Inconsistent treatment of identifiers and keywords

A few identifiers and keywords are presented with varied visual treatment:

$ git rev-parse HEAD
8344d2ff9b26571ad58c73e68b2d62625ce3bfab
$ grep -Eo '(<[^>]+>)?(true|false|undefined|null)' index.html | sort | uniq -c
     88 <a>null
     37 <a>undefined
      8 <code>false
      1 <code>null
      7 <code>true
      3 <dfn>null
     25 false
      4 null
     26 true
      3 undefined
      4 <var>false
      2 <var>null

Some of the unadorned usages may be correct (e.g. when they are used for their plain-English meaning in informative text like "the UA's behavior is undefined"), but many seem to be unintentional omissions.

I would be happy to submit a patch to correct this, but I could use guidance on which treatment is preferred for each word.

Also, we've identified and fixed this once before, so linting for this may be warranted.

Start by reviewing the identifier and keyword occurrences in index.html, using the grep output in the issue to locate the inconsistent treatments. Read w3c/webdriver#760 and pull request #771 for the earlier approach before deciding which usages need changes. Done means the intended presentation is consistent and any linting direction is documented or implemented.