Dear Mats and Bill,
Many thanks for valuable comments.
Though I am very much convinced to go for MSR 2 approach, keeping in view large number of languages which may take years to complete the task. Furthermore, in case of non-availability of some code point in Unicode / MSR, our exercise will not give desired results.
However, I agree with the proposal to start with 2-3 large languages (excluding English) and based on the experience, we will proceed further.
In this case our principles will also be changed, as we have to look for all available codes in a particular language.
On Mon, 5 Jun 2017 at 9:33 pm, Bill Jouris
<bill.jouris@insidethestack.com> wrote: