From: James Simmons <jsimmons@edgeglobal.com>
  To  : Jason McMullan <jmcc@ontv.com>
  Date: Tue, 30 Mar 1999 09:32:07 -0500 (EST)

Re: Brain Fell Out

On 30 Mar 1999, Jason McMullan wrote:

>   Hiya there. Here's Jason's Status:
> 
> 	My brain fell out. There has been no progress on GGI Console
> for over three months, I seem to have lost interest, mainly due to 
> Steffen's new KGI work [which makes GGI Console practically redundant]
> and the fact that Geert and the FBCon folk are FINALLY grasping the need
> for unified input drivers. 

Yep. I have been working with Geert and the guys to fix up the console
code. My patch is begining to seperate the vt switching and the
console code. 

> 
> 	A couple of people from FBCon have looked at GGI Console and
> have a few salient points:
> 
> 	1) The concept of a unified input event mechanism is needed
> 	   in the kernel.
> 

Yep. Being worked on. Hey i haven't heard from you on this. We need to
join the kernel-input mailing list to see whats going on.

> 	2) The console emulation (VT) and framebuffer (/dev/fb)
> 	   must be independent code.

I'm working on this right now. Also the kernel guys actually like me so 
I can present these ideas to them. Also take a look at the vger tree. I
heard that some improvements where done to the VT code. I haven't looked
at it yet.

> 
> 	3) GGI Console solves more problems that it's worth. If
> 	   people actually USED their consoles 99% of the time,
> 	   it would be work the kernel internal API bloat.

The ideas behind GGI consoel where great. Yes the code wasn't so hot.
Please take no offence. My ggi doom code sucked big time. You where
exploring new ground as I was with ggi programming.

> 	Oh well. This is _officially_ the end of my involvement 
> in the GGI Console project. If anyone wants to take it up, go
> ahead - I'll provide all the support you want. But I'm done with
> the beast for now. 

How about helping out the kernel guys. 

> 	BTW: I still have a Dec Alpha Multia and a 603e based 
> PowerPC, and I don't mind testing ports, or even writing 
> display/input drivers. The core, however, has me licked.

Cool. Machines to test the future developement kernels on. Steffen I like
your ideas. One point is to be made. Instead of developing on our own we
should go to the kernel developers and help work all the ideas together.
Its bad enough that I'm developing a fbcon patch, Vojtech Pavlik 
developing a input patch and Jakub Jelinek is working on a new VT API. Yet
none of us really get together to make sure are patches work together.

James Simmons
System Admin EdgeNet Inc 
Linux Kernel Developer
jsimmons@edgeglobal.com
http://www.edgeglobal.com/~jsimmons