contemporary games usually boast 3D graphics and full motion video using assets developed by professional graphic artists.
The desire for adding more depth and assets to games necessitated a division of labor. Initially, art production was relegated to fulltime artists. Next game programming became a separate discipline from game design. Now, only some games, such as the puzzle game Bejeweled, are simple enough to require just one fulltime programmer. Despite this division, however, most game developers (artists, programmers and even producers) have some say in the final design of contemporary games.
Since processing cycles are always at a premium, physics programmers may employ "shortcuts" that are computationally inexpensive, but look and act "good enough" for the game in question. Sometimes, a specific subset of situations is specified and the physical outcome of such situations are stored in a record of some sort and are never computed at runtime at all.
Some physics programmers may even delve into the difficult tasks of inverse kinematics and other motions attributed to game characters, but increasingly these motions are assigned via motion capture libraries so as not to overload the CPU with complex calculations.
See also: dynamical simulation
Some games, such as strategy games like Civilization III or role-playing games such as The Elder Scrolls IV: Oblivion, use AI heavily, while others, such as puzzle games, use it sparingly or not at all. Many game developers have created entire languages that can be used to program their own AI for games via scripts. These languages are typically less technical than the language used to implement the game, and will often be used by the game or level designers to implement the world of the game. Many studios also make their games' scripting available to players, and it is often used extensively by third party mod developers.
Game AI programming should not be confused with academic AI programming and research: game programming has little use for developments in this area of study. Although both areas do borrow from each other from time to time, they are usually considered distinct disciplines.
Programmers specializing in this area of game development can demand high wages and are usually a scarce commodity. Their skills can be used for computer games as well as games for game consoles such as the PlayStation 3, Wii and the Xbox 360.
Many games use advanced techniques such as 3D positional sound, making audio programming a non-trivial matter. With these games, one or two programmers may dedicate all their time to building and refining the game's sound engine, and sound programmers may be trained or have a formal background in digital signal processing.
Scripting tools are often created and/or maintained by sound programmers for use by sound designers. These tools allow designers to associate sounds with characters, actions, objects and events while also assigning music or atmospheric sounds for game environments (levels or areas) and setting environmental variables such as reverberation.
This programmer may implement strategy tables, tweak input code, or adjust other factors that alter the game. Many of these aspects may be altered by programmers who specialize in these areas, however (for example, strategy tables may be implemented by AI programmers).
More often today the core game engine is usually separated from gameplay programming. This has several development advantages. The game engine deals with graphics rendering, sound, physics and so on while a scripting language deals with things like cinematic events, enemy behavior and game objectives. Large game projects can have a team of scripters to implement these sorts of game content.
Scripters usually are also game designers. It is often easier to find a qualified game designer who can be taught a script language as opposed to finding a qualified game designer who has mastered C++.
Input programming, while usually not a job title, or even a full-time position on a particular game project, is still an important task. This programmer writes the code specifying how input devices such as a keyboard, mouse or joystick affect the game. These routines are typically developed early in production and are continually tweaked during development. Normally, one programmer does not need to dedicate his entire time to developing these systems. A first person shooter such as Quake may need a very complex and low latency input system, while the needs of a turn-based strategy game such as Heroes of Might and Magic are much lower.
It is the tools programmer's job to write the tools that handle these game-specific tasks. Some tools will be included with the game, but most will not. Most tools evolve with the game and can easily consume all of several programmers' time. Well written and fairly bug-free tools make everyone's development tasks easier. Poorly written or poorly documented ones can seriously hamper development and jeopardize the project. Due to time constraints, however, many tools are not carefully implemented.
This programmer must be familiar with both the original and target operating systems and languages (for example, converting a game originally written in C++ to Java), convert assets, such as artwork and sounds or rewrite code for low memory phones. This programmer may also have to side-step buggy language implementations, some with little documentation, refactor code, oversee multiple branches of code, rewrite code to scale for wide variety of screen sizes and implement special operator guidelines. They may also have to fix bugs that were not discovered in the original release of a game.
Increasingly, universities are starting to offer courses and degrees in game programming. Any such degrees should offer a well-rounded education that includes all the fundamentals of software engineering.
Generally, lead programmers are the most well compensated, though some 3D graphics programmers may challenge or surpass their salaries. According to the same survey above, lead programmers on average earn $103,409 annually.
A controversial subject in the game industry recently has been the issue of unpaid overtime, particularly as highlighted in a blog entry by EA Spouse, who described some of the darker aspects of working in the industry. At some companies (the blog entry targeted Electronic Arts specifically), "crunch time" (a period where programmers will often work around the clock to finish a project) has become the expected norm through the whole development cycle. This is not true of all companies however, and many game studios provide excellent working environments and conditions.
Third-party developers, the most common type of video game developers, depend upon a steady influx of funds from the video game publisher. If a milestone or deadline is not met (or for a host of other reasons, like the game is cancelled), funds may become short and the developer may be forced to retrench employees or declare bankruptcy and go out of business. Game programmers who work for large publishers are somewhat insulated from these circumstances, but even the large game publishers can go out of business (as when Hasbro Interactive was sold to Infogrames and several projects were cancelled; or when The 3DO Company went bankrupt in 2003 and ceased all operations). Some game programmers' resumes consist of short stints lasting no more than a year as they're forced to leap from one doomed studio to another. This is why some prefer to consult and are therefore somewhat shielded from the effects of the fates of individual studios.
Most commercial computer and video games are written primarily in C++, C, and some assembly language. Many games, especially those with complex interactive gameplay mechanics, tax hardware to its limit. As such, highly optimized code is required for these games to run at an acceptable frame rate. Because of this, compiled code is typically used for performance-critical components, such as visual rendering and physics calculations. Almost all PC games also use either the DirectX or OpenGL APIs (or SDL) to interface with hardware devices. Xbox games use an API similar to DirectX.
Various script languages are also used for the generation of content such as artwork and especially AI. Scripts are generally parsed at load time (when the game or level is loaded into main memory) and then executed at runtime (via logic branches or other such mechanisms). They are generally not executed by an interpreter, which would result in much slower execution. Scripts tend to be used selectively, often for content design and high-level game logic. Some games are designed with high depedency on scripts and some scripts are compiled to binary format before game execution. In the optimization phase of development, some script functions will often be rewritten in a compiled language.
Java is used for many web browser based games because it is cross-platform, does not usually require installation by the user, and does not pose security risks, as may a downloaded executable program. Java is also the dominant language for mobile phone based games. Adobe Flash is also a popular development tool for browser-based games using the ActionScript language (there is some debate as to whether Flash actually constitutes a "language," but it is a popular web-based game development medium).
As games have grown in size and complexity, middleware is becoming increasingly popular with the industry. Middleware provides greater and higher level functionality and larger feature sets than the standard lower level APIs such as DirectX and OpenGL, such as skeletal animation. In addition to providing more complex technologies, some middleware also makes reasonable attempts to be platform independent, making common conversions from, for example, PC to PS2 much easier. Essentially, middleware is aimed at cutting out as much of the redundancy in the development cycle as possible (for example, writing new animation systems for each game a studio produces), allowing programmers to focus on new content.
Other tools are also essential to game developers: 2D and 3D packages (for example Photoshop, Maya or 3D Studio Max) enable programmers to view and modify assets generated by artists or other production personnel. Source control systems keep source code safe, secure and optimize merging. IDEs with debuggers (such as Visual Studio) make writing code and tracking down bugs a less painful experience.