I put the statements in priority order for me.
1- libtheora optimized for sse/sse2 , actually encoder is too slow.
2- Two pass encoder , i think gets better quality doing two pass instead of one pass.
3- ffmpeg2theora should have a delay(+,-) for vorbis audio output adding silence or not.
4- Need a good visual editor like Avidemux that support ogg Theora input and ogg Theora output.
5- oggztools need a program to split ogg files by size or time , this program can cut the file for the last keyframe near to the time or size input.
6- oggzmerge needs an option to merge vorbis audio with delay(+,-) to mux with Theora video.