I'm going to uncheck this to see if it helps anyone. I've been aware of conversations about curly quotes but wasn't aware of this setting. I don't believe it should cause problems but let me know.
3 Likes
The option does improve readability overall, but any time someone pastes code outside of a code block (which is pretty often) then it's bad. I could see going either way with it.
5 Likes
It's also good for me to be aware of it now if people have problems sharing code bits. I always thought the code block tags would avoid odd formatting, but I never paid attention to it enough Testing...
" " " " "quotation"
" " " " "quotation"
2 Likes
I’ve turned it back on…
" " " " “quotation”
" " " " "quotation"
Okay, just wanted to make sure, the forum (or rather the post entry text editor) isn’t converting curly quotes inside code blocks. Thank you for your indulgence! 
I’m wondering whether certain themes might have a slightly different monospace font set that might include curly things, but doesn’t that defeat the concept of “monospace”? I might just just be stupid and need more coffee.
3 Likes
em–dash ™ (c) – hm no symbol for copyright…
That’s odd…
2 Likes
em–dash ™ (c) – hm no symbol for copyright…
This all looks right as I’m composing, but if I put code formatting elsewhere in the message:
em--dash (tm) (c) -- hm no symbol for copyright...
em--dash (tm) (c) -- hm no symbol for copyright...
Okay, now it’s not doing it - I think when I was typing before I had a line of three back-ticks for code and it undid that formatting of the top line, but now that I’m doing it more carefully it’s not doing it.
just ignore Hanon scooting around on the roll-y platform under the engine
4 Likes
I like the curly quotes for readability; some of the other improvements, like — and …, I’ve configured my keyboard to let me type on my own. But the curly quotes just take too many keypresses too frequently for me to do them manually.
3 Likes
What I want to know is, when are programming languages going to admit that it’s no longer 1990 and start accepting curly quotes? Not like it’s hard, or ambiguous, or anything.
3 Likes
They can even be nested safely! Though I wouldn’t want to rely on that.
1 Like
Yeah, when you start looking at the details it is kind of ambiguous.
Inform 7 accepts curly quotes! But should it also accept «angle quotes»? (It doesn’t currently.) What about other Unicode forms like ❝these❞?
Should it be a syntax error to use curly quotes in “unbalanced pairs“? (Inform doesn’t care, but that means it can’t use that information to accept nested double quotes.) (If you did accept nested quotes, what happens if you need to nest unbalanced quotes?)
Etc.
4 Likes
Eh. I thought about mentioning that, but I feel like those are pretty basic implementation details. A first undergrad compiler-construction course is pretty cramped anyway, so you might not discuss those sort of questions there, but you pick it up pretty quickly when playing around with compilers or taking a second course.
I think the clear general-purpose answers are…
If you’re doing it, of course you should accept other forms that written languages use as quotes. You could argue about some of the edge cases but IIRC there are, what, like a dozen or two? Just accept all of them.
It should absolutely be a syntax error to use curly quotes in unbalanced pairs. I don’t think I’ve ever seen a good argument for needing to use them in an unbalanced form, and disallowing it dodges all sorts of unnecessary questions like whether »this is a valid string❝.
You shouldn’t worry about nested quotes unless you really have to (string interpolation with full programming-language expressions?). Otherwise you should just require escaping the starting and ending character(s).
Etc.
Of course, it’s easy to argue for doing things differently in a specific language, but in general… I wouldn’t go so far as to call it ambiguous, any more than any other programming-language feature.
2 Likes
My assumption is it used to be there was one “character” for quotation marks (the straight ones) with I assume a typography code. When word processors came into use, I would guess that the right and left curly quotes were two completely new characters created for prettiness that automatically swapped in via auto-correct magic with two different character codes and didn’t register as the same typography to programs that were looking for quoted text between specifically straight quotes?
That’s my impression as to why they cause problems. Plus sometimes in code you’re quoting odd things that may not register to a word processor as quotable, and you may get two left curly quotes - such as in a case where Inform 7 is programmatically assembling a sentence and the specific quoted text is only supplying a closing quote - it confuses the formatting switch which assumes that straight quotes will be in pairs to get converted.
1 Like
That’s pretty much it. Most of ASCII is based on what typewriters used, which is why there’s e.g. a single - character instead of all the different ones used in proper typography (hyphen, minus sign, en dash, em dash, etc). Or why there are ` and ^ characters with no obvious use. (On a typewriter you could backspace to combine them with something else to get à , â, etc.)
And now since ASCII is what everyone supports, that’s what keyboard layouts continue to offer.
3 Likes