Since those can appear almost everywhere handling is likely to be problematic. This is just to keep track of the my discussion with gleki about this.
gleki: I forgot what was the solution for {ba'e}. Did we decide to strip it out for now? In future we might want to consider a chain of parsers where the second parser would deal with this {ba'e}. The same is applied to "indicators". They are more powerful than "free" (where as I remember we put it). In some places "free" can't be used but "indicators" can. In such cases we need to fallback to "the rule "indicators" modifies the preceding word only" rule.
(ilmentufa doesn't do that, btw, it simply thinks that "indicators" modifies the preceding word thus losing semantic hierarchy like in {lo ui klama ku} = {lo klama ku ui} equation)
Since those can appear almost everywhere handling is likely to be problematic. This is just to keep track of the my discussion with gleki about this.
gleki: I forgot what was the solution for {ba'e}. Did we decide to strip it out for now? In future we might want to consider a chain of parsers where the second parser would deal with this {ba'e}. The same is applied to "indicators". They are more powerful than "free" (where as I remember we put it). In some places "free" can't be used but "indicators" can. In such cases we need to fallback to "the rule "indicators" modifies the preceding word only" rule.
(ilmentufa doesn't do that, btw, it simply thinks that "indicators" modifies the preceding word thus losing semantic hierarchy like in {lo ui klama ku} = {lo klama ku ui} equation)