Hi Gordon, I'm aware of the sweeping changes, but I don't believe it requires a fundamental change to the way a scene graph can use it. The backend implementation details will be very different, and the public interface will need to evolve to better expose this, but the general concept of the scene graph and the majority of the class decomposition remains with similar granularity. Supporting OGL3.0 in a OSG 2.x will not require sweeping changes to the scene graph, it won't need to be a fundamentally different beast.
However, there is an opportunity in embracing OGL3.0 and starting from a clean slate and building a scene graph around it from the ground up. OGLES2.0+ could also fit in this this new scene graph. Thanks to programmability one could dispense with much of the scene graphs state classes, so it could be a really slimmed down scene graph. Much more of what the scene graph does will become part of the art path. Such a clean slate might also be worth addressing the issues of scripting at a lower level than we do at present, as shaders themselves are effectively scripted too so there could be a nice pairing. Robert. On 8/11/07, Gordon Tomlinson <[EMAIL PROTECTED]> wrote: > Hi Y'all > > Having attended the SIGGRAPH OpenGL Bof the first thing you have to realize > about OpenGL 3.0 is that it is NOT an extension to 2.x it's a very new beast > with some cross over but the changes to OGL3.0 are very very intrusive to > the way OSG or any Scenegraph works or even OpenGL program, and the > scenegraph that supports OGL3.0 will be a very different beast to what we > have today > > OGL 3.0 is not backwards compatible with previous versions, > > >From what I have seen and what I understand, IMO the OSG version that will > embrace OGl3.0 will be a very new Scenegraph, the changes are simply too > many and fundamental to the way things currently work. If you we don't > approach this in a new clean slate way then we will spend most of our time, > going down the road of this is how it used to work and we will mimic that > and spend most of our time trying to replace OGL2x in OGL3.0 > > Once the spec is finalized hopefully shortly and folks get a read and start > to see what was talked about at the BOF and how fundamentally different > OpenGL 3.0 is it will be much clearer on why any Scenegraph that supports > Opengl3.0 will be different > > Simple things like glbegin() and glEnd() are dead, no clients side array's > any more, only 2 draw Functions, many uniforms currently supported in > Shaders are gone, a new program stage placed between Vertex and Pixel > stages, many things made immutable > > Basically we have a new language to learn > > You can see many of the bullet points ar > http://www.khronos.org/library/detail/siggraph_2007_opengl_birds_of_a_feathe > r_bof_presentation/ > > > Just my 2 cents > > __________________________________________________________ > Gordon Tomlinson > > Email : [EMAIL PROTECTED] > YIM/AIM : gordon3dBrit > MSN IM : [EMAIL PROTECTED] > Website : www.vis-sim.com www.gordontomlinson.com > > __________________________________________________________ > "Self defence is not a function of learning tricks > but is a function of how quickly and intensely one > can arouse one's instinct for survival" > -Master Tambo Tetsura > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Robert > Osfield > Sent: Friday, August 10, 2007 5:51 AM > To: [email protected] > Subject: Re: [osg-users] OpenGL 3 announced > > On 8/10/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > > someone wanted to know when will osg support opengl 3? yes i know the > > specs aren't even finalized, but i want to give the guy an answer and > > not let him pass in a forum with foul arguments against osg, > > especially not at heise.de. ;) > > There is no way I can given a specific time to support 3.0, since it's not > finalised nor any implementations available. I can say that I'm a committed > supporter of OpenGL in general and the path the ARB are taking to OpenGL > 3.0, so I'm following developments closely with the intention of supporting > it. > > There are some major technical issues to resolve w.r.t how we go about > integration. In some ways a clean slate, OSG 3.0 mapping to OGL 3.0 would > be ideal, with the both general hardware, OSG and OGL all becoming ever more > multi-threaded, and with new object models that perhaps suit a different > means of encapsulation at the scene graph end. With a clean slate OSG > 3.0/OGL3.0 would be trim down the scene graph substantially too. Perhaps > also an opportunity to improve the scriptability of the scene graph too. > > A clean slate approach is not good for existing users, not good for getting > out version of the OSG with OGL3.0 in a timely manner. Right now we need to > support the latest OGL extensions, then look at how one might integrate > OGL3.0 features in a relative non intrusive way. The OSG has already coped > surprisingly with evolution from OGL1.1 to OGL2.1, so we can probably handle > OGL3.0 too given a willingness to refactor parts that need it. Maintaining > backwards compatibility with older versions of OGL is what will complicate > things, I see its part of the OSG's role to hide this complexity of managing > multiple OGL versions so I'd like to see a graceful way of handling the new > object model. > > Robert. > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
