[Freeciv-Dev] Re: (PR#4111) Client-side AI

["Raimar Falke" <rf13@xxxxxxxxxxxxxxxxx>]

On Wed, Apr 30, 2003 at 02:30:05AM -0700, banjo@xxxxxxxxxx wrote:
> On Wed, 2003-04-30 at 10:23, ChrisK@xxxxxxxx wrote:
> > On Tue, Apr 29, 2003 at 03:03:51PM -0700, Gregory Berkolaiko wrote:
> > > 
> > > The purpose of this thread is to collect the arguments for and against the
> > > client-side AI and to discuss possible dangers and advantages.
> > > 
> > > PRO:
> > > 
> > > 1. The cheating is absolutely controlled by what server is sending to the
> > > client.
> > > 
> > > 2. The introduction of such AI will promote the distillation of the rules
> > > into a module separate from the server.
> > > 
> > > 3. AI will no longr be in danger of mucking the server up by its buggy
> > > actions.
> > > 
> > > 4. It will create infrastructure for writing AIs in other (possibly more
> > > specialised) languages.
> > 
> > 5. The same infrastructure can be used for player controlled modules which
> >    support different parts of the game (think CMA) (CAF Computer Aided
> >    Freeciv).
> > 
> > 6. Users can write AI modules.
> > 
> 
> One place to start coding for this is packets.c
> 
> protocolize the network chat
>   - reduce sending redundant info
>   - formalize and cut down on what clients can see
>   - reduce client side cheating
>   - make it easy for client side AI interfaces to connect
> 
> make a client lib that perl, lisp, python, ruby, java, ..
> can connect to, and let the ai hackers loose.

A formalization of the network protocol as well as a reduction of the
bandwidth was done by the delta patch. It should be included in the
next release.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "When C++ is your hammer, everything looks like a thumb."
    -- Steven M. Haflich