Know any good legibility tests for typefaces?

[Member ebe…]

Bill, don't get me wrong here, I think readability is a viable issue to explore through humanist efforts. I just don't think science will study it directly. I don't say science couldn't impact our understanding of it though. Nick still might. Despite his praise for Kevin, he hasn't said - yet.

Nick, While your list was a interesting one it also shows how very very little was been done. Not as you would have it how impotent science must therefore be. With so little science done I think it is presumptive in the extreme to suggest what the outcome might or might not be. Moreover much of what you referred to was not science. It was theory. But not scientific theory.

So rather than a horse riding in the right or wrong direction I would say you have at best, a tiny little mouse baby to ride. Don't ride it - it's too small to take your weight.

Nick, I assume you are not saying that readability as an idea has no legs whatsoever. Am I correct? You are also not saying that all faces have more or less equal readability. Correct? You are just saying that it is a bad idea to take two faces that are quite close in readability and make superficial and importantly reductionist observations about them. No? Who would disagree with this?

[Member Wil…]

>I just don’t think science will study it directly.

Huh? I guess I misunderstood you then.

Do you think that scientific tests like my proposed one above, comparing serif and sans fonts at given spacing, size and leading can't ever be informative about readability?

Nick, Eben, what's wrong with my specific proposal? Is it unscientific? Doesn't it test readability? Nick you complain about my being general, but you don't acknowledge or address my specific proposal.

Readability means generally speaking that text can be read with comfort, speed and comprehension. It may be that we have to disaggregate these three factors, but that's really no problem conceptually. These will simply be aspects of the general concept of readability, and they may have some complex relationship--which can be tested for.

The distinction between legibility--ability to distinguish letters readily from on another--and readability was I think to contend that there is more to readability than simple recognition of individual letters. By the way this has yet to be shown in any crucial experiment, so far as I know.

[Member ebe…]

I guess I misunderstood you then. Maybe. Go back & look at what Peter wrote about Readability. I thought it was very helpful.

And here is an analogy while I am at it. Peter, Kevin, feel free to ding this up if you like...

Suppose you wanted to study Cuteness scientifically. You can't say scientifically what Cute is right off the bat so how do you study it? Are you blocked forever? No way! What can yo do?

You could create an objective set of criteria describing perhaps not even the thing that was meant to be cute or not cute but the reactions you thought were indicative of a person coming into contact with the "Cute". Get the right reactions ( whatever you have decided they are ) and you know what is cute. Bring subject & target together and report what you get. What is you are looking at? Reactions.

Or you could expose people to stimuli designed by you and ask them to report what was cute and so test hypotheses about what is needed for say a human face, or dog, or cartoon to be cute. Now it's reported reactions.

Or you could track eye movements. Whatever.

No matter how much you test you will not describe the whole phenomena! But you could start to figure out things about the impact of eye to mouth & eye to nose ratios ( or whatever objective phenomena you choose) on what people are willing or not willing to report as cute. Not too bad really.

Doing this you yourself could report that in a rough and casual manner "I am studying cute". You wouldn't be lying exactly. On the other hand you wouldn't be completely accurate either. You would be fudging because on some deeper level it is the atomized stand-ins for cuteness, the objective* phenomena to gauge that are being studied.

But back to Readability. If comfort, speed and comprehension are your standards for "Readability" then you face some problems. You have to define comfort. You have to define how you calculate speed. You have to define how you gauge comprehension. This is atomizing**. It's not common sense. But then it isn't meant to be. It's meant to be a reproducible scientific test.

This is what I mean by indirect. This is what I man by "contribute to". It isn't that there cannot be a relationship. It's that it has to be a certain kind of relationship. Not a bad kind. Just a different kind.

So this not lying thing is tricky. Nick wants to spin things one way so he says simply " You cannot study readability with Science". It's sort of true. In a specific way. But he is leaving out some important stuff in a statement like that. You can say "You can study readability with Science" and you are doing the same thing. Leaving out something.

Obviously we can't get into infinite detail all day so we fudge like this about lots of stuff. And most of the time it's okay. But on a thread like this it makes for a very long thread and talk about hobby horses.

So it isn't so much that what you described before was unscientific perhaps; but that it was under-scientific.

Also, I & Nick ( I think...) are not saying these is no such thing as Readability. (Or if Nick is he may be the only one.)

I hope this is helpful.

* As objective as you can imperfectly make them.
** Atomizing sounds a bit like all the life has gone out of it but I don't think thats accurate either.

