Since deciding to make a career from type design, I’ve had many discussions with graphic designers that have led to the subject of their own type experiments. They mainly fit on the modular/geometric branch of type design. There is an endless fascination with alphabets composed of squares, circles, triangles or angles. This fascination led me through the same journey as I endlessly twisted and distorted each character to fit into one simple idea. Invariably this produced a range of exotic letterforms, some more successful than others.
The point of trying to create an entire alphabet from a few simple rules is a design challenge, problem-solving at its purest. Creating all the letters of our alphabet from a few simple shapes appeals to the modernist in us, reducing everything down to its functional minimum. With software such as FonstructFont Constructor, which allows the user to quickly assemble a font from a set of geometric elements – this approach is now easier than ever.
Our alphabet isn’t quite as simple as that. A purely geometric solution may look fine in a word or a handful of characters, but set in a block of text, really shows its faults. A typeface composed of strict geometric rules can lose those subtle details and relationships of white space and stroke that have developed over centuries. Characters with quirky details that looked great in isolation, snag the eye when repeated in text.
Attempting to apply exactly the same set of rules to each letter is similar to handing out the same size clothes to a random selection of adults. Some will have excess baggy sleeves, others will be skin tight, and some will barely squeeze over their heads. The pattern has to be adjusted for each character, without losing sight of the overall design. This may sound like admitting defeat, but this is simply about redefining the rules. As you make adjustments for new characters, this echoes back through the characters already designed.
Take an example of a font created from a simple set of circles and lines. This works great for a, b, c, d, e, but then throw in a v or z, or even an s, and you meet a dilemma. Should the letter be squeezed into the current template or the template be adjusted for the new letter? It’s best to start with a group used frequently within the English language such as a, d, e, i, n, p & s, then later throw in a diagonal such as a v or x to test the formula.
My argument isn’t against geometric or modular types. Despite their faults, they offer alternative solutions when used as illustration or when the content is short or relatively unimportant. But to make them a little more readable, there are a few tweaks that can make a strictly governed modular set of characters work better as a more consistent typeface.
Balance
This is an example of a typeface I created some years ago. This is based around a very strict grid of squares and circles. Many characters look quite presentable, but these characters look particularly top heavy. Despite both counters of the 8 being identical in size, optically the top looks bigger. The 5 has a squared off counter on the top half which exaggerates the white space on top – making it look ridiculously unstable.
Widths
By cutting and pasting modular elements it’s easy to make many characters the same width. This creates widely different white spaces in each character. Take the b and h for example, the squared off counter in the h makes it seem much large than the b’s.
The joins
Where strokes meet or cross the join is liable to ‘clog up’. A typical example above, shows a circle attached to a line to create a b. This creates a heavy area where the curve tries to pull away from the straight. By trimming a little from this area, optically it pushes the direction of the curve back in the right direction.
The ‘S’
The s is an awkward character. It’s a fine balance of two open counters that have to balance horizontally and vertically. The classic ‘cut and shut’ technique of pushing two semi-circles together gives that tell-tale kink in the middle.
Stroke widths
The horizontal and vertical strokes should not be the same thickness. If they are, the horizontal strokes will look heavier. The example about shows how a visually mono-linear typeface such as Futura, has subtle adjustments to the horizontal strokes, to make it appear even.
These examples are merely the tip of the issues you will discover when designing type. A strict set of rules at the beginning can produce some very interesting ideas, but they need to be flexible. This will not only to make your type work better, but also differentiate from those being churned out in Illustrator every day. The most valuable thing I’ve learnt is learning to trust my eye more than the grid.
Your Voice
What do you think? Fill in the form below to submit any discussion points you like to raise and we'll do our best to get a debate going, in the immortal words of Bob Hoskins, it's good to talk.
Very informative and useful post, thank you.
[...] its design style seems to exude über-English contemporary-ness. Whatever the case, his article on making geometric type work is worth reading, and worth recommending to all your graphic designer friends! (just like I’m [...]
