Consider renaming ByteString to IsomorphicString

[annevk]

Context: w3ctag/design-principles#454 (comment).

[domenic]

I'm -0 on this. It's not immediately clear from the name what that means, whereas ByteString seems pretty clear: it's strings where each code point/code unit fits into a byte, to be used in contexts like HTTP where strings are always treated as having byte-sized code points/code units.

[annevk]

Fair. I tried to consider renaming isomorphic string to byte string instead, but I'm not entirely sure about byte string encode and byte string decode.

[nektro]

Latin1String / UL1String ?

[domenic]

Latin1 is a pretty bad choice, because on the web it means windows-1252 and on other platforms it means something else. (And the actual spec that defines it leaves its behavior undefined for most bytes.) See also nodejs/node#56542 (comment) .

[nektro]

got Latin1/UL1 from "Basic Latin" and "Latin-1 Supplement" being the block names for U+0000-00FF

[annevk]

I suspect when most people hear "latin1" they think of the encoding, not the Unicode block. And while it's not technically an encoding on the web, it's a label for the windows-1252 encoding as Domenic mentioned, which is very different from the actual encoding.

And the Unicode block only defines the non-ASCII part of what this string type can contain, so it's not entirely fitting either.

I'm going to close this for now, but I'm open to revisiting this if we have a much better idea at some point.