From: Taura M <le...@fu...> - 2001-06-08 03:24:44
|
On Thu, 07 Jun 2001 16:36:58 -0600, Brian Paul said: > > How would a new glEnable/Disable for this particular attribute of > > polygon offsetting be benifit? It seems redundant since the proposal was > > to add an option to specify otherwise from the current GL standard > > of writing the offsetted depth. > > Maybe I don't understand what you want. The change in behaviour that > you want (as I understand it) can be expressed by a boolean. It > seems a lot simpler to implement it as a glEnable/Disable() rather > than invent a new function which only takes two values. You'll find > that adding a new enable/disable flag is a lot simpler than adding a > new API function (consider the dispatch code, display lists, etc). Oh I understand what yer trying to do now! I don't think this `sits right' from a design stand point but you obviously have more experiance with GL. I've written/designed a lot of libraries myself and the way I felt was different I guess. So a new glEnable(GL_POLYGON_OFFSET_WRITE_UNBIASED_DEPTH) it is then, where it is disabled by default? Does that sound okay? > > For example a wave simulation program, multiple levels of ocean currents > > would have to be generated and as waves originate farther from the beach > > their interpolation of height may be nearly identical at some points > > which makes distinction difficult without polygon offsetting. > > > > Also this situation could not be stored in a display list since the height > > values are constantly changing. > > I was assuming that your object/geometry data was stored in a > data structure that could easily be traversed and rendered twice > per frame (like many OpenGL apps). If your computation and rendering > code are intermingled then a two-pass algorithm may indeed be > impractical. You never explained this when I asked before. Sorry, i'll be in more depth next time, I didn't want to clutter your brain with our petty design issues. I know developers have enough in their brains already. ;) Yes we do have the drawing routine in a separate itteration function that we can call with ease and the GL rendering calls outside, the two way pass can be done... but this program has to run on a variety of platforms and not all have graphics acceleration. Speed is an issue so calling the function twice isn't really an option. -- -- Sincerely, ,"-_ \|/ -Capt. Taura M. , O=__ --X-- .__ ,_JNMNNEO=_ /|\ OMNOUMmnne. {OMMNNNEEEEOO=_ UOOOBIOOOEOMMn. 'LONMMMMNNEEEOOO=.__..,,.. UUOOEUUOOOOOOOObe '"=OMMMMWNEEEOOOOO,"=OEEEOO=,._ OOUUUIEEIOONNOIUbe. "7OMMMMNNNNNWWEEEEOOOOOO" "'. EEBNNMMMNWNWWEEIMMNe. __ 7EMMMNNNNNWWWEEEEEEEOO. " . NNMMMMWWWMMMWEINMMMNn "=BBEEEEMMMMMMMMNNNWWWEEOOOOO=._ . http://furry.ao.net/~learfox/ |