from Hacker News

The Upside-Down "e" - an Editor's Nightmare (1993)

by rcrowell on 4/9/10, 10:42 PM with 15 comments

• by weilawei on 4/10/10, 1:43 AM

  "You're right. Man should not be a slave to technology. The 'tail shouldn't wag the dog'. But that's idealistic and unreal. Technology always has its own limitations and parameters. It's not the first time technology has shaped the alphabet. The same story has been repeating itself since the origin of the alphabet. Greek and Latin were clearly determined by the hammer and chisel against marble - it's easier to carve straight lines in stone than rounded ones. That's why so many of our capitals are based on straight lines. The cursive Arabic was influenced by pen; cuneiform, by clay and sharp stick. Other alphabets were created by carving on bamboo. Simply today, the determining tools are fonts, computer keyboards, and satellite transmissions - the only difference being that they're a bit new and unfamiliar in the alphabet designers' hands."

• by quant18 on 4/10/10, 9:29 AM

  Yes, when you ask a committee of linguists to design an orthography for you --- you get a godawful almost-IPA "one phoneme = one grapheme" academic exercise which is utterly impractical for everyday use. (Again, an object lesson in the robustness of top-down-designed systems vs. ones which evolved from centuries of use). Azerbaijan would have had the same problem even if they had stayed with the Cyrillic alphabet. E.g. after Kazakhstan became independent, they had lots of trouble trying to use Kazakh as the administrative language due to a shortage of typewriters with the appropriate letters.

  Uzbekistan is greatly underappreciated for having done something really neat after the Soviet breakup --- designing an orthography with no "funny letters". They use context to distinguish the "back" and "front" versions of phonemes like i [1]. And in the two cases where that would be too confusing (o and u), they put an apostrophe after the letter --- i.e. u', instead of something like ü.

  [1] Turkey, which speaks a closely-related language, solved the "front i" vs "back i" problem by making one dotted and the other dotless --- those of us here probably know all of the toLowerCase-related bugs that caused, I think articles about this have been posted a couple of months ago

• by j-g-faustus on 4/10/10, 3:45 AM

  I remember we had problems with Norwegian fonts in the early 1990s - printing, copying between computers and email would invariably mess up the æ, ø and å. Even the computer science department of the university I attended in the mid-90s didn't get it right. I think it took close to ten years before the majority of printouts and emails were correct.

  But compared to Azerbaijan, I understand we were lucky - at least we didn't have to mess around with creating our own fonts :)

  Good point on how tools put constraints on expression, BTW. I guess ASCII art is another example, and the ASCII smiley :-) is still with us.

  Even today mixing text and freehand drawing or images is troublesome enough that we rarely bother. In a sense it is somewhat curious that 60 years of computer science has not been enough to replicate the convenience of a handwritten note...

• by techiferous on 4/10/10, 5:06 AM

  Also relevant: "The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets"

  http://www.joelonsoftware.com/articles/Unicode.html

• by chiquita on 4/10/10, 9:48 PM

  Aren't they talking about U+018F and U+0259?

  Nowadays, adding that glyph to an existing font could be a very easy thing to do. It depends a lot on the font design you're after...

  Fonts released under a free license would have helped.

  A nice font editor: http://fontforge.sourceforge.net/

• by theschwa on 4/10/10, 5:11 AM

  Who'd have known I'd cause so much trouble.

• by wendroid on 4/9/10, 11:18 PM

  Painful to read something easily solved by free software, even in 1993