Re: gem.sys sources

[hohmuth@inf.tu-dresden.de (Michael Hohmuth)]

Eric R. Smith wrote:

> I also wish that we could have gem.sys source code. Unfortunately, this is
> almost certainly *never* going to happen. GEM is the property of
> Digital Research (now owned by Novell, I suppose) not Atari, and the license
> agreement that Atari signed is extremely restrictive. I heard that we
> couldn't even send source code to Atari subsidiaries -- only Atari Sunnyvale
> is allowed to use it!

I would have thought that Atari has re-implemented most of GEM when
they worked on MultiTOS, and that the new implementation doesn't fall
under DR's license?

> What we (users: I'm speaking here as an ordinary user, definitely *not*
> as an Atari employee) really need is a freeware GEM replacement. I can
> see two distinct possibilities: (1) PRAX (whatever became of that? Michael?),
> or (2) a GEM server running on top of X11. (2) would also have the
> advantage of perhaps being portable to Linux/68K.

I had to stop working on PRAX due to increasing time constraints.
(For those who don't know: It was intended to be a free AES clone.)

When I stopped, I asked several people whether they would like to take
the project over.  However, nobody wanted to.

The sources for PRAX AES server (whose job is comparable with that of
an X server) are partly finished, but in a bad shape; they would need
a major clean-up.  Today, I wouldn't recommend using them as a
starting point for a freeware AES; I'd rather go with option (2) and
build something on the top of X11 (the PRAX server already used most
of X11's Region implementation anyway).  (However, if whoever starts a
new freeware AES project opts for building a server of his own and
wants to see my old sources, I'd go through the hassles and bring them
in a state that I can give away without making a fool of myself. ;-)

As Ulrich already explained, the AES client library (sitting in the
AES trap #2) was mostly finished.  This library implemented most of
the AES calls and communicated with the AES server with a special
protocol over pipes.  I've never seen Ulrich's implementation, but I
guess that this library could be a good starting point for a similiar
library interfacing to Xlib instead of the PRAX protocol.  However, I
don't know whether Ulrich's sources are available for a freeware
project (he once intended to make the client library shareware, an
idea I never quite liked :-).

I still like the idea of a free AES close; just can't do any
implementation work on it.  (However, I would join your mailing list;
I could even create one for you if you want. ;-)

Michael
-- 
Email: hohmuth@inf.tu-dresden.de
WWW:   http://www.inf.tu-dresden.de/~mh1/