A misconception regarding the PDF format (Re: Text processing in Unicode

[Dominik Wujastyk]

Very impressionistically - I haven't done any real test testing - my
experience is that if I use Unicode for my source file, then I get a Unicode
PDF.  So I can cut-and-paste and get all the diacritics.  And if I do "save
as" plain text from PDF, I get a plain text file that's correctly Unicode
too.

I'm using XeTeX.

Best,
Dominik

PS Zdenek Wagner has done successful but still experimental work on getting
TeX + Velthuis Devnag => searchable Devanagari PDFs.
Cf. http://sarovar.org/projects/devnag/