Hacker News new | past | comments | ask | show | jobs | submit login

I.. agree. After all, even before ASCII people already used the English character set and punctuation to type out many non-English Latin characters on typewriters (did typesetters do that with movable type, ever? I dunno, but I imagine so). Just as ASCII was intended to work with overstrike for internationalization, i can imagine combining codepoints having been a thing even earlier.

OTOH, it wouldn't have been UTF-8 -- it would have been an EBCDIC-8 thing, and probably not good :)

There actually is a rarely used, but standardized because it needed to be, UTF-16 variant UTF-EBCDIC, if you needed nightmares about it. [0]

In some small decisions EBCDIC makes as much or more sense than ASCII; the decades of problems have been that ASCII and EBCDIC coexisted from basically the beginning. (IBM could have delayed the System/360 while ASCII was standardized and likely have saved decades of programmer grief.) The reasons that UTF-EBCDIC is so bad (such as that it is always a 16-bit encoding) could likely have been avoided had IBM awareness of UTF-8 ahead of time.

Maybe if IBM had something like UTF-8 as far back as the 1930s, AT&T needing backward compatibility with their teleprinters might not have been as big of a deal and ASCII might have been more IBM dominated. Or since this is a sci-fi scenario, you just impress on IBM that they need to build a telegraph compatible teleprinter model or three in addition to all their focus on punch cards, and maybe they'd get all of it interoperable themselves ahead of ASCII.

Though that starts to ask about the scenario what happens if you give UTF-8 to early Baudot code developers in the telegraph world. You might have a hard time to convince them they need more than 5-bits, but if you could accomplish that, imagine where telegraphy could have gone. Winking face emoji full stop

[0] https://en.wikipedia.org/wiki/UTF-EBCDIC

> Maybe if IBM had something like UTF-8 as far back as the 1930s, AT&T needing backward compatibility with their teleprinters might not have been as big of a deal

I think this points to why the science fiction scenario really is a science fiction scenario -- I think decoding and interpreting UTF-8, using it to control, say, a teleprompter, is probably significantly enough more expensive than ASCII that it would have been a no go, too hard/expensive or entirely implausible to implement in a teleprompter using even 1960s digital technology.

Yeah, and I was thinking about it this morning the reliance of FF-FD for surrogate pairs would shred punch cards (too many holes) and probably be a big reason for IBM to dismiss it when they were hugely punch card dependent and hadn't made advances like the smaller square hole punches that could pack holes more densely and with better surrounding integrity.

"Sigh, another swiss cheese card jammed the reader from all these emojis."

Yeah, no thanks to that nightmare, or to UTF-5. Did you know that UTF-7 is not a joke??

Yeah, email is one of those things that in a sci-fi scenario should probably be invented after UTF-8, and hopefully too with a better plan for binary attachments. uuencode/uudecode, MIME, base64, UTF-7, so many fun hacks on top of each other on top of a communications system restricted to ASCII to play things safe.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact