David Braben's Raspberry Pi PC launched

[Greyth]

Ticks at 16mhz

[BF Bullpup]

What a misleading thread title... Braben didn't actually develop that PC; he is merely one of the lightning rods attracting attention to that little thing (gotta do something while he waits for more people to work on Elite 4 with. ) I have never heard of the Raspberry Pi until reading this thread but it sounds like a cool little piece of technology and a tempting one at US $35.

I guess Oolite in that machine is out of the question and the only way to play Elite without using an emulator would be the Linux port of Elite TNK... assuming it can be found online.

[aegidian]

A version of Oolite for RaspberryPi should really be a priority - in fact I'm quite tempted to come out of retirement to have a bash at it myself.

The key problem is that Oolite uses features of OpenGL that aren't available in OpenGLES - however OpenGLES has alternatives to them that wouldn't be an impossible task to implement. Since OpenGLES is used by many mobile devices this would also open up the chance of an iPad or Android Oolite, which would be a killer app.

I propose we begin an OpenGLES branch now and work on it a bit.

[Smivs]

A version of Oolite for RaspberryPi should really be a priority - in fact I'm quite tempted to come out of retirement to have a bash at it myself.

The key problem is that Oolite uses features of OpenGL that aren't available in OpenGLES - however OpenGLES has alternatives to them that wouldn't be an impossible task to implement. Since OpenGLES is used by many mobile devices this would also open up the chance of an iPad or Android Oolite, which would be a killer app.

I propose we begin an OpenGLES branch now and work on it a bit.

Awesome!

[maik]

If just it was but a joke...

[Gimi]

Ooooh.
Oolite permanently connected to my telly. I can't code, but if there is anything else, I'll help.

[Selezen]

Damn, my stomach actually turned a flip at the thought of a raspberry pi/phone version of Oolite...

I for one would contribute towards a payslip for Giles to do this...

If there's anything I can do to help, please let me know.

[JensAyton]

I propose we begin an OpenGLES branch now and work on it a bit.

I for one intend to stick to my plan of working on one major feature at a time, but it seems to me that the easiest path to GLES support would be to dump support for OpenGL 1.x and modernize the desktop renderer first. A GL 2.0/GLES renderer with some conditional cases would be much less of a headache than two completely different renderers. (This would drop support from Intel 900 series GPUs and those SIS abominations, but that was always on the table for post-1.76.)

Alternatively, it might be worth backporting the partially-complete Oolite 2 graphics code, which is already designed around modern GL pipelines with VBOs and no legacy attributes and all that jazz.

If you seriously intend to work on any flavour of this I’d appreciate knowing, since it will inform my work on materials.

[aegidian]

... it seems to me that the easiest path to GLES support would be to dump support for OpenGL 1.x and modernize the desktop renderer first. A GL 2.0/GLES renderer with some conditional cases would be much less of a headache than two completely different renderers. (This would drop support from Intel 900 series GPUs and those SIS abominations, but that was always on the table for post-1.76.)

The two aren't incompatible working solutions, but it would be easier for me (I suspect) to rewrite for GLES first and then use that experience to inform modernising the GL renderer.

Alternatively, it might be worth backporting the partially-complete Oolite 2 graphics code, which is already designed around modern GL pipelines with VBOs and no legacy attributes and all that jazz.

The biggest stumbling block for me ATM is not having a GLES renderer to test on. I can cure that by buying myself a Model B when they come out and in the meantime start on some serious GL -> GLES
research.

Yes, I'm pretty serious about wanting to do this. There are other major hurdles in the way of a PadDroid Oolite, fun ones like touch controls and very difficult ones like OXP support, so it's no small undertaking, but I think it's something I should try to do.

[DaddyHoggy]

Refuses to cheer just in case the sonic shockwave shatters the delicate bubble of happiness I see forming...

Elite running on a "Model B" (again) after all these years...

[maik]

That is the best piece of news I have heard today

[Selezen]

I'm so happy right now that I could burst.

[SandJ]

a raspberry pi version of Oolite...

Poolite?

[BF Bullpup]

That would be very cool. I can see the wireframe graphics Oolite uses being used for Raspberry and anything more than that would be icing on the cake!

[JensAyton]

The biggest stumbling block for me ATM is not having a GLES renderer to test on.

WebGL is a reasonable proxy. It has most of the same restrictions as ES 2.0 – specifically, everything must use vertex buffers and shaders – and a similar GLSL dialect. (One difference is that WebGL always supports floats and never fixed-point numbers, while ES requires fixed-point with float as an extension.)