[Member enn…]

Bill, here is my reaction to your talk of factors, and to your question about whether your proposal tests readibility:

Suppose we agree readability simply means it can be read. That that which can be read, can be read with comfort, speed and understanding, indicates that comfort speed and understanding are affordance-related, but not affordance-constitutive. Specifically, your proposed test tests "how the sustainability of immersion with comprehension is modulated (if at all) in tightly line-spaced settings by the presence or abscence of serifs of a certain dimension at the terminations of strokes.” That, literally, is what it tests. The readability is assumed. It doesn't anymore require testing, but the experience of reading can be enhanced. In this sense there is more to readibility than just that it can be read. But let's be pointed and specific about what it tests

The results of your test will tell us about exactly the same thing: how the sustainability of immersion with comprehension is modulated (if at all) in tightly line-spaced settings by the presence or abscence of serifs of a certain dimension at the terminations of strokes. This tells us about how affordances can presumably be enhanced.

If we want to know why, we will have to entertain hypothesis about how perceptual processing in reading works, and part of the reason why, might very well be that there is more to perceptual processing in reading than simple recognition of individual letters.

Unlike comfort, speed and understanding, criterial cue-values and proper salience of role architectural parts and their evoked forms are affordance-constitutive. Criterial cue-values and proper salience are the very things that make it possible to make out the identity of the letters.

I guess we are all trying here to find a language that is comfortable for us all. I'm not sure though that we are advancing in leaps and bounds. A way of speaking about or in a certain domain is in part a social contract. Maybe I've ridden my own little pony long enough.

[Member Wil…]

>If comfort, speed and comprehension are your standards for “Readability” then you face some problems. You have to define comfort. You have to define how you calculate speed. You have to define how you gauge comprehension. This is atomizing**. It’s not common sense. But then it isn’t meant to be. It’s meant to be a reproducible scientific test.

Eben, I don't understand what is bothering you about 'atomizing'. When you have subcategories of a general category, the general category doesn't vanish. Because people can be 'atomized' into male and female do human beings become impossible to study scientifically? You can't study feet scientifically because there are male feet and female feet? No, you study both, the similarities and differences. But they're still feet, and you can study them with observation and experiment.

Similarly, because you can distinguish different aspects of readability, and characteristics of type that are more readable in different circumstances, it doesn't invalidate the concept of readability or make it only possible to study 'humanistically'. You just study the different aspects.

Comfort, speed, and comprehension can and have been already be measured scientifically.

I would get at "comfort" first of all by self-report. In Kevin's report at ATypi I believe he said Sheedy had a number of questions, such as complaints of eyes burning, blurring, etc. as discomfort. Then he has identifies physiological analogues he can measure that correlate with these. I would be looking more at mental fatigue, but there have been reports like that also--that people are tired of reading and don't want to go on, etc. Any study of perception and cognition is going to have to start with self-report, but that doesn't make it invalid.

Any such study is a process of refining, but it doesn't erase the concept of comfort, it just makes it more clear and precise for testing. Again, I don't see that problem. Comfort is not gone, and the whole thing is conceptually pretty straight forward. Finding good questions and good measures requires creativity, but scientists, including Kevin and his colleagues, have already started. And the tests are all described in a way that they can be reproduced. No impossibility.

On reading speed, this is words per minute. What is your problem about this? How do words per minute become only 'humanistic'? Doyouthinkicantmaketextthatslowsyoudown? Or speeds you up? Is that not objective and repeatable?

On comprehension, there are oodles of standardized tests of comprehension. Conceptually, they you just ask questions about the content, and how many answers are right or wrong. People will do better and worse, and under different conditions. Standardizing the difficulty of tests is tricky, but it's been done including in SAT reading tests.

So comfort, speed and comprehension can all be measured by objective, reproducible tests. The test I proposed is scientific--how good it is is another question.

Peter, you are saying that 'affordance' is a better focus for research. I would say that 'affordance' is a concept to help explain better and worse readability--and other aspects of reading. And it may be a very useful concept, in which case we should use it.

But readability shouldn't then become a taboo word. You design text that is more rather than less readable as one of your goals, I'm sure. And I am trying to do the same for a typeface.

Science attempts to explain the world as we experience it. In the process it refines the concepts, but it doesn't vanish our experience. Because Newton distinguished weight and mass doesn't mean that we no longer experience gravity. The refinements in terminology can help, but they are no reason to say that we can't use ordinary English concepts, or to say that gravity or readability can only be studied 'humanistically'. So I don't think there is any need to avoid the word 'readability', and *only* use technical terms like 'affordance'.

