Font verdict

Thank you, readers of this weblog, for your advice as to which typographic font I should use in a book.

There were 22 responses (including two via email).  If this were a simple vote, the result would be:

ITC New Baskerville 11-point: 8.
Optimum 11-point: 4.
Arial 12-point: 4.
Arial 11 point: 3.

(For a couple of reasons, those don’t add to 22.)

As soon as responses started coming, which was almost immediately, I was delighted by the thought that went into them.  It says something for civilization that people care about the aesthetics of the printed page.  Several respondents have relevant professional experience; but they were not the only ones with sensitivity to the shape and arrangement of letters.

Some opinions conflict, and when some say “Arial definitely!” and some “Definitely NO Arial!” I’m not going to be able to please everybody.  In fact I risk not pleasing anybody.  Despite your opinions and my delight at them – but as a result of thinking about them – I’ve decided on an option that I hadn’t even shown you, but will.

About fonts there are interesting sub-topics, brought out by one commenter or several.

– Aging eyes.  This must be why I used to assume the superiority of the classic typefaces enriched with serifs (the little thorns at the corners of letters), and believed them to be readable at greater distance, whereas recently I’ve been finding not only that they are less so but that even Optimum, non-serif but with variation of thick and thin, is less legible for me than the featureless Arial.

– Monitor versus paper.  For computers, a font generally has a separate file for the characters to be displayed on the screen.  Arial, super-plain, seems to have become the default font for the internet.  I imagine that details like serifs flicker in and out of visibility at slight differences in the resolution being used from moment to moment – the number of pixels available for the detail.  So a font that performs poorly on the screen may be fine in ink.

– Spacing within words.  If it’s too small, or not well calculated, text can look “jumbled.”  If spacing between words is too small, I am slowed by barely being able to find the breaks in sequences such as “minimum commemoration”; I complained of this to the editors of the Guardian‘s online edition (they haven’t heeded me).  But this may be a trouble of the screen and not the paper version.

– Spacing between lines (called “leading” because of the slivers of lead between lines of the old metal type).  We talk about “Optimum 10 on 11,” meaning 10 points (1/72-inches) from the line base to the tops of the capitals and 11 points between the line bases.  Text can seem larger even if only the leading is increased.  I couldn’t inflict too many more variations on you by giving other leadings, so when you said that one of the versions looked crowded, it may have been because the leading wasn’t proportionately increased.  There is “auto” leading, convenient because you don’t have to change it if you change the point size.  It seemed to be a larger percentage of the point size than I wanted, but now I’ve discovered that 11 points with auto leading is barely larger than 11 on 13, so I’ve gone over to it.

=     – Spacing between sentences (mentioned by Hector).  My private habit, in writing all sources files, is to give two spaces, so as to be able easily to break paragraphs into sentences and think about their order; however, publishing software may treat the double spaces as single.

– Width of the column of text.  As Geoffrey remarked, this causes more hyphenated words in some fonts than in others.

– Shiny (coated) paper versus matte.  Coating improves the saturation of colors, but I’ve now noticed that it makes black type paler.

– Other fonts or combinations that people suggested: Arial 10, Verdana, Georgia, Caslon Old Face (I have only Caslon Open Face, which has hollow letters).  Not mentioned was the common Times Roman, a traditional serif face but unusually compact.

– Similarity of I, l, 1. and of O, 0 – one of the drawbacks of the Roman- and Arabic-derived alphanumeric character set we are stuck with.  I agree with Kelly that Verdana better distinguishes 1 from l, but the serif fonts do quite well with 0 and O.

– Distinction between body text and captions.  I shall continue to keep it clear, by continuing to use Optimum bold as the caption style.  And, as some noticed, Optimum is what is used inside my illustrations.  There are also the styles to be chosen for headings and sub-headings.

– And in tables.  Courier is a common font of the kind that is fixed-width – also called monospace, or typewriter-like.  (It is slightly “Egyptian”: with square ugly serifs.)  I use it in tables and other things that may need to line up by columns, so as not to have to fiddle with tab calculations.  (I also use it in the plain-text source documents for everything I write, but that’s another subject.)  Unfortunately it is so spidery as to be, to me, unreadable unless all boldfaced.  I learned of another fixed-width font, Lucida Console, but haven’t yet tested it.

And there were other insights, such as Rick’s that large open type creates a 3-D effect – “floating” close up in contrast with the astronomical depths of the pictures.

A last and perhaps fundamental consideration is: literary versus technical writing.  Should refined fonts be kept for novels and skeletal ones for science?  I think the divide can be straddled.  If you’re compiling a textbook you may have to keep a lid on humor and imagination, but in a small self-published book about planets I’m free to indulge in digressions or poke fun at authorities.  What delighted me from the first response (Howard’s) and onward was not only the discernment but the possibility that a serif font such as Baskerville may be, as I had rather hoped, the way to go.

Once I noticed that one of the books I had (America and the Survivors of the Holocaust, by Leonard Dinnerstein) had such pleasing type that I wanted to stroke it.  The publisher didn’t say what the type was, but I got as similar to it as I could by finding modern versions of two famous old fonts, Garamond and Baskerville.

I oscillated between them – Baskerville 11 on 13 for my Troy Town Tale, Garamond 11 on 13 for Berenice’s Hair and 10 on 12 for The Under-Standing of Eclipses.

Though the Troy Town Tale looks “handsomely produced” (as Geoffrey is kind enough to say), Garamond is the more luscious, and I now see that, when they are side by side, it is to my eyes the more clear.  This is only partly because the paper on which TTT was printed is slightly yellower and shinier than that of Berenice.  Garamond seems larger, as indeed it is for the same point size – it is less compact laterally, so that it can flow to more lines.  Space doesn’t need to be saved so rigorously as it had to in the Astronomical Calendar.  It may be that Garamond keeps some of what we like about the non-serif Optimum.

I hesitate to impose another printing task, but this time it needs only one piece or paper, with short sample text in five fonts: Baskerville, Garamond, Times Roman, Arial, and Verdana, all at 11-point size with auto leading.

Verdana is included for showing how the fonts deal with 1, l, 0, O; Times to show its compactness.  The word “equator” is underlined because, being the last tthat sqieezes into all of the samples, it serves as a marker of compactness (there are zero words after it in Verdana, 4 in Garamond, 5 in Arial, 6 in Baskerville, and 8 in Times).

I’ve re-set my normal style for the book to ITC Garamond and am setting about dealing with the changes this makes throughout.  You still have a chance to scream “No!”