Teaching on a postgraduate course feels very much like a spiral: the annual repetition of projects, each a vehicle for a journey of education and discovery for the student, blurs into cyclical clouds of shapes, paragraphs, and personalities. There seems to be little opportunity for reflection across student cohorts, and yet it is only this process that improves the process from one year to the next. Having passed the tenth anniversary of the MA Typeface Design programme was as good an opportunity as any to reflect, and ILT's offer to publish the result an ideal environment to get some ideas out in the open. Although my perspective is unavoidably linked to the course at Reading, I think that the points I make have wider relevance.

Our students, both young and mature, often find themselves for the first time in an environment where research and rigorous discussion inform design practice. The strong focus on identifying user needs and designing within a rigorous methodology is often at odds with past experiences of design as a self-expressive enterprise: in other words, design with both feet on the ground, in response to real-world briefs. In addition, students are expected to immerse themselves in the literature of the field, and, as much as possible, contribute to the emerging discourse. (There are many more books and articles on typeface design than people generally think; some are not worth the paper they're printed on, but some are real gems.) I shouldn't need to argue that research, experimentation, and reflection on the design process lead not only to better designs, but better designers.

In recent years, two significant factors have started influencing attitudes to design. Firstly, as generations grow up using computers from primary school onwards, it is more difficult to identify the influence of the computer as a tool for making design decisions, rather than implementing specifications. Secondly, the trend in higher education to restructure courses as collections of discrete modules results in a compartmentalisation of students' skills and knowledge: it is becoming more difficult for the experience in one class to have an impact on the work done in another. (A third, less ubiquitous, factor would be the diminishing importance of manual skills in rendering and form-making in design Foundation and BA/BFA courses, a subject worthy of discussion in itself.)

So, repeating the caveat that these observations are strictly personal, I offer them in the hope they will prove interesting at least to the people setting up and running new courses in typeface design, and the many designers teaching themselves.

1. Design has memory (even if many designers don't)

Typography and typeface design are essentially founded on a four-way dialogue between the desire for identity and originality within each brief ("I want mine to be different, better, more beautiful"), the constraints of the type-making and type-setting technology, the characteristics of the rendering process (printing or illuminating), and the responses to similar conditions given by countless designers already, from centuries ago to this day. Typographic design never happens in a vacuum. A recent example is Emigre magazine: can its early period be seen without reference to the sea-change in type-making and typesetting tools of the mid-eighties? and is not its middle period a mark of emerging maturity and focusing, critically and selectively, on those conventions worth preserving in a digital domain? Emigre is important as a mirror to our responses to new conditions and opportunities, and cannot be fully appreciated just by looking at the issues. (Especially if you look at scaled-down images, rather than the poster-like original sizes!). At a more subtle level, the basic pattern of black and white, foreground and background, for "readable text" sizes has been pretty stable for centuries, and pretty impervious to stylistic treatments. Does not a type designer gain by studying how this pattern survives the rendering environments and the differentiation imposed by genre and style?

And yet, many designers have a very patchy knowledge of the history of typography and letterforms. More worryingly, students and designers alike have little opportunity to experience genre-defining objects in reality (imagine discussing a building looking only at the blueprints for building it, not walking up to it, and through its rooms). It is perhaps not surprising that the wide but shallow knowledge gained from online sources is dominant; there seems also to be little discrimination between sources that employ review and editorial mechanisms, and those that are open to wide, unchecked contributions. This shallow approach to reading and investigating results in a lack of coherent narratives, not only about how things happened, but also why. And how were similar design problems addressed under different design and production environments? What can artifacts tell us about how people made decisions in similar situations before? How did changing conditions give rise to new solutions? To paraphrase Goudy, the problem is not any more that the old-timers stole all the best ideas, but that the old ideas are in danger of being re-discovered from scratch. (Just look at the web designers rediscovering the basic principles of text typography and information design, as if these were newly-found disciplines.)

Michael Hochleitner's Ingeborg, an award-winning typeface that revisits Modern conventions with originality and humour.2. Design is iterative, and improved by dialogue

The process of typeface design is, in essence, a reductive refinement of ever smaller details. First ideas are just that: sketches that may offer starting points, but have to be followed by a clear methodology of structured changes, reviews, testing – and repetition of the whole process. The attention of the typeface designer must progress in ever decreasing scales of focus: from paragraph-level values on the overall density of a design, to the fundamental interplay of space and main strokes, to elements within a typeform that ensure consistency and homogeneity, and those that impart individuality and character. At the heart of this process is dialogue with the brief: what conditions of use are imposed on the new design, and what are the criteria to determine excellence in responding to the brief? (For example, how will the end users make value associations with the typeface?)

