Know any good legibility tests for typefaces?

[ebensorkin]

Carl, great post.

Nick you said. Eventually, the ergonomic project bogs down in demographics.

You are asserting that demographic & regional differences might be significant?

you have to target a sector

What sectors would those be do you suppose?

I agree that this might be so, I would even guess that it is, but I am not aware of anything that shows that it is. It would be all too easy to mistake regional taste for one style or another with an objective test of what they read best. They might overlap in practice - but I think they are not the same thing.

I don't think that reducing things to simply "trust the craftman (or artist)" & waving hands helps us. Sorry, that's probably too much rhetorical flourish - but I hope you see what I mean. Obviously the craft has gotten plenty far already. But being luddite re: Science seem counter productive to me.

Your sense that science is bought off by the big guys and is in their camp; ( unless I have you wrong ) may be so. In fact it seems probable. But I think that is a separate cultural issue and should be dealt with as such. Put another way; intellectually I don't like to see them conflated. And I think it would be disingenuous to deny that science has anything to offer just because I don't stand to benefit from the admission. Or maybe I will... depending on what researchers and their employers ( or universities) decide to do.

I have faith that the best ideas for improving type will come from type makers be they indie or otherwise, and with or without research to inspire or gird their results. Maybe in that sense I agree with you. And maybe if people (type makers and scientists alike) notice useful stuff they will share that useful stuff. I plan to. You have. ( Thanks BTW! ) I think it's very promising.

And lastly I think that Fonts will probably always get their the vast majority of the economic value from their style. The utility differences are a kind of economic sideshow for the most part even if I think that culturally it is important enough to spend time on. Do you agree or not so much?

[Randy]

And lastly I think that Fonts will probably always get their the vast majority of the economic value from their style. The utility differences are a kind of economic sideshow for the most part even if I think that culturally it is important enough to spend time on. Do you agree or not so much?

I agree. The sideshow would include much of the commission work that goes on for sake of utility, often with real economic benefits (agate comes to mind). Meanwhile, brush scripts sell like hotcakes.

[ebensorkin]

Randy would you even agree that text faces are sold on primarily on style too - with utility most assumed? eg It's a newspaper face it's a book face ( as in suitable for novels etc), it's a book face specifically with this flavor or that flavor.

Are Agate's economic benefits to do with space saving or reading utility or both? If both; which do you think was the dominant issue?

[Nick Shinn]

You are asserting that demographic & regional differences might be significant?

They are. One font for the lion and ox is oppression.

What sectors would those be do you suppose?

That is determined by the nature of a publication's readership.
So an Index of Typeface Readability that didn't have a demographic breakdown wouldn't be much use.

But being luddite re: Science seem counter productive to me.

Luddite? I'm all for science--in its place. We can investigate how reading works, but readability is not and never will be a scientific concept.

[ebensorkin]

One font for the lion and ox is oppression.
hee hee! That is pretty purple prose but I basically do agree.

So an Index of Typeface Readability that didn’t have a demographic breakdown wouldn’t be much use.
Sounds like a better kind of science to me. You might make a heck of a good advisor to a scientist making a study of readability. ;-)

readability is not and never will be a scientific concept
No more so that love is or even he far simpler "sweet" is. So I agree - but in a specific way. Consider; it is possible to study what happens when people are in love and compare heart rates and variety of other data. It is also possible to see what reactions lead to the perception of "sweet". If you look at the studies and see how the scientists couch their work I doubt that the scientists were claiming to be able to in any way encompass love or to be equivalent to experiencing "sweet" for yourself. So the two are not synonymous exactly. What the scientist would say is that they are making falling in love the subject of their study - as a expedient way of putting it. Or making a study of how we perceive the taste "sweet". And doing so in a specific a specific, probably narrow, and hopefully repeatable way.

I suggest that we could learn things; maybe interesting or important things; or even just useful things, from studies of readability. And I think I can say this without doing any violence to idea of people "falling in love" or experiencing "sweet" or good bad or indifferent "readability".

[William Berkson]

>readability is not and never will be a scientific concept.

Within 5 years you will be proven wrong, is my bet.

[Nick Shinn]

Is that the result of scientific studies of predictability, or just a hunch?

[William Berkson]

Hunch

[Nick Shinn]

You luddite! :-)

[John Hudson]

Nick: One font for the lion and ox is oppression.

Ironically, almost all modern readers encounter Blake's 'writings' in typeset form, rather than the etched illustrated form that he intended. Doubtless the typeset text is more legible/readable than Blake's deliberately naïve etched writing, but the latter demonstrates nicely the irreducibility of 'a good reading experience' to a set of legibility metrics. This is not to say that the scientific study of reading as phenomena (not phenomenon) is without value, but there is no typography without text and text is anything but uniform.

