She smiled, and brushed some loose stands of hair away from her face.The editor caught "stands" in there, and suggested "strands". That's amazing. Stands is a real word, right? Yet by context it somehow knew it was wrong. I'm assuming it's informed by the contextual engine they built for voice recognition?
On the other hand, to test whether is was something as stupid as "stands" not being in their dictionary, I typed this:
So it seems that it needs some context to evaluate a word. Note I made that expanded sentence so it could be presumed that Fibberhta was a proper noun, perhaps someone's strange name, but Docs did still flag it. What does Google suggest instead of Fibberhta? Why "Fibber Hta", of course.
Not perfect, certainly useful and incredibly impressive. If Blogger used this proofreading engine, my blog posts would be way cleaner.
Anyway, just an FYI.