I think you should ...

make a series of tutorials on AI

from basic enemy movement to advanced topics like decision making based on npc observations, this would both be fun and enlightening...

32 votes
Vote
Sign in
Check!
(thinking…)
Reset
or sign in with
  • facebook
  • google
    Password icon
    I agree to the terms of service
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    Ahmet Mete ÇOM shared this idea  ·   ·  Admin →

    4 comments

    Sign in
    Check!
    (thinking…)
    Reset
    or sign in with
    • facebook
    • google
      Password icon
      I agree to the terms of service
      Signed in as (Sign out)
      Submitting...
      • Marcus commented  · 

        Ditto for learning a lot from RB, and I'm glad you replied, Ahmet Mete ÇOM.

        As for what you said (about how you weren't so much asking how to make specific AI, but more the foundations of AI programming), I hadn't thought of that. That's a good idea. It still seems like it would be kind of tricky, though, figuring out what would be a good "foundation" tutorial, and what would be getting too specific. There's that one suggestion going on right now to make a small game and put it up for download so people can explore it and learm how to put everything together, and I think that a tutorial on AI to accompany it could be good. Maybe something about how to get started making AI, and good tricks to use, or things to keep in mind.

        What do you mean by "scripts," though?

      • Ahmet Mete ÇOM commented  · 

        Fair enough points, I agree. However, any game beyond extremely basic ones, namely any game with an NPC in it that is actively doing something, will need an AI. I agree there are other priorities, but the day will come we will need to know more about AI.

        When I say AI I do not mean a chess game's AI or a real time strategy AI, but about how we can form a basis for structural thinking of any NPC that will have to move, shoot, run for cover etc. We will need help both with a efficient way of thinking how to structure an AI AND how to do it with our tools. Like, Is it better to do it with scripts? How do we integrate these scripts to our code? Is it better to do it with a solid list of variables to which the npc class keeps track of and check every x time units via some in-code functions?

        I am an amateur, in the end, and I believe I came to the point where I am ready to tackle some AI. And we didn't even talk about it until now. I understand if it's not a high priority, but I have learnt alot from RB and would like to learn AI from him too, when he has the time for it, so here ise the entry and I stand by it. :)

      • AdminRB Whitaker (Supreme Allied Commander, RB Whitaker's Wiki) commented  · 

        Marcus brings up lots of good points.

        Unlike, say, 3D graphics, it will be difficult to write AI tutorials that will be of general use. But I've thought about it a lot, and I think there might be a few good "guiding principles" or "common tricks" that might be worth discussing. As you mention, it would be hard to cover all possible scenarios effectively, but I think there's still room for something of value in there anyway.

        As for the second point, I tend to agree with Marcus. My gut tells me that there are plenty of other ideas that would be meaningful to more people. But part of the purpose of this site is to get a feel for what's actually meaningful to people. I'm learning really quickly from the feedback that I've received so far is that I don't necessarily have that great of an idea of what's truly going to be the most helpful. (Case in point: I'm not surprised to see the "Upload a small but complete game" item getting votes, but I'm surprised it is getting *so many*.) So I'm definitely happy to see AI tutorials on here for people to vote on. Without it, it might be hard to judge just how useful it will be to people.

        To Marcus's point, though, when/if you vote on this, I'd love for you to leave a comment with some ideas on what types of AI you'd like to see covered. When it comes time to making these tutorials, I'd love to have a clearer picture of what people would like addressed to maximize my time.

      • Marcus commented  · 

        I thought a lot about whether or not to vote for this, but came to the conclusion not to (I'll explain why). Also, do not take any of this as criticism, because that's not what it is.
        While it does seem like common sense to have tutorials for programming AI, I don't think it's a particularly good idea because
        A) AI, it's purpose, it's execution, and how it fits in with the rest of the game is so individual and unique, there could be tons of AI tutorials and still there would be people who didn't get anything out of it because they had something else in mind. Also, it would be really tricky deciding what kind of AI to make tutorials for first, and the depth of the AI - does it simply follow you around and shoot at you, does it follow you around and help you fight, does it try to avoid you, does it wander around randomly and attack anything it sees, and so on.
        B) AI is really complex, which is the reason you requested a tutorial for it. However, it would also take a ton of time to write a tutorial for, especially if it was a series of tutorials, which it would need to be. Considering point A, I would much rather that huge chunk of time spent on making AI tutorials go towards other projects, namely finishing the 100 Experience Points project (http://100experiencepoints.com/), but also doing some of the other suggestions on here that would help a larger amount of the community (http://rbwhitaker.uservoice.com/forums/216936-general/suggestions/4252266-provide-downloadable-complete-projects-for-each-tu).
        This is just my opinion, though, and would love to hear a counter-argument if someone disagrees with me. =]

      Feedback and Knowledge Base