Typographic quotes are left- and right-handed, vs. the ASCII double quote which is just a single character:

   “quoted”

   "quoted"

Who in the blazing highs of techno-utopianism fervor thought it was a good idea to automatically translate the latter to the former we'll never know.

Why not stylize it as typographic but store it as a "regular" double quote though?

One problem with that is that the correct start/end quotes to use depend on language which is not always know for all text snippets.

How do you handle non-paired ones then?

Render them as standard double quotes? The same way Markdown renders a single backtick as just a backtick, but text surrounded by backticks becomes code.

Nice try, but the problem with this is that typography is _really_ complex. For example, there is a rule in English typography (I'm not sure if it's often used today though) that when you have a quotation spanning several paragraphs, you should put an _opening_ quote at the beginning of each paragraph – but only one _closing_ one at the end of the quote.

We've let go of other historical typographic and spelling conventions, it's time to let go of that one too.

Computers exist to represent, transmit, and display _human_ communications. If a computer system cannot represent a human utterance we should extend the representational capacity of the computer system, not force the human to conform to the limitations of the technology.

Computers also generally don't support scribbles in the margins of your document. Or stylized glyph variants except a few that have been grandfathered in, e.g. ꙮ. Neither does the computer support whatever ligatures you might make up on the spot.

One thing that makes computers more powerful than analog tools is that they are much much much more structured.

So while we should not take computer limits for granted I think that the opposite extreme is just as if not even more absurd.

I'm also all for eliminating overly complicated typographical conventions from places where they are technologically possible.

It would be quite sad. Besides, it is not an arbitrary convention, it serves a purpose.

No it doesn't. Unbalanced quotation marks are a crime against humanity.

I would base direction on adjacent whitespace/punctuation instead of trying to do pairing.

Nice try, too. How would you handle Spanish quotation marks at the beginning of the sentence, then? And what about French quotes which are separated from their contents by a thin space? All of that is possible, of course, with large tables of special cases.

Or, you know, you could just have separate characters for opening/closing quotation marks. ;-)

> All of that is possible, of course, with large tables of special cases.

Yes, that's how ligatures work. Pairing across characters is the one I'm not sure I've heard of before in fonts.

Edit: Besides, how do you think these are typed on keyboards without separate keys? The software already detects and replaces them based on context.

I'm french and I actually really like the auto-translation. This way software that does not care about which quote those are (mail, web, etc) can swap them, and where it matters then it does not translates it (vim, etc).

Sadly the new official french azerty keyboard has dedicated keys for both opening and closing quotes, and the good ol' simple quote tucked away behind modifier keys. As a dev I hate it. (arguably I should not even use azerty for development but that's another issue)

> I never understood why a "stylistic" choice requires separate characters.

I don’t think it’s a purely stylistic choice, there is actually semantics to it:

    “ opening quotation mark, i.e.: starts the quote

    ” closing quotation mark, i.e.: ends the quote

You could otherwise make the same point about parenthesis: why not just |do this| instead (of this)?

Double quotes can be unambiguously autodetected by checking word boundaries, but separate single quotes are needed since apostrophe is the same character as a close quote, but can appear at the start of a word. Note that most smart-quote-generating software does fail badly at this.

Some common ambiguous pairs:

’bout - abbreviation for "about"

‘bout - starting a quote about a round of a fight

’cause - abbreviation for "because"

‘cause - starting a quote about a reason or ideal

’em - abbreviation for "them"

‘em - starting a quote about typographical units

’n’ - abbreviation for "and"

‘n’ - speaking of the letter itself

It's not a stylistic choice. Opening quotes and closing quotes are different things, and it isn't possible to tell the difference (when not already provided) without parsing the language in which they're used. That's why in TeX you have to manually specify which kind you want, and in software like Wordpress that just guesses, the guess is usually wrong and your published text looks ridiculous.

The complaint was purely about software replacing normal quotes with "smart quotes", not that it did it wrong.

But the question was why there's more than one code point between normal quotes, left quotes, and right quotes, and the answer is that it isn't possible for left and right quotes to share a code point.

"quote unification'':

  most typefaces: “Hello” „Hallo“

Verdana: “Hello” „Hallo“