This popped up in #90. Relaxing the grammar of identifiers is a harder task but it may be worth doing in Fluent primarily for attributes of terms, and in the future—for private terms (i.e. terms which are not checked for presence against the source language). The asymmetry of Fluent makes it a natural choice to spell out these identifiers using the native language.
There's UAX31 which should be a good start and base for the research here.
This popped up in #90. Relaxing the grammar of identifiers is a harder task but it may be worth doing in Fluent primarily for attributes of terms, and in the future—for private terms (i.e. terms which are not checked for presence against the source language). The asymmetry of Fluent makes it a natural choice to spell out these identifiers using the native language.
There's UAX31 which should be a good start and base for the research here.