🚀 go-pugleaf

I have an idea.  Its probably already been thought of.  Why
don't we as developers come up with a POSIX/ISO/IEEE portable
class library for GUI's.  

Our company uses two different portability packages ZINC &
XVT (both rather expensive) and it seems that unless developers
are going return to being slaves of the operating systems
wars its time to take a stand on this issue.

I know I'm probably sticking my neck out here but it seems
that we've come all the way embracing open standards/unix/
C/etc. etc. and now we are back to deciding which platform
to develop on based on what the predominant GUI is.

Its time for developers to take the situation in their own
hands.

I propose the following...

ISO/IEEE/POSIX/ACM come up with an elegant OO library 
standard for GUIs .

Several different groups work concurrently on their library
and then we can adopt the one with the best design.  This
would avoid stratification of the design by committee.

Various manufacturers agree to support the library.
This would be a consortium much like VESA et al.

I suggest the name POOL (Portable Object Oriented Library).

POOL would have various levels of conformance related to:

(U)ser Interface
(G)raphics
(F)ile System/Path names
(N)etworking/Distributed Processing
(D)river level
   (The driver level would define a POOL/D interface and the
    system manufacturers would tie that into their driver
    level software. Board manufacturers would then jump
    on the bandwagon and make sure their boards spoke
    POOL/D to the OS)

The manufacturer would write a run-time implementation of POOL
for the native interface.  They would then use the POOL conformance
to advertise the openness of their computers.
I.E.: POOL/U,G,F,N,D conformance

Benefits:

1. POOL could be implemented in C++,OO Lisp (xlisp etc),Smalltalk,
   OO Pascal, and others.  Obviously C++ would be an easy first choice.

2. POOL would encourage manufacturers and developers of windowing systems
   to compete to make the best system possible.  Upgrades to OS's and
   windowing systems, networking etc. would be made faster to keep up
   with competition.  Prices for system software would be lower because
   competition is much higher.

3. A greater diversity of software would be available to everyone.

4. Developers would not have to go to wasted effort to develop for a
   new platform.

5. Developers could write tools/controls/objects in C++ (or whatever)
   to contribute to the productivity of other programmers.  I envision
   something similar to the Visual Basic market where anyone who wants
   a spreadsheet or specialized 3D or high level multimedia support or
   business graphics can simply buy their favorite control off the 
   shelf and drop it into their product.  The result would be much higher
   quality software being released at a much faster rate.

I'm talking about nothing less than software developers set free
in our lifetime.

My only problem would be how to get it available to everyone on the
consumer side.  In other words the typical problems of binary compatibility
arise.  If this issue cant be resolved then the developers would have
to recompile for every new platform.

This might not be a big deal if its pretty much a transparent recompile.

I'm actively soliciting comments on this proposition.  Feel free to
comment on problems we might have or suggestions for new ideas.  No
flames please.  I'll digest all the responses and repost them to the
above newsgroups.  If this process is already under way let me know.
If someone has already tried this and failed let me know why it failed.

Suggest away...

Jim Burnes
jburnes@compusci.com