What I mean is, define your wish list here regarding development tools, APIs, etc. Do not take an existing product and say: "Well, I wish such and such product had this feature..." I would like to see a little more thinking outside of the box. Try to visualize languages, content creation programs, interactive scripting tools, new methodologies, KB builders, or whatever would enable you to not only to be productive, but would allow you to easily define and create the product you would like to create. Be vague or be explicit, assume it must use today''s computing power or instead, tomorrow''s computing power.

Look, if the existing tools were adequate, or even better than adequate, there would be more productivity and less of a failure rate with regard to game design and game development. Individuals with less resources than larger companies would be able to compete.

Graphics engines are one move in this direction, but in many ways they seem more constraining than expressive. OpenGL is a wonderful API, but it is very low-level. And among all these tools, there is no unifying glue except that of some language which is usually C or C++.

There is room out there for new paradigms of development.

For example, instead of desiging architecture in 3d studio, imagine a tool could be given a ''style'' of architecture and a set of constraints defining the structure''s footprint and functionality, and the program could create architecture for you.

Instead of using OpenGL, a 3d engine, or customized wrappers for OpenGl which you have written, imagine a new expressive language which compiles into a C program comprised of OpenGL calls.

Instead of programming your AI in a language like C++, imagine a development environment which lets you declare the knowledge of your agents, maintains logical consistency, and contains a mechanism to do goal satisfaction given knowledge about the environment.

How about a tool which enables you to storyboard ideas, but is unified with other tools to tap into the storyboard via some unified framework. In other words, the storyboard program is sort of a director''s link to the elements of the game.

Does anybody have any ideas?

The problem with such high level tools is, without customization (which would need to be low level), they only can produce limited results. This stunts creativity and innovation (always a bad thing), so we don''t use tools like this.

quote: Original post by Impossible

The problem with such high level tools is, without customization (which would need to be low level), they only can produce limited results. This stunts creativity and innovation (always a bad thing), so we don''t use tools like this.

I don''t think that''s an objective observation. If your statement were true, we wouldn''t use C compilers, function definitions, 3d modeling programs, 3d engines, paint programs, terrain generators, physics modelers, etc.

As a counter-example to your implication, I am currently attempting to develop a programmable agent interface that would increase expressivness, rather than reduce it. If you accept that more expressivness means reducing the limitations of current tools, than you must accept that some higher level tools can actually increase the flexibility and descrease the limitations.

quote: Original post by bishop_pass
Individuals with less resources than larger companies would be able to compete.

No, the big companies would just be able to afford a $70,000 license for Microsoft Galactacor, the talking game programming robot/MMORPG mainframe, and the little guys would still use the same things they always have.

quote:
Instead of using OpenGL, a 3d engine, or customized wrappers for OpenGl which you have written, imagine a new expressive language which compiles into a C program comprised of OpenGL calls.

Great idea, let''s name it "COCOA". It''ll be the fastest thing out there! Better yet, how about a language for the language itself? Sort of like a... Beginner''s all-purpose symbolic instruction code! But what will we call it?

I like writing my own code. Really. It''s faster, I have more control over what goes on, and I learn a lot more from it. For that reason, I wouldn''t be interested in most of the things you''ve suggested.

However, I *DO* like the idea about the AI development environment. It would be very interesting to have tools centered around AI development - it''s an underused feature in many games, it makes a world of difference, and (pathfinding aside) it usually consumes very little processing power, so designing it through third party tools rather than hand-crafting it wouldn''t cause much of a problem.
I think that''s a very interesting idea. I also think this is a very interesting topic, although it''s hard to come up with ideas for things that would enhance game development without serious negative side effects. The majority of potential tools for game developers would either result in extremely inefficient code, or make games into clones of each other.

As a side note, I believe DirectMusic currently has the ability to dynamically change the "mood" of your music according to certain parameters. That''s pretty cool.

I don''t think that higher level tools are necessarily bad, just some of the tools you suggested. The higher level language that converts to C++ code could work for some projects (you could even think of it as a scripting language.) The same goes with the AI tool. The architecture generator, however, is an example of something that would not work. Although a specific game, or even set of games may use a level generator, I don''t believe it would be a good industry wide tool.

quote: Original post by Carnivorous Duck

No, the big companies would just be able to afford a $70,000 license for Microsoft Galactacor, the talking game programming robot/MMORPG mainframe, and the little guys would still use the same things they always have.

Not necessarily. What would likely happen is game development would become easier, more games would be developed, they would become more of an individual statement from its creator, less money would be made from game development, and more money would be made developing tools to create games.

quote:

Great idea, let''s name it "COCOA". It''ll be the fastest thing out there! Better yet, how about a language for the language itself? Sort of like a... Beginner''s all-purpose symbolic instruction code! But what will we call it?

Your version no doubt. If all you want to do is change the syntax, then their is no purpose to creating a new language. This was just an example to get people thinking. To do it right, a new paradigm would have to be developed. New paradigms can change things dramatically.

quote:

However, I *DO* like the idea about the AI development environment. It would be very interesting to have tools centered around AI development - it''s an underused feature in many games, it makes a world of difference, and (pathfinding aside) it usually consumes very little processing power, so designing it through third party tools rather than hand-crafting it wouldn''t cause much of a problem.
I think that''s a very interesting idea. I also think this is a very interesting topic, although it''s hard to come up with ideas for things that would enhance game development without serious negative side effects. The majority of potential tools for game developers would either result in extremely inefficient code, or make games into clones of each other.

Read this post where I discuss an algorithm for such a tool. As for consuming very little processing power, that is wrong. Good AI will ultimately require a great deal of a CPU''s resources. If not, then you aren''t maximising the potential of AI. But processors are getting faster, the graphics pipeline is being offloaded to the video card, and the time is ripe to take AI to the next level.

As for games being clones of each other, this is an artifact of using tools which are not expressive or expandable. It is not because tools are used. In fact, diversity arises out of productively using the appropriate tools. However, an example of where this fails is harnessing an engine like those used for Quake or Unreal Tournament. These tools aren''t really expressive, they are more constraining.

Well, if you''re taking orders, I want:

A gui-interface, drag-n-drop tool for building a 3-d world that you can walk through in real-time, view in 360 degrees, pick up objects from, etc. You can import several layers of flat graphics to it and it will warp them into round or square room perspectives. You can look down on your whole network of rooms and passages, or stand within a room and see each passage and object marked.

quote: Original post by sunandshadow

Well, if you''re taking orders, I want:

A gui-interface, drag-n-drop tool for building a 3-d world that you can walk through in real-time, view in 360 degrees, pick up objects from, etc. You can import several layers of flat graphics to it and it will warp them into round or square room perspectives. You can look down on your whole network of rooms and passages, or stand within a room and see each passage and object marked.

This is good. I would liken it to playing within the game-world as you are building the game-world. In other words, the 3d game-engine should actually be running as you walk through your world in development. You walk across the floor and say: "put a wall here, 10 inches thick, 20 feet long, 10 feet high." You then rise up like a bird and view your master plan from above. As you create your world, the physics models actually work. You wave your magic wand, get a vase and set it down on the ground next to the wall. You then whip out your magic 3d bump map brush and paint a rune on the wall just created. Later, you add another wall, and the necessary intersection routines carve up the geometry in the most efficient manner possible. Add to the modeler contextual knowledge of basic things. Where a typical modeler only knows the geometry, a more sophisticated modeler should know what a wall is, and how it comprises one barrier of a room.

Yes, exactly. Also, we should be able to draw motion paths for NPCs and animated backgrounds right onto it.