Know any good legibility tests for typefaces?

[crossgrove]

"We all contribute a paragraph of gobbledygook type-speak strung together into nearly incomprehensible sentences"

See above.

"The people who post questions like Ryan’s at Typophile aren’t paranoid, they want numbers that can be used to reassure clients (or themselves) about typeface choices."

Now we're in upside-down land. You're reiterating what I said days ago. WTF? I'm not talking about Ryan.

You are the one who is presupposing all kinds of unrealistic, irrelevant and detached scientific processes in the discussion, which was originally about assessing the legibility of a typeface or some typefaces, and not, as you say above, about applying readability standards to typefaces, or measuring random micro features of letters, or isolating single words in artificial reading scenarios, or giving all typefaces ratings or rankings of some kind. Seeing all that as inevitable, as you seem to, is what I'm calling paranoid. You seem to be saying that the testing done to date proves that there is nothing useful to be learned by additional testing, because it also seems that you're saying that there is no possibility that there will ever be tests that account for the variables that we've all identified...... Even though we're here virtually designing just such tests.

[Nick Shinn]

Seeing all that as inevitable, as you seem to, is what I’m calling paranoid.

If I'm paranoid, you're naive.

it also seems that you’re saying that there is no possibility that there will ever be tests that account for the variables that we’ve all identified...... Even though we’re here virtually designing just such tests.

Who's this "we"? I'm not designing tests. Are you working on any such projects as a typographic consultant, or are any other typographers?

No, I don't believe that it's possible to account for all, or even very many of the variables; typographic, demographic, and textual.

I agree with you that it is better to at least have some typographic intelligence in reading research, but that's not my bone of contention. It's the absurd notion that readability is an inherent quality of typefaces, which can be measured. IMO, readability does not begin to exist until the graphic designer, art director or typographer begins to work on a layout. For analogy (this is for William and Eben), as Maria Muldaur once put it, "It ain't the meat it's the motion".

[ebensorkin]

Nick, in fairness I think you have climbed down slightly from your original doom and gloom outlook on all science/font nexus'.

I’m not designing tests

Not designing them soup to nuts - no. But you have begun to state what might start to look like a reasonable set of parameters to be considered if you were attempting to examine one face vs another. You wroteAll in all, for an index of typeface readability to be meaningful and useful, a standards manual would have to be published in the form of a type specimen, showing the exact setting which had been tested.

When you say No, I don’t believe that it’s possible to account for all, or even very many of the variables; typographic, demographic, and textual. I can help but think that okay not at once... but in small steps, bit by bit a picture might emerge of what is going on & eventually testing methodologies can get better & better, and meaningfully include more & more variables ( typographic, demographic, and textual and others too perhaps). This might sound like I am building to the dramatic "all" of your statement but I am not. Instead it's the bit by bit aspect of building an increasingly cogent but always incomplete view over time that matters. That is why an absolute/authoritative list can't really be drawn up by scientists or anyone. Opinions we can always have though! And how much better to have them be increasingly informed opinions?

I don't think that anybody would disagree with your notion that type outside of the context of use cannot be considered fully/properly. That said I don't think that you have to test the full battery of possible uses & possible demographics etc ( which is infinite ) to notice meaningful things. If you did you could never do meaningful research.

So instead of saying "readability is an inherent quality of typefaces" which is a bit too 2-d for reality (I agree) you might well do better to admit "readability is a potential quality of typefaces which could be described albeit imperfectly with some decent scientific tests". No?

Dan you wrote “What typeface and Why, and How to effectively convince your client” Would you make the book you describe given a 1k advance? 5k? 10k? In other words; "what price self respect?". ;-)

[John Boardley]

Carl, I don’t worry about shampoo anymore.

Priceless :)

I think mention of ranking, rating, or establishing specific, literal metrics that define readability are all figments of paranoia.

But how else would this 'science' be used?

[Dan Gayle]

@Eben
"Would you make the book you describe given a 1k advance? 5k? 10k? In other words; “what price self respect?”. ;-)"

Dude, sign me up. I have school debt I need to kill.

[abattis]

dberlow wrote:
> Sheedy, J., Y.-C. Tai, et al. (2008). “
>
> Where, please? Must keep up to data.

Visit http://www.sciencedirect.com and search for author:sheedy :-)

--
Regards,
Dave

[John Hudson]

Eben: “readability is a potential quality of typefaces which could be described albeit imperfectly with some decent scientific tests”. No?

