If the "6" and "8" look identical don't you have a more serious problem? :-/
hhp
I have default oldstyle figures and lining figures in a font. Due to its particular constructed nature, the glyphs for 6 and 8 are each identical in both styles. Can anyone think of a reason why I shouldn't leave out six.lnum and eight.lnum altogether, and omit their substitutions from the feature?
If the "6" and "8" look identical don't you have a more serious problem? :-/
hhp
The 6 does not look like the 8. Ha ha. No, six looks like six.lnum, and eight looks like eight.lnum, and I don't think that's outrageous.
Oh, OK! Duh. Sorry.
hhp
Just make components, instead of omitting them. It is always nice to have a clean logical structure in the font file.
With the caveat that I think Frode is right, no. You write the lnum and onum GSUB code. As long as everything is in the default number style, it should work. But why not own a copy of, say, InDesign and MS Word, so you can check such things out?
Odd that an eight (in particular) is the same. Usually the old-style 8 has to be condensed a bit more than the full-size lining, to fit with the other os numbers.
In the end though, what is saved? Which is at least one more reason Frode nailed it.
Can anyone think of a reason why I shouldn't leave out six.lnum and eight.lnum altogether, and omit their substitutions from the feature?
If shapes happen to be identical, a single occurrence in the font is enough.
A good question, in my opinion. It is adding redundant info to fonts that asks for a damn good reason, not omitting redundant info.
On the other hand, by enforcing the "elegance" of redundancy for the sake of consistency, you might for example save yourself trouble if you ever changed your mind about the shapes being the same.
hhp
Captured from commoncrawl.org on 27 Apr 2015. original URL