New GUI tool for Linux. [idea]

New GUI tool for Linux. [idea]

Post by harmsNOS.. » Thu, 23 Apr 1998 04:00:00





> Hello all,

>    I have an idea for a new windowing system for Linux to be developed in
> SVGAlib & OpenGL. Lindows or something ;-)

> The final product will (hopefully) include a good GUI builder, a 3D look,
> lots and lots of nifty little useless Win95 like (except different, I have
> hundreds of ideas, but only tens of hours to code and design) widgets.

> The whole thing will hopefully make Linux very easy to use for people to
> develope on, and to use. The difference between it and XWindows is that all
> the widgets will be standard. They better be pretty nice looking ;)

> Why not use XWindows? XWindows is big and bloated. It doesn't run well on
> the throwaway systems that stray Windows users put Linux on to see what it
> is. Its too complex. I'm thinking of developing something simple to develope
> in (if that made sense).

> If interested, follow up, but try to nudge the cc button over to my email
> address too ;-)

>    Matt (sighing as the spambots are taking aim on him as he hits ctrl-k-x-y
> from JOE to send his news posting..)
> --
>                         -=-=-=-=-=-=-=-=-=-
>                            Matt Busigin

>                        OmegA MUD System Coder

Yes, and I keep thinking it should have the ease of use of the old Appware
(from Novell) with the ability to wrap existing widgets into the
equivalent of the (Appware) Loadable Modules!

Unfortunately the company that bought Appware (Network Multimedia) is not
doing too much recently.  The website is down and the change of name
(MIcrobrew) turns any websearch for it into a nightmare.

However the ease of use is astonishing. A tool like that would let
beginners like me contribute much more.

Stefan Harms

--
remove NOSPAM to reach me

 
 
 

New GUI tool for Linux. [idea]

Post by Alan Shutk » Thu, 23 Apr 1998 04:00:00


M> Hello all, I have an idea for a new windowing system for Linux to
M> be developed in SVGAlib & OpenGL. Lindows or something ;-)

Is there a version of OpenGL which will run on SVGAlib, or are you
writing your own?

Personally, I think it's all a waste of time, as SVGAlib is dead and
should remain that way.  You might talk to the berlin group, who have
the same idea as you, but runnng on top of GGI.

--

An adequate bootstrap is a contradiction in terms.

 
 
 

New GUI tool for Linux. [idea]

Post by Jonathan Apfelker » Thu, 23 Apr 1998 04:00:00



> Hello all,

>    I have an idea for a new windowing system for Linux to be developed in
> SVGAlib & OpenGL. Lindows or something ;-)

Well, sounds like being obsolete already ;)) SVGAlib is not developed
anymore,
but instead GGI tries to be it's replacement. On top of GGI you might
implement
whatever you want, having the drivers already at hand.
http://www.ggi-project.org
See also http://www.debian.org/berlin/

Quote:> The final product will (hopefully) include a good GUI builder, a 3D look,
> lots and lots of nifty little useless Win95 like (except different, I have
> hundreds of ideas, but only tens of hours to code and design) widgets.

How about NOT making a specific gui but instead give hints (say,
fileselector
of 70% of the window, then 15% inputline, then button "ok" left and
button
"cancel" right) so ANY widgetsets may be applied? See the
nirvana-project
at http://tek.flynet.de/linuxsite/nirvana.

Quote:> The whole thing will hopefully make Linux very easy to use for people to
> develope on, and to use. The difference between it and XWindows is that all
> the widgets will be standard. They better be pretty nice looking ;)

Easy. Install only KDE apps. http://www.kde.org

Quote:> Why not use XWindows? XWindows is big and bloated. It doesn't run well on
> the throwaway systems that stray Windows users put Linux on to see what it
> is. Its too complex. I'm thinking of developing something simple to develope
> in (if that made sense).

Yes. It's braindead. So don't reinvent it. ;)

Cheers
Jonathan
--
"...one of the main causes of the fall of the Roman Empire was that,
lacking
 zero, they had no way to indicate successful termination of their C
programs."
-- Robert Firth

 
 
 

New GUI tool for Linux. [idea]

Post by Christopher Brow » Fri, 24 Apr 1998 04:00:00



Quote:>   I have an idea for a new windowing system for Linux to be developed in
>SVGAlib & OpenGL. Lindows or something ;-)

>The final product will (hopefully) include a good GUI builder, a 3D look,
>lots and lots of nifty little useless Win95 like (except different, I have
>hundreds of ideas, but only tens of hours to code and design) widgets.

>The whole thing will hopefully make Linux very easy to use for people to
>develope on, and to use. The difference between it and XWindows is that all
>the widgets will be standard. They better be pretty nice looking ;)

>Why not use XWindows? XWindows is big and bloated. It doesn't run well on
>the throwaway systems that stray Windows users put Linux on to see what it
>is. Its too complex. I'm thinking of developing something simple to develope
>in (if that made sense).