Either you can measure something or you can't, so 'describing something imperfectly' isn't really science, although it may be anecdotally useful (and the plural of anecdote is data). I think legibility is probably an inherent quality of a typeface, determined by the legibility of individual glyphs, as individual forms and as distinct from each other. But readability is a measure of a document: we read typesetting, not typefaces. Choice of typeface is one variable in the readability of a document. If one were able to isolate and make equivalent all the other variables, then perhaps one could measure the degree to which the typeface contributes to the overall readability of the document. But this would still be something other than a readability index for the typeface, because I'm pretty sure that if the other variables were different, even though equal for the purposes of comparison, the contribution of the typeface to readability would also vary. For example, typeface X might be determined to be more readable than typeface Y in a document set in 12pt, but typeface Y might be determined to be more readable than typeface X in a document set at 8pt. Isolating the other variables in a single document only indicates which typeface is more readable in the context of those variables, not which typeface is inherently more readable.

[dberlowgone]

"Visit http://www.sciencedirect.com and search for author:sheedy :-)"
Thanks!

"Within 5 years you will be proven wrong, is my bet."
How much is the bet?

Cheers!

[eliason]

Either you can measure something or you can’t, so ’describing something imperfectly’ isn’t really science

Seems to me like Eben's talking about a kind of "proxy" measurement. For example, there's no clear way to measure the effectiveness of my teaching directly, but one could make inferences about it from evaluating student evaluations of it. The assumption is that there is some reasonable (albeit "imperfect") correlation between the student evaluations and my unmeasurable (but not ipso facto nonexistent) teaching effectiveness.

[ebensorkin]

Describing something incompletely is might be said to be imperfect. Incompletely & precisely are not at odds. For instance I can measure very precisely the lumens a light bulb gives off. This doesn't mean I have perfectly described the light bulb. Or even the light. So how about "readability is a potential quality of typefaces which could be described albeit incompletely with some decent scientific tests”?

Isolating the other variables in a single document only indicates which typeface is more readable in the context of those variables, not which typeface is inherently more readable.

Yes. Absolutely. But, what if you do a wide range of tests? You could see trends. Or maybe not I suppose depending on the shape of reality. But I bet you would; and it would be the shape of those trends across a wide enough range of variables that would start to show what works better & in what contexts. Sort of like digital photo is made up of layers of RGB or CMYK color. One dot is pretty meaningless. 12 megapixels may not be.

So while I agree it is stupid or perhaps more charitably overly simple to say in a blanket way one face has better potential readability than the other, it would not be nearly so silly to describe a pattern of relatively greater or lesser success in a series of specific contexts and then interpret the data to help choose a type face for some project or purpose ( say way finding signage or a novel ) or to take conclusions away in order to try to design a better font for contexts tested.

John your post suggest that you doubt that data you would get back could be meaningfully interpreted. Is that correct? Also, is this formulation at odds with your legibility/readability distinction? I don't think it is but I am interested to hear what you would say.

If you can test cars & cameras and draw meaningful but specific conclusions you can take things a few steps further and test a typeface.

[John Hudson]

Eben: So while I agree it is stupid or perhaps more charitably overly simple to say in a blanket way one face has better potential readability than the other, it would not be nearly so silly to describe a pattern of relatively greater or lesser success in a series of specific contexts and then interpret the data to help choose a type face for some project or purpose ( say way finding signage or a novel ) or to take conclusions away in order to try to design a better font for contexts tested.

But this is what we've been doing for five hundred years, only we called it typography, not reading science. We have been building the data set, and we have these meat machines call typographers who interpret the data and develop new hypotheses and experiments based on that interpretation.

[ebensorkin]

John, of course I agree. :-)

We have to trust ourselves; especially as we start out with a new design.

On the other hand, I think it's not a bad idea to see what other meat machines say about comfort - especially if they will actually be using our design. That could be anecdotal or scientifically measured. A spirit of service need not belittle ourselves. Agreed?

Also, what do you make of my point about many small specific tests potentially building up an incomplete but maybe still very useful picture of a font's potential performance?

[John Hudson]

That could be anecdotal or scientifically measured.

Practically, I think the science, at least what I have seen to date, is itself anecdotal. That is, the only way to relate the science to design is to treat it as further anecdotal data, to combine with existing practices. It is, perhaps, a matter of how we interpret the data: in a hermeneutic of continuity, in which scientific knowledge of reading is interpreted in combination with typographic experience and practice, or in a hermeneutic of rupture, in which scientific knowledge of reading is interpreted as overthrowing and replacing typographic experience and practice.

Peter Enneson's comments in the legibility and comfort thread are worth reading again. Peter has probably read more of the actual science than the rest of us, with the exception of Kevin, and I think he has a good handle on the limitations of generalising from the data.

