Bibtex and unicode: my understanding is that bibtex assumes single-byte encodings (Philipp Lehman explains on comp.text.tex).

The nasty bit is that you can’t tell that bibtex is doing anything wrong in the vast majority of cases: if you’re just moving around strings it doesn’t matter if it has the right idea or the wrong idea about their internals. The error only shows up (afaik) if the bibtex style counts characters. (Of course I prefer a style that happens to do this.)

On the upside, this means that if it’s working for you there’s no need to change. And for Polish and I guess Norwegian there is always the fallback to (La)TeX macros for the occasional non-ascii characters (only needed in the bibliography, and only in fields that the style will chop out characters from). For Greek that’s not an option.

now about unicode: I’m actually using latex-ucs package and it works fine, also for bibtex. perhaps it doesn’t work for Greek for some reason, but it sure does work for Polish and Norwegian.