In this case, more levels of abstraction is definitely better.

On Thu, 3 Feb 2005 12:55:51 -0500, Cody Brocious
<[EMAIL PROTECTED]> wrote:
> I apologize for talking about this here, and also apologise for coming
> off as sounding like I wanted him to rewrite it, and also for
> forgetting one of the larger points I wanted to make.
> 
> If we seperate the backend from the frontend, it'll make it easier to
> tie into mesa, in theory.  I'd love to help in any way possible, and I
> was mainly asking for opinions on whether this method would be useful
> or even work, as I'm entirely new to the low-level portions of 3d.
> I've never so much as read the source of a video card driver;  the
> closest I've come is reading the source for the simulator.
> 
> If you think that my idea for abstraction might be useful in the
> least, I'll take a shot at it, as I am very interested in contributing
> to this project, even though I'm not very knowledgable in the
> low-level details.
> 
> 
> On Thu, 3 Feb 2005 12:49:07 -0500, Timothy Miller <[EMAIL PROTECTED]> wrote:
> > This is ABSOLUTELY NOT the place to be having this discussion.  Ogsim
> > uses Qt because that's how Nicolai wrote it.  He wrote it, therefore
> > he gets to say how it's written.  If you want to rewrite what he did
> > to use GTK or something else, go right ahead.  That's what free
> > software is all about.  But in the mean time, I suggest you install
> > Linux and Qt if you want to use it.
> >
> > If you want argue with people about off-list, that is also up to you.
> > But trying to convince someone to rewrite their program just for a
> > minority of users isn't likely to be fruitful.
> >
> > Furthermore, we are going to be working on a way of splicing the model
> > into Mesa and that should be toolkit-agnostic (since it doesn't
> > require a UI, we can just use Xlib directly).
> >
> >
> > On Thu, 3 Feb 2005 12:24:44 -0500, Cody Brocious
> > <[EMAIL PROTECTED]> wrote:
> > > I do agree that Qt is nice, but I don't think that it's neccesarily
> > > the right tool for this, or at least it's not the _only_ right tool
> > > for this.  Ogsim is not a UI-heavy system.  If the backend were to be
> > > seperated entirely from the front-end in that we pass in what we would
> > > to a video card and get the result back over a unix socket the
> > > front-end could be designed in python and wxPython/pygame if we really
> > > wanted, and I think that's what I may do.  Qt is nice, but it's
> > > definitely not the only thing capable of running this sort of
> > > software, especially at the level it's at now.
> > >
> > > (Sorry about sending this directly to you, Nicolai, the first time.
> > > Sometimes gmail does the recipient right, sometimes it doesn't.  I'll
> > > have to check by hand next time)
> > >
> > >
> > > > On Thu, 3 Feb 2005 17:41:10 +0100, Nicolai Haehnle <[EMAIL PROTECTED]> 
> > > > wrote:
> > > > > On Thursday 03 February 2005 16:56, Cody Brocious wrote:
> > > > > > I've taken a look at ogsim and I'm quite impressed.  The problem 
> > > > > > is, I
> > > > > > don't use Qt (it doesn't like to build on my system for some 
> > > > > > reason).
> > > > > > I'm wondering how hard it'd be to latch into the backend with a 
> > > > > > custom
> > > > > > client written in something a bit more lightweight (perhaps even 
> > > > > > more
> > > > > > portable, since Qt for win32 costs quite a bit)
> > > > > >
> > > > > > I don't want to do anything that might be construed as trying to 
> > > > > > take
> > > > > > over your project, either, as I know how much it sucks when someone
> > > > > > does that.
> > > > > >
> > > > > > If you have any input on this, I'd love to hear it.  I'm out of 
> > > > > > school
> > > > > > for quite a while due to getting my appendix out, so I've got 
> > > > > > nothing
> > > > > > but time to work on stuff :P
> > > > >
> > > > > Well, Qt/KDE is the right tool for the job. We could get into a 
> > > > > toolkit war
> > > > > here, but the fact is that I've written programs in all of the major 
> > > > > C++
> > > > > GUI toolkits, and Qt really the only one I'd consider when licensing 
> > > > > issues
> > > > > aren't a problem. Or, to put it another way: I really don't like 
> > > > > writing UI
> > > > > code on a normal day. I *hate* writing UI code in toolkits that suck 
> > > > > (i.e.
> > > > > basically everything but Qt).
> > > > >
> > > > > Having said that, the simulator core, which consists of simulator.h,
> > > > > simulator.cpp and render.inc as well as the Python script to generate 
> > > > > the
> > > > > register header are all very UI-agnostic. They do use Qt core classes 
> > > > > for
> > > > > threading.
> > > > >
> > > > > cu,
> > > > > Nicolai
> > > > >
> > > > >
> > > > >
> > >
> > > --
> > > ... And not to pull your halo down around your neck and tug you from
> > > your cloud...
> > > - Maynard James Keenan - A Perfect Circle-  "The Noose"
> > > _______________________________________________
> > > Open-graphics mailing list
> > > [email protected]
> > > http://lists.duskglow.com/mailman/listinfo/open-graphics
> > > List service provided by Duskglow Consulting, LLC (www.duskglow.com)
> > >
> >
> 
> --
> ... And not to pull your halo down around your neck and tug you from
> your cloud...
> - Maynard James Keenan - A Perfect Circle-  "The Noose"
> _______________________________________________
> Open-graphics mailing list
> [email protected]
> http://lists.duskglow.com/mailman/listinfo/open-graphics
> List service provided by Duskglow Consulting, LLC (www.duskglow.com)
>
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to