Also, what do you make of my point about many small specific tests potentially building up an incomplete but maybe still very useful picture of a font’s potential performance?

I think that's what typography is. If you can quantify the data, I suppose you have a mechanism to help non-typographers do typography or, rather, to do a kind of typography minus inspiration. That's not a bad thing, because a lot of text is being produced by non-typographers these days, and if it can be made more pleasant to read with some help from data collection and quantification that's welcome. But that's not the same thing as turning a non-typographer into a typographer.

[ebensorkin]

But that’s not the same thing as turning a non-typographer into a typographer. Agreed. Let alone a designer of type... But you seem to be saying that a built up picture made from many tests cannot help a font designer. Am I reading into what you are saying accurately? I think I am but I would like to be clear about it.

Practically, I think the science, at least what I have seen to date, is itself anecdotal. Would you expand on this? There are waaay too many ways of reading this. When you say I think that’s what typography is. it does make me laugh a bit because of course yes on some level that's right. We try things. We observe the result. We keep going. We gain experience. We start to generalize & so on. This has some properties in common with science to be sure. But saying that they are much of a muchness is simply not accurate. There are important differences. These differences are both limitations & strengths.

I think he has a good handle on the limitations of generalising from the data. Very probably yes. I don't feel qualified to say despite my respect. But that is maybe a side issue because it's today's data. We are talking about what might happen in the future. What might be possible.

John B, you said, Attach ’scientific’ to anything and it all to quickly evolves into that ugliest of beasts, dogma.

Must it? Is in inevitable? I don't think so at all. Maybe with people who don't know what Science is. But those people will do that with anything. Arts, Etiquette, Politics, Religion and yes, Science. It's that is the fault of rigid and small minds. It's not the fault of Science per se. It seems like you have a small axe to grid here. What's it all about?

Dogma is best avoided. That we agree on.

[eliason]

Dogmatism and skepticism are both, in a sense, absolute philosophies; one is certain of knowing, the other of not knowing. What philosophy should dissipate is certainty, whether of knowledge or ignorance.
-- Bertrand Russell

Not sure where that leaves us, but it seems apropos...

[John Boardley]

Also, what do you make of my point about many small specific tests potentially building up an incomplete but maybe still very useful picture of a font’s potential performance?

Yes, I have no problems with tests, whatever their form, whatever their foundation; what troubles me is how the results of those tests will be misused and misapplied. Attach 'scientific' to anything and it all too quickly evolves into that ugliest of beasts, dogma. However, I do think there's something in John's comment,

...because a lot of text is being produced by non-typographers these days, and if it can be made more pleasant to read with some help from data collection and quantification that’s welcome.

That sounds reasonable to me--tempered by the above caveat.

[John Boardley]

Eben
No, no axe to grind. I had a feeling that this thread would come to this: science vs whatever. Science is not the problem; attempting to scientize typography is the problem. Therefore, the only axe I'm grinding is the one that defends art from science. If 99% of all texts produced in the past 500 years were unreadable and illegible, then yes, bring on the scientists, but...

So, to reiterate, I am most certainly not anti-science; I just don't like this obsession with attempting to codify and rationalise and pigeon-hole everything--just employ a good typographer.

Must it? Is in inevitable? I don’t think so at all. Maybe with people who don’t know what Science is.

As the majority know little or nothing about science and its methods, then I fear that it is indeed inevitable; men in white coats selling shampoo and diet pills comes to mind yet again.

By all means test away, but doesn't science best begin with a theory--a testable falsifiable theory? It's the theory that you need to concentrate on; not the testing (that's step 2).

[Dan Gayle]

The fact is, science can study anything and produce useful results, because science does not start with a capital S. It is nothing more than observing cause and effect, and trying to make a clear distinction between the two.

Employing the scientific method isn't the opposite of fine craftsmanship. In fact, it might be an inherent, if seldom recognized or acknowledged quality of a good craftsman.
Whether or not Big Science, PhD can publish an article about readability and legibility has no impact on the careful observations of centuries of typographers.

And vice-versa.

[John Hudson]

Eben: But you seem to be saying that a built up picture made from many tests cannot help a font designer. Am I reading into what you are saying accurately?

No, I don't think I every said that. What I'm saying is that this gradually built, imperfect, incomplete picture, insofar as it is useful in making decisions about type design, has to be made up of a lot more than what specifically scientific methodologies provide. This is what I mean when I say that at the point of usefulness scientific data and other kinds of information are basically at par, i.e. I think it would be a mistake to privilege new scientific data in the context of a humanistic craft with a developed wisdom of several hundred years. Instead, one would need to interpret that data within a hermeneutic of continuity with that wisdom, which means putting the scientific data and anecdotal data and experience and inspiration -- and everything else that makes up typographic wisdom -- into the same bucket and giving it a good stir to see what floats to the top.