[Randy]

Randy would you even agree that text faces are sold on primarily on style too - with utility most assumed?
I would guess so, but it will depend on the designer and an the application. A book designer might choose differently than another designer.

Are Agate’s economic benefits to do with space saving or reading utility or both?
There are lots of ways custom (newspaper) type can "save money." I'm sure someone with more experience in custom work can elaborate. Space savings and reading utility yes, but what if the new type allows you to run on cheaper paper? Saves money. What if the new type has all weights based around the same metrics so there is no text reflow for stylistic changes? Saves money. Small type is used in more than the financial section, there's tv listings, sports scores and schedules, weather forecasts, classifieds etc. A typeface tailored for these specific situations could beneficial in many ways. There are brand and sometimes license benefits too. H&FJ printed a booklet on their agate face Retina, that goes into detail about it's benefits. Font Bureau may have something similar for Poynter Agate.

John that is beautiful.

[John Boardley]

...but typefaces should not be tagged with readability standards, which is where this is leading.

Exactly. The day scientists hijack typography, then all is lost. There is something rather noble sounding in a scientific method for testing type; however, ultimately such a system would only do a disservice to type and to readers. This scene came to me in a dream:

The year is 2025. My client proposes that we go with WhatevernextSuperLegible Sans because it has a 5-star legibility rating. Cut to me, pan right as I slit my throat with a rusty razor blade. The End.

The problem I have with the whole science and type thing is that legibility is just one part of the equation. A good typographer (in fact, even an average one) knows whether a text is legible (and whether the typography is an aid to readability). We don't need science to confirm what we already know. Do we?

I think the proposition is that “the world’s most eminent scientists” would build their tests around.... The Reader’s Opinion.

But that doesn't sound like good science. 10,000 subjective opinions does not sum to universal truth; so, either a falsifiable theory of legibility is proposed; or, we simply do as we have done for the past 500 years.

I'm reminded of those TV commercials for shampoo and skin care treatments: "clinically proven", "scientifically tested", et al. The day a typeface is accompanied by something similar, I'll be reaching for that razor blade.

[Nick Shinn]

Right John, text (content) is a variable that influences readability, so add that to demographics (who's doing the reading) and typography (how the typeface is positioned and scaled on the page), as factors that effect readability, as well as typeface.

The effect of text on readability has been measured, as mentioned earlier, in the Flesch index.

I don't think typographers, in designing a page, consider those four in isolation from one another, so my primary objection to readability research is that it does. Kevin is doing the right thing in trying to introduce typeface as a variable in reading theory, but my concern is that an Index of Readability for typefaces that doesn't factor in, especially, typographic variables, will be used by the communications industry as if it does. So the client will tell the designer that they must use the "most readable" typeface. Or will foundries have to apply to a standards body to have their types tested and categorized by readability grade?

Pardon me for dragging this out, I've spent too much time on the 'phile today!

[eliason]

I’m reminded of those TV commercials for shampoo and skin care treatments: “clinically proven”, “scientifically tested”, et al. The day a typeface is accompanied by something similar, I’ll be reaching for that razor blade.

I hate to be an accomplice to your suicide, but Christopher Burke records that this kind of pitch was already being made in 1931 for Futura.

[ebensorkin]

John, Nick, so it sounds like it isn't the science per se. It's the use (or misuse) that it it's put to. Fair enough!

What we might differ over is the inevitability of science being put to misleading and counter productive use.I really doubt that a 'standards body' is going to be set up. Nobody would want to pay for it.

Your previous scenario where only big type houses get the benefit of research seems a bit more likely although maaaaaybe that too can be avoided. I sure hope so. Certainly being interested and encouraging with researchers might help that to happen.

We don’t need science to confirm what we already know. Do we?

If we know it; then no. On the other hand some of what we 'know' is still subject to lively debate. So maybe we don't know it exactly. But even if we do; what could be exciting, and the purpose of science, is to learn something new!

Believe me, I don't think science is the only model of knowing that is worth while. If I did I would be a scientist. I just think it's worthwhile to notice the difference between science as a way of considering something and other quite different things - marketing, PR, hype, client idiocy, simple minded interpretation of science & so on. In fact, if you want to effectively debunk false claims this kind of separation is in my view, a prerequisite!

Craig, nice point!!

[enne_son]

On account of my reading of the literature on ‘crowding’ I’ve come to think of affordance as a curve with an initial steep climb, than a wide plateau and eventually a gradual decline. Proposing that affordance is a curve with a distinct shape is more useful than thinking of readability as a simple equation of more versus less. And it is a perspective that has it's source in science.

