Contextual raising of ¡ and ¿

guifa's picture

In my font, I have rather large (or, more specifically to this question, long and to the left) descenders. While I've figured most of the time if, for instance, a j comes right after a g (which would cause overlap), I'd just have an alternate j, I figured since Spanish has a typographic tradition both of a lowered ¡ (as a descender character) and a raised (as an ascender), I'd have it be lowered normally, since i prefer this style, but if a descender would cross over, to have it push up the more modern raised style. (incidentally, this would also provide a nice alternate glyph for people who prefer it raised). After ransacking my brain for an hour or two why I couldn't get it to work, I realised that neither TextEdit nor Pages nor any of the Cocoa apps seem to support chain substitution on latin fonts (which Font Forge mentions somewhere in the docs that some apps won't because of some technicality in the opentype format). While I could just do "ligatures" that would cause the substitution, eventually that could start to build up in combinations. Is there any more compatible way to do this?

(the image shows traditional placement on ¡ay!, then crossover on ¡jóder! and then the modernised variant ... no kerning for it ... to prevent crossover also on ¡jóder!)

AttachmentSize
context_placement_ayjoderjoder.png10.79 KB
Nick Shinn's picture

My current practice:

1. Default: solve the problem by kerning
2. Alternate (raised): include in "case" feature.

I suppose it could also be placed in a stylistic set, which is what some foundries are doing these days.

Don't you think it would be strange to have both versions in the same piece of typesetting? For instance, a succession of questions.

guifa's picture

Kerning was my first thought -- but in this font, a ¿y combination (which happens with relatively high frequency in Spanish) would be kerned so far apart as to be distracting.

I don't think it would be too weird, at least, certainly not more than having a full em of extra kerning or crossover. Whilst I've never seen them in the same document before, some major publishers like Cátedra are still a little confused as to which they use, with about half their books now using the raised, and half the lowered, although the rest of the font is the same, or barely noticeably different.

I actually forgot I was going to place the raised in a stylistic set, thanks for reminding me about that. I guess since the only two languages that use ¡¿ are Spanish and Galician I can plan for the only collisions (y, y-acute for old Spanish, j, and j-dotless for transcriptions old books). I guess I was hoping for a "more correct" solution in case it started coming up a lot more with some other characters later on in the design process.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

Nick Shinn's picture

There is related issue: when a "j" with such built-in kerning (i.e. negative left sidebearing) begins a line, it protrudes substantially into the margin.

hrant's picture

And the primary location of the "j" is as the first letter of a word.
But a strong tail on the "j" is of large benefit to readability.

hhp

Mark Simonson's picture

I think this is why you don't often see a long tail on the j in fonts. Same thing with the f, but less so if you can have f-ligatures.

hrant's picture

Well, the main old (metal) reason was to reduce the chance of breakage. In digital the main reason has been to reduce unsightly overlaps. But now with OT the reasons are mostly -if not completely- gone.

For the "¡j" I wouldn't raise the "¡" - way too much potential for confusion with the "i". I would instead give it a shorter descender when it occurs before a "j", and in this (exteme) case maybe not even kern it when OT support is lacking - the huge gap would be worse than the touching. Anyway unsophisticated typesetters are unlikely to mind unsophisticated typesetting!

hhp

guifa's picture

Mark: you bring up another good point:

Is there an OpenType method for defining glyph changes at the beginning or end of lines? (not that I would expect such a feature to be supported by any piece of software). There are some camps that prefer punctuation outside of the main block of text when it comes up on the margin. I figure that would generally be considered a program feature, but is it possible to encode it within the font (let me first say I'm not planning on doing it, just a theoretically example, à la encoding proper quotes substitution in font)

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

Mark Simonson's picture

There's the Optical Bounds feature (optb) but I don't think it's been implemented yet in any current applications.

charles ellertson's picture

It is interesting how different people come up with different solutions. My first inclination would be to make a ligature for gj, and to use an alternate "j" when it follows an inverted question mark. The "y" is a tougher case, but reminds me of some old metal fonts like Monotype Garamond italic, where you both have to make some extra ligatures, and an alternate "y" is helpful to boot -- sometimes to be used in making the ligature.

I think a reader far less likely to complain about -- or even notice -- alternate letter forms than to complain about letters that overlap.

In passing, I think you may have the sidebearings too tight on the j -- how does the word "Kilimanjaro" look?

guifa's picture

Incidentally I noticed even some pretty standard and quite refined fonts still do crossover on the ¡j ¿j combination (Times). I'll try out hrants solution and see how it looks. The annoying thing is that while these combinations will come up in no languages but Spanish and Galician (and more so the former), they come up all the time. Kilimanjaro looks okay to me, maybe a tad more space on the left but only a hairline, although it might be more I need less space on the left of the a.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

hrant's picture

Your spacing in "kilimanjaro" is all over the place dude!

hhp

guifa's picture

Let me post an updated version and we can talk about it in the critique thread ;-) I do need to get some people to opine on my numerals.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

charles ellertson's picture

Kilimanjaro is a tough word -- I usually test it. Also Egypt.

But your "a" needs work on the sidebearings, and the "i" is generally too tight -- "silliness" is a good word to test, esp. in combination with another word with some half rounds -- m's and n's -- Another test word would be one ending in "ing" (which comes up a lot in text), and really shows when the "i" doesn't have enough sidebearing space, particularly when "ing" follows a "t" or "r" or "c". Even so, "ng" usually needs a small kern.

I never visit the critique forum, so I'll put these here & stop now.

Syndicate content Syndicate content