X is too "big and bloated" for the throwaway systems of five years ago.
Anything of "PCI" description is generally well enough endowed that X is
not normally overly constricted.

And of course the assertion that you can build something substantially
better is somewhat presumptuous, as X has involved hundreds and possibly
thousands of person-years of design and implementation effort.

I am starting to suspect that there would be some value to building a
"legacy X" implementation that just implements X10R4 along with a
"generic SVGA" driver.

- At that point, X was significantly smaller and simpler.

- X10R4 nonetheless contains substantially the same core functionality
that there is in newer versions.

- By starting with a *working* codebase that has been used to develop
and support substantial applications, you have the benefit that you know
that the system will at least potentially work.

I expect that by the time you add back in all the things that you think
are missing and need to be 'beefed up,' you will find yourself with a
remarkably similar memory footprint to what we have today in X11R6.3,
but to each his own...

I think this can start to be a "standard response" to proposals to
create a "less bloated" alternative to X...  "So why don't you go back
to X10R4, and build something better?"
--
linux: the choice of a GNU generation


 
 
 

New GUI tool for Linux. [idea]

Post by nathan wagn » Fri, 24 Apr 1998 04:00:00




> X is too "big and bloated" for the throwaway systems of five years ago.

This is something i've never understood.  What's big and bloated about X to
begin with?  X is a protocol, and having recently read over book zero, i
don't see the problem.

The *libraries* i'd agree are mostly junk, but the protocol itself strikes
me as fine.  Why not just write new libraries that use the same protocol?

--
nathan wagner             "People with neckwear always have money."

-- looking for a *NIX sysadmin job within driving distance of Madison WI
   resume at http://granicus.if.org/~nw/resume.html

 
 
 

New GUI tool for Linux. [idea]

Post by Matthew Busigi » Fri, 24 Apr 1998 04:00:00


: Hello all,

:    I have an idea for a new windowing system for Linux to be developed in
: SVGAlib & OpenGL. Lindows or something ;-)

: The final product will (hopefully) include a good GUI builder, a 3D look,
: lots and lots of nifty little useless Win95 like (except different, I have
: hundreds of ideas, but only tens of hours to code and design) widgets.

: The whole thing will hopefully make Linux very easy to use for people to
: develope on, and to use. The difference between it and XWindows is that all
: the widgets will be standard. They better be pretty nice looking ;)

: Why not use XWindows? XWindows is big and bloated. It doesn't run well on
: the throwaway systems that stray Windows users put Linux on to see what it
: is. Its too complex. I'm thinking of developing something simple to develope
: in (if that made sense).

I've heard many points of views over this, and I think I forgot to mention
the main reason for it;

Easy of use to code and run games. X11 runs applications really well, since
most applications don't use a tremendous amount of graphic intensive things.
Ease of use to code applications might also draw people to it as well, but
it is *not* meant to repalce X11. I am not thinking of designing with the
same networking concepts as X (althrough doing a simple implementation of
such wouldn't take too long. I am thinking of a very very small GUI that
will let larger, really graphic intensive games run on. The whole system
will be catered (sp?) to game design, development, implementation & play.

I know this seems like a change of direction (it sort of is, I never really
said that in my original posting), but I sort of wanted something like this
the whole way around.

Game creation with RAD tools (good RAD tools) would really kick ass, and
that's what I'm thinking.. second opinions? Flames?
--
                        -=-=-=-=-=-=-=-=-=-
                           Matt Busigin

                       OmegA MUD System Coder

 
 
 

New GUI tool for Linux. [idea]

Post by Tim Teulin » Sat, 25 Apr 1998 04:00:00


Hallo!

Quote:>   I have an idea for a new windowing system for Linux to be developed in
>SVGAlib & OpenGL. Lindows or something ;-)

SVGALib does not work on a number of new graphics cards. Seems like
developement slowed down or even stoped.

Quote:>Why not use XWindows? XWindows is big and bloated. It doesn't run well on
>the throwaway systems that stray Windows users put Linux on to see what it
>is. Its too complex. I'm thinking of developing something simple to develope
>in (if that made sense).

Use VisualOberon (http://www.ping.de/sites/edge/VisualOberon.html). It uses an
OS abstracting graphics core and already runs under X11. Port it to SVGALib or
better GGI and get a fullblown widget set. A port to Windows in future maybe
possible, too.

--
Gru?...
       Tim.

 
 
 

New GUI tool for Linux. [idea]

Post by Geoff La » Sat, 25 Apr 1998 04:00:00




Quote:> Personally, I think it's all a waste of time, as SVGAlib is dead and
> should remain that way.  

Hope not - it's the only resource available for creating very light weight
semi-portable Linux GUIs.  

