Perhaps this would have been already well known, but I happened to come
across "CSets: Supplemental Unicode Mapping Tables".
<http://crl.nmsu.edu/~mleisher/csets.html>
where, among others, you'll find mapping tables for Iran System, IRNA,
ISIRI 2900 and ISIRI 3342, which are not supported in Encode 1.97.
seemingly due to lack of documentation.
While ISIRI 3342 is a logical encoding, Iran System, IRNA and ISIRI
2900 are visual ones, so would need some algorithm in conversion into
them.
In conversion from them, I think isolated, initial, medial and final
forms should be replaced with a basic glyph omitting U+200C and U+200D.
For example, in Iran System, both
0x90 0x200C 0x0627 # ARABIC LETTER ALEF, isolated form
0x91 0x200D 0x0627 # ARABIC LETTER ALEF, final form
should be converted into just U+0627. Moreover, it would be necessary
to rearrange characters in reverse order per logical line, I guess.
Kino