However, the user interface part is - probably because of the way ChoX11 works (or better: has to work) - non-responsive. As we all know, non-responsiveness of the user interface is perceived as "slowness". Plain drawing speed on the screen is not too bad and certainly not the problem.
So overall I would suspect that the parsing/layouting/drawing part would run OK-ish on a StrongARM Risc PC - after all, there are people who used Webster XL on an ARM610. But without optimizations in the user interface part (e.g. keyboard entry, scrollbar handling, dialogue operation, font rendering), there is really no point in providing a Risc PC port.
I guess Peter's first priority is ensuring correct operation above fast operation of ChoX11, and I agree with this priority. It remains to be seen how much optimization can be done on the ChoX11 layer - I am not really qualified to comment on this.