But in any case, I think we are in agreement that greater and lesser ease of reading is something that scientific research can shed light on.

[Member enn…]

Bill, there isn't a taboo. I don't *only* want to use my terms.

‘Legibility’ and ‘readibility’ have a well-established and perfectly acceptable use within the public sphere and the type domain. But they’re are poly-valent and multi-referential, and their usefulness in the public sphere and type domain is in a large part dependant on their not being tied too tightly to specific measures in distinct functional domains. This doesn’t make these judgements crude or inexact. The eyes of a capable typographer are more discriminating than the tests.

Sometimes science fails to honour this.

Yet, when I am rummaging around in distinct functional domains I do feel better served with new hand-crafted constructs. That’s all.

We need a transcendental critique of judgements of legibility and readability, a sustainable and typographically attuned theory of perceptual processing in reading, measures of performance that are geared to thresholds and ranges in specific contexts of use, relative to a range of measures, and a theory about how to put this information into practice.

[Member dbe…]

John: "For the typical typographic Latin script I’m pretty sure the atom is the letter..."
Boooooo-ooo-oo-o. See these things at the end of each sentence? . . . that is three atoms separated by two atoms, when . isn't an atom sitting beneath another atom or a molecule ! ? Ooopah, there's them now. When we read we swing along on things that are not letters. Else, the analogy fails any next step along the atomic analogy path, as we've seen, when applied to Latin text, as de-atomized by ClearType and displayed on lower screen resolutions.

Place a piece of paper, or, another window, over the lower 1/2 of the x-height of this Latin, and read on. There are few atoms you need that disappear, . being one of them. Then, think the 1/2 test again, the other way though — is the letter, horizontally as a unit, and as part of a text, more important, than a unit that begins at the mathematical center of one letter, and ends horizontally in the mathematical center of the next letter? These two units, a whole letter and it's space, and 1/2 of two adjacent letters and the space in between, are of equal value.

Kevlar: "..all [filtering] options that have been available in MS Reader for many years..."
Yes but you've seen as well as the rest of us, that it's not your reader running on the average Windows platform that makes a long a read a swing. How many exodus' need to be executed from the MS environment for higher resolution devices, before a study is done of why the tizzy over text so fizzy?

John: "Bringhurst wonders if people read differently when all literate people were writers as well as readers"
There is testament to important literate people who were only readers. But even then, 'History' of anything, is so funny! It makes one seem like an idiot until one states clearly, who one is, then what one's calling people, places and things, of a particular time or span of time, with time being measured in a particular way, along with defining everything else one wants to measure in such a History. We can't do it now...understand how people read, that is, so wondering back...is not completely unlike seeking a black dog on a moonless summer night...in glacier goggles...with a bag overhead and no bacon. But thanks, the unknowable history of readability, makes what some are doing today, more possible to imagine.

Cheers!

[Member eli…]

If Science ever wants to study metaphors, I hope it starts with this thread!

[Member Joh…]

David, the 1/2 test doesn't prove anything about how we normally read: it only proves that we're able to recognise letters from parts of letters. It doesn't tell us anything about the significance of that ability to reading: just because we can recognise letters from partial information doesn't mean that we use such partial information during reading, and if we do -- as seems likely to me* -- is it a direct contribution to word recognition or indirect via letter recognition. Being able to recognise a pair of letters from two halves and the space in between them doesn't mean that the basic unit of reading is not, still, the letters. Being able to recognise fragments of letters in sequence doesn't imply that we build up to word recognition through recognising half letter + space + half letter sequences; it implies only that we can recognise more than one letter at a time in a fixation, which everyone knows and no one contests.

* You skipped a key aspect of my (first) hypothesis, which is that while the letter may be the atom, we make free use of both subatomic particles and molecules as we need during reading. But the atom sits in the middle and is what is built up to and built up from. I think the letter plays that role in typical typographic Latin text.

But I don't think we should generalise from such text. If I consider all the other writing systems I know about, it seems to me that the atom is what we make it, when we write or set type or create text in any other way. There are too many writing systems in which the individual identity of the letter is subsumed into a graphical molecule for the letter to be considered always the atomic unit. I think we make it the atomic unit by disconnecting it from adjacent letters and giving it a static form: we emphasise its atomicity.

[Member Nic…]

the atom is what we make it,

So, in reading studies that focus on types with discrete letter forms, the outcome of letters-as-atoms would be expected.

writing systems in which the individual identity of the letter is subsumed into a graphical molecule

Including Latin script script, i.e. cursive writing/fonts.

