Like top-hb instead of just top for Hebrew.
In the case when the font contains several different scripts, each of which have an anchors named like top, bottom, etc. Even though the anchors have the same names, each script has its own marks for these anchors. It make sense for sharing the marks like between Latin and Cyrillic. But I’m not sure how it work between LTR and RTL scripts.
Does it mean that, in mark and m2mk features, Latin letters will have Latin marks and also all the marks from other scripts (like RTL)? I mean that Latin letters don’t need Hebrew marks or Arabic marks.
Thinking about this, I have some automation to rename marks like top to top-hb or top-ar, to make them script dependent only. Does it make a practical sense for the output? Like … will it reduce the GPOS table, or any other profit?