heya frog? what's up?

Hey timm.

I'm 25, and in the second week of doing a Grad Dip in IT. I've done a BSc with honours, and pulled out of a Phd in pure maths. I've returned to my first main interest (I was doing assembly language programming on the Tandy color computer in grade 7).

It's yet another bot for Quake.

I really only play lan Quake, but when I used to play against bots I played the reaperbot. Unfortunately, it didn't quite feel like playing against real players. I wanted to make a bot that felt like multiplayer quake lan (as opposed to internet quake which is an entirely different experience). I thought I would just tinker with the reaper, but after reading some of the code I put it away and decided to start afresh.

First thing I did was the Frogbot physics mod (the code may still be at the botshop). This is a manual control bot. It is the only bot to my knowledge that basically emulates player physics exactly. I started this physics patch last November and basically finished it in January. At the time of was also testing out various waypoint approaches.

When I released the Frogbot physics code (the Frogbot was taking too long so I wanted to release something) I decided to start afresh on the Frogbot. I scrapped all the dynamic waypoint stuff I was doing because I couldn't get a desirable framerate with it and it was generally far more unworkable than using static waypoints (which is all that the Frogbot supports).

Well I can see at least a couple more months yet on the Frogbot, if I can squeeze in the time.

In terms of the ideas I have for it at this stage, I'd say about 25%.

Hey this is the time for me to make big ego boosting comments. :)

Currently the bots have the same physics model as a player, which would seem easy except id has deliberately or by accident obstructed that (it would take minor changes to quake.exe to allow this, but I have been forced to do the changes in QuakeC).

I'm pretty proud of the mouse movement of the bots (you can become a bot). It hopefully is getting very playerlike.

The bots are quite good at hogging stuff on the level and generally move around fluently (though the omicron is quite fluent - I've seen it navigate well on dm6).

Generally all that stuff you hear that they can now do with Quake2 because it uses C (which of course is better to use) I know can be done with little old QuakeC. (of course without configuring from .ini files, etc - ed)

Hopefully the Frogbot will appear to be a thoughtful (at least on higher skill levels) and realistic opponent.

Well free for all is generally the easiest for a bot (ie. a high player density).

I hope the bot will do all types of games well. Some of the code I plan to do has h2h (1 on 1) games in mind.

Though eventually I want the bot to play team games, which are my favourite type of game (teamplay 2 of course, then you can abuse your teammates for shooting you :).

Well I could release it pretty soon. If people would understand that it is not a complete product and that I am continually working on it, then I would risk releasing it in a premature state. It's hard to say how much initial impressions matter in these type of things.

That's hard to say. I could release the Frogbot and people may think it sucks. I guess you could poll people if you were really concerned about such things. :)

I see better as meaning being more like a player - with skill 0 being like a beginner, and skill 100 being like an expert.

So hopefully the bot movement, aiming and strategy will be more like a player.

I'm not just sitting down typing code out. There is a lot of thought that has to go into decision making algorthms. The big headache is when you know the best theoretic algorithm for something, but it is computationally too slow. So you have to bang your head against a wall and try to find a cpu friendly approach that approximates the best algorithm.

Well lets see. I must get at least 8 hours sleep a night. I lose hours eating and stuffing around. But on a lot of days the main thing I'm doing is working on the Frogbot. Though now I'm back at uni so that may eat into my "Frogbot time".

Well back when I started doing the Frogbot I knew that I could do it. Once I start something I feel is important I get strongly motivated. It's like when you get that feeling that you can do anything with sheer will-power.

I don't get tired of important stuff or stupid questions because I am bullet-proof. :)

If I start to get frustrated with some code I try to simplify what I was trying to do, but keep an eye I where I want that "something" to go (so as not to turn what I was doing into a completely unworkable hack).

I don't plan to scramble the progs.dat so basically anyone will be able to get their hands on the code.

I won't release the source code originally because I don't want people to see my poor commenting - good commenting is what they are supposed to be teaching me this year at uni. :) (w0w, now that should be good news! -ed)

I'll have to see when the time comes. The future is a hard thing to predict. Quake is definitely my favourite FPS. Hey I like the weapon imbalance!

Yeah I watched you massacring the Frogbot in that demo you sent me. Don't worry, the Frogbot won't forget. :)

Cool, no worries.