Emojis: cute but deeply flawed

There is nowhere on the planet like Akihabara in Tokyo. Most big cities now have electronics districts, but all pale by comparison. Its size, density, and intensity are so extreme that exploring its warren of showrooms behind the facade of neon ads is the twenty-first century equivalent of visiting temples, as you might in Kamakura, on the coast to the south of Tokyo.

One unique feature of Akihabara shops is that they are used to test the market for future products. So not only will you see product ranges larger than in manufacturers’ catalogues, some of those products may not appear on the general market for another year or more, or may never go any further.

Akihabara is ideally fertile territory for the likes of emoji: exploratory, driven by some bright ideas, but far from complete or mature. They’re fun, chic, and excitingly modern, thoroughly mobile, language-free, but deeply flawed.

We have millenia of experience in communicating with text, whether you use an alphabet or a character set. The forms of each character or letter have evolved, and design disciplines have developed to help us get the most out of them. We have progressed from handwriting and carved inscriptions in stone, to movable type, carved from wood blocks or cast in molten metal, and resulting ink on paper prints, to increasingly fine pixels on high-resolution displays.

All the lessons learned during those long journeys have been ignored in the rush to incorporate the tiny coloured images of stylised faces, people, and objects, with which to spatter the social media. Long, and often bitter, experience gained from the design and use of icons, symbols, and signs have also been ignored.

Emoji have not been designed for use at the same size as regular text. You can either set display font size so that the text is readable but compact, leaving the emoji pitifully small and indistinct, or you can increase the font size to the point where the emoji are legible, but accompanying text too large and wasteful of screen space.

Excerpts from the emoji epic translation of Moby Dick, known as Emoji Dick, illustrate this well. When the PDF is viewed at 100%, its text remains quite legible (particularly for the myopic) but it is impossible to decipher the emoji.


At 150% the text is starting to get a bit wastefully large, and many of the emoji remain impossible to read.


At 300% the emoji are much clearer, but the whole content, particularly accompanying text, is far too large.


Even standard tools for using emoji struggle: here the subtle differences between emoji of very different meaning are only seen when magnified.


This results, of course, from the lack of good design in emoji characters themselves. Many mimic the face, with its subtle changes in features, such as the angles of the mouth pointing up or down, to impart opposite meanings. Although emoji faces are simplified and stylised, such important differences only appear in very small parts of the emoji’s image.

In text, contrasting words such as good and bad, happy and sad, are visually easy to distinguish. You can blur the letters, cover much of them up, but still their meaning is discernible. In emoji, slightly smiling face πŸ™‚ is so similar to slightly frowning face πŸ™ that – unless set at about 18 points – they appear almost identical. Even if you can detect a difference, you have to look carefully to see it.

This is because they fail on one fundamental principle: design for the size that the icon will be used at. I don’t want to have to set Tweetbot to use a ‘huge’ font size, so most of the emoji that I see in tweets will appear incomprehensible.

Another fundamental principle of icon design is to keep them simple and, well, iconic, and free of references to specific cultures. Although clapping hands sign πŸ‘ is now fairly international, and it is neat to be able to customise the hue of those hands from βœ‹πŸ» (so pale as to be almost invisible) to πŸ‘πŸΏ (which could be an owl in flight), I am not sure how multicultural πŸ–– (raised hand with part between middle and ring fingers) might be.

As the design adage goes, realism has its limits, and symbolism is often much clearer. So full marks for πŸ‘€ if only the rings around those eyes had been much darker so they showed more clearly, but πŸ‘Ž for πŸ’… πŸ’ƒβ˜„πŸŒ―πŸŽ«, and many others which far exceed those limits of realism. In many cases, emoji seem to have been designed on the assumption that they will be viewed at 256 by 256 pixels.

Whether emoji will ever prove to be more than marginal decoration is another matter. Efforts by Fred Benenson seem to be pointing the way: following a successful Kickstarter campaign, he put the task of translating the whole of Moby Dick into emojis out to the sweated labour of the Amazon Mechanical Turk. The resulting book Emoji Dick;, available for free from archive.org, is completely and tediously incomprehensible unless its full text appears interleaved, as the excerpts shown above demonstrate. If you would prefer to pay $200 for a hardbound printed copy, perhaps it would restore the sense of Akihabara.

Take note that Benenson’s latest book, How to Speak Emoji (Ebury Press, Β£9.99, ISBN 978 1 785 03202 8), is classified by its publisher as humour.

It is hard to know whether emoji will prove like ASCII art, telegraphic Q codes, or CBer slang, and all but vanish, or whether they will evolve into a proper pictographic language. For the moment they are too flawed to be anything more than passing eye-candy.