[Freeciv-Dev] Re: PATCH split-up of caravan capabilities v2 (PR#1148)

[Jason Short <vze2zq63@xxxxxxxxxxx>]

Gregor Zeitlinger wrote:

> On Wed, 19 Dec 2001, Petr Baudis wrote:
> Another thing: while scimming over the ai code I discovered that the ai
> can be confitured with a handicap bitfield but it's hardcoded. What about
> moving into the config files. Example
>
> add_to_size_limit = 8
> add_to_size_limit.ai = 4
> add_to_size_limit.ai.hard = 9
>
> this would mean that in the hard mode the value is 9 and for easy and
> medium 4. Those postfixes could be applied to about anything, but it would
> require to move all options to files (as opposed to the ones that are
> configuable by command line). What do you think of that? 
By "handicap bidfield", I assume you mean enum handicap_type in player.h?

This bitfield looks like it's set up based on the skill level of the ai 
in handicap_of_skill_level().  It's checked with ai_handicap(), which 
deals with human players as well.
The problem, as I understand it, is that the handicaps aren't checked 
very much.  Most code just unhandicaps the AI in every way.  This can be 
fixed gradually, but it'll take time and a willingness to make changes 
to the code.  Then we can start think about adding more options to the 
setting up of handicaps.
In the more general case, I don't think the handicap should be something 
specific to the AI.  AI and human players (and client-side civbots) 
should all be handicappable.  There was a short discussion on this a few 
weeks ago.
jason