Crimes against readers — bad typography

A Thought…

Fanaticism consists of redoubling your efforts when you have forgotten your aim.

— George Santayana

Crimes against readers — bad typography

Or, To Hell with Helvetica

Posted on 2017-Jan-06 at 15:39:48 by Phil

People in the computer world, especially those preparing documentation, seem to be in love with Sans-Serif fonts such as Helvetica. This is a very bad practice, and such fonts’ usage should be minimized.

Sans-Serif fonts^[1] are the bane of reading. I don’t know how much time I’ve wasted over the years trying to make sense of ambiguous characters because they were written in a Sans-Serif font (typeface) such as Helvetica. It is bad when a character printed on a page (or on a display screen) can be interpreted in two or more ways. Instead of zipping right through some text, I have to actually stop and puzzle out a word or symbol. That is not good for reading speed and comprehension, and I find it personally frustrating. And no, I’m not “slow” or dyslexic. I can read like the wind when a good typeface is used.

Say I’m reading along and encounter “lll”. I have to stop at this point. “What is that? Is it three hash marks? Is it a Roman numeral III? Is it an abbreviation for Illinois?” I have to look around at the context and try to figure out what they’re saying. It’s bad enough (that I had to stop in the first place) when I see Chicago just before it (ah, it’s Illinois), or I see something suggesting a Roman Numeral would go there (Army Corps following). It’s far worse when I’m looking at computer source code and those three strokes could be any combination of capital “I”s, lower case “els”, or “ones”. Oh, let’s not forget “logical OR” (pipe) symbols too! I can, and have, spent many minutes at a time puzzling out just what was written in a manual.

Serifs on letters were invented by the Romans to improve the look of negative relief chiseled lettering. Over the millenia, they have come to serve very useful purposes in helping to distinguish among different characters as well as in guiding the eye along a line of type. Why are modern authors and their cohorts in crime, text designers^[2] so anxious to make it hard to read printed (or displayed) text? Sure, it’s good to use different typefaces to distinguish among different kinds of things in technical writing, but please, don’t choose a Sans-Serif font for any of them!

If Sans-Serif fonts, such as Helvetica, are banned from general writing, should they be used for anything? Well, if you treat them as decorative typefaces, to be used for special effect in specially chosen places, there’s no harm in using them—with care! Unless it introduces ambiguity, section headings are generally an acceptable use, as are “dropped caps” at the beginning of certain section levels. Just as only a sadist would write a block of text in italics^[3], only that type would write a block of text in Sans-Serif. You would not use a decorative typeface such as Blackadder ITC font sample or Braggadocio font sample for large amounts of text, would you? It’s just as cruel to be subtle in torturing your reader as it is to go “over the top”. The smaller the chunk of text your reader is dealing with (book title, chapter headings, dropped caps), the fancier the typeface you can get away with. The more the set-off text is likely to disrupt the smooth reading process, the more restraint should be exercised.

Are you a Dolt?

A famous example of confusion induced by the use of Sans Serif fonts occurred years ago at Xerox PARC ^[4]. It seems that researchers were testing a new Graphical User Interface on “civilian” users. Everything went smoothly until the very end of each trial, when users would become very upset and refuse to continue. What happened? The last thing presented was a big button to click, labeled “Do It!”. Actually, the words were run together (DoIt!) and the font was Helvetica, where the capital “I” looked like a lowercase “el”. What these people saw was that the computer was calling them a dolt! (DoIt!).

One last thing before I step down from my soapbox and put on my nice jacket with the extra-extra long sleeves—it’s possible to go awry in other ways. Be careful with selecting a particular typeface to get a certain effect. Just last night, I was inspired to put these thoughts down on... phosphor by a frustrating experience reading an otherwise fine JavaScript manual^[5]. The text designer had chosen a Sans-Serif font for code samples that, while bad enough, had a fatal flaw. Opening (left) parentheses looked almost identical to capital C’s. I was looking at variables like “imgCt” and they looked like “img(t”. I was reading “image Count” as “image open-parenthesis sub-t” and wondering how they had overlooked the missing closing parenthesis. I initially dismissed it as an isolated typo, but soon came to realize that it was happening consistently. This was quite frustrating. In ordinary prose (say, a novel), this would not be a serious problem, but in technical material it is, because parentheses are used so heavily. If only the designer had taken care to see that the parentheses were distinct from other characters! Parentheses should be noticeably shallower than C’s, should extend below the baseline and as high as possible, and should be clearly distinct from curly braces {}. That’s my gripe with other manuals I’ve read—they use a typeface for code samples where the curly braces “{” are almost indistinguishable from parentheses “(”. This isn’t much of a problem in ordinary prose, but it’s a critical flaw in in programming text!

Posted 2018-May-13 at 10:00:05 by Phil
Last update on 2022-Mar-17 at 15:37:00 by Phil

Sort of related… https://caselaw.findlaw.com/mi-supreme-court/1608291.html arguing over typographic minutiae [facepalm].

TL;DR A petition signed by 200k people was rejected because it didn’t exactly meet the requirement that the heading text be 14 point bold.

Read it and weep: “We do not share Justice Mary Beth Kelly’s technical literacy: we lack the necessary knowledge of printing and, more important, the programming necessary to create computer fonts. Therefore, we respectfully dissent from her conclusion that plaintiff actually complied with the type size requirements in MCL 168.482(2).” Yet they decided that they could rule on a very technical dispute over print size…

Posted 2018-Oct-23 at 10:33:00 by Phil
Last update 2018-Oct-23 at 12:19:59 by Phil

Another font for the Hall of Shame: Georgia. While a nice looking serif typeface, most of the digits are 1ex in height, and full size digits are dropped (sometimes referred to as “old-fashioned numerals”). While acceptable for a mystery novel or a bodice-ripper, this is not good for technical material where single digits may appear, e.g., default '0' looking like default 'o'.

I was using Georgia for this website, and noticed some 0’s had turned into o’s. I couldn’t figure out for the life of me what was going on, until I happened to see a multidigit number, and realized the elegant digits in Georgia made single digits look like letters!

^[1] Despite myself, I will probably use “font” and “typeface” interchangeably. Technically, a typeface is a specific design, such as “Times Roman” or “Helvetica Oblique”. A type family is a collection of related typefaces, such as “Times” consisting of “Times Roman”, “Times Bold”, and “Times Italic”. A font is actually a typeface at a specific point size. In the old days, type came in predefined, fixed sizes, but today a given typeface can be scaled over a wide range of sizes, and is usually simply referred to as a “font”.
^[2] The people who pick typefaces, line lengths, and other stylings, not the people who design the typefaces themselves.
^[3] Although I have seen publication abstracts done this way...yecch!
^[4] Palo Alto Research Center
^[5] JavaScript for the World Wide Web, 3^rd ed., by Tom Negrino and Dori Smith, Peach Pit Press, 1999