> From: Curtis Stanford <curtis@...>
> Date: Wed, 04 Sep 2002 13:11:54 -0600
> To: Chad Hardin <cehardin@...>
> Subject: Re: Awesome idea!
> Thanks for the response. What is your goal for the display backend? I
> know GNUstep has a pluggable backend that uses X right now.
The goal is to have a fast, anti-aliased, alpha-capable, general purpose
postscript interpreter which can be used for the display and printer. The
EXACT same code will be used for both. It will be shared lib which
executes postscript onto a generic in memory pixmap. The memory can be
shared so the WindowServer can grab it and draw it to screen quickly
> I notice=20
> your webpage says you're working toward using a framebuffer type
> display. What about using something like GGI or SDL? They seem to have
> abstracted the differences between video cards and different
> acceleration models. It sure would be nice to drop X-windows some day.
I'm leaning towards DirectFB, they have support for the newest card and are
This is how I'm thinking it will be as of right now:
*Right when init starts try to figure out what the Vid Card is(are).
*Set up the proper kernel framebuffer modules and attach it (them)to the
*also set up some type of input managemnt to handle things like multiple
mice or keyboards.
*start the WindowServer, which is very lightweight and can handle n-display=
and works with the input manager
*we are now graphical and can boot up as such!
I want to write a basic postscript interpreter, I know a full blown one is
no easy feat, but I've been reading the specs and I think I can write a goo=
'nuff one for now. Anyhow, this interpretor will not be an executable, it
will be a shared lib and the only thing it will know how to draw to will be
into a memory pixmap. This pixmap can be transferred to the display manage=
or Print Manager for output via shared mem.
A new backend will need to be created. It will be a bit diff then the
others since clients will need to draw their own window decorators.
> I've been trying to figure out why Apple went with a Display PDF scheme.
> PDF is an OK format for describing platform independent documents but
> seems a little complicated and verbose for a display server. Same with
> Display Postscript I suppose. It does have the advantage of matching
> printed output I guess.
That's a very good thing to get real that WYSIWYG. I admit that postscript
code has the possibility to get really complicated (anything goes in that
language). But as long as simple commands are sent it shouldn=B9t be a big
> I've always thought that the smartest thing Apple could do is release
> OSX for existing 386 platforms. If they really want to compete with
> Microsoft and take away market share, that's what they should do.
I don't think they really want to be what Microsoft is. Trying to support
everything with the ultimate product being crap=81. If they switch from
Power, I think it will to AMD x64 and their machines will be propietary.
> are hundreds of millions of PC's out there begging for a better O/S. I
> guess that's where Simply GNUstep comes in eh?
> Chad Hardin wrote:
>>> From: Curtis Stanford <curtis@...>
>>> Date: Tue, 03 Sep 2002 09:50:28 -0700
>>> To: cehardin@...
>>> Subject: Awesome idea!
>>> Chad, this whole Simply GNUstep thing in an excellent idea.
>>> I've been thinking along these lines for a while now. All
>>> the existing Linux GUIs pale in comparison to OSX. I've been
>>> a fan of the NextStep API ever since it came out I don't
>>> know how many years ago. Kudos to you for actually doing
>>> something about it. A few questions if I may:
>>> 1. Are you worried about the rumors that Apple has OSX for
>>> i386 platforms in its back pocket, ready to spring upon the
>>> world when the right time comes?
>> Nope. I'm certain Apple would use some proprietary components in th=
>> x86 systems (especially the BIOS) in order to make them easy to use and =
>> (booting from cd-rom by simply pressing 'c', booting from firewire,
>> etc,etc). If apple ever makes these machines, I'm sure these machines w=
>> be priced above generic x86 hardware.
>> More importantly, I think that if Apple does move away from PowerPC
>> they will most likely make the jump to 64bit as well while they at it. =
>> guess would be the AMD 64bit arch, which I can't even think about target=
>> right now anyhow.
>> SimplyGNUstep is targeted at the x86 32bit arch (right now). That means=
>> can get an old dirt-cheap 500MHZ K6-2 box and it will probably run faste=
>> than an iMac and at a fraction of the price. If Apple does go 64bit x86
>> then SimplyGNUstep will have the niche in the 32bit x86 area.
>>> 2. Will you be able to re-compile OSX applications on
>>> GNUstep and have them work?
>> Yes, but bot straight out. The Interface Builder files (the files that
>> define how the User Interface is laid out) are not compatible. They hav=
>> be translated to the GNUstep version of the file. Also, the Makefiles a=
>> not structured the same.
>> Take a look at the GNUMail project. It is a great example. The app ca=
>> be compiled on either platform.
>>> 3. Is anyone working on a decent skin for GNUstep? I dig the
>>> basic NextStep look but it sure is boring compared to Aqua.
>> It has been brought up and some quick hacks have been made. Nothing ser=
>> is being developed though (AFAIK). The Nextstep look is not beautiful b=
>> any means. It is highly functional, easy to understand, and fast to dra=
>> though. While Aqua is visually appealing it also very resource intensiv=
>> Could somebody please explain to me why my old 25MHz Nextstation running
>> OpenStep 4 boots faster than my 500MHz iMac running Os X?
>>> Thanks. I don't have a lot of time but if you need a little
>>> help here and there, I can try to help out. I'm a long time
>>> C, C++, Java, UNIX Programmer.
>> Awesome, I need some good hardware detection and module loading!