Explain searching for   and &nbsp in the DOM

[hsivonen]

https://wicg.github.io/scroll-to-text-fragment/#next-non-whitespace-position searches for literal   and &nbps appearing in the DOM. This is strange enough that it could use an informative note explaining why these strings are special-cased. That is, why are these expected to appear in the post-parse DOM and often enough to require special handling?

[bokand]

I think this is a carry-over from an early iteration of the spec and likely a bug. Will address.