Edge: so actually, it's really gone as I see in R3 GUI that you need to manually write code for drawing edges. Anyone knows the reason of this change from R2? I always found face/edge very convenient in R2.
R2's edge may be convenient but is too simplistic also. You have just few presets available. In R3 GOB system, you would need to implemented it using special GOB or special DRAW block which would be appened/inserted to main DRAW block (but taht is something that is possible with small effort with current implementation).
What do you want to know about draw blocks in R3 GUI? You can have one or more of them, each for different state (focused/non-focused state, etc.). Some variables are bound so you can change colors globally and the block is composed before use, so it's possible to use some math.
You can pass a DRAW block to R2's edge when you need more complex layouts.
I think there was some talk to add something like EDGE to R3GUI, to simplify things like focus, but noone ever implemented it.
Thanks for the info. I will most probably support EDGE facets in Red/View.
I liked EDGE too, and images could be used for it too, though I never did more than play with that, I thought that was the original intended way to do more advanced edges, if not as eifficient as DRAW. I do remember Carl saying something about not like EDGE himself, so it went away.
I suppose using edge was somewhat redundant if you used an image with 'extend to make your shapes. Chris R-G often made very nice looking styles with R2, so worth getting his input as a designer as well.
R3-GUI was supposed to be mostly a draw based, no bitmaps ....
EDGE is fine for some widgets. If you have for example circular radio button, EDGE won't help you very much.
Vector is the way of the world for many things now, because of varying display densities and screen sizes, but EDGE can still get you a long way IMO. At least for squares like me. ;-)
Actually rendering bitmaps is stil the most efficient way and so is used on mobile devices. You just must do it in the way how GPU likes it, for example use texture atlases, multiple assets per resolution, etc. Drawing gui using CPU is a no-go on mobiles and OSes too... no wonder that Vista was so bad OS, when it had only CPU gui rendering. Using pure vectors on GPU is also not that good, if you don't need just rectangles... because of expensive anti-aliasing.
If Vector was "the way of the world", Flash would not be banned from most mobiles.
And although browsers did big improvements since last years with SVG rendering, it still sucks a lot with performance.
I have to agree on bitmaps. Also, I can't imagine designers ever spending more than a few moments on DRAW before they throw up their hands and go back to Photoshop, as it's just much easier to work out skins there.