So you are misunderstanding what I'm doing due to your admitted not
reading what I wrote at the top of this thread.  Please read before you
argue then.  :-P

I AM reading what you are writing, but it's obvious you are not
talking about what I'm talking about.  What you are saying implies
primitive 3D, as you actually pointed out in the quoted bits below.  I'm
not doing primitive 3D, as I pointed out to begin with.

On Sun, 12 Jan 2014 11:49:24 +0900 Carsten Haitzler (The Rasterman)
<[email protected]> wrote:

> second life is nor hard core 3d.
> it's primitive 3d. hard core 3d is very far from what you are doing.

Second life has progressed from primitive 3D.  When I go to great length
to to explain how what I am doing is more Skyrim than Second Life from
ten years ago, you could at least read it.  B-)

Specifically what I am doing is making SL type technology more hard
core.  As hardcore as Skyrim, that's why I mentioned Skyrim.  I also
want to make SL type technology more generic, and bust out of SL's
walled garden.

> hard core 3d has shaders on everything with those shaders being 100's
> or 1000's of lines of GLSL,

Yep SL has that.  I've not actually counted the lines, but there's a
big directory tree with hundreds of GLSL files, I suspect they come to
thousands of lines easily.

> often runtime generated GLSL from templates,

Nope not got that, but I want to add something like it.  Not hard to add
though.  It's just text manipulation before you pass the resulting GLSL
to the compiler.  Been there, done that.

> that provide specular lighting, bump/height maps, with
> multiple rendering passes for multiple shadow maps (self shadowing
> too) with filtering, bloom filters for glow, 

Yep, SL has all of those.

> and other camera effects, motion blur, 

Other camera effects yes, motion blur no.

> physics simulation with water, clouds (volumetric rendering) 

Yep, SL has those.  Well, no physics in the water, but there are water
waves, plus reflection and refraction.  Physics they got, SL uses
Havok, OpenSim uses a home grown physics engine, but is in the middle
of switching to Bullet.  Water physics has been simulated via the built
in SL scripting engine, I wrote one myself.

> and insanely high resolution textures where
> you're pushing a 1gb of texture memory far beyond its limits.

SL has a limit of 1024x1024 on textures, a limit set many years ago.  I
want to increase that.  They used to have a limit of 2048x2048, long
ago, but they cut it down further for their own reasons.  Apparently
there's still some ancient content with the higher resolution hidden in
their content servers that slipped past the cut off.  They also have an
artificial limit of half a GB of texture memory usage, again I'd like
to remove that limit.  In fact, I want to remove all of the SL imposed
artificial limits.

> if your engine is anything close to second life, your 3d is far from
> hard core. if it's close to what i just described, then you are
> impressive as you're in the ballpark of unreal, latest id tech
> engines etc. which are massive undertakings.

So yes, the current SL engine is close to what you describe.  Pretty
much the only thing missing is motion blur, but that's likely coz SL
didn't think to add it.  Likely it could be added relatively easily to
the current SL engine.

The entire first post on this thread was equating SL style virtual
worlds with Skyrim, this is why, they are both hard core 3D.  By your
definition they are, and by my definition they are.  I guess you have
not looked at Second Life or OpenSim recently.  Even if SL itself was
not that hard core, it's my goal to make it even more hard core than it
is.  That's the entire point of all of this.  B-)

The SL engine is crap, hobbled by lack of imagination, the need to
monetize every little thing, and a long history of bad development
decisions.  I want to replace it with a real engine that does all the
extra hard core stuff that SL doesn't do, and does things more
generically, and easily grows in directions SL either doesn't want to
go, or that they didn't think of.

And yes, a massive undertaking, which is what I have been saying all
along.  This is why I want to use some one else's open source engine,
much less work for me.  I just have to get this engine to cooperate
with EFL in a way that EFL developers and the engine developers are
mutually happy with.  I want to push what ever changes I have to make
to their engine upstream and hope they accept it.  That's why I'm really
reluctant to make massive changes to their source code.  There's enough
flexibility in their system that I might be able to, but I don't know
yet.  I also don't want major changes to EFL just to suit my whims.

I'll have a look at your suggestions for changing the Irrlicht engine,
and see if I can do that in a way that is acceptable by upstream.
Before I do that though, I'm going to see if I can get my playground
app to share the window with EFL and Irrlicht like it did last year
with EFL 1.7.  Coz that's all least a minimal change to Irrlicht that I
know they'll accept, since it pretty much brings Linux inline with
their existing Windows support.  I'm not so sure the Irrlicht
developers will accept your suggested EFL porting.  Maintaining my own
private fork of a major sub system is not an option, part of the reason
I want a real 3D graphics engine written by others is so that they do
all the hard work of updating it with new features and bug fixes.

Keep in mind please that my major goal is itself a massive undertaking
that might take me very many years.  There are many big parts to this
that I really would love to just become a systems integration problem
for me instead of having to write those big parts from scratch.  So
yes, I'll fight to keep things at a systems integration level instead
of a write or rewrite major systems level.  Coz I'd like to get major
parts of my goal up and at least limping along before the end of the
decade.  Hell, I want to get a basic "here's all the major parts working
together" test thing running by the end of this year.

In this case the major parts include EFL and Irrlicht.  Last years EFL
1.7 + Irrlicht 1.8 based "they share a window" thing fitted my "all
parts working together" sub goal.  I'm jinda pissed it bit rotted,
but that's life.  Deeper integration involving sharing GL contexts is
for the future, so I know I can go there when it comes time to tweak
performance.  Specifically your constant mentions of how much of a bitch
it is to have multiple contexts makes me worry about that more, and
makes me try to make sure I can do it, keeping in mind my future goals.

I'm between a rock and a hard place here, trying to find wriggle room.
There's a big lack of docs and examples for doing this sort of thing.
Once I get it working, and it stays stable, I'll document it all, and
push docs plus examples to our git.

If after twenty years I'm happy with the bolted together frankenstein
system I produce to suit my goal, then I'll take time to actually write
an EFL based 3D graphics engine in C, I promise.  B-)

-- 
A big old stinking pile of genius that no one wants
coz there are too many silver coated monkeys in the world.

Attachment: signature.asc
Description: PGP signature

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to