How can we get small capitals to appear in our answer posts? There is no guidance that I can find anywhere.

Here's some that I've copied and pasted, but I can't actually get small caps to appear in the dialogue box myself.

2 : to put on the appearance of : sɪᴍᴜʟᴀᴛᴇ


It's possible to get a REASONABLE APPROXIMATION of small caps by enclosing a string of capitals in both <sub> and <sup> (in that order). As you see, the baselines don't line up exactly, although the size is quite a good match for the x-height of normal text. True small caps should be slightly wider than a simple optical reduction, though.

<sub><sup>REASONABLE APPROXIMATION</sup></sub>

This method is easy and will display something intelligible in browsers that can't handle the full Unicode set. If that doesn't appeal, then there's either tchrist's method, or using the Windows Character Map application to find, copy and paste real small caps. That app does not make it easy to find small capitals in fonts, though, and there's no guarantee that other systems will be able to display them (it depends on browser capabilities and the fonts installed).

  • My own aversion to the optical reductions, an aversion which I fully recognize to be a problem peculiar to me and which I do not assume is one shared by the general public at large, derives principally from the mockery made of the font’s stroke width, which should match that of the other lowercase letters at that size but does not. Of course you’re right about the baseline and letterfit, too; the baseline should be identical and the letterfit should be spaced differently than used for the corresponding full caps. There may also be x-height discrepancies when simulated, depending on the face. – tchrist Mod Nov 15 '16 at 13:56
  • 1
    Exactly. When I'm forced to use Word, whose small-caps are simply optical reductions, I generally increase the width to 115% to compensate for the stroke width. It's a reasonable compromise. Can't easily do that in a browser with the facilities we have here. – Andrew Leach Mod Nov 15 '16 at 14:04
  • 1
    That reasonable approximation seems awfully hard to read, though. Just a bit too small for my eyes. – J.R. Nov 17 '16 at 23:08

I’m delighted you asked...

...but as I am unclear on the clusivity of we, I shall answer not in the semantically ambiguous first person plural but in the far more concrete first person singular, explaining what I myself do. I merely run whatever lowercased text string through the following one-to-one transliteration set via a simple little ᴄʟɪ tool written in Perl, wherein I map each code point from the first set to the one that figures immediately below it:


This does not always produce the most appealing of results in our selected type face, but in the fullness of time this too shall pass.

This is currently less that idea, given its current incompleteness: there’s nothing for x or q, and s and f are so newly defined that support for them in your computer’s implementation of our selected Georgia face may be poor, triggering an ugly font-substitution algorithm to kick in.

Because these are actual code points not out-of-band formatting mark-up, this has the advantage of surviving reuse of the content which loses formatting. However, it has the disadvantage of being at the mercy of whoever cut the type, er, created the electronic font.

If the CSS style sheet for our site could be modified so that when someone used **bold** here it produced a span with this CSS style hint:

font-feature-settings: "smcp" 1

Then we could in the most trivial of ways employ the actual small capitals in our own font, allowing us to produce text like this via simple markdown(content credit to the ᴏᴇᴅ, typesetting my own):

example of small caps done right
                  click here to see properly rendered without compression artifacts

Isn’t that elegant? It’s how real dictionaries do things, you know, and quite useful in scholarly text. Alas, my ancient entreaties that this be done fall ever on ears deaf to the request and eyes blind to the aesthetics of type well set.

If I had my druthers, all inline **bold markup** on ᴇʟᴜ would magically work this way, with the font’s actual bold reserved for headers where it belongs, not used as inline weighting where specifying bold would map to native small capitals.

Plus it would give Professor Lawler’s weighting score another ᴄᴏɴᴛʀᴏʟ-ᴍᴇᴛᴀ-ᴄᴏᴋᴇ-ʙᴏᴛᴛʟᴇ bucky bit to dress up one’s postings with. :)

  • T'would be good, but for the moment AL's suggestion's a bit more user friendly. Specially for a technophobe like me. – Araucaria - Not here any more. Nov 15 '16 at 14:09
  • 2
    @Araucaria Beyond a doubt, which is why my original supplication counselled the use of <sup><sub>YOUR SMALL CAPS HERE</sub></sup> in the event that we were unable to get at the font’s “real” ones. That particular approach to emulation did seem to suffer somewhat when our custom CSS style pages changed back during the big changeover to a common network-wide system; I feel that approach may no longer look quite as good (well, “good”) as once it did. I know no perfect solution for all this. – tchrist Mod Nov 15 '16 at 14:54
  • So much code mark-up from tchrist :D – Helmar Nov 16 '16 at 19:31
  • @Helmar Because it's actual code, mirabile visu. – tchrist Mod Nov 16 '16 at 19:48
  • How does the funky unicode work with screen readers? And what about search engines? Unless it works perfectly we should not encourage the use of it. – curiousdannii Nov 17 '16 at 4:59
  • @curiousdannii You judge. – tchrist Mod Nov 17 '16 at 5:00
  • @tchrist So SE's search works okay. Google doesn't. (Compare). Chrome's in-page search doesn't like it either. But that still leaves my question of screen readers, which I can't test myself. – curiousdannii Nov 17 '16 at 5:04
  • @curiousdannii Google should know better. Meanwhile, do please feel free to lobby ᴛᴘᴛʙ to change our bold markdown to use the font’s actual small caps via CSS fiddling font-feature-settings: "smcp" 1;. – tchrist Mod Nov 17 '16 at 5:13
  • 1
    The sample layout page looks beautiful. Wow. – Sven Yargs Nov 17 '16 at 8:20

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .