漢字 · かんじ
Aule Kanji Pages · Kanji Recog PagesThe image below is of a PDF open in the latest Adobe Reader ; as you see, the text copy selection also grabs the furigana, producing slightly problematic results as plain text.
Web page copies tend to place the furigana in-line, which is even worse!
A typical result from Adobe Reader follows:
夜の
ちやう
帳にささめき尽きし星の今を
げかい
下界 の人の鬢のほつれ
よ
Is the last item furigana or not?
Compare my Curl markup example in this blog post. The MIT Curl browser plugin is available at the Tokyo SCSK site.