Short of rewriting the copy (which they could have done, of course), dealing with orphans on the web is kind of impossible. It's a valid point, but I wouldn't let that stop you from taking the rest of the article at face value.

[...] dealing with orphans on the web is kind of impossible.

Not at all! Just:

  html.replace(/\s+(\S+)$/mg, " $1")

EDIT: Bugfix, don't replace whitespace without any characters after, replace all occurrences in a multi-line text

Well, yeah, but anything is possible with regex. (EXCEPT PARSING HTML, I KNOW.) I meant more along the lines of your typical HTML/CSS markup.