We discussed Letter-fitting from two points of view: the relation between legibility and readability, the expected use of a typeface, identifying two general categories: Text and Display.
Together they may define a micro-typographical context, where a choice becomes a point in a two-dimensional diagram with axes: Text-Display and Legible-Readable bias.
The Text-Display axis has two extremes that correspond to two opposite and very different ways of typesetting. In so trying to include any possible one. On the Text side a loose and balanced setting, suitable for a novel. On the Display side a very compact interpenetrated one, like so-called “tight but not touching”, suitable for a logo.
The Legible-Readable bias axis has limited excursion, instead. First of all, this axis is not meant to describe features, but the preference between features that should coexist when they become mutually self-excluding. So that both readability and legibility may increase along their direction, but just to a certain extent. And eventually negatively beyond that.
The limited excursion of these axes means that choices that make more sense from a typographical point of view are centred on the chart.
This micro-typographical space may be more efficiently mapped with a different pair of (rotated) axes: Tight-Loose and Rhythmic-Textured. This way, a choice may be defined in terms of concepts before recognised as to form a «vocabulary of Letter-fitting»: Tightness, Rhythm and Texture. It’s a conceptual step beyond because it allows us to move from qualitative to potentially quantitative properties.
Each end of axes represents a sort of extreme approach. «A sort of» because near axes are not orthogonal, so their properties are not decoupled. «Extreme» because any real setting may be described by a point inside the chart, preferably in the central portion, just being a combination of extremes:
The remaining end of axes represent choices that make less sense typographically speaking. Eventually more sense from a merely graphical point of view.
This micro-typographical space’s two-dimensionality implies that one can set the same typeface in many different ways as long as there are stylistic choices to support them. But it doesn’t mean freedom of choice. As we have seen, when readability and legibility become more and more self-excluding, a compromise gets needed. The term compromise may seem to carry a negative connotation. But it’s not if we acknowledge that it’s a necessary consequence of glyphs’ interaction. And that this interaction is the matter of letter-fitting, in the end. The need to find the best compromise reduces that freedom, as it acts as a further constraint.
To get to a certain compromise among all the possible compromises, the one that may be defined the best, is a matter of strategy. Strategies can be praxes, rules, preferences, styles or any other choice. What matters is that the result gets pleasing to the eye, in the end.
A strategy consists of assigning costs to losses in legibility and readability verified during the search for equilibrium, in abstract terms. And then taking decisions that minimise those costs. A controlled way to move from an ideal – the mere equilibrium – to a real working setting tailored for a certain context. Critical combinations, the ones more prone to produce overlapping or huge interpenetrations, become the direct bench tests. In themselves and inside a longer series of glyphs. Especially those forming disfavoured asymmetric configurations and triggering potential conflicts.
The need for a strategy means that perfection is inherently impossible. But not even theoretically desirable: by definition, readers/observers would not notice perfect combinations anyway – whatever that means. They would notice excessively wrong ones instead—those going beyond a certain tolerance in perceiving equilibrium. As we have seen: balancing is not relevant in itself but as lack of unbalancing (visual noise). A strategy, then, becomes the way to exploit that tolerance finding its maximum range: all the combinations of glyphs should work as much as possible averagely well enough the same in all the potential cases. Like for slalom skiers seeking the ideal trajectory in a path determined by opposed and irregular constraints, knowing the risks they can take, conceiving controlled deviations from the straight line to maximise their average speed.