Editorial: Replace non-ASCII quotes #2149
Conversation
|
I'm definitely in favor of getting rid of the smart quotes, but discussion on ecmarkup suggested that some people would prefer to use the actual codepoint rather than HTML entities for other non-ASCII characters like |
|
Either way, it seems desirable to use the same patterns throughout the spec. This PR makes the minimum necessary changes to achieve that. |
|
I'm fine temporarily using HTML entities for consistency, but I am not fine with using straight quotes in prose, which are always typographically incorrect. I used curly quotes in Annex E quite on purpose. |
Annex E contains no raw curly quotes; literally the only use of them is around "nullish" in the introduction. And there are several phrases wrapped in U+0022 QUOTATION MARK where curly quotes would be appropriateβwhy should that one be special? |
|
Oh, I misinterpreted your comment. Updating now. |
gibson042
force-pushed
the
2020-08-ascii-consistency
branch
from
3efd8bb to
ede94de
Compare
- For consistency, replace non-ASCII "curly quotes" with "&[lr]dquo;" character references - For consistency, replace non-ASCII guillemets with "&[lr]aquo;" character references
ljharb
force-pushed
the
2020-08-ascii-consistency
branch
from
ede94de to
fb22c2b
Compare
The resolution of #1287 established a UTF-8 encoding of spec.html, but in practice the file contents are almost exclusively ASCII with the exception of
π½andβ. This PR remedies the handful of outliers (a pair ofββ¦βleft/right quotes and some literalΒ«β¦Β»guillemets), although I would not object to a subsequent sweep that replaces character references and/or "dumb" quotes.