Re: FYI: Linux Game development

[Christian Reiniger <warewolf@mayn.de>]

B. Kreimeier wrote:
>Two remarks (nope, not going to join the mailing list
>any time soon).
>
> a) I have been working for a commercial game company
>    that intended to use Java with their next game.
>    Partly decause of the insufficient support for Java under
>    Linux (which is Sun's fault, and occured despite the
>    great efforts of the Blackdown Group), this was
>    postponed.
>
>    IMO it is crucical to publicly urge Sun for better Linux 
>    support, including distribution of a Linux reference
>    JDK (which is being worked upon by Blackdown with
>    Sun support already) from their web site, on their 
>    CD's, announced by Sun as an *official* *SMI* reference
>    implementation.
>
>    Games like RBR's "PraxWar" are already written in Java,
>    but will always rely on JNI (and, in some cases, Invocation).
>    Failure to deliver a decent Win32 JDK will force developers
>    to use MS Java, which does not support JNI, but RNI,
>    which hampers ports to non-Win32 OS. Vice versa, if a
>    stable JDK reference is available on Win32, Linux,
>    and maybe other server platforms like Solaris and OS/2,
>    that *does support JNI and Inovcation reliably*, ports
>    of such state of the art games using Java will be possible
>    at diminishing costs.

Good point. I think we should place this post from you in the
"Discussions | Feedback" section on the homepage once it's up. Is that ok
with you?

This also sounds like a good topic for a linuxgames.org
editorial:

Writing cross-platform games in Java
* Current Java situation
* Problem areas
* Java and speed
* ...

BTW (mainly to the list): What about starting a "promising/interesting
editorial topics" list on the homepage?

>    It is also essiential to urge Sun to publicly endorse and
>    support the ARB's Java bindings for OpenGL, work in
>    progress that has been delayed and hampered by Sun
>    politcking and API inflation (Java3D) for years. Again,
>    these bindings (and portable, non-Direct3D games relying
>    on them) will require a performant, portable JNI.
>
> 
> b) There is a subset of resource-intensive games that will
>    preferably be run full screen, and w/o any overhead
>    running. Gary McTaggart (of 3Dfx fame) once created a
>    DOOM engine demo using a resque boot disk and a root
>    disk - circumventing DOS/Win32 implementation completely.
>    It got him his job.
>
>    There have been various attempts to create a GameOS, some
>    of which Linux based. Shooter games increasingly rely on
>    Internet support, and fast sampling of user inputs, and
>    could benefit from Linux features like modules and efficient
>    drivers, and lean kernels. Ideally, a bootable CD ROM that
>    identifies as many hardware components as possible w/o any
>    setup, and proceeds automatically into a fullscreen app,
>    providing a Realtime capable kernel to the developer, would
>    make an excellent Game-only base. There is no huge market,
>    but the support advantages and a significant number of
>    hardcore gamers could provide this approach some leverage.

Well, there are both advantages and disadvantages of this.
Having a bootable CD wit both OS ad game on it can (theoretically) make PC
gaming just as easy as console gaming, but it also makes it very hard to
impossible to e.g. open a web browser to look for cheats while running the
game.

I think having a good (un)installer and console gfx support for Linux is a
better thing even for these extremely resource-intensive games. Linux -
even a standard multi purpose installation can be very slim at runtime.


Cu	
	Christian
--

AAAAA - American Association Against Acronym Abuse