to recognise text within a document a search utility would need to know how it was encoded, otherwise it is just the equivalent of trying to read a foreign language you have never learned, quite possibly in a script you don't recognise, with no idea of whether to read left/right right/left top/down centre-out.....

I think the chances of windows 10 doing this are nil

I use agent ransack - it will search word [but not xara]; word formatting is not proprietary in the way xara is - interestingly sometimes you can get a 'result' because the search finds a string of characters that 'spell out' a word, but are in fact encoding something else - this often happens with jpg/png files for example when doing a wildcard search