I think your on to something here. What would be really useful is to have a very limited set (say 10 or less) of re-usable chunks of text (the two line pairs say).
Then you can put that re-usable chunks of text as variables (?) to the text effect say.
WHat I've found is that with say a "tune to" sign on the matrix ... would be nice to put it one place, then re-use like crazy. Then if you ended up needing to change it, then also change one place. I know you would need some way to regenerate the .xseq afterwards to reflect the text changes for affected .xseq files.