While everybody is building huge mounds of software to rival Windows, what
*I* need is a GUI that is _fast_, simple and small.  If anyone here has used
Acorn's WIMP GUI they will know that kind of thing I'm talking about.

--
Geoff. Lane.                                    Manchester Computing

"Cool Brittannia"
Track 1 from Gorilla by Vivian Stanshall + Bonzo Dog Doo/Dah Band, 1967

 
 
 

New GUI tool for Linux. [idea]

Post by tcklnbr » Sat, 25 Apr 1998 04:00:00


No flames, I am favorably inclined to OpenGL myself, but some
considerations:

OpenGL has lacked some critical feature:  a persistent state mechanism
and meta-file format.  As in, OpenGL can't save or recreate a drawing
path.  "Undo" is impossible. This has been remedied by third parties as
C++ libraries but their solutions lack standardization.  Some people
propose using VRML---but VRML describes what you want something to look
like, not how to draw it., ie, also lacks necessary OS-level primitives.

OpenGL is a purely, in memory, drawing engine---it doesn't have a clue
as to how to draw even on-screen!  The OS's graphics engine puts the
OpenGL "drawing pipeline" to the display or printer.  OpenGL
implementations are really matrix math libraries.  

While conventional graphics is built on matrix-math, its easy to show
that matrix math is inefficient for the normal use:  Newer, more
efficient concepts are focusing on polar coordinates and rotations.  Do
we wanna go with the old stuff---"Its a mess, but its a mess I have
learned how to live with" or try a new approach for the future?

Also, OpenGL only does bitmap fonts, per se.  It is awfully primitive
for rendering glyphs (line-based, Vector fonts) even if you have data
sets for the outlines. Compare the OpenGL API to PostScript to get an
idea.

Postscript exists for a reason. Correctly scaling a character glyph is
quite different from scaling a solid.  If you just primitively
"multiply" the glyph in all directions you can wind up distorting the
character's look.  Also, an accurate set of data for an outline of a
glyph is larger than a Postscript instruction set and slower to render
and scale.  Also a glyph is defined exclusively as 2D.  If you extrude a
glyph to make it 3D, you now have a solid, alright. but no scaling hints
for the third dimension---if it matters.

OpenGL is great but it doesn't contain enough stuff to be a system-level
drawing engine, in itself.

You probably need something like a basic 2D system-level drawing engine,
Postscript (or something else? what??? for glyphs and other "line-based"
drawing, and then Open Gl for solids.  Something that had all these
capabilities built-in would truly be awesome.

/**/

(How do I know all this---because I spent 2 months trying to write a 3D
type 1 font app for OpenGL under OS/2.  The learning curve ended in
failure.  Might have succeed though, if IBM and Adobe would have agreed
to expose the ATM's API or put the actual glyph drawing instructions in
the metafile.)

(OS/2 had a real gap in this area---IBM liscensed the ATM and whenever a
call to render a type 1 font was made, the meta-file didn't contain the
drawing instructions for the font, but rather a paramaterized call to
the ATM which returned a hidden bitmap.  This meant that you couldn't
use the "bought-and-paid-for" ATM drawing instructions inside OpenGL for
font rendering, either. )

(If you want to see a good, clean, textbook implementation of a 2D
engine with nice device independent drawing, metafile format,
etc.---look at OS/2's Presentation Manager, IMHO. OS/2's technology
didn't fail--->IBM failed. Pitiful really!)

All of this made be appreciate Postscript's particular magic.
/**/

Best Wishes,

 
 
 

New GUI tool for Linux. [idea]

Post by Alan Shutk » Sat, 25 Apr 1998 04:00:00




>> Personally, I think it's all a waste of time, as SVGAlib is dead
>> and should remain that way.

G> Hope not - it's the only resource available for creating very light
G> weight semi-portable Linux GUIs.

Well, fact is, it basically is dead.

I also mentioned GGI in my post, which has the advantage of people
actually working on it (as well as intended stability and other useful
thing).  What's wrong with it?

--

Happiness is just an illusion, filled with sadness and confusion.

 
 
 

New GUI tool for Linux. [idea]

Post by David E. Fo » Sat, 25 Apr 1998 04:00:00



>The point is that it is simply *too large* for a 386 system with 8MB of
>RAM to happily run along with any interestingly complex applications.
>And that is is the sort of "throwaway system of five years ago" that
>one sees.

I started running Linux on such a throwaway system, by the way. It was a
386sx/16.  Originally it had 4 megs of RAM, which i later upgraded to 8. I ran
(and still do, only on faster hardware these days) Accelerated X, which (in
those days) could eat up a rather large chunk of memory. I also ran it on el-
cheapo 14" monitors and graphics cards that didn't have a chance in hell of
being able to go past the VGA (640x480x16, or maybe 256) standard.

