One of the first, and most successful, examples of corporate identity was the commissioning and use of a sans serif typeface from Edward Johnston by Frank Pick for the London Underground in 1913. Johnston was an highly original and influential type designer, who taught Eric Gill (and Gill Sans bears more than a passing resemblance to Johnston Sans). This typeface is ideal for purpose, which is for signage: it is easy to read station names and “Way Out” signs even in the poor light of the underground. The typeface was re-designed by the Japanese type designer Eiichi Kono, with extra symbols added, and is used for information posters by the Underground now. The redesign, in my opinion, keeps the spirit of the original very well. A splendid account of the history of this typeface can be found here on the London Reconnections blog.

In 2007, I was nominated by some students for one of the Teaching Prizes given by the Drapers’ Company. I was informed of the nomination by someone in the College registry, who, after congratulating me on the nomination and quoting what the students said about my teaching, went on to invite me to submit a teaching statement in 11-point Arial(sic), with widths of top, bottom and side margins specified to three significant figures (specifically, “2.54cm margins left and right and 3.17cm margins top and bottom” – I invite you to convert these to Imperial units!). I was sufficiently upset that I seriously considered ignoring the invitation. In the end, I wrote something, and set it in Palatino, with the correct margins – I do not have Arial available, not being a user of Microsoft products. What happened next is interesting, and perhaps I will tell it later; but suffice to say here that my statement was not rejected on the grounds of typeface.

Perhaps I would not be so lucky another time. I understand that the College now requires the use of Arial in all its documentation. The reason is an interesting example of how the tail wags the dog nowadays. Apparently some survey (I cannot give documentation since nobody has been able to point me to the source) showed that a certain group of dyslexic students are confused by serifs on letters. (This is still no good reason for using Arial rather than a better sans serif face, but let that pass.) I believe I would have no difficulty in pointing to research showing that serifs actually aid readability for the majority of us, especially in body text. Sans serif fonts are ideal for station names but turgid 40-page policy documents are slightly less indigestible in a serif font. Indeed, Times Roman is widely recognised as the most easily readable font.

Mathematics, as in so many things, is a special case. Mathematicians need a wide variety of symbols in their writing. It is not uncommon to use the same letter in several different cases with different meanings in the same document.

When I was a student, I spent some time reading papers of Richard Brauer. For him, 𝔊 (your browser probably can’t read that – mine can’t – it is Fraktur capital G) was a finite group; its order was italic lower case *g*, while italic capital *G* was a typical element of the group. Apart from the fact that I couldn’t tell the Fraktur capital letters apart (I still have trouble with this), I found this confusing because conventions have changed: now *G* is a group and *g* one of its elements, and its order, if we need it, is probably *n*. Worse is to be found by going back further. Sylow, in the paper proving the most important theorem about finite groups, used *n* for a typical prime number, and *p* for the part of the group order not divisible by *n*.

(By the way, I frequently write or lecture about graphs and groups. Both graph theorists and group theorists like to denote their objects of study by *G*, so I have to upset at least one camp. I reckon that, since “graph” is a Greek word and “group” a German, the logical approach would be to write Γ (Greek capital Gamma) for a graph and 𝔊 (Fraktur capital G) for a group. But I am afraid I am incompetent at writing Fraktur letters on the blackboard, so my groups remain as *G*.)

The point is that reading mathematics is hard enough, it is important that the typesetting should help you rather than distracting you!

In any discussion of typesetting, the mathematician in the company will always look smug. The reason is that, for about 40 years, we have been the beneficiaries of Donald Knuth, a mathematician turned computer scientist and pioneer of computerised typesetting, who devised the wonderful system `TeX` (pronounced “tech”: the program name τεχ (tau epsilon chi) is made up of the first three letters of the Greek word “techne”, which combines the senses of “technology” and “art”. For mathematical typesetting is an art. In the 1990s, the last time I had any experience with Microsoft Word, it allowed you to include a mathematical formula in a document by selecting symbols from a menu and placing them by trial and error until you had something looking approximately correct; then you would find that the printout looked quite different from what you had on the screen. `TeX`, on the other hand, took the rules evolved by hot-metal typesetters over many decades for spacing (a bit more space around a relation like = than around an operation like +, for example – I wish I knew how to do that nicely in HTML), shrinking size of subscripts, and so on, and built them in. The result meant that an ordinary mathematician could produce work comparable to that of a professional typesetter. Also the output was device-independent, to the limit of resolution of the device (screen, printer, typesetting machine).

Knuth also designed fonts to go with `TeX`, with the mathematical symbols carefully matched to the letter forms. His Computer Modern fonts were not to everyone’s taste (“modern” here is a technical term, referring to fonts such as Bodoni which were developed in the eighteenth century, lighter in weight than the earlier “old-style” fonts); but there is no doubt that his aim, of enabling us to produce beautiful books, especially books containing some mathematics, was realised. Now, it is completely straightforward to use, say, Times Roman or Palatino instead of Computer Modern (though the symbols are not integrated as successfully with the letters in either of these). Finally, he gave the copyright to the American Mathematical Society.

We need all possible clues, including familiarity with the typeface, consistent design of symbols, and serifs, to distinguish letters. I fear the day when the College tries to force us to set our exam papers in Arial. Hopefully I will have retired before this happens! Since we all produce course material using `TeX`, students get to know what proper printed mathematics looks like. Changing the system for the exam would seriously disadvantage everybody, even (I believe) the students it is designed to help.

Pingback: Excellent teaching? « Peter Cameron’s Blog

Well, I have succeeded in retiring before exam papers in Arial became compulsory …