Should CASE switch standard-TF into LF?

If tabular figures are standard (assumed additional lining, oldstyle and tabular-oldstyle figures) Glyphs’s automatic feature generator writes
sub @TF by @LF;
into the case feature what I find inappropriate here.

I think case should only switch (tabular) oldstyle to (tabular) lining.
But it should not question the proportional or tabular quality.
sub @OSF by @LF;
sub @TOSF by @TF;

Anyone supporting that?

Good question, the spec says whatever ‘works better with all-capital sequences or sets of lining figures’. It is about shifting or substituting punctuation, but also other subs if they prove to be more fitting for all-cap setting. It doesn’t say anything about tabularity, so I tend to agree with you.

It can be argued though that proportionality may make more sense in all caps. But that would depend on the design too. And what if the user specifically picked tabular figures…? So I guess you’re right. :slight_smile:

Yes, definitely. CASE should not touch the tabular-proportional property of the glyphs.

What also matters here is the order of the features, I guess. If CASE comes first then TNUM would/could override it. What is the default order generated by Glyphs?

By default, case comes very late. So that shouldn’t be a problem.

Any other supporters or opposite views?

@GeorgSeifert,
Otherwise I’d like to turn this into a feature request! :slight_smile:
Thanks!

If case comes late then isn’t that exactly why it is a problem? Late means it overrides the others.

Late also means it is less likely to find the glyphs it needs anymore because they have been substituted by an earlier feature.