Great article, Ian. As many designers, I had a few (very short) takes on modular types but these were rather unsuccessful. Now I know why.
I’d like to pick out one sentence which made me think: “[...] simple shapes appeals to the modernist in us, reducing everything down to its functional minimum”
I was wondering, what the functional minimum of a typeface is? It can’t be the reduction to geometric shapes. This was the case for modernist architecture (i.e. Bauhaus), where simple shapes were required in order to be easily mass produced.
These limitations are, and never have been applicable to typefaces, so using these principles, is purely a visual imitation.
Maybe Matthew Carter’s ‘Bell Centennial’ for AT&T is a good example of a typeface reduced to a functional minimum.
Ralf – I was really referring to modernist tendency to reduce everything to a minimum amount of detail, whether it’s furniture, buildings, or type. But you’re right, it’s not really ‘functional’ it’s merely the novelty of reducing our alphabet to geometric components.
A case where modular type has been necessary is electronics. Where characters had to be conveyed in a small number of pixels – there was no room for different widths, curves or contrast.
What is a functional minimum for a typeface anyway? Bell Centennial is a great example of a typeface designed for a particular purpose (i.e. small sizes and ink spread), but it has plenty of character which has nothing to do with function.
I wouldn’t say that Bell Centennial is a minimal typeface at all. Instead of being made of simplified forms, it’s actually quite detailed. It’s just that those details are optimized for minimal sizes.
Nice article, thank you!
One direction to go with minimum amount of detail would be this:
http://typographica.org/001145.php
Everything irrelevant on a point of defineing a character is dropped
[...] lot of thought goes into the design of fonts. Here’s a little write up describing design techniques that we probably all take for [...]
[...] article on Design Assembly. Since deciding to make a career from type design, I’ve had many discussions with graphic [...]
great arcticle, thank you. I tried a modulor typeface that i thought looked great using around 5 letters for a particuluar word, trying to expand that to encompass the entire alphabet did however create a dilemma. Whether to honour the purity of the initial idea or allow adjustments for practicality? I’m sorry to admit in the end i shelved the idea, and i think it’s fair to say that meant i gave up!
I know realise i was missing some of the subtleties required.
[...] Moore of Design Assembly has a great post on his blog about making geometric type work. Attempting to apply exactly the same set of rules to each letter is similar to handing out the [...]
Great notes! Very useful since i’m working on a geometric typeface project. And I agree that every designer should trust more in the eye rather than in the grid.
Thank you!
@Ralf Metzger : as far as I know (i’m not a graphic designer) the idea was that reducing the shapes would increase legibility. All the “superfluous stuff ” e.g. serifs and such, were hinder reading. I think I’ve read that this theory is somewhat outdated, because we read word shapes rather than individual letters, and the “superfluous” stuff can make the word image stronger. Something like that….
[...] of ‘geometric type’ check out the article that I linked to on the 11/13 post at Design Assembly. This will be an ongoing project of mine, so I made a new page here called ‘GTE’ which [...]
[...] A brief, though excellent little article, Making geometric type work: [...]
[...] dimension—the third—you might enjoy this list: A brief, though excellent little article, Making geometric type work: Attempting to apply exactly the same set of rules to each letter is similar to handing out the [...]
[...] Design Assembly » Making geometric type workgreat article on designing type that works, from scratch [...]
[...] A brief, though excellent little article, Making geometric type work: [...]
Hello webmaster
I would like to share with you a link to your site
write me here preonrelt@mail.ru
Wow, this article literally set a weeks worth of typo course straight in my head. Thanks very much!!!
[...] There’s also some clogging in the joins of the As. This might help you understand it a little more: Design Assembly
[...] There’s also some clogging in the joins of the As. This might help you understand it a little more: Design Assembly
An updated and expanded version of this post is now on Typographica: http://typographica.org/2010/on-typography/making-geometric-type-work/