We are talking about what might happen in the future. What might be possible..

We are? I'm not big on speculation, so I'd rather look at where we are now and consider what might be likely, not what might be only possible. Of course, I'm always ready to revise my analysis....

The way I see it now: reading science is providing some good insight into how we read, but I'm looking for a lot more, especially with regard to readerability, i.e. insight into why we're so darned good at reading such a diversity of systems, styles, etc. and what the tolerances or limits are. I don't yet see a complete description of how we read, and that makes me very wary about trying to draw practical conclusions in terms of 'doing things differently' in type design, which is what we're all concerned about isn't it? If the science ends up confirming everything that we're already doing, that's interesting and reassuring, but it doesn't have a positive impact on the development of new type design. The tantalising notion is that the science might one day give us a clue how to make type better, and that's what makes it sexy or scary depending on one's proclivities. At the moment, I don't see anything in the science that is being done that suggests I should be doing anything different. Put another way, if you put today's science into the bucket with accumulated typographic wisdom it just gets absorbed: it isn't significant enough to contribute to the mix.

There are narrowly defined areas in which scientific study is presently useful, e.g. reading at low resolution, but the same limitations that make scientific study useful make type design correspondingly less significant. The science is more useful for rendering engine developers, screen manufacturers and document creators than for type designers.

[ebensorkin]

John H,

I think you have sunmed up the state of affairs extremely well. And Maybe now your Paragraph should be read & re-read as well. Specifically:

At the moment, I don’t see anything in the science that is being done that suggests I should be doing anything different. I haven't followed the science closely enough to suggest otherwise. And certainly I am not suggesting that you should.

which means putting the scientific data and anecdotal data and experience and inspiration — and everything else that makes up typographic wisdom — into the same bucket and giving it a good stir to see what floats to the top. Yes. Absolutely.

The science is more useful for rendering engine developers, screen manufacturers and document creators than for type designers. And maybe given time they can be useful to Typographers and type designers as well. We shall see.

Craig,

If what you mean by your quote is that we should remain curious then I certainly agree.

John B,

I agree that there is a part of our culture that would just assume as you put it attempt to codify and rationalise and pigeon-hole everything that is certainly anti-craft, anti-art and incidentally; anti-science. You cannot defend art from science because art does not need defending - from science. It needs defending from that culture of pigeon holes. And incidentally science needs protection from them too. And as a side note i am not sure typography, font making or even letter making is just an art. I think you are still so to speak setting aside intellectual homelands. This canton for Typographers that one for Scientists... Is that correct? If it is, I can't think that's healthy. Different ways of looking at things is a bit like different tools; my saw should have no cause to envy my hammer even if they both work on wood.

As the majority know little or nothing about science and its methods, then I fear that it is indeed inevitable; men in white coats selling shampoo and diet pills comes to mind yet again. It's time for you to stop holding the wrong party accountable. This is marketing.

It’s the theory that you need to concentrate on; I am busy working on this now. And I am keen on it's being tested in the fullness of time. And in the meantime if the culture of type wants to bury it's head in sand that won't help one bit.

the past 500 years Or even longer; 5000 years or more if you care to admit scribes and stone carvers to the group!

Dan,

It is nothing more I think it's a little more...

Whether or not Big Science, PhD can publish an article about readability and legibility has no impact on the careful observations of centuries of typographers. If by this you mean that science must remain irrelevant then I cannot agree. If you mean that science is not a threat but merely adds it's offering to their observations then I agree.

[enne_son]

”…has no impact…”?

Science does more than measure. In my reading I encounter terms like “salience,” “response bias,” and “cue value,” and because I am unhappy with the cognitive processing connotations of the term “word recognition” I propose “visual wordform resolution” to describe the perceptual processing component in reading.

My aim in doing this is to give type designers new and useful ways of seeing what they are doing when they manipulate proportions, contrast, weight and construction. That is, I want to expand the repertoire of constructs by which we channel our actions, or according to which we make our assessments. What I think we are doing (when we say we are improving legibility or enhancing readability) is manipulating cue-values, strengthening response bias, managing salience, improving perceptual discrimination affordances in such a way that the ease and automaticity of visual wordform resolution is enhanced. I want to expand — or diversify beyond the conventional, rather contentious ones we use — the repertoire of personal constructs we use to channel are actions or make our assessments. Metanoia

(I am also interest in “crowding” and “interfacilitation” because they seem to hold a key, but that’s anther story.)