X ran, but it didn't do all that well on the older video hardware -- it wasn't
until I got a better monitor / video card before most apps would run
more or less correctly, and display nicely.

The same machine I used to run Linux on used to run Windows 3.1, before I
switched over to Linux. And, at least as far as displays go, Windows did a
better job on that class of hardware than X did.  X apps across the board pretty
much assume that one has a nice video card with a good monitor, and can do at
least 1024x768. (Try running something like xmahjongg in 640x480x16, using the
VGA16 X server, and you'll see what I'm talking about -- one can only see a
sixth or so of the playing field.)

In contrast, (at least some) Windows programs would display better under such
extremely confined circumstances.

Quote:>If you want to have 8 bitplanes by 640x480, that works out to something
>like 2457600 bits, which is a significant little chunk of RAM, which when

Sure. The memory for the display must needs be allocated on Windows as well.
X is going to have to slog these bits around the system just like Windows is
doing.  

--
------------------------------------------------------------------------
David E. Fox                 Tax              Thanks for letting me


-----------------------------------------------------------------------

 
 
 

New GUI tool for Linux. [idea]

Post by Christopher B. Brow » Sun, 26 Apr 1998 04:00:00





>> X is too "big and bloated" for the throwaway systems of five years ago.

>This is something i've never understood.  What's big and bloated about X to
>begin with?  X is a protocol, and having recently read over book zero, i
>don't see the problem.

>The *libraries* i'd agree are mostly junk, but the protocol itself strikes
>me as fine.  Why not just write new libraries that use the same protocol?

Sigh...

Yes, "X" can assortedly mean:
a) Protocol, or
b) Implementation of software that uses that protocol.

An X system is devoid of value if it is merely a protocol sitting in
books on the shelf; it must be expressed as implementation in order to
be of any use.  Calling X a protocol may be correct, but ignoring the
software is rather silly.

All known implementations (and I'd be entertained to hear of any
counterexamples) are based at some point on MIT and/or X Consortium
code.

I would claim that X *isn't* particularly bloated (see:
<http://www.hex.net/~cbbrowne/xwindows.html> for material detailing
such claims)

The point is that it is simply *too large* for a 386 system with 8MB of
RAM to happily run along with any interestingly complex applications.
And that is is the sort of "throwaway system of five years ago" that
one sees.

If you want to have 8 bitplanes by 640x480, that works out to something
like 2457600 bits, which is a significant little chunk of RAM, which when
added to code to implement protocol as well as something to represent
any other required data structures (buffers, anyone?) will consume most
certainly a couple megabytes of memory at the very least.  Not gonna work
well on the "throwaway" system.
--
Those who do not understand Unix are condemned to reinvent it, poorly.  
-- Henry Spencer          <http://www.hex.net/~cbbrowne/lsf.html>

 
 
 

New GUI tool for Linux. [idea]

Post by Geoff La » Tue, 28 Apr 1998 04:00:00




Quote:> I also mentioned GGI in my post, which has the advantage of people
> actually working on it (as well as intended stability and other useful
> thing).  What's wrong with it?

I too had high hopes for GGI but until it's accepted as a standard part of
the kernel it's not suitable.

--
Geoff. Lane.                                    Manchester Computing

"Cool Brittannia"
Track 1 from Gorilla by Vivian Stanshall + Bonzo Dog Doo/Dah Band, 1967

 
 
 

1. New GUI tool for Linux. [idea]


No hope *YET.*

There are two problems with putting KGI into the kernel:

a) It's not robust enough yet, and

b) There is disagreement as to the merit/quality of the architecture.
(Linus thinks that the GGI people are trying to push some dumb things
into the kernel; I prefer to stay out of that fight...)

Both problems are solvable.  In particular, if the GGI coders "build a
mousetrap" that is so good that everyone starts lobbying to have this
*USEFUL CODE* in the kernel, then this is the point at which
disagreement may be (perhaps reluctantly) set aside.

In other words, if it gets really robust and really functional, then b)
may be overcome by GGI's other merits.  It's Not There yet.

The Right Answer may be to also fix whatever is "broken" about the GGI
architecture; no doubt that is malleable enough that there will be some
changes made by the time the overall system starts to get robust enough
to consider moving it into "production" kernels.

--
We are MicroSoft.  You will be assimilated.  Resistance is futile.
(Attributed to B.G., Gill Bates)

2. About partitioning

3. Help with WAV files

4. New tool idea?

5. DVD-Writers and Linux

6. New REXX GUI Front End TOOL for Win95/NT & OS/2

7. my email address from tin

8. Gui tool similar to "Motif Tools"

9. Invite ideas on implementing a packet generation tool for linux...

10. GUI based embedded Linux config tool - correction

11. What GUI development tools are there for Linux?

12. Looking for GUI - TOOL (DOS and LINUX)