Today's Obsession: 65,536 Characters

This, to the casual viewer, is a video of random characters flashing by. To me, it’s a triumph of automated communication—it’s every character in a comprehensive typeface. Technically: it’s every displayable character in a Unicode double-byte face from data positions 0 to 65,536.

This is the current standard of typeface production which became formally available to typographers in 1992, but didn’t see much use until later in the decade.

Before that, typographers were forced to isolate their typefaces into little dribs and drabs of language containing only a few hundred characters. 256 if memory serves. (So the number stated above, 65,536 is 256 × 256).

Now, there’s room in the formal Unicode specification for over a million different characters in a single typeface file, including things as esoteric as Greek musical notation, Western European musical notation, and Egyptian heiroglyphs. It’s beyond cool to me that there’s actually a specification for human language so that devices worldwide can display and perceive it.

The video’s from Jörg Piringer, and was pointed out by my friend Miguel Carvalhais via his Twitter feed. Both are brilliant; spend some time with their sites.

Related Articles:



  1. actually, mindy, type designers aren’t required by anything other than market forces to have any particula rnumber of characters. most times, the only things that sell well with less than 256 characters are nicely-made display faces or novelty typefaces. but there is no enforced requirement.

  2. Thank you for the answers, Patric. I’ve noticed that some of the fonts that type designers created may not have all the unicode characters needed under basic latin. So we try to be careful with what fonts we can use dynamically on our direct mail templates. If one unicode doesn’t exist that our customers happen to enter such as an asterisk, the PDF will not be able to open to preview, resulting in error. That will prompt in a customer calling our customer service for help; it’s the reason why we always must test our templates before they go live. I’ve had faced a few templates in the past that had that font issue in which we were forced to change the font to a similar font that has most unicode characters. Ridiculous, isn’t it? So, from what I understand, you’re saying that current fonts including new fonts are (or will be?) required to have the basic latin unicode characters? I’m not sure why not all type designers are willing to include unicodes especially when it’s a basic latin. I hope that will change one day because technology is booming than it has in the past. 

  3. mindy, the short answer is: they are, but it’s time-consuming. for example: the FontFont library recently announced that it is fully opentype compatible with varying degrees of language support. that’s a big step forward.
    longer answer: the current modus operandi is to create a standard version for those of us who need western-only languages available, then a pro set for additional european characters.
    asian and devanagari characters added beyond that are a whole ‘nother ball of wax requiring familiarity with many, many languages. the nature (and market) of typographic design is rapidly changing from a single artist creating something to an entire multilingual team creating them.

  4. If in that case, why don’t current typefaces get updated with more unicode characters? I think it would be nice to see more unicode characters for each font especially when using them dynamically through PDFlib. Some fonts will get weird boxes because that character do not exist while other fonts has that character. Because of that, we have to limit fonts we can use that contains more unicode characters.