I think of type design as both a fine art and an exact science. It is an exact science because, beyond the simple requirement of making a recognizable letter ”m,” manipulating salience and cue values is a game of tiny incremental adjustments in conformity with the laws of gestalt vision. The fine art is in knowing where to make the adjustments.

[William Berkson]

I'm back, so some responses:

>Bill, reading is not an ailment.

I didn't say it was. That's a red herring. But there are indeed reading ailments, known as dyslexia. Your view here seems to be that scientists should not research reading, because their work is never going to be of help. And that I think that any restriction on scientific research--except for ethical issues of ill treatment of human research subjects, etc.--is a bad idea. It is opposed to the growth of knowledge, which can help us all.

>How much is the bet?

David, I would say a good 'stakes' would be a good lunch for all the participants of this thread at a type conference five years from now. And say Eben could be the referee on whether there has been any significant advance on readability.

I personally doubt that future discoveries in reading are going to discover that classic type faces, printed on paper at the usual sizes, have something fundamentally wrong with them. However, I do think that the progress will be able to tell us more about the limits of readable type: what screen resolution is needed, what spacing becomes dysfunctional, and so on. So I do think they will be able to guide the creation of new type faces, though more than readability will always be involved.

>manipulating salience and cue values is a game of tiny incremental adjustments in conformity with the laws of gestalt vision.

I very much like Peter's phrase here "the laws of gestalt vision". We are able to "resolve", as he puts it, marks on paper into meaningful words, and I am sure there is law-like process in our brain that good theory can describe and good testing reveal in the future. My bet, as I said is five years. We'll see.

[Nick Shinn]

Your view here seems to be that scientists should not research reading, because their work is never going to be of help.

I'm all for the science of reading, and for incorporating typographic expertise into that research.
But I don't believe that readability is a scientific concept.
It's too soft. Which is to say that there are too many cultural variables: typographic, environmental, demographic and textual.
There is also the issue of what yardstick to use: speed is too trite, while comprehension and retention of anything other than simple grammar and facts is the preserve of the humanities, is it not?

Scientizing typography will do more harm than good: it's a job for designers, not technicians.

[ebensorkin]

But I don’t believe that readability is a scientific concept.

I don't believe that concepts are there to be balkanized; in other words they don't necessarily belong exclusively to one area of information or study rather than another.

I am unclear on what you mean by Scientizing but it does sound catastrophic. Even without know yet I suspect that scientizing typography isn't going to happen. I am fairly certain that's a straw dog.

Instead; if we are lucky we typographers & type makers might have a situation a bit more like another deeply human and sensorially rich activity: cooking. For the cook there are nutrtionists, agonomists, biologists, and culinary anthopologists. None of which stop me from cooking any way I like. On the other hand I do draw on their observations from time to time. How is food any less complex than Typography? Moreover there is no rush to displace cooks from their jobs by the dreaded white coats...

As you know I don't think that Science is all good all the time. It can be used for "bad stuff" the same way Typography can. So I am happy to point out the obvious counter-argument of all the heavily processed food made with the help of "food scientists". It's a hell of an counter example. Maybe this is the kind of thing you mean when you say Scientizing.

But the counter-counter argument might be increasing use of composting and also the food experiments of Ferran Adrià. Having eaten at the restaurant of one of his proteges, Jordi Butrón's Espai Sucre in Barcelona I can say that I was definitely getting the sense that in this case science was being used to enhance the humanities.

[William Berkson]

>Which is to say that there are too many cultural variables: typographic, environmental, demographic and textual.

Um, we are talking about type here. Yes, one can write unreadable prose, but we are talking about the contribution of type. Good experimentation separates the influence of different variables. The idea you seem to be assuming--that in principle the influence of different variables can't be separated--is just wrong. It isn't easy, but it it is done all the time in science. If you couldn't separate the influence of different variables there would be no successful scientific research, which is clearly not the case.

>while comprehension and retention of anything other than simple grammar and facts is the preserve of the humanities, is it not?

No, it's not, in so far as typography is also an influence. You can have a text book on a difficult subject, and a good typeface and good typography can help make it more readable. Of course, how well the writer writes is critical.

Eben's analogy to the influence of science on cooking is superb: science helps the cook, but isn't going to make a mediocre home cook a great chef. It can help both, though. For example, Julia Child turned to food scientist Shirley Corriher. And I cook a little better because of her science-based advice also.

Same with advances in readability of type and typography. It would help the amateur and the expert, but not turn one into the other, because so much more is involved.