UTF-8 Unicode operators vs digraphs

[Richard (Rikki) Andrew Cattermole]

On 14/08/2026 8:09 PM, Dlighted wrote:
> On Friday, 14 August 2026 at 06:41:45 UTC, Walter Bright wrote:
>> Some good observations!
>>
>> If there is a specific code point for é, then the combining character 
>> alternative should be illegal.
> 
> Don't quote me on this but it (NFC) already seems to be convention. I 
> don't know how many input methode there are that produce decomposed 
> forms (NFD) with separate accent and how many of those decomposed 
> letters are actually in circulation today outside documents explicitely 
> talking about them. I wish I'd see some hard data on this.
> Exceptions often named on the internet is Apple filesystem metadata and 
> filenames and copy pasting from it using Mac Finder file manager. In 
> theory anything can make them but in practice little does.
> However if I copy paste an ö I see in the wild anywhere else, it's 
> almost certainly precomposed (NFC).
> 
> I pasted some German Wikipedia articles in this converter to check I'm 
> not talking bullshit and it told me they were already in NFC.
> https://onlineminitools.com/normalize-unicode-text

Strictly speaking html should be NFC. 
https://www.w3.org/TR/charmod-norm/#unicodeNormalization