Ugly-Ass Math

Although I’m open to serendipity (or pragmatism) when it comes to choosing a thesis topic for next year, I’m leaning very heavily in the direction of working out a type family with closely integrated Greeks and math characters.

It’s a topic that’s been particularly prominent in my thoughts for years, since my job has involved lots (lots!) of work integrating math and other technical notation into text for engineering books. My hands have been largely tied when it comes to making real improvements, since the publisher has been so resistant to substantive changes, but it’s given me lots of time to analyze the problems:

  • Bad typeface choices, especially considering the poor quality of printing used, the increasing distribution via PDF of the books, and the tendency for pages to get photocopied again and again. I’ve certainly spent a lot of time trying to figure out exactly what design qualities a typeface needs to function under these conditions.

  • If you’re going to use a typeface for math, you need glyphs that will be completely legible at very small sizes, sometimes below the proportions used for normal superiors and inferiors. This would also apply to the Greeks and symbols.

  • Greek fonts present all kinds of problems. Even when you get a Greek font that works well enough on its own, the Latinization of the design tends to make it difficult to distinguish some of the individual characters (Α, Β, Κ, et al.) from the normal Latin ones around them. This happens a lot in math, where Greek characters are used in the midst of a dizzying array of other glyphs. The reader is often expected to distinguish Greeks either from their context, the slight design discrepancies between the Greek typeface and the regular text face (since they rarely match), or the use of artificial obliquing applied to the Greek characters. A total disaster.

  • Math fonts are a total mess. Since they’re always designed separately from text fonts, they don’t integrate in any way when it comes to proportions, stroke widths, overall color, etc. Plus, I have yet to encounter a family of math fonts that supports the full set of Unicode positions for math and technical characters. When they do, they often include a number of poorly documented alternate characters. (I’ve spent a lot of time lately mapping the characters in old PostScript math fonts to their Unicode positions. It’s awful.)

So, there are a few key elements that lend themselves to a meaty solution: a Latin face that’s clear and robust enough to withstand low quality display and reproduction, a really full assortment of glyphs to be used at small sizes, Greeks that are designed to be easily distinguished from the Latins instead of designed to match, and a full set of Unicode-ready math and technical symbols designed to integrate with the basic text.

I know: sounds like a huge undertaking, right? Maybe I can save it for a PhD instead.

The need for someone to do this well is really clear, though, if you look at what’s out there. The available options are an unsatisfying assortment of incomplete character sets and poor design. A consortium of publishers is working on the STIX Project, whose goal is to build a set of fonts that will include the full Unicode character sets for Latin, Greek, math, and symbols, but it doesn’t look like they’re going to release OpenType versions. Also, judging from the existing work from the company designing the fonts, the quality of the type design is still going to be problematic. Although I’m all for the work the STIX Project is doing (“preparation of a comprehensive set of fonts that serve the scientific and engineering community in the process from manuscript creation through final publication”), I cringe at the thought of their low-quality, free fonts being the only option out there (“the STIX fonts will be made available, under royalty-free license, to anyone, including publishers, software developers, scientists, students, and the general public”), since they will really result in even lower standards for technical publishing.