"Mike A. Harris" wrote:
> On Tue, 5 Jun 2001, Digital Z-Man wrote:
>
>
> Well, you haven't looked very hard my friend. ;o)
>
> The design document is included in the XFree86 source code, and
> always has been.
>
> xc/programs/Xserver/hw/xfree86/doc/DESIGN
>
> In fairness to you though, it isn't like someone would
> accidentally stumble across it easily when it is buried 6 levels
> deep.
Maybe that should be relocated to a more visible spot, or a symlink to
it
in the rootdir? I guess
you could 'find' it, but again, you kind of have to know what you
are looking for.
>
>
> I also applaud Matrox for their friendliness to open source, but
> there still are no G450 docs available at least last time I
> checked.
There will be, when Matrox makes a new chipset, or it stands to reason,
at least.
>
>
> If one is interested in becoming an XFree86 developer, it makes
> sense to apply first, and at least _see_ the agreement before
> commenting on it being some secret magic ring.
I'll read the agreement, but your statement about "applying" to join
the development team gives me the impression that there is a secret
magic ring. Just my observation.
>
> There is a project on sourceforge to create a new X server from
> scratch. "linuxgfx"
>
> While it is a cool idea, it will take 10 years to complete.
> XFree86 wont be sitting idle for that time. It is easy to say
> "scrap XFree86", and I agree that it is a huge amount of code,
> and that a lot of it is a mess, but nobody is going to just dump
> it and start rewriting something from scratch all of a sudden
> and start using it.
Ok, maybe I was exaggerating a bit here...I meant scrap XFree86
in its present form. Keep the server, the networking code, and 2D
primitives. The fact that the newbie has to understand the X11 Server
codebase IN ADDITION to the DRM tells me there's alot of coupling
going on. Maybe this was done for the sake of efficiency, but it just
makes the newcomer shun away from helping out.
>
>
> We are the largest Linux distributor out here, and we definitely
> don't have the resources to put into something 1/100 the size of
> such a project, even in combination with resources from many
> other vendors. Not to mention that for such a project to be
> truely successful, it would require a lot of talent from those
> who have been doing this stuff for 15+ years. Those people are
> unlikely to dump XFree86 and start a new project that has little
> chance of getting off the ground.
Aren't we tooting our own horn just a wee bit, here? ;o)
I'm glad that RH is blessed with so many capable developers, but
remember that they need not shoulder the burden alone within
the confines of the company. There's a bit of M$ mentality in
this statement. There are other people who can help. It's not just
up to the (RH/VA/PI guys) to redesign XFree to be more modular wrt DRI.
Anyway, this kind of inertial thinking is not good for the project in
the longrun.
>
>
> Ever seen the Berlin project? Nice idea 6 years ago. Where is
> it today though?
It was at the 2000 Linux Conference in Ottawa, Canada...very
interesting.
>
> Rather than redesigning a new X, what would be more useful is to
> find bottlenecks and optimize them. The biggest bottleneck is
> the X11 protocol and friends itself. If the client and server
> are on the same machine, it doesn't make sense to have xlib
> encode a protocol stream, connect a socket to a server process,
> and then have the server decode that, and then make it happen.
It's not X I'd like to see redesigned, at least not majorly. The DRM
is ok as a driver layer. But to add a large software module that
entails
significant changes to X11 just to optimize things locally is baaad
news...
I know this is the way industry does its programming...believe me I
think
we all see this all the time.
>
>
> 329Mb+ of source code however doesn't get written overnight
> though, so I'm skeptical as to how far it will get.
No, but it's discouraging to hear you say that. I will look into
linuxgfx...as soon as I get a good broadband connection ;0P
Peace, Ice.
--
"Klingon program function calls do NOT have 'parameters' - they have
'arguments' - and they ALWAYS WIN THEM."
"You question the worthiness of my code?
I should kill you where you stand!"
-- K'Pleth Klingon Programmer Squadron
NetZero Platinum
No Banner Ads and Unlimited Access
Sign Up Today - Only $9.95 per month!
http://www.netzero.net
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/dri-devel
