I don't know, I think you should make 2 separate fonts, but I guess it depends on whether it is a text face or a display face. Because, theoretically, you won't mix long and short descenders in a text block. But that kind of variation may be very interesting in display fonts.

The problem with using <calt> to this kind of thing is that you won't be able to include 'real' contextual alternates. Or, you can include contextual alts, but they'll only work with the alternate letters; and they'll always get activated with the alternate letters.

Thanks for your comments, Eduardo. I guess I technically COULD make two separate fonts, but I thought it would work better as an OT feature. The face I'm working on is a text face, not display. I guess my logic in justifying the "calt" feature is 1) i have no other contextual alternates that these descender variants would interfere with. 2) ideally these alternates would be truly contextual if features could rely on leading and switch to the short descenders when leading is decreased below a certain level, say below 120% of the point size.

Contextual alternates won't work for this because you can't define the context in OpenType terms (nearby glyphs).

Personally, I'd use stylistic sets. It's my firm belief that shoehorning glyphs into inappropriate features because of what apps happen to support this month is a bad idea. Stylistic sets are pretty much at the top of our list for next features to support.

Failing that, make separate fonts.

My two cents.

T

what apps happen to support this month

is adobe putting out updates that often? how have i been missing out on those???

but seriously, i understand your concerns. thanks for your 2 cents.

Paul,

put the alternates into "ss01" and into "salt".

Illustrator CS (all versions), InDesign CS ME and Photoshop CS ME all support "salt". The upcoming Microsoft WinFX architecture with support both "salt" and "ssXX", I'm sure other applications will add "ssXX" support as well.

Regards,
Adam

>> I guess I technically COULD make two separate fonts, but I thought it would work better as an OT feature. The face I'm working on is a text face, not display. I guess my logic in justifying the "calt" feature is 1) i have no other contextual alternates that these descender variants would interfere with. 2) ideally these alternates would be truly contextual if features could rely on leading and switch to the short descenders when leading is decreased below a certain level, say below 120% of the point size. >>

As a user of type, rather than a type designer, I

Charles knows better than I, but I do agree with him about usage. Since I am pretty sure you are talking about your Lanston Goudy Old Style, I would think that people are not likely to mix the regular and full descender versions in the same paragraph. So with two fonts they can just set two styles--such as for body text and footnotes--and go with it.