Re: UI proposal

[Eric McDonald <mcdonald@phy.cmich.edu>]

Hi Bruno,

On Wed, 19 Nov 2003, Bruno Boettcher wrote:

> but i thought that adding other interfaces was hard due to the deep
> interwindling of xconq game machine and UI, the stuff that prevented a
> clean client server design IIRC?

I am no authority on adding interfaces to Xconq as I haven't spent 
much time with the UI code, but the kernel is essentially 
abstracted from the interfaces. I also believe that Stan(?) 
pointed this out recently, and he said that the ability to add 
new interfaces was the primary motivator for this. From a kernel 
hacker's perspective, I can tell you that I don't see alot of 
evidence of deep intertwining.

However, I will agree that a client-server model would be nice. I 
currently don't feel ambitious enough to attempt the separation, 
and dropping in the clisrv communication layer in between. I don't 
plan on doing this in the 7.6 release cycle either. But perhaps 
others are more ambitious....

The reason I have thought about a clisrv architecture in the past 
was not for the GUI's, but for the AI's.

  Regards,
   Eric