I'm looking for information on how the apostrophe "should" be kerned in the case of contractions and possessives, like don't, he'd, Pete's and James'. In particular, I'm interested to hear what people think of the very tight kerning in Minion Pro.
The most extreme combination that I have found in Minion Pro is e'd, which is set almost exactly as ed with the apostrophe added afterwards, that is, no extra space whatsoever. To me, this doesn't look very good: for instance, it almost turns he'd/she'd into hed/shed. I'd looks similar.
I've attached a small sample for reference (original font size is 10pt), keep in mind that I'm talking about running text here, not headings and so on. The sample shows metric and optical kerning in InDesign CS3. I've also included four lines set using LaTeX, showing how the different groupings look. (I couldn't do it in InDesign: it seems as if InDesign ignores zero-width spaces when it does metric kerning?) Please ignore that the InDesign text is grey and not black, a mistake on my part.
To me, InDesign's optical kerning looks better in the case of he'd, but not necessarily so in the case of quotation marks. An easy way to differentiate the two would be if different code points were used, for example U+02BC (MODIFIER LETTER APOSTROPHE) in he'd and U+2019 (RIGHT SINGLE QUOTATION MARK) in quotes. However, the Unicode Consortium thinks otherwise, stating that U+2019 should be used for he'd and so on, leaving U+02BC for glottal stops and such; see The Unicode Standard 5.0, section 6.2, "Apostrophes".
I'm not sure, but I don't think it is possible using OpenType and metric kerning to encode "conditional" kerning, where the apostrophe is kerned differently depending on whether or not it is a quote, as instructed by the software. Of course, it would be possible to add a'd, b'd, c'd, ..., to the kerning table, leaving the rest as kern-as-in-quote, but that would really bloat the kerning table (or the other way around, if all characters preceding a quote are known).
Obviously, this must have been a conscious design decision by Slimbach, so I really don't feel comfortable just changing it to something of my liking without a second opinion. Having said that, I'd be happy if you could tell me which of the four LaTeX samples (He'd, He{'d}, H{e'}d, He{'}d) you think looks best.
Perhaps the tight kerning is great, and I just have to learn how to read it...
Tried searching the archives, but I didn't find anything.






