Contextual variants to facilitate readability

I want to know what people think about contextual variants, as seen in arabic, in latin scripts. I am trying to design a typeface that consists of different letterforms according to their context. In otherwords, an a would look different depending if it was interacting with an a straight edge letterform, a curved letterform etc... My hypothesis is that if we can eliminate some of the unnecessary positive and negative space between letters, we can facilitate a more comfortable reading experience. I was wondering if people had any insight? resources that I might consider looking into? Or any examples where this has already been done.

thanks!

George Thomas's picture

My thought is that it would not make for a smooth reading experience because people haven't been trained to read that way so it would take quite a bit of exposure -- and retraining -- to get them used to it. That shouldn't be taken to mean I think it would not work or that it shouldn't be done.

A good way to find out would be to actually make such a font or a small family, then perhaps have a study done by researchers or scientists to validate the merits of such an approach.

cerulean's picture

Many fonts have contextual alternates, mostly to avoid unsightly collisions. A recent success is Ernestine, which makes pairs like rv, rw, ry look better by trimming the left serif of the latter glyph so that the bodies can be moved closer together. But you need to keep in mind that the different spaces between letters can be a big part of what facilitates reading. Quickly differentiating letter shapes is important to reading, and just like ascenders and descenders help the reader recognize word shapes, so do things like the space in the lower right of the r.

ralf h.'s picture

an a would look different depending if it was interacting with an a straight edge letterform

But an a is read, because it looks like an a. How would different a shapes be an improvement? Connected scripts have different letter shapes to facilitate the connections itself, not to improve the legibility of the single letter (I think). I am not convinced that structurally different shapes as contextual alternatives or forced ligatures (people have tried this before) can be an improvement for Latin.
At least this is how I understand your approach. Small corrections for certain letter combinations are already getting more and more common in modern OpenType fonts.

But maybe you can create some sketches of that approach first? Would probably be easier to talk about it, if we can actually see and read it ourselves.

John Hudson's picture

This is an idea I have played around with, aiming to regularise the shape of smaller white areas between terminals and adjacent shapes. I sometimes find the angularity of triangular shapes that occur when oblique terminals are adjacent to vertical stems, or vertical terminals are adjacent to curves, visually distracting. I've no idea whether his has any impact on readability, and am not claiming this design approach as aiding reading. I just wanted to explore the potential for producing a more even pattern of white space between the letters. Obviously, this approach is primarily relevent for sans serif types, and perhaps more so for display settings than for text (where the particular white space areas may be too small to have significant impact).

hrant's picture

Shoshana, your idea is hard to grasp - I can imagine ways it can be quite revolutionary, but also ways it can be very banal. So I second Ralf that we need at least some minimal visuals.

hhp

Nick Shinn's picture

I agree with Ralf.
About Vesper, from my essay Engaging contextuality at I Love Typography:

Rob Keller’s Vesper (2009) has pushed the envelope furthest in using a large set of alternates to improve the fit of text type. By retracting or extending serifs, clashes and gaps are smoothed away. This is extremely practical in many character combinations, but as with kerning one wonders at what point the effect becomes counterproductive, and where the threshold is between smooth and bland. The question arises of whether such a process might dilute the personality of a typeface—or disguise its faults—and more profoundly, to what extent does the physiological process of reading type depend on consistency of letterform? Vesper raises these questions, but one theoretical typeface cannot provide the answer. Don’t ever expect a categoric decision, the best one can hope for is a consensus derived from practical experience with this and other types, yet to be designed, which implement contextuality as a readability strategy.

Being even more discriminating, I would say that if the principle of improved readability through profuse contextuality is to be proven useful, it’s not enough just to implement it and say problem solved (which has been the case with all those godawful-looking types for dyslexics), it has to actually be a type design, and designed brilliantly enough to be a game changer, a genre spawner. Like Jenson, Baskerville, Futura and Frutiger. That’s a tall order. Go for it.

Nick Shinn's picture

I agree with Ralf.
About Vesper, from my essay Engaging contextuality at I Love Typography:

Rob Keller’s Vesper (2009) has pushed the envelope furthest in using a large set of alternates to improve the fit of text type. By retracting or extending serifs, clashes and gaps are smoothed away. This is extremely practical in many character combinations, but as with kerning one wonders at what point the effect becomes counterproductive, and where the threshold is between smooth and bland. The question arises of whether such a process might dilute the personality of a typeface—or disguise its faults—and more profoundly, to what extent does the physiological process of reading type depend on consistency of letterform? Vesper raises these questions, but one theoretical typeface cannot provide the answer. Don’t ever expect a categoric decision, the best one can hope for is a consensus derived from practical experience with this and other types, yet to be designed, which implement contextuality as a readability strategy.

Being even more discriminating, I would say that if the principle of improved readability through profuse contextuality is to be proven useful, it’s not enough just to implement it and say problem solved (which has been the case with all those godawful-looking types for dyslexics), it has to actually be a type design, and designed brilliantly enough to be a game changer, a genre spawner. Like Jenson, Baskerville, Futura and Frutiger. That’s a tall order. Go for it.

Nick Shinn's picture

I agree with Ralf.
About Vesper, from my essay Engaging contextuality at I Love Typography:

Rob Keller’s Vesper (2009) has pushed the envelope furthest in using a large set of alternates to improve the fit of text type. By retracting or extending serifs, clashes and gaps are smoothed away. This is extremely practical in many character combinations, but as with kerning one wonders at what point the effect becomes counterproductive, and where the threshold is between smooth and bland. The question arises of whether such a process might dilute the personality of a typeface—or disguise its faults—and more profoundly, to what extent does the physiological process of reading type depend on consistency of letterform? Vesper raises these questions, but one theoretical typeface cannot provide the answer. Don’t ever expect a categoric decision, the best one can hope for is a consensus derived from practical experience with this and other types, yet to be designed, which implement contextuality as a readability strategy.

Being even more discriminating, I would say that if the principle of improved readability through profuse contextuality is to be proven useful, it’s not enough just to implement it and say problem solved (which has been the case with all those godawful-looking types for dyslexics), it has to actually be a type design, and designed brilliantly enough to be a game changer, a genre spawner. Like Jenson, Baskerville, Futura and Frutiger. That’s a tall order. Go for it!

quadibloc's picture

@Nick Shinn:
That’s a tall order. Go for it.

Oh, indeed. But to simply design a typeface in which this feature is present, and where it turns out to be unobtrusive, and at least not impair readability, is a first step that should not be dismissed as worthless.

The example shown of changing the shape of the letter "c" shows that the benefit to be expected from this is the same as that from good kerning - a more even color of the text.

People have lived with typographical methods that permitted essentially no kerning, from which we can conclude the effect of good kerning is subtle. Sometimes, doing it right is its own reward, and designing types as they should be designed won't have enough of an impact to bring one fame and fortune - or even significant recognition from one's peers.

Syndicate content Syndicate content