Where, in terms of type-size the climb begins or the plateau sits; how wide the plateau is are issues involving thresholds and ranges and vary from type to type. I suspect an analysis of variance will show that, for most standard text fonts the thresholds and ranges don't vary dramatically under normal conditions of use or covary linearly. In other words, a change that affects a threshold positively might have a negative effect on where the plateau sits. Perhaps the styles of termination, or types of contrast can be an example. But thresholds and ranges do nevertheless, I think we'll find, interact with the traditional typographical parameters. And I suspect they do so in demonstrably more complicated ways than we have seen codified so far.

Simple affordance has to be balanced with issues of comfort in large texts, with concerns about heirarchical transparency, surveyability and navigability in instructional manuals, and with the cultural-historical positioning or affiliation statements a document has to make. Comfort in large texts, positioning, and the demands for heirarchical transparency, surveyability and navigability can exert countervailing pressures on the choice of type. This makes a decontextualized descision based soley on simple or generic affordance micro-advantages short-sighted.

Many clients groups are aware of these subtlities in how affordance works and of countervailing pressures. Many others willingly learn. What seems to count is explicit evidence of well-motivated and knowledge-based decision-making.

[ebensorkin]

Nice see your ideas again Peter!

Would you expand on & give some context for this phrase?
In other words, a change that affects a threshold positively might have a negative effect on where the plateau sits.

What is affordance? Surveyability and navigability would have to do with layout rather than anything to do with font choice - yes? I am guessing thats right except in as much as they relate to having Italics or Small caps or Weights as options.

What seems to count is explicit evidence of well-motivated and knowledge-based decision-making. That has been my experience as well.

Having said all this though; you would not I suspect say that generic affordance micro-advantages are not worth thinking about. Just that they have to be seen in a broader context. And this was Nick's point as well.

[russellm]

If those doing the research don't refer to and include other relevant specializationst, it all can get to be a bit like those blind sages trying to describe an elephant (It's like a tree. No, it's like a snake...) (Why didn't they just ask the mahout?)

It might be difficult to design a font that is completely illegible and still include identifiable letters, and it's true, the most legible font in the world can be mangled by bad typography. The fact there is bad typography though, suggests that starting with a highly legible typeface should mitigate the awfulness somewhat. But, what's a legible typeface 'sposed to look like?

The rules of what make a legible typeface are in my opinion are quite simple and straightforward and allow for a great deal of flexibility. I don't think is is even possible to design a definitively "most legible" typeface. I think in any fare test of comparable fonts, you'll find that differences in performance can be so slight as not to matter.

I really don't think legibility standards need to be anything to get knickers in a twist about. It's mostly common sense to anyone with some expertise in typography. I think where that kind of standardization would apply would be on wayfinding signage, medicine bottles, captioning on TV, operating instructions that come with power tools and the TV guide in the newspaper. (OK that last one is personal, but Really! Would it be possible to make it worse?) and things like government and public transit websites and other printed materials, where accessibility is a right that these organization should make honest efforts to honour.

This sort of graphic work seldom actually gets designed in any real way. The objective in producing a set of instructions may turn out to be "How many tiny black words can I fit onto this tiny piece of paper that will be packaged with these potentially deadly drugs" instead of what percentage of people who should read these instructions will not be able to?". Even if the designer conscientiously designed this sheet to be as readable as possible, he or she would have no way of telling you what percent of people could read it. That should be something that could be determined. At the very least, there should be some set of guidelines that can be referred to, that are a bit more sophisticated than, "use Arial and make it big" (the real purpose of such a spec is to keep people who know less than the spec writer from using some crazy Spenserian script, and Arial is a font they can name.)

Taking the example of Erik Spiekermann's work for the Financial times, where he improved the legibility of their typeface by cleaning up the shapes somewhat, increasing tracking and leading and (counter intuitively to most) slightly decreasing the font size.

I think that the form such "standards" might take need not be unreasonably restricting at all and would have no impact on what the vast majority of designers do.

Hellbox, email me (russ.mcg at gmail.com), and I'll tell some of what I've done to test legibility. It was for signage, but some of it might apply to what you are doing.

-=®=-

[enne_son]

Eben, I felt the need to say something in this thread, but this reply will have to suffice for the time being.

[Affordance] (I explained it elswhere some time ago on typophile — I think it was the rule or law thread — affordance is the relationship between an object or artifact and what we choose to do with it or use it for): legible writing affords or promotes or lends itself to effective perceptual processing in reading, so ease of reading is an affordance of legible writing or its typographical correlatives or surrogates.

[Expanding / giving some context] I'll have to ask you to think about how styles of termination, or the presence or abscence of contrast, or wider spacing, or the sizes of counters affect performance at thresholds of size or distance, versus how they effect performance at normal type sizes and distances where reading speed and comfort are at a plateau.

