…in that neither of them exists online, but we love to talk about them in a kind of girlish and dreamy way.

To continue, then, our discussion of footnotes on Web pages:

Incorrect semantics

Gruber marks up a footnote reference like this:

<sup id='fnref1-2005-07-20'><a href="#fn1-2005-07-20">1</a></sup>

(Single or double quotes? British or American? Single quotes are more apt to be damaged by automatic smartening.)

I suppose you could use numerical footnote references if you had to, but they only make sense in documents with large numbers of footnotes (more than nine, I’d say). The more-typographic option is to use traditional footnote characters, namely asterisk *, then dagger (Option-t on a U.S.[-equivalent] Mac keyboard), then double dagger (Shift-Option-7). For references four through six, double them in the same order (** †† ‡‡), and for seven through nine, triple them in the same order (*** ††† ‡‡‡).

You don’t need the sup element, meaning you don’t need to style it to avoid disrupting your lead. Nor do you need to use Unicode superscript numerals, though that horse has already left the barn given our reliance on arrow characters.

More importantly, the anchor element can be its own named anchor. You don’t have to add an id to sup (or, if you don’t use sup, insert a superfluous span). Hence:

<a href="#fn1-2005-07-20" id="fn1-2005-07-20-a">1</a>

Use something consistent in your anchor reference, as by adding -a or -ref to the named destination (Cf. longdesc filenames). Adding characters to the beginning is a worse idea because it obscures the sort order (as when showing internal anchors in BBEdit).

In this method, we use the same element to send the reader somewhere else and bring him or her back. Curious – but kosher.


small is passably tolerable as a method to mark up footnote text. I may be the only standardista who uses big and small anyway.

My arrow is better than your arrow

A left-pointed arrow is, sadly, still incorrect.

Gruber now:

Only here must I disagree completely with Clark. I chose ↩ not because it implies moving left, but because it implies moving back. And a left-pointing arrow has meant “back” ever since Mosaic. I chose the hooked arrow, instead of, say, ← or ⇠, simply because it felt right.

(I know we sometimes have to approximate in semantic markup, but can we be less dreamy and go on something a bit firmer than “feel,” please? Like the names and directions of arrows, especially when the arrow has three directions? In any event, we have experience with well-meaning people trying to tart up something simple, as with incorrect staffnote characters in captioning.)

Gruber previously (double hyphen as failed en dash sic; emphasis added):

The problem with this is that a weblog post -- especially longish ones like those here at Daring Fireball -- is very different from a printed page in a book. When you, i.e. the reader, encounter a footnote reference in a book, you can glance down to the bottom of the current page to read the note, then glance back up to your original position in the text. You can do this because you have some short-term spatial memory of where the footnote reference was -- and all you have to do is get close, visually, and your eyes will quickly spot the superscript footnote reference. [Debatable! – Ed.]

But with a weblog post, there are no “pages.” It’s just one long article that scrolls down continuously. By placing the notes at the bottom of the article, they’re in some way more like endnotes than footnotes. Assuming there’s a hyperlink from the superscript footnote reference to the note itself, how do you get back to where you were when you’re done reading the note?

You go straight up. (No, Richard and Peter-Paul, indeed not “to top.” Nonetheless, there is no difference in markup.) You do not do a U-turn (two right turns to get you pointing left).

I distinctly remember when Netscape finally learned how to handle the browser back button on pages with named anchors. Before they fixed it, you’d go to the previous page. After they smartened up, you returned to the previous location on the current page that had sent you to the named anchor. Necessary to avoid confusion at the browser level, surely – but I doubt that anyone can really misunderstand uparrow ↑, while many will misinterpret the hooked left arrow. Frankly, it reminds me of an autoreverse function on a tape deck.

What do the experts say?

I asked John Hudson and the very severe Michael Everson what they thought, and they both replied inscrutably.

  • Everson: “It depends on the site. A straight arrow might lead right to the next room [indeed “room” – sic]. A curved arrow might suggest motion or turning.”
  • Hudson: “I was amused by his comment that ‘a left-pointing arrow has meant “back” ever since Mosaic,’ because in my sleepy state I thought for a moment he was referring to Moses. Surely for Moses the back arrow would have hooked right, not left. And there is part of the problem with the left arrow: it is culturally specific and, in a well-localised [Hebrew? Arabic?] UI, it would be a right arrow. The up arrow does seem to me to be less problematic in that regard at least.”

Should we really be using arrows to return the reader to a footnote reference? What’s wrong with the word “back”?

The foregoing posting appeared on Joe Clark’s personal Weblog on 2005.08.02 13:31. This presentation was designed for printing and omits components that make sense only onscreen. (If you are seeing this on a screen, then the page stylesheet was not loaded or not loaded properly.) The permanent link is:

(Values you enter are stored and may be published)



None. I quit.

Copyright © 2004–2024