Motivation

We all know the “perfect game.” We play a game and think “Wow, that’s cool, *but*…” If you are like me, then thousand of ideas pop up in your head – how you would do this better or that differently. Once the brain starts spinning, it produces step by step a blueprint of “the perfect game” – the ultimate experience (at least how far you’re able to envision it).

Building a real game take of course lots of work. Don’t ask me how many person-hours (which I don’t have) are involved, let alone the specialized skills (which I don’t have) required for each sub-task. So I won’t even try to do it. Instead I just pick some semi-random topics which look sexy to me and explore them. Write down some thoughts, perhaps even create some code around it. But I consider this whole thing as a “research project,” there is no product at the end of the tunnel, it is an open ended journey and it will take me where ever it leads me.

My main interest is on the technical side, I can hardly claim to be a skilled game designer or artist. So the questions that I might ask myself look more like “How would I design [insert subsystem here]?” or “How should [insert problem here] be solved?” Although I probably won’t have the time to dive deep into coding.

Possible areas of interest

…that jump to my mind right now.

Each one of those might turn out to be already well covered, completely over my head or technically impossible in the foreseeable future, and there might be nothing I could add to it. Also my knowledge of the current state-of-the-art is very limited (I just recently started playing current games again), so some stuff might turn out as “but we’re already doing that.” I simply haven’t done any research yet. Being a braindump for that kind of research is the whole purpose of this site.

Dynamic dialog system

Use speech synthesis instead of prerecorded text. Increases the possibilities of the character’s AI at the expense of less expressive voices. This would involve programmatic generation of text, having some model of the language and modeling the state/emotions of a character so it can influence the way they express themselves.

Procedural story telling

Script based story systems have various limitations, especially in an open game universe. A player may complete tasks in unexpected order, get stuck on sub-tasks or choose unplanned actions. Or stuff like “Rescue the princess now, that’s the only hope for the empire!” And then you walk away for a week of looting dungeons and leveling up, just to find the princess still waiting for you and you rescue her dramatically in the last second before she’s sacrificed on the altar. A “super AI” which governs the game, lays out the story plot and constantly adjusts it to the current state of the universe and the players actions seems to the next step in the evolution.

Highly detailed game universe

The capitol of a big country that consists of a palace, cathedral and two hands full of shops and houses? All inhabited but a two digit number of people whose primary occupation seems to be walking on the street, sitting in a tavern or standing at the city gate 24/7, waiting for you to talk to them? Does not feel very realistic. What would it take to create (and run) a universe of the size of a real world continent (or bigger), with hundreds of cities, towns and villages? And hundreds of thousands people roaming that world, each one having a real live? That is something that cannot be handcrafted, but has to be generated algorithmically.

Building a motion capture system

I guess (that’s a pure guess) that motion capture systems are highly expensive, proprietary systems which are provided by specialized companies. Trying to build such a system with consumer grade hardware and open software would be an interesting challenge.