Thread: Future plans
View Single Post
Re: Future plans
Old
  (#7)
CrapShoot
A Monkey
 
CrapShoot's Avatar
 
Status: Offline
Posts: 386
Join Date: Jan 2005
Default Re: Future plans - 15-02-2009

Having spent considerable time working for both omnibot and fritzbot, I'll add my two cents. Firstly, comparing the two is a bit unfair since Fritzbot was made for ET, while omnibot is a bot framework that can be used for any game that has a 'friendly' SDK.. I understand that to an ET end user this is irrelevant, but it may have some relevance to people interested in expanding on the feature set of the bot specific to one game.

One of the big differences between the two bots for ET I think is the quality of waypoints and map scripts. The Fritzbot waypointing team (TomTom, the BindleStiff, 420Blunt, and myself) did a great job of packaging up completed waypoints that were passed around for testing, critique, etc before TomTom or the BindleStiff packaged them up for a 'release'. This guaranteed a certain level of quality where as the omnibot approach of releasing beta quality and below waypoints significantly impacted many first time users impressions of the bot.

Omnibot isn't open source, but DrEvil was very open to granting me access to the code for adding rtcw support. It sounds to me like some of you are at the point where I was with rtcw fritzbot; I had nothing against rtcw fritzbots, but wanted to see some more features in an rtcw bot. It may be something to look at if any of you are seriously considering building a bot from the ground up; especially at this stage of ET's lifecycle. You may even be a bit surprised at the current scripting capabilities in which you can basically define bot behaviors, create custom goals, etc.

For clarity about a few things, i'll respond to some of TomTom's comments.

Quote:
Originally Posted by TomTom View Post
Well the question to my mind is would the bots still have the best qualities of FritzBots. To my mind their better qualities are;
(1) They are less taxing on the CPU than Bobots, Omnibots or EastFront NPCs. You can run a good number of them on a old PIII.
Nothing to debate here. Maleficus was extremely careful with resources. Although some limitations were a bit restricting (32 max bots, 4 max connections per node, etc), at the end of the day it's a great functioning bot with minimal hardware requirements.

Quote:
Originally Posted by TomTom View Post
(2) The Cvop behaviour is decent enough (though Mal's cvops work in ETQW is much better)
This is an example of something that could be completely done in script with omnibot.

Quote:
Originally Posted by TomTom View Post
(3) The map aiscript botsight setting allows for a more realistic feel than bot mods that use a global or fixed value. Bot skills are not overpowering (nobody likes bots that pwn you).
This isn't unique to Fritzbot. Omnibot also has the ability to change sight distance in a map script.

Quote:
Originally Posted by TomTom View Post
(4) The waypointing editor can be run while play testing, and basic waypointing is easy enough to do after an evening of learning curve.
Can be done with omnibots as well.

Quote:
Originally Posted by TomTom View Post
Really I have had no plans on doing modding. Mapping or modelling I have thought about doing but not starting a mod.
So I doubt I could put much time to such a project, but getting the WET source to compile in Gcc and MS VC++ and merging in Chrukers "project fix" might not be too much time off hand (though I'll need a newer linux box ).
Pretty easy to do really. The sdk has projects for windows and build scripts for linux (and i think mac). It would just be a matter of applying the patch. If this is really something that you guys are going to pursue, I think I still have the code snippets I sent Mal for adding ETPro scripting support.

Quote:
Originally Posted by TomTom View Post
The open source concept in an active bot project would be novel enough (BTW DrEvil of Omnibot has posted the source of his foxbot project).
foxbot is a team fortress classic bot and i believe that is really old code.

To be clear, the point of my post isn't to say one bot is better or worse than the other. I've invested time in both and it would be silly for me to say anything bad about either . I just think that if you are seriously considering creating a brand new bot for ET, you may want to consider talking with DrEvil first to get an idea of the frameworks capabilities (or even access to the code).
  
Reply With Quote