[William Berkson]

Arguing against studying readability scientifically is to me like arguing against studying disease scientifically. You can make the same arguments, and they will be just as wrong-headed and against the progress of humanity.

For example, the issue of text size has an obvious and measurable affect on readability. You can make text small enough that people can't read it. And you can measure how small that is, how it varies with age and acuity, and so on. There is no reason to think that less obvious things can't be studied also, with sufficient imagination in theorizing and sufficient cleverness in experimentation.

That doesn't mean that everything to do with readability, particularly the aesthetic side, is likely to be measured soon or ever. But to just wave aside scientific study is highly misguided.

As Peter explains, a successful science of readability is likely to come up with theories that are as complex and subtle as we know reading to be. And such insights may well help us to improve reading on screens, to help those with limited vision, with forms of dyslexia, and so on.

[Nick Shinn]

Bill, reading is not an ailment.
I'm wary of analogies used in discussing type and reading, because there's nothing quite like it.
(I'm sure I use analogies though from time to time, it's hard to avoid.)
So I won't develop your health analogy by going into the perils of big pharma.

And you can measure how small that is, how it varies with age and acuity, and so on.

You can measure text size, but not as a scientific absolute, in the way that an object's dimensions or mass may be determined.
Size is a nominal value, determined by the foundry.
Don't you recall the thread here last year on the definition of the em square?!

A typeface does not have size.
A scientific parallel would be fractals, as in "how long is the perimeter of the Jenson "a"?
It would depend on which particular "a" one chose to measure, the resolution of the medium in which it's set, and the resolution of the measuring device.

Of course, it's possible to take a particular setting of a particular typeface, vary its size, and see how the variations perform with different age groups.

Whatever knowledge this reveals, as Peter says, "a decontextualized descision based soley on simple or generic affordance micro-advantages [is] short-sighted."

Consider these specimens from The Designer's Guide to Text Type (King and Esposito, Van Nostrand Rheinhold).

I put it to you that the readability of Janson and Kennerley will vary against one another according to leading, amongst other factors.
Also consider paper stock and, of course, tracking (this book was published in 1980, and man, is that 8 pt. type tight!)
Then there is line length and paragraph length. And might not one type perform better for rag right setting, the other for justified? And how would they fare when one varies the Flesch index? Would Janson's readability index be better for longer words?

All 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. For online use, the monitor and operating system would have to be specified, as well as pixel-height and application or file format (hinting varies).

However, AFAIK, reading tests in laboratories are not conducted using "real" text. For instance, the Moving Window technique restricts the amount of text that is visible to a certain number of letter around a fixation point, and replaces all of the other letters on the page with the letter x.

I don't believe there would be anything particularly scientific about indexing the readability of typefaces. It would be a simple form of market research. Investigating how reading works is one thing, and of course I support it, but applying readability standards to typefaces (the subject of this thread) is something else.

[Nick Shinn]

Neural networks may be developed to simulate a reader, in which case a piece of text in a particular setting might be run through a software application to measure its readability.

Compare with recent developments in the shampoo industry. It so happens that CGI has become so proficient at rendering the effect of moving hair (thanks to Hollywood money)--at a microscopic resolution--that it is now possible for chemists to apply a virtual shampooing of a chemical formula to a virtual hair type, and see how it will effect body and lustre.

Ultimately, this kind of virtuality will be great for product development (laboratory animals especially), but would be problematic if an "independent" organization used software to grade products by chemical formula. IMO, that's a bit too virtual, but I guess it's coming.

[crossgrove]

Nick,

Your comments are confusing, to put it mildly. What comes thru is your irrational fear of science, or maybe laboratories, or possibly shampoo.

Previous testing has indeed been less than satisfactory; often irrelevant and misleading. It does not follow that all testing forevermore has to be similarly flawed. We're already at the stage of discussing designs for tests that intentionally account for the variables you, and others have named. In fact, you list some techniques or features of such tests that would be useful. We know the problem is that typographers have been left out of such testing. So, here we are.....

Ryan asked for tests to assess the legibility of typefaces. This does not presuppose excluding, ranking, or naming names. I think Ryan should reappear and clarify, and maybe comment on the discussion. I think mention of ranking, rating, or establishing specific, literal metrics that define readability are all figments of paranoia.

[Nick Shinn]

Carl, I don't worry about shampoo anymore.

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

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.

[Dan Gayle]

Here's the plan:
We all contribute a paragraph of gobbledygook type-speak strung together into nearly incomprehensible sentences that we can then index into a master "Use this to convince your client about your choice of typeface" book.

Why not? There are scads and scads of books on color combinations with explanations that designers can thumb through. Why not a book about "What typeface and Why, and How to effectively convince your client"?

Best seller, I tell ya.