Print

Print


   > Hey, I'd be happy to undertake the production of all these Ada
   > utilities/bindings/etc.  If arrangements were made to continue
   > my annual salary I'd gladly work out of my downstairs office :-)

I join Marc in this sentiment. Further, I do about 5 hours Ada
development at home each week. I have posted all of my code,
either to the Net or to the Magnus Archives of Group Theory. It
would be nice if the volunteers could be coralled and work towards
common goals, such as a working DOS version of gnat, the Booch
components, components beyond Booch, working ASIS global
cross reference list, Ada MIDI player, Ada MIDI composition
(musician makes MIDI file), Ada conversions for every graphical
and text compression type, and Ada methods of doing everything
available to awk, perl, grep, cshell, commandDOTcom, SVGA (say VESA),
RPC, Visual XXXX, Turbo Pascal, and Delphi users.

It seems the priority would be the following:

    (1) make a gnat 3.11 version for DOS, but upgrade only the
        following parts of the runtime library, worry about
        the tasking problems and the delay problems much later.

    (2) fix the runtimes of gnat across all systems (dos, linux,
        nt, sparc, mips, etc.) to consistently and correctly
        handle get_immediate, stream_io to and from standard
        input and standard output, access I/O ports,
        correctly handle the difference between virtual and
        absolute (memory-mapped) addresses, read directly,
        set/get environmental variables, rename files,
        get mouse coordinates, light dots on the screen in SVGA modes,
        read and write serial and parallel ports, and
        submit a text string as if the user typed it into the
        command line processor character for character (that is
        without cracking off the args first) (and, that is,
        returning both statuses: errors if it could not
        submit the command and errors if the command was
        submitted but did not execute with return code 0).

    (3) brutally, one by one, take every public domain language
        like perl, awk, gcc, and make each feature in their
        libraries available to Ada programmers, starting with
        awk regular expressions, awk system function, etc.

    (4) gently, with permission, or by redeveloping completely
        in a different way, take visual basic, turbo pascal,
        delphi, etc., and start making those features available
        to Ada programmers

This corral is not closed, but open on two sides. You can join
this development sequence without giving up personal projects.
And you can pick and choose projects that you have expertise in
without worrying about higher priority projects that you cannot
do. Eventhough the DOS implementation is obviously highest
priority to get in the people with cheap machines and memory mapped
DOS editors that work many times better and faster than windows
editors -- it is probably wisest to wait a few months before
starting it so that the good combined C library becomes
available first.

Mike Brenner
.


Mike Brenner