The wider the typeface family, the deeper the need to test conclusively, not only with documents that highlight the qualities of the typeface, but also with documents that approximate a wide range of possible uses. Even in cases of very tight briefs (as in the case of bespoke typefaces for corporate clients), the range of uses can be extremely broad. But good designers are also aware of the constraints of their testing environment. The misleading impression of transparency and fidelity that computer applications give, and the limitations of laser-printer output, obstruct trustworthy decisions. Designers must be aware of how looking at medium resolution printouts in dark toner on highly bleached paper can bias their decisions.

We are also seeing a gradual return to typeface design being a team enterprise, drawing on the expertise of a group rather than an individual. This, of course, is not new: typeface design in the hot-metal and phototype eras was very much a team product. But just as the digital, platform-independent formats enabled designers to function outside a heavy engineering world, so it enabled the explosion of character sets and families to unprecedented levels. The necessary skills and the sheer volume of work required for text typefaces have driven a growth of mid-size foundries, where people with complementary skills collaborate in a single product. The corollary is a rise in the need for documentation and explanation to a community of fellows. The short-lived "creative hermit" model is giving way to new models of work.

Eben Sorkin's Arrotino, a contemporary typeface with deep roots in fifteenth-century typography.3. Scale effects are not intuitive

The conventional curriculum for design education rarely tackles scales smaller than a postcard. More importantly, the compositional aspects of design tend to take precedence over details at the level of the paragraph, let alone the word. Typeforms for continuous reading are designed at fairly large sizes (on paper or, more usually, occupying most of a computer screen) but are experienced in much smaller sizes where their features have cumulative effects, weighted by the frequency with which specific combinations occur. These conditions arise in every text setting, be it for prose read forty centimetres away, or a sign viewed from a distance of tens of metres.

Of all the skills typeface designers need to develop, understanding how to make shapes at one scale behave a particular way in another scale is the most troublesome one. Imagining the difference that a small change in a single letter will have in a line or paragraph of typeset text is not an innate skill: it is entirely the result of practice. The best designers are the ones who will naturally ask "why does this paragraph look this way?" and try to connect the answer to specific design choices.

A common example of problems connected to scale effects arises whenever a student follows a writing tool too closely as a guide for designing typeforms: whereas the ductus (the movement of the stroke) and and the modulation can be preserved across scales without much difficulty, the details of stroke endings and joints cannot; typographic scales demand a sensitivity to optical effects that simply do not apply at writing scales. The best examples come from typefaces designed for the extremes of text scales: for telephone directories (famously by Ladislas Mandel and Matthew Carter), Agate sizes for listings, and early typefaces for screen rendering. The smaller the size (or the coarser the rendering resolution), the more the designer primarily separates blobs and bars of white space, and only secondarily deals with style and detail.

Alice Savoie's Capucine: an award-winning typeface in a fluid modulated style that successfully integrates Latin and Greek in magazines.4. Tools are concepts

Regardless of the scale effects mentioned above, there is a requirement to appreciate the link between typeface design and writing, and the tools used for writing. To be clear: I am not talking about calligraphy, but writing in the widest possible sense, from graffiti, a hasty 'back in five minutes' sign, to the most elaborate piece of public lettering. More than the specific forms of letters, the process of writing illuminates the patterns and combinations we are used to seeing, and gives insights into the balance of shapes and the space between them. The relationship of writing tools to the marks they make has been discussed in some depth (for the Latin script by Noordzij and Smeijers, most importantly), but the transformation of these marks through the computer much less so. (There are some texts, but mostly they focus on specific cases, rather than general principles; the notable exception is Richard Southall.)

And yet, since the early days of punchcutting, type-making involves a process of fracturing the typeforms, modularizing and looking for patterns. Later on, when the roles of designer and maker began to be distinguished (most emblematically with the Romain du Roi, like the Encyclopédie a true product of the Age of Reason) typeface design became programmatic, each typeface an instance of a class of objects, rooted in a theory of letter construction – however sensitive to human practice or aloof that may be. Later, the hot metal "pattern libraries" and the rubylith cutouts of shapes to be photographically scaled and distorted for phototype point to the same process, of abstracting the typographic shapes into elements that have little to do with the movements of a tool. As for the digital domain, deconstruction and repeatability remain key aspects of the design process.

