next »

[Danmark]

Game implementation is in such a sorry state right now. This kind of thing has been my wet dream for a few years. In no particular order:

Generalizability:
E.g. a "real" type that may be easily switched between fixed-point on the decimal level, fixed-point on the binary level, and floating point implementations. Basically the capability of specifying the high-level structure of everything, as opposed to always specifying the low-level structure upfront, which can waste tons of time.

Determinism:
The runtime always behaves the same on a given platform, allowing piss-simple implementation of (e.g.) replays and lockstep multiplayer.

Efficiency:
Every program made with this hypothetical tool is cache-friendly and fairly easy to parralelize. This allows developers to make the best use of processing power available to them, thereby allowing the maximal expressivity on any hardware.

Visual programming:
Programming by entering arbitrary text strings makes no sense whatsoever; dispense with it. In a given execution context there is only ever a relatively small set of valid statements, so visual programming allows the correctness of a program to be enforced more easily and earlier on. In theory, it also enables faster programming, because the user doesn't have to hammer out the entirety of their statements (no, code completion doesn't cut it) or keep referring to their libraries' documentation; they can just select them from some kind of palette. However shite every VPL I've tried is, the paradigm is fundamentally superior to text-based programming. On a related note,

Self-documentation at every stage:
Support for the arbitrary visualization any type, so one can watch the program state change as they debug, in more depth than current debuggers allow. Also, the elimination of such ancient notions of software implementation as whitespace-free variable names.

Real-time compilation:
Programs built in the hypothetical tool assemble themselves into testably efficient executables as you change their code. The computer is generally way faster than the programmer, so compile downtime is a waste.

Huge standard library:
One that contains a majority of the tools frequently used to implement games, from the abstract to the concrete.

Community-driven reusability:
A massive, easy to browse library of snippets and tools created and uploaded by the hypothetical tool's users.

A tool for making tools:
Unity gets halfway there then trips up. Encourage people to make bitmap and vector graphics and 3D model editors, animation editors, scene editors, AI designers, all sorts of shit, and fully equip them to be able to create arbitrary GUI-driven editors.

[agj]

Assuming you mean a game making software, I'd like one whose interface is highly scriptable. That would be beautiful.

[John Nesky]

Baiame: That's a really good list. I'm putting it in my pie-in-the-sky to-do list.

[Danmark]

Thanks John. It's in mine as well, the problem is that actually kicking off such a project would undoubtedly leave one with zero time to actually keep making games, until said project was in a fairly advanced stage. And that could be anytime or no time at all...

[BoxedLunch]

a game design program where all i have to do is the graphics and level design, the music and programming would make themselves. 

[JackieJay]

the title of this topic is misleading. I thought it was about game design programs and not about game making software. You dicks.

[TobiasW]

a game design program where all i have to do is the graphics and level design, the music and programming would make themselves. 

I heard there's something called "collaboration" (also known as "team") that can accomplish that!

[AndrewFM]

game maker, but with a better room editor and the ability to play the game while you edit it instead of forcing you to re-compile it every time you change something in order to see the changes. real-time game editing, in other words.

If you do use Game Maker, this technically can be done to some extent, with some interesting usage of the available functions. I made this video a while ago that shows how you can pull this off:





Though, I figured out a much better way to do this. Like, don't even use the method in that video; it's horrible.

You can use GM's object_add(), and object_event_add() functions. Add the code from a file into a new object, and update it (making use of object_event_clear()) each time a certain button is pressed, instead of every frame. This way, you completely avoid the massive lag from reading a long file 30+ times per second . It reads and compiles the code only when the button is pressed, and runs it internally at all other times.

[hexageek]

well no one said we should be realistic so here it is:

*one that could produce games that would work on all platforms: consoles, PCs and handsets
*one that works on all OS.
*one that is open source
*one that doesn't limit the developer
*one that can use python, lua, javascript and other popular languages for scripting.

[Ntero]

well no one said we should be realistic so here it is:

*one that could produce games that would work on all platforms: consoles, PCs and handsets
*one that works on all OS.
*one that is open source
*one that doesn't limit the developer
*one that can use python, lua, javascript and other popular languages for scripting.

Interestingly enough, if you have enough cash to pay for an open source version, that sounds like Unity's next milestone release.

[hexageek]

except that it doesn't produce games that work on all platforms, that it doesn't work on all OS and that it's not open source

[Ntero]

That's why I said their next milestone.

Though it's not all Platforms it's planning to contain: Windows, Mac, iPhone, Wii, Xbox 360, PS3, as well as browser based, which can extend it's support much further.  I also believe there was some discussion about Linux as well.

It works on Both Windows and Mac for coding.  Sorry Linux, maybe soon.

It has an available Open Source License.

So really the only things I see it missing for this generation is Linux(building anddeployment), Android and Windows 7 Mobile (deployment)(the 2 mobile platforms being quite new platforms).

[hexageek]

It has an available Open Source License.

what do you mean by this? unity is not currently open source. do you mean it will be in the next milestone? if so, could you provide any links?

[gunswordfist]

Okay, I'll tell you my secret. I wish I had a program that took ideas from my head and implemented them perfectly into a game world. But that'll never happen so I'll have to do they hard part by spending frustrating times trying to actually get those ideas to work myself. 

[gimymblert]

Okay, I'll tell you my secret. I wish I had a program that took ideas from my head and implemented them perfectly into a game world. But that'll never happen so I'll have to do they hard part by spending frustrating times trying to actually get those ideas to work myself. 

It can read your mind, but it can ask you crucial question to create a model of what you want (something along eliza and dramatica).

[gunswordfist]

Okay, I'll tell you my secret. I wish I had a program that took ideas from my head and implemented them perfectly into a game world. But that'll never happen so I'll have to do they hard part by spending frustrating times trying to actually get those ideas to work myself. 

It can read your mind, but it can ask you crucial question to create a model of what you want (something along eliza and dramatica).

Elaborate please.

[nihilocrat]

*one that could produce games that would work on: Wii, NDS, and any other system I can get my hands on with homebrew
*one that works on Windows, Mac, and Linux
*one that is open source
*one that doesn't limit the developer
*one that can use lua for scripting.

This is what I'm hoping to make right now. However, I think the best language for the task is C++, and I don't get along too well with it. I'm hoping that if I can't make a full blown engine+tool I can make a library for game logic that lets the developer decide what to do about graphics, input, etc. There are tons of libraries out there that do the latter but none that I know of that just handle logic/object management/simple collisions/events and nothing else.

[gimymblert]

Okay, I'll tell you my secret. I wish I had a program that took ideas from my head and implemented them perfectly into a game world. But that'll never happen so I'll have to do they hard part by spending frustrating times trying to actually get those ideas to work myself. 

It can read your mind, but it can ask you crucial question to create a model of what you want (something along eliza and dramatica).

Elaborate please.

I'm elaborating this pretty concretly right now as a side project 
As a part of a reasearch on procedural story and gameplay. IT may not produce top par gameplay or story but it can provide a methodologie for some aspects for some kind of game i think.

[nihilocrat]

A program that basically makes a designer talk through their process is actually pretty clever, at least to get them to really think about the game.

[jisou-kishi]

ill go with the three Fs
Fast Free and Friendly