Hacker News Reader

Comment by kbenson | Hacker News Reader

What about the difference in use? Do you know of any work that examines whether there is a difference in the importance of attributes depending on whether it's for programming or reading?

I know different things are important to me when viewing programming text than reading. A simple, somewhat contrived example would be whether 'r' and 'n' together are easily discernible as 'rn' or 'm' (which is mostly a non-problem with mon-spaced fonts). When reading text, I can infer from context what the word most likely should be, but when programming, that's an error waiting to be found at some point.

To expand on this, different programming languages may stress different qualities of a font. From experience, I find when parenthesis look like curly braces in a font, it increases my frustration as it causes ambiguity where there should be none. This may be less of a problem in a language that uses fewer braces.

View on HN

You read code when you're programming. The only difference is the mix of characters you'll see near each other. General readability matters -- when is it OK for a regular font to be unclear about which character (a zero or 'O') you're looking at?

The 'rn' case is one of kerning, and is an issue independent of the mono- or proportional-ness of the font; it's one of kerning. You can have 'r' and 'n' at full width for the font, and they'll look very 'm'-ish. If you have 'rn' looking like 'm', it's just as awful when you're reading a novel as when you're reading code.

> You read code when you're programming. The only difference is the mix of characters you'll see near each other. General readability matters -- when is it OK for a regular font to be unclear about which character (a zero or 'O') you're looking at?

The difference in mix of characters reflects on the design decisions made for the font type. Target use is important. Some types are great for titles, some for text bodies, some for logos. It makes sense to me that code is another category.

Yes, readability is important, but it's not binary. The level and aspects of readability required for a novel body are not the same as for an article title.

Differentiating between zero 0 and uppercase O is critical for code (and work terminals, and perhaps data tables), but IMHO isn't interesting when designing for text bodies. Same goes for 'rn' and other issues that have ever annoyed only programmers.

> when is it OK for a regular font to be unclear about which character (a zero or 'O') you're looking at?

This is OK virtually 100% of the time, in that there are no circumstances where you'd be confused over which was meant. In contrast, I write 1 and l exactly the same way under most circumstances. That could be pretty annoying if, say, I was doing a geometry problem involving a line named l, since numbers frequently appear in math problems. In fact, I suspect that many other people have had that same problem, since high school geometry textbooks, while they still like to name lines l, use a cursive font for it.

It's very rare for a literary work to feature e.g. one character named Oscar and one named 0scar.

"The 'rn' case is one of kerning, and is an issue independent of the mono- or proportional-ness of the font; it's one of kerning."

No. Kerning applies to proportionally spaced fonts, not monospace fonts. See, e.g., http://en.wikipedia.org/wiki/Kerning Designing monospace fonts so that character combinations are more readable has nothing to do with kerning.

> You read code when you're programming.

Right, but not for comprehension the way you read prose, but for technical accuracy of an input to a mechanical system that will read it and which is intolerant of errors even when the meaning is clear from context. For human consumption of prose, fonts that can sometimes create ambiguities of individual characters that are easily resolved from context but which keep the overall shape of morphemes recognizable are mostly fine, for creation of content -- particular content to be read by a machine, they are not.

> General readability matters -- when is it OK for a regular font to be unclear about which character (a zero or 'O') you're looking at?

For readers, quite a bit, because for most purposes, the meaning is clear from context (same with 1/I/l), and because people with more than marginal literacy mostly are recognizing the visual shape of (approximately) whole morphemes, not individual characters.

> If you have 'rn' looking like 'm', it's just as awful when you're reading a novel as when you're reading code.

For the reasons stated above, that's not really true.

Reply on news.ycombinator.com