[Member dbe…]

John, I don't think anything from here is going to seem likely to you for long times, and then, poof.
"...the 1/2 test doesn’t prove anything about how we normally read."
"...it only proves that we’re able to recognize letters from parts of letters."
We’re able to read words and ideas from parts of letters, skipper, not letters.

"...we make free use of both subatomic particles and molecules as we need during reading..."
Agreed. But then the reader, again, by making use of less or more than a letter, is not using your atom of reading. You need a typoscope? There is another proof that employs 1, I and l, shows in a, 'total' sans. There is yet a much longer proof that shambles through written vs. mechanized Greek. But I don't think that's gonna persuade, here. :)

Cheers!

[Member ebe…]

Bill, I think you are glossing over what is an essential difference. The objective neural reaction to smelling saffron isn't the smell. The molecules involved in sensing saffron are not the smell. The smell is a subjective* experience. This distinction isn't the same thing as the specific derived from the general. Obviously these things are not unrelated. Far from it. The smell is not the main thing from which descends the specifics. There isn't a hierarchy. These things are aspects. Aspects with differences.

I admit a smell is not a good pure analogue to reading. Reading is more complicated. But it has aspects too. Some aspects can be studied well with science. Science is in effect a good tool. And some of them - not so much.

* A reportable experience. But it's not the experience collected then. It's the report of the experience.

[Member Wil…]

>It’s the report of the experience.

Science as a social enterprise--involving many researchers--basically deals with reports of experience, or "observation reports" and tries to explain them. This is true whether you are looking at the position of a star or reading or smelling saffron. Reports of being tired reading are perfectly acceptable data for scientific explanation. The explanations may be complex, but there is no bar to scientific method here, if that is what you are arguing. For example there is food science, and that involves hard science, such as chemistry, and tasting and reports of tasting.

[Member ebe…]

but there is no bar to scientific method here We don't disagree about this. My point is separate.

[Member Wil…]

Eben, if you are saying there are limits to scientific explanation, I certainly agree.

I just don't think that there is any inherent barrier to a deeper understanding what makes type more and less readable, and that there is such a barrier seems to be Nick's position.

I don't expect a complete theory, especially of the aesthetic aspects, but I do think we can get a lot more insight than we have so far into what causes fatigue, and what promotes the combination of good speed and good comprehension under different circumstances.

[Member Joh…]

David: But then the reader, again, by making use of less or more than a letter, is not using your atom of reading.

Unless the information gained from what I characterised as subatomic and molecular levels were used precisely for the purpose of identifying letters and, hence, recognising words. That seems to me at least as viable a model as proceeding directly to word recognition from the sub-word and/or inter-word information.

Let's say that the atom is, by definition, what you need to be able to identify in order to recognise the larger semantic building block, the word. [Furthering the metaphor: words are the complex molecules, the DNA of text.] That's the main value of the general hypothesis, and I'm specifically questioning what constitutes the atom in different scripts, and suggesting that it might vary even within a script depending on the writing style: i.e. that how we read is not a generalised process independent of what we read.

[Member ebe…]

Unless the information gained from what I characterised as subatomic and molecular levels were used precisely for the purpose of identifying letters and, hence, recognising words.

The striking thing about these ideas is that with just a pinch of magic cleverness they spring into really really complex theoretical structures. See above. My feeling ( and this is all gut - sorry ) is that the notion of the atom may not be useful because it's creation stems from the human urge to simplify. It's not a bad urge at all. It's a hugely useful one when you can make it work. This is why scientists cherish the simple equations that actually work - they are like refuges from what is ordinarily a torrent of complexity. So if the "atom" turns out to be an apt metaphor then that's great. But my guess is that reading is more like weather with everything impacting everything else and making prediction better done by hugely complex modeling than by equation or any other simpler model. Still, we don't know and that's the persistent reality.

Bill, it sounds like we agree now for the most part. But it isn't key to my regard for you or Nick or anybody else that we do. The main thing I wanted to accomplish ( why did I write so many posts?) was to combat what sounded like the all too common morbid fear of science is in our culture. I don't think it's healthy - in lots and lots of ways. And that is why I have asserted as you have, that there is no inherent barrier to a deeper understanding of Type/fonts/scripts etc via science. And that this science need not be a threat to the work typographers do; and quite importantly, that it is certainly no substitute for it either.

[Member enn…]

Bill, you say, “[s]cience as a social enterprise—involving many researchers—basically deals with reports of experience, or “observation reports” and tries to explain them.”

