There's an Apple II clone from Taiwan which apparently forgoes the text mode and just has a graphics mode because it's simpler that way to render Chinese writing.

It had a 16K ROM, of which 12K is a BASIC interpreter, which leaves 4K for whatever else. How did those 4K accommodate a Chinese font on top of whatever else it needed to do?

(I have the idea it might've done some kind of compression, for example 說, 話 and 語 all have identical left-hand sides. Hard to say if that's worth doing though.)

  • 1
    No computer of that era supported anything even close to a full set of Chinese or Japanese characters. Both languages, however, had phonetic alphabets available which could be used to form almost any word. I suspect that computers would likely have had some characters defined in addition to those of the phonetic alphabet, but I don't know how many. 2K of ROM is enough for 256 7x8 characters or 64 14x16 characters; 4K would be enough for twice that.
    – supercat
    Commented Jun 1, 2018 at 17:33
  • 2
    Not using text mode also means that the font doesn't need to be in ROM.
    – user722
    Commented Jun 1, 2018 at 17:50
  • @RossRidge: Having the BASIC interpreter in ROM wouldn't seem terribly useful if the fonts weren't also in ROM. Perhaps they could be compressed in ROM using an approach which could be unpacked to RAM on startup, but wouldn't be practical to process directly while drawing characters, but I wouldn't expect such a design to have been used back then.
    – supercat
    Commented Jun 1, 2018 at 19:10
  • @supercat BASIC just uses normal ASCII characters. An application that wanted to display Chinese characters could just keep the subset of characters it needs in RAM.
    – user722
    Commented Jun 1, 2018 at 19:45
  • @Ross except Chinese BASIC. That's actually got tokens based on Chinese words. Like 如果 instead of IF. Commented Jun 1, 2018 at 20:29

2 Answers 2


Check out this page https://classictech.wordpress.com/computer-companies/acer-groupmultitech-electronics-inc-sunnyvale-calif/

Esp. the PDF at the end of the article, dated September 14th 1982: https://classictech.files.wordpress.com/2010/03/1982-multitech-presskit-introduces-first-microcomputer-to-handle-chinese-9-14-82.pdf

It says:

The MPF-II-C takes advantage of a new method of coding Chinese characters to decrease the amount of memory space used to store the complex symbols. Before this "Dragon" method of character coding was developed by Multitech, one average Chinese character would use 32 bytes of memory space. The MPF-II-C uses only 64K bytes of memory to store 22,000 characters.


The Micro-Professor II-C is comprised of a Chinese Character Controller (CCC) and the Micro-Professor II, a 6502-based, Apple-compatible microcomputer also manufactured by Multitech. Two special systems--the Dragon Chinese Alphabet Coding System and the Dragon Chinese Character Generating System are incorporated into the CCC. The MPF-II-C can process either Roman or Chinese characters, and it also can program in Chinese Basic computer language.

By using the 24 keys on the MPF-II keyboard, the user can access the 24 Chinese "alphabets". Each character requires between one to five keystrokes to be input. With one month of training, an operator can input 30 to 50 characters per minute, according to Chang.

  • 2
    I think this explains a lot. The MPF-II with only 16K of ROM couldn't display Chinese characters. It was the MPF-II-C that could, and it apparently had 64K of memory on the Chinese Character Controller card for storing Chinese characters.
    – user722
    Commented Jul 18, 2018 at 21:44
  • Wow! Next time I'm in Taiwan I'm gonna hunt for one of these! Commented Jun 6, 2020 at 4:56

The Wikipedia page on 'Chinese BASIC' (which cites the Multitech Microprofessor as the article's most prominent example) depicts a traditional Chinese keyboard of the time, showing only 113 Chinese symbols on offer, plus 68 you'd expect on an English keyboard.

Even at 32 bytes per symbol (i.e. 14x16 pixels on an Apple II if there's no reprocessing in between) I make that only 3,616 bytes for the Chinese symbols. If you used 28 bytes per symbol (i.e. 14x16 with the expectation of unpacking during drawing) then that's only 3,164 bytes, leaving more than enough space for the western 68 at 7x8 without overflowing 4kb.

The article mentions that all of the radicals of the Cangjie input method are represented so I guess another way around of looking at it is that there was some kind of compression at play, but it's the sort that requires the human to map between characters and components.

  • 1
    They keyboard layout in the picture is for at least two IMEs. You don't enter one character with one keypress. Cangjie is uses the symbols in the bottom-left or bottom-right of the keycaps, or both. Cangjie is very difficult and requires training. Each character requires a set sequence of keypresses based to a degree on what components make up the positions within the character, with more common characters requiring fewer keypresses. The symbols in the top-right are Zhuyin Fuhao a.k.a BoPoMoFo, which is easy to learn if you know the sounds of Standard Mandarin Chinese. Can't find a real photo. Commented Jun 6, 2020 at 5:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .