Like Andy, I'm hesitant about changing the gc of NNBSP, because of backwards compatibility concerns.
I'm also starting to think that scoping the wb change to Mongolian may not be a bad thing. We might want to explore what it would look like, since it would preserve the maximum compatibility for current use of NNBSP with French and other languages. (The use of NNBSP in French, although not all that common, I suspect would swamp—in terms of frequency of usage—the use with Mongolian, simply because the amount of text worldwide in French is so much greater.) Context The proposed WB change is from XX to EX Old relevant props: WB ; EX ; ExtendNumLet WB ; LE ; ALetter WB ; XX ; Other Old rules with EX: WB13a (AHLetter | Numeric | Katakana | ExtendNumLet) × ExtendNumLet WB13b ExtendNumLet × (AHLetter | Numeric | Katakana) ==== Off of the top of my head, perhaps something like: We add: WB ; ML ; Mongolian_Letter WB ; NN ; NNBSP // maybe different name We change the contents of LE and XX to move characters to the two new value sets. Eg, MN gets http://unicode.org/cldr/utility/list-unicodeset.jsp?a= [:scx=/Mong/:]&[:wb=ALetter:] We change the "macro" AHLetter(ALetter | Hebrew_Letter | Mongolian_Letter) *At this point, all behaves the same; that is just a 'refactoring'.* Now we can modify the behavior for sequences with NN adjacent to ML. We add: WB13c Mongolian_Letter × NNBSP WB13d NNBSP × Mongolian_Letter *If* we want to also change behavior on the other side of the NNBSP, whenever the Mongolian_Letter and NNBSP occur in sequence, we add 2 additional rules (with the appropriate values for ..., like Numeric) WB13c Mongolian_Letter NNBSP × (...) WB13d (...) × NNBSP Mongolian_Letter