To ensure a typeface built with fragmentary processes has internal consistency, the designer needs to develop a mental model of a tool that may follow the tracks of a writing tool, but may include mark-making and movement behaviours quite distinct from anything that is possible to render with a real writing tool. (Easy example: the parallelogram-like serifs of a slab, on a typeface with a pen-like modulation.) Such mental models for typemaking are increasingly important as type families expand into extremes of weight and width, where any relationship with a writing tool quickly evaporates. So, an invented tool that, for example, makes incised vertical strokes and pen-like bowls, can become the basis for a wide range of styles, ensuring consistency without the limitations of a specific tool; at the same time, because the model is agnostic of weight and width, it does not hinder the generation of large families with overall consistency but local richness. (Compare this approach with a wide family developed through extremes of multiple master outlines, where consistency relies on the details of typeforms having close correspondences.)

A small part of Jérémie Hornus' analysis of the Amharic script in preparation for developing his own successful typeface family.5. The Latin script is the odd one out

The demand for typefaces with extended character sets has been growing steadily for many years. OEM and branding typefaces are expected to cover more than one script, and often three or more. Beyond the obvious scripts of the wider European region (Cyrillic, Greek, and Latin), the interest has shifted strongly towards Arabic and the Indian scripts. But there are two key differences between the Latin typographic script, and pretty much everything else: firstly, that the type-making and typesetting equipment were developed for a simple alphabetic left-to-right model that would have to be adapted and extended to work with the complexities of the non-Latins. Although rectangular sorts will work sufficiently for the simple structure of western European languages, the model strains at the seams when the diacritics start multiplying, and pretty much collapses when the shapes people use do not fit in neat boxes, or change shape in ways that are not easy to describe algorithmically. No surprise that most non-Latin typesetting implementations make use of compromises and technical hacks to get the script to work. The second factor is that most non-Latin scripts did not experience the full profusion in styles that arises from a competitive publications market, as well as a culture of constant text production. (It's no surprise that the language of display typography first developed in nineteenth-century Britain, in parallel with the Industrial Revolution: urbanization, rising literacy, and trade in goods and services go hand in hand with the need for typographic richness and differentiation.)

Many students (indeed, many professionals) will ask 'Can a non-speaker design a script well for a language they do not read?' But a typeface arises in response to a brief, which by definition taps into wider design problems. For example, many of the conventions surrounding newspapers apply regardless of the market; the constraints on the typographic specification can be deduced from the general qualities of the script and the language (e.g. can you hyphenate? how long are the words and sentences? with what range of word lengths? what is the editorial practice in the region in terms of article structure, levels of hierarchy, and headline composition?). Having established the typographic environment, we can examine the written forms of the language, and the tools that have determined the key shapes. In this matter most scripts other than the Latin (and to some degree Cyrillic) maintain a very close relationship between writing and typographic forms. Writing exercises and a structural analysis of examples can help the designer develop a feel for the script, before reading the words. More importantly, in their non-Latin work, analysis of the script's structure and the relationship between mark-making tools and typeforms can help the designers to develop criteria for evaluating quality.

Typographic history is well populated with designers excelling in the design of scripts they could not read – indeed, the examples are so numerous that it would be difficult to choose. Encouraging students to address the complicated design problems inherent in non-Latin scripts is not only a way of enriching the global typographic environment, it is also a superb means of producing designers who can tackle a higher level of difficulty in any aspect of their design.

Fernando Mello's Frida: an award-winning typeface that redefined what is possible in Latin and Tamil typeface design.6. And finally…

The final lesson for students of typeface design is that a formal environment can teach the functional aspects of design, but can only help them at a distance to develop the aesthetic qualities of their typefaces. Especially when they are working in categories already heavily populated with typefaces, the distinctions between the simply good and the superb will be very refined. And when the consideration turns to originality, inventiveness, and how much a particular design causes us to rethink our responses to typeset text, then teachers have little input. The student, balancing between the deep knowledge of the specialist and the broad curiosity of the generalist, must develop, largely on their own, their capacity to be conscious of past and emerging idioms, to see their own work in the context of developing styles, and – most difficult of all – to identify how their own personal style can co-exist with the restrictions of utility and the conventions of genre.

About the author

Gerry Leonidas is a Senior Lecturer in Typography at the University of Reading (UK) and Programme Director of the MA in Typeface Design. He spends most of his time talking and writing about typeface and document design, and is frequently invited to speak, teach, and review work.