I thnk that's right, but, what is being resisted is an explaination of the form: “x or y is actually objectively more readable because reading using x or y is faster or less tiring." What we or I don't resist is explainations of the form: "judgements of readability and legibility are possible because affordance varies complexly across many factors, and yes some typefaces compromise discrimination affordance of criterial role-architectural parts in certain conditions of use by reducing the formal grammer through a reduction of stroke contrast or the size and the distinctiveness of counters, for instance."

[Member enn…]

“…at least as viable a model…”

John, it may seem so as a matter of surface-level type-analytic* common sense and from a basically complicit reading of actually inconclusive tests. But does it have the same disclosive power or practical relevance in terms of expert-level attunements**? Can it help us see why, in perceptual processing terms, inter-letter spacing is important, or the synchronization of intra and inter-letter whites is vital.

A critical reading of non-dissenting studies reveals caveats, actual and potential anomalies, historical inattention, and only partially warranted inferential leaps, as well as truncated knowledge or jaded characterizations of competing hypothesis.

* changed “analytic” to ”surface-level type-analytic”
** added “in terms of expert-level attunements”

[Member Nic…]

the all too common morbid fear of science is in our culture.

It's the culture that's pathological, not me.
Science in the thrall of industrial capitalism.
That's what scares me, as I contemplate climate change.
I'd prefer a little less science, a little less religion, and a bit more art and humanity.

[Member ebe…]

Science in the thrall of industrial capitalism

I agree that is too true today and has been. I share the thrust of your sentiment. But that is a political and a social issue. Throwing the baby of technology & science out with the bath water of industry is no solution. The attitude of less science please and a bit more art and humanity is just not that far from Bush's suppression of climate science out of NOAA. A bit more art and humanity can be easily exchanged for a bit more "God & country" or more accurately "$" or whatever you like best just as your mood changes. Don't get me wrong; I am all about a bit more art and humanity. Nevertheless distinctions need to be made. Exchange that broad brush for a more narrow one.

[Member ebe…]

Peter if you don't mind would you critique this explanation? I think it's a similar idea but it is also less reliant on a specific hypothesis:

You can run speed tests just as one example; and that can tell you which is faster - but it doesn't tell you why one was faster or begin to illuminate what mechanisms are responsible or the ways in which those mechanisms interact or the conditions in which they will or the degree to which they will & so on. To do good reading science you need to concern yourself first, not with superlatives, but with underlying mechanisms.

[Member ebe…]

I think that Bill's urge can't be completely dismissed however not least because the science of reading as practiced by Peter isn't going to help a type maker with their design of a font for long texts right away. Rather than take apart Bill's example which was in fairness probably a little spur of the moment in it's construction let's consider James Montalbano's Highway font. I quote his site:

Based on extensive human factors research ClearviewHwy was designed to reduce the halation experienced by older drivers with increased contrast sensitivity, when reading road signs at night, while increasing legibility and recognition of the sign displays. The results of this project were significant, with various tests showing ClearviewHwy to be between 12% and 29% more legible and easier to read than the current font used on highway signs.

http://www.terminaldesign.com/fonts/sets/?CID=3

None of us has an issue with this. Correct?

And if so, how reasonable would it be to do similar work for a newspaper font you are designing, or one for novels? This isn't creating Readability science - but it is an attempt to use some scientific techniques to advance the craft. That cannot be too terrible can it? Comments?

[Member enn…]

Eben, your synopsis is in line with my assertions.

[Member dbe…]

Nick: "So, in reading studies that focus on types with discrete letter forms, the outcome of letters-as-atoms would be expected."

Yes, and perhaps unexpected, when a discrete letter is studied. An example is the study of letters, including zeros, where a lone character, shorn via isolation, i.e. mis-atomized, is graded for legibility by the subjects. Their results, as I recall, showed that 9 of 10 subjects could not identify the zero of a popular font of the CT collection correctly. I recall the number of correct subject responses was high (1 in10), only because the subjects did not know they could guess that the zero could also be misread under such circumstances as a degree sign, a ring accent, or a dehydrated spittle mark, if it had the right stuff in it to begin with.

Eben: "None of us has an issue with this. Correct?"
I only have issues with the tests and the results Eben. Otherwise, I think James and his crew have done a remarkable job of designing & selling the concept.

"Throwing the baby of technology & science out with the bath water of industry is no solution."
Actually, it is the perfect solution, because it tests the resolve of those who wish to hoodwink you.

But now, the dreamer is coming down from the highlands, to have a 'blog war' on the topic.

Cheers!