This Typeface’s Letters Are the Average of the World’s Handwriting | Hacker News Reader

This Typeface’s Letters Are the Average of the World’s Handwriting

wired.com

80 pointsshill27 comments

Why would you have an image with "The quick brown fox" without "jumps over the lazy dog"? The whole point of the phrase is to use every letter.

Something suggests it's an attempt to keep the font from being appropriated. Though if it's used, that won't last long. And letterforms aren't copyright-protected (though the software by which they're derived may be: the basis of font copyrights in software).

Reminds me of Averia, a font made by averaging fonts. http://iotic.com/averia/ https://news.ycombinator.com/item?id=3877025

That's actually a surprisingly-good-looking font.

Meh, it looks pretty average.

badum TSSH

This reminds me very strongly of older moveable-type fonts. No, not the blogging engine, but physical moveable type made of melted tin and lead hand-set in a press.

The letterforms are influenced by the dies from which they're cast, and are further worn and damaged with age, both softening the overall shape and introducing quirks to them. You'll find print from older books -- generally 19th century and prior, sometimes much prior, tend to look like this.

I like the effect.

With my own handwriting, once described in school as "bovine diarrhea committed to paper" (a rough translation) I have come to admire people's handwriting in a sort of wistful way.

I think a discussion around this post would be incomplete without reference to the font made from Dijkstra's beautiful and clear hand writing. http://lucacardelli.name/indexartifacts.html (search for Dijkstra) or if you prefer a pdf preview http://lucacardelli.name/Artifacts/Fonts/Dijkstra.pdf

I wonder if Ramanujam's handwriting has received similar treatment.

A sidenote: while searching for Dijkstra's font on Google had an experience that had but disappeared from my memory. It has been a while since I have been in a situation where I am unable to google to the the page that I know exists on the web. This is the second time in the week that I felt this, something afoot ? In any case wanted to find the letter to Dijkstra asking for permission to use his handwriting for a font.

I wonder how many misunderstandings and frustrations occurred because of that "z" which looks like a "2". Any of my math professors would have marked anything with that 2 wrong.

Weirdly, it's only capitals, no numbers or punctuation. Not sure what I would do with such a typeface! Also it's more like an average of the world's finger-painting styles, since you have to draw the letters super big and on a touchscreen. Still, everyone does have their own style, and there are some interesting stats on the "left" of that nearly un-navigable website, letter averages by country or profession and such.

Limited typefaces can still be useful for logos or titles/headers. For instance, it's very common for stylized Japanese typefaces to only support a small fraction of common characters.

Certainly true. In this case it seems like since it's meant to be an "everyman" font it would be more useful to have it be in lowercase too, and include numerals etc.

I'm guessing they decided writing 26 letters was about the limit of the average attention span online - having to do it twice, along with 0-9 and a basket full of symbols probably would result in lots of incomplete forms.

Yeah, I tried it out, and while it's a fun little thing, drawing with one's finger is very different than drawing with a pen (even a bic! :), and the results are inherently sort of clumsy. Something based on real handwriting (as is often done for individuals) would be vastly more interesting.

Also of course, the idea that this is anywhere near "universal" is pretty silly...

> Not sure what I would do with such a typeface!

WHAT DO YOU MEAN? WHO DOESN'T LOVE ALLCAPS? I DON'T NEED NUMBERS, EITHER. X EQUALS FIVE MILLION ONE HUNDRED EIGHTY SEVEN THOUSAND TWO HUNDRED NINETY NINE POINT ONE FIVE FIVE SIX THREE. SEE?

IT'S LIKE I'M WRITING IN COBOL. THIS IS FUN.

Where did you get those question marks and periods from?

the link to the experiment: http://theuniversaltypeface.com/home - a download apparently follows in august

Does anyone know what kind of algorithm they'd be using to "average" all of the letters? If you look at the I's and J's, you can see that though some people use horizontal crosses, they aren't included in the universal letter. Anyone have any insight here?

I'd similarly be curious about letters that have multiple ways of forming them: for example, capital G's and the "serif"

I haven't seen anything specific about how they are doing it, but I can give you a few guesses as to what common approaches would be.

The most straightforward way to model the shapes themselves is to put some points on each letter in corresponding locations on each letter. Every sample of each letter would need the same number of points, and each point would need to be in roughly the same spatial location on the letter (so, for a letter S, the first point in every sample would be at one end, the last at the other, and the ones between at some kind of identifying landmarks).

Given these collections of sampled curves, the simplest thing to do is to just compute the Euclidean mean, treating each of them as a point in a high-dimensional space. You could go farther and do PCA, giving you not only a mean but modes of variation. Using this you could examine the most common ways in which each letter varies in the population, which can be an interesting thing to study.

What I've described is building a statistical shape model from a boundary point distribution model (PDM) of an object. This is typically done in the context where you want to fit your model to new instances of the object, not just for finding a mean, and are known as Active Shape Models. You can check it out on wikipedia (en.wikipedia.org/wiki/Active_shape_model). Here is one (among many possible choices) describing generally how this is done:

http://tinyurl.com/nxjwygn

There are other techniques for representing the shapes or computing statistics that can produce better models, both from theoretical and practical points of view, but this is generally the most common and would be my first choice if I was going to do something like this. Of course, they could be doing something much simpler like a simple averaging of the images (assuming the letters are all in roughly the same place) as well.

I'm reminded of Kevan Davis' "Smaller Picture" web experiment that has been running since 2002. Using a grid of random bits, readers could choose to vote to flip a bit forming a given letter or picture. You can also view a time-lapse animation. Unfortunately, the letters ("Typophile") are no longer online. :(

http://kevan.org/smaller.cgi

It uses your finger though, not the right source material.

Wouldn't most people using the Latin alphabet be using cursive?

In the US (no clue about the rest of the world) cursive has gone out of style. Some schools don't even teach it anymore.

This does look more like "This Typeface's Letters Are the Average of America's Handwriting" or perhaps the researchers asked for only upper-case letters.

Does this average include all the 6 year olds of the world?

Like my buddy Leo, I write in mirror image. Should I...

Oh, neat. Are you left handed by any chance? Have you written about that decision/habit anywhere that I could read? It sounds somewhat interesting. e.g. Has it been inconvenient in many situations? Should I assume that that implies that you can read mirror image just as well? Why/at what point did you decide to start doing that?

Sorry if this sounds nosy, but I'm curious as to whether that would help avoid smudging more easily for people who are left handed (which I think I've heard can lead to avoiding certain types of pens?). (I personally am not left handed)

Of course, don't feel obligated to answer.

Reply on news.ycombinator.com