[kern] Subtable breaks in FL: right, wrong, abusive?

Jean-Baptiste Levee's picture

Hello dear typophiles,
I'm working on a font that has 731 glyphs, 1275 "real" kerning pairs, all are class-based, (which, once expanded, generate some 27510 pairs).
When generating/updating kern feature, FL adds a subtable after the 1258th pair, which is kinda close to the end of the kern listing.
Considering the following:
1 I think I do not have *that* much kerns in comparison to other fonts
2 There's a discussion going on about FL possibly adding unnecessary subtables
3 removing this sole subtable does not prevent font from compiling and working once installed

I'd like your opinion about this. Could it be something wrong with my kerns or do I worry too much or is it perfectly normal and I should leave it?

All input appreciated :)

FLSM5.0.2 OSX.4.8 on PPCG5

Miss Tiffany's picture


Miguel Sousa's picture

If the features still compile after you remove the subtable break, then it's fine to remove it.

Jean-Baptiste Levee's picture

Thanks, Miguel.
I'm also worrying about the popping-up of this subtable, since the kern isn't that big. Could it be somehow related to "integrity" and consistency of my kerning logic?
(I mean, the kerning might be not as clean and logic as I'd like, but FL allow some deviance and adds a subtable or something like that)
Or, FontLab is just adding a subtable because the amount of kerns justifies it to its eyes…?

twardoch's picture

Jean Baptiste,

enough to say that the subtable break stuff is complicated and that there is currently no satisfactory automated mechanism for generating an optimal form of the "kern" feature. I hope this will improve in future versions of our applications.


Jean-Baptiste Levee's picture

Hope I made myself clear… This wasn't an attack on FL's consistency, as I'm not 100% sure of my kerning technical quality and logic as well.

Syndicate content Syndicate content