I've implemented PDF support.
After a lengthy and careful evaluation of the current Java libraries used for PDF parsing, I've decided to choose Adobe's own one.
To use it,
1. navigate to
http://menneisyys.s156.eatj.com/pdf.html and paste the URL to the PDF there.
2. if someone else has some seconds ago checked translated the document, you'll be presented its textual contents, in the usual multiple text areas (see
http://menneisyys.s156.eatj.com/i2.html ) right away.
3. if it's the first time in the last few seconds that the document has been converted, you'll be dispatched to Adobe's page. After some 10-20-30 seconds of waiting, the conversion is finished and you'll be shown the textual version of the page.
4. NOW, just press Back until you get back to
http://menneisyys.s156.eatj.com/pdf.html . Press the Submit button again. Now that it was just some seconds ago that you made Adobe convert your document, you'll be presented it, already in the text areas.
It's a bit convoluted but "hiding" the Adobe interface and make it even easier would have taken me too much time / effort.
Hope you'll like it :)