Uploaded a new Konoko Payne version. Nothing really new here, just bug-fixes, “new” characters (Elite Striker, Ninja…), fixed motion scripts, an automatic camera for the demo-mode, etc. The previously mentioned “blood puddles” are in there, but it’s otherwise basically the same as last time, slightly improved. I just don’t have any time to work on a real “game” with a proper scenario, cutscenes, different levels, nice level design, etc. Or voice acting. Some guy once sent me an email complaining about the MGS2-like dialogs in KP, saying that “it would be better with recorded voices”. Yeah, right. What about this: I hire Amanda Winn-Lee and you pay the bill. Deal?
I’m considering dropping all the commercial libraries used out there (mainly Novodex & Pathengine) to make a proper open source project, but even this takes forever. I don’t know any good open source pathfinding library to replace Pathengine, so I’m kind of stuck here. I could re-implement it myself using John Ratcliff’s AI Wisdom 4 article, but I admit I’m not very motivated for this. It’s always painful to re-do something that already works, it doesn’t feel like making progress. On the other hand, the whole KP project doesn’t evolve very quickly anyway, and an open source version could boost this. Or ruin it completely.
I think I’m reaching a point where I can’t make that project a lot better just by myself. The code is pretty much “done”. Sure the AI could be improved, things could be more polished and more optimized, but the hand-to-hand combat, gun fights, character control, etc, everything’s working well and feels good. Despite this, I have an insanely long “TODO” list for Konoko Payne. Let me check… 93 Kbytes! It’s a vanilla ASCII text file with something like one task per line, and ~2000 lines. So that’s ~2000 things to do. If I do one thing per day, I’d need 2000/365 = 5.47 years to complete the project. This is so depressing. In comparison, I have another “DONE” file listing the things already implemented. I move the tasks from the TODO file to the the DONE file when they’re done. Today the DONE file is 62 KBytes and a little bit more than 1000 lines. Not bad for a hobby project.
Anyway. Better, later.
February 23rd, 2008 at 2:07 pm
>>to make a proper open source project
Sounds good!
It would be nice if you can adapt the ZLib license with all your open source code. It makes legal departments of companies who consider using your work happier. Also it makes it easier to combine your collision/rigid body developments with Bullet and some other projects.
Cheers,
Erwin
February 23rd, 2008 at 10:47 pm
>>I don’t know any good open source pathfinding library to replace Pathengine
Try PathFind, from http://www.cs.ualberta.ca/~games/pathfind/.
It’s not complete framework like Pathengine, but it has various search algo & search environment + utility classes (stats, etc)
February 24th, 2008 at 5:59 am
Erwin, I started using something like the ZLib license in my recent stuff. Check the License.txt in the previous convex-convex SAT experiment. I think it’s quite similar to ZLib.
February 28th, 2008 at 8:08 am
Hello,
Have been following your website for a while now. Glad to see things are still progressing well.
Would love to see a project like this as an open source collaboration. I am highly interested in character animation and control, as well as physics and path finding.
Anything interesting in particular on your current task list?
Always interested in learning more about how other choose to tackle solutions to various problems.