[HECnet] Arabic / Unicode support on VMS?

Are these encoded as four separate characters or is the renderer meant to figure out the ligatures and render the correct form of MEEM when presented with it from memory? In both 8859-6 and unicode, there is one code point for the letter, and the renderer is to figure out which form belongs at that place in the text. It appears that xterm isn't really able to do this. It may also screw up the Unicode bi-directional algorithm, since I see the Arabic characters in L-R order. Not sure if one could sort this out by embedding RLE/PDF characters; my unicode entry setup is in a state of disarray due to upgrades. De