Re: XBOX Linux Parralel project

Selon Karim Liman-Tinguiri <[email protected]>:

> Greetings,
>
> I'm a developer interested in launching an alternative to the OpenXDK project
> while using Linux.

Great, one more fork, let's spread our efforts a bit more...

> My goal is to create a compiling toolchain that would create a single XBE
> file with (a slightly modified version of) Xromwell inside + (towards the end
> of the file) a minimum filesystem, a kernel and X (no need for a complete
> windows manager, just graphic card support; i know there's no accelerated
> drvier yet).

What's the point in providing such a toolchain (mostly aimed at game
developpement) as long as you can (and surelly will never be legally able to)
provide 3d accelerated graphic support?

> At boot time, the following would occur:
> [... how does any live distribution works ...]
>
> The most important thing within the XBE is the "user program". The user
> program is what we actually want to boot. The kernel, X and the drivers are
> just there to provide support for the program to be able to access XBOX
> hardware without relying on the microsoft kernel. That way, XBOX developers
> can code app that'll run on a sort of micro-distro bundled within a single
> XBE file.

Homebrew developpers are mostly interested in building on a widelly used SDK to
get hired by a real software company. You're aiming at the wrong target anyway.

> The main obstacle I'll have to face within this project is to reduce to the
> minimum possible the XBE overhead (the Xbeboot + the kernel + the filesystem
> + X) for example by compiling the kernel only with XBOX-specific supports,
> using agressive kernel optimization (I'll strip out lot of hardware-safe
> procedures since we know precisely on what hardware the program will run),
> rewriting lots of routines in the drivers for exemple in assembly, optimizing
> X only for the XBOX hardware etc...

Oh, so you want to fork the kernel and xfree too? Isn't it a pure waste of time
as a standard kernel built with modules takes less than 1.5MB? And why would
you use X? SDL on FrameBuffer is barelly enough for what you want to do.

> That way we could ship small versions of mplayer for example in a single
> small XBE that doesn't need any Linux installed on the box to run since it
> comes packed with it. We could finally have some legal binaries for
> media-players, dashboards (amongst other)...

As far as it rellies on a specific BIOS (I mean for game developpers), on a
single platform (Why would you only provide xbox support? what about PS2, PC,
...), this project cannot be attractive enough to hire developpers and other
staff to work on it imho.

> Sincerly yours,
>
> PS: What I'm looking for are coders, webmasters, documentalists and anybody
> who can help throughout the achievement of this project. I aim to have it run
> in parallel with the xbox-linux project. It could be called for example the
> LinuXDK - or LinXDK to show it's really binded with Xbox-Linux.

Please try to think in a rationnal way, evaluate your target, then think your
project to be the killer SDK for any game developper. That way you'll find
people to work on it...

Digitally yours,

Jérôme


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/