whatwg
Escaped characters in source
The current source file has a large number of encoded entities. This makes it rather hard to edit and read. As UTF-8 is everywhere, is it time to replace these with their Unicode representation?
For example:
<li value="9"><cite lang="sh">Црна мачка, бели мачор</cite>, 1998</li>
Becomes:
<li value="9"><cite lang="sh">Црна мачка, бели мачор</cite>, 1998</li>
And
<p w-nodev>In an algorithm, steps in <span data-x="synchronous section">synchronous
sections</span> are marked with ⌛.</p>
Could be changed to:
<p w-nodev>In an algorithm, steps in <span data-x="synchronous section">synchronous
sections</span> are marked with ⌛.</p>
There is one obvious exception - invisible / non-printing characters.
Would you be interested in a pull request to transform all the &#x... references to decoded equivalent?
This builds upon the HTML5.3 work done in https://github.com/w3c/html/pull/1280
I think that'd be fine. We already adopted UTF-8 to some extent as per 0b37b531bb19f83c3039aa1e00bddd6657758905. It'd be good if the PR message includes the methodology as this might be somewhat error prone.
I agree with the idea of avoiding escapes unless necessary. Fwiw, there are some other situations where escapes can occasionally be useful, although i doubt there are many of those in the html spec:
- some bidi examples, esp including markup or punctuation, so that the sequence doesn't get messed up and difficult to read in the source (although straighforward monodirectional sequences of rtl scripts are usually best stored as unicode characters, as they don't cause confusion)
- any place you don't want normalisation to affect the character sequence
- sometimes isolated combining characters are easier to manage as escapes.
(A lot of people seem to find the utility at https://r12a.github.io/app-conversion/ useful for converting to/from escapes. I was just wondering whether it would be useful to point to it or something similar for the benefit of people writing source code contributions.)
