A Look at Early Japanese Typewriters (2016) | Hacker News Reader

A Look at Early Japanese Typewriters (2016)

filthyplaten.com

60 pointscaust1c19 comments

For more on the topic of tyewriters in non-alphabetic languages, I recommend The Chinese Typewriter (https://mitpress.mit.edu/books/chinese-typewriter) from MIT Press.

Awesome! Thanks for sharing!

Came to the comments to recommend this as well! Fantastic read.

+1, and I’m eagerly awaiting Mullaney’s follow-on (in progress) about post-typewriter text entry.

An odd but interesting related book is Unger’s _Fifth Generation Fallacy_. If you ignore the dubious framing about 1980s Japanese AI, there’s a long discussion of kanji text entry (including computer text entry) with some kanji-specific references that _Chinese Typewriter_ necessarily skips.

It bothers me so much that the pairing of katakana and alphabetical characters has nothing to do with pronunciation.

>The positive thing about Kanji is that it allows for quite an efficient expression with more nuance.

As a person only a year and a half into learning Japanese, the positive thing is clearly marking separate words and particles. Japanese only has a few unique syllables and rarely uses spaces or other separators. I've repeatedly had trouble parsing loan words written in katakana, especially when it's more than one word smushed together.

What are you talking about? Have you used a proper Japanese input on Android or Google?

They are paired, vowels first and then phonetically.

On Android and iOS,

あうえおい A, u, e, o, i かくけこき Ka, ku, me, ko, ki さすせそし Sa, su, se, so, shi

The alphabet has no real existence in the Japanese language, it is a different language and romaji is bad to pick up, but the characters are grouped together to enable fast typing and auto detection of grammar and Kanji.

Then with any proper keyboard, you can change the input setting so you do not need a Japanese keyboard, you can write solely phonetically via a English keyboard and have the software correct it , which I think is annoying because you have to type two keyboard inputs for one character.

GP is referring to the first typewriter in the article, where the katakana has no relation to the English letters they share a key with.

Such a mapping would not be feasible, for several reasons.

* There are more kana than English letters. You can see that there are numbers paired with kana, keys in the bottom-right with two kana, and on this particular keyboard, ヲ (wo) appears to be omitted.

* Broadly speaking, each kana represents a pair of sounds: one of ten consonants plus one of five vowels. You will have English consonants with no real Japanese counterpart, and Japanese consonants with fewer than five appropriate English counterparts.

The comment was referring to the keyboard of the Olivetti typewriter. The Katakana are arranged without regard to the Roman letters on the same keys. There's some amount of logical organization, vowels are on the top row, the タ group is on QWASZ. And next to it on DXZ is シサソ, and ス not too far away on R, but what is セ doing all the way on P? At a glance it doesn't seem to match a recent frequency analysis[1], so I have no idea what motivated this particular arrangement.

[1] https://link.springer.com/article/10.3758/BF03195600

Japanese actually have intonations (like Chinese) to differentiate between words that have the same hiragana construction.

For example “hashi” can mean bridge or chopsticks depending on the up and down tones.

>> The positive thing about Kanji is that it allows for quite an efficient expression with more nuance.

> As a person only a year and a half into learning Japanese, the positive thing is clearly marking separate words and particles.

The positive thing about kanji is a function kanji don't perform?

Very interesting, thanks for posting!

Reminds me of the articles I saw about 2-3 years ago on the difficulty of creating new fonts for Asian languages compared to Roman alphabets. The sheer number of characters required to build up a base of shapes is huge, compared to our 26x2 or whatever variation on that, plus accented characters and symbols.

And yet despite this, graphic design in Asian countries still manages to use a greater variety of fonts than in Western ones. For example compare this[0] advertisement for curry to this advertisement for fries[1].

[0] https://chancurry.com/wp-content/uploads/2018/11/9c0f748dfcd...

[1] http://i.huffpost.com/gen/1965451/thumbs/o-SATISFRY-570.jpg?...

The font helps to reinforce a branding identity. This is why the fries example has a unified face. The copy is also seconday to the brand, which is the first thing the eye sees. Contrast with he curry one where the branding is taking a back seat to the messaging.

These are two different styles of advertisement.

Most Asian languages are alphabetic and/or syllabic (all the south Asian languages, Mongolian, Khmer, Thai, Korean, etc) quite apart from those that use the Roman alphabet like Malay.

Only a few have the difficulty you describe though it Is a big one.

I used to see a lot of those last machines (the one with the pointer and the ray of type) back in the 80s. The blog author says he bought it off a journalist, perhaps thinking it was for foreigners. Presumably it was light because the journo needed something portable.

The reason for the type case is that companies would insert the specific characters they needed to conduct their business.

Sort of related, but I remember visiting a ham radio store in Akihabara and finding out that Japan has its own Morse code: Wabun code (https://en.wikipedia.org/wiki/Wabun_code)

Obviously, I have no idea what this was used for but I expected to see a katakana/hiragana combo where one of the two alphabets is activated via a shift-like mechanism. Are such typewriters a thing?

That’s basically how JIS works, which is the layout that dominates the market later on, and what you’d find on most PCs. The keys has hiragana on them, and a special key is used to switch input type.

Reply on news.ycombinator.com