On 19 Feb 2010, at 21:58, erik quanstrom wrote:
In this model the client doesnt actually draw an image, but just
operates on an (changeable) vector graphic. Things like scaling,
resizing (even funny effects like deformed windows while moving)
all are done by the display server - the client just sees object
vectors in an continous 2D space.
that model expired in the 60s.
if you want to support scaling, you'll likely need to support
something
more complicated than vectors. (vectors scale trivially.)
once you branch beyond vectors, life gets a bit complicated.
and generally one jumps immediately to 3d.
I have some ideas for a 3D draw device where the client may build and
work with a tree of 2D and 3D objects. The plan was to make
"deepterm", a drop-in replacement for drawterm with additional
capability implemented with aid of the host's OpenGL. Beyond that my
ideas are fairly vague. I was hoping to work them out as I developed a
proof-of-concept server, but the project has been on ice for a couple
of months and I barely remember what I was doing.
