New Goals doc

[Christian Reiniger <warewolf@mayn.de>]

Ok, the 2 (two!) responses I got up to now regardibg this topic have been
relatively positive on the "distro + own code" model, so I'll detail a bit
more on that:


Abstract:
---------

PenguinPlay aims at creating a comprehensive, universal, efficient an
cross-platform set of tools for game developers, taking as much work off
their shoulders as possible.


Details
-------

We want to create a package of everything game developers will need - the
finest libraries available, combined with our own code to form one
comprehensive toolchest. This includes:

  (1) Collecting the best existing libraries (e.g. the low level GGI
    graphics, input and sound libs, the MESA OpenGL implementation, the
    CrystalSpace 3D engine, the libsigc++ event distribution library),
    ensuring that they properly interoperate (expr?), helping to port them
    to other platforms etc.
    [Note: A style of organization similar to the one of Debian - having
     dedicated "maintainers" for each of these packages - should work best
     for this]

  (2) Creating a "PenguinPlay conformance" specification describing what a
    system needs to contain (in terms of libs) in order to be able to run
    PPlay games [sort of LSB-like, but only for this one aspect]

  (3) Identifying what areas are not covered by existing libraries and
    writing own solutions for those. Also identifying weak spots in
    existing libs and either helpinh to improve those packages or replacing
    them with own code.

  (4) Create a list of recommended applications for game developmemt, such
    as compilers, IDEs, image manipulation software etc (-> LGDC ?)


Target Group: [unchanged]
-------------

PenguinPlay primarily targets hobbyist game developers for now. Thatīs
not because we were not serious about delivering the highest possible
quality and power, but because professional game developers usually already
have their good collection of code and applications and thus wonīt be
interested in PenguinPlay.

Until PenguinPlay is significantly better than what they have that is ;)


Notes
-----

  * This is very close to what Charles Durst proposed as "clearing house
    project" (if you don't know what I'm talking about see the LGDC history
    pages). Back then the primary concern against turning PPlay into that
    clearing house thing was the fear that we could unconsciously favor
    PPlay code over "foreign" one. I don't think that still applies in such
    a setting. Adrian, what do you think?

  * API consistency across contained libraries is of course not maintained.
    It is anyway better to have the best possible API for the respective
    job than to have one that "feels" the same for all tasks yet is fat /
    mediocre / whatever because it has to be suited for all tasks alike.

  * I wouldn't keep the "hard" distinction between Layers P and O in such a
    setting. Simply stating that "all base functionality will be useable
    from plain C and much added functionality will be present for use with
    C++" is IMHO better.



	Christian

PS: GCC 2.95 is out! Hoooray!

--

Drive A: not responding...Formatting C: instead