Is there any chance to reorganise viewer code to relies on externaly provided library for rendering? Right now I have created my own VNC viewer based on OpenGL rendering API. I can connect to several VNC servers at same time and place their desktops on 3D... see pictures:


I did try to cleanup viewer code to achive such feature but now code too differ from original. Bacically, i changed viewer to deliver remote desktop as OpenGL texture. With additional code to handle mouse in 3d I build fully functional 3d VNC viewer. Because it uses hw acceleration scaling is "free" and viewer runs much faster than GDI version.
Now.. the question (for developers) is.. Do you want to spent some time to cleanup viewer code and use externaly provided rendering interface? Even more.. if you are intresting in this, I can join to dev team and handle this part of project.
Another succestion is to update server to use externally provided JPG encoder and viewer to use externaly provided JPG decoder. Intel's IJL is ~10x faster than JPEGLIB. User that have Intel IPP licence can build .dll/.so with JPEG codec and setup vnc serer and client to use it. It will speedup image transfer.