PDF Pubmed Finder

As announced earlier this week, here is a first version of the PDF Pubmed Finder.The Finder helps to identify an unknown paper within seconds and rename it within milliseconds. Although the Finder consists of a single page only it uses several areas to update the data. I have written the code for my local installation but you may try it also online with your own PDFs either by uploading them with the available form or by appending the PDF URL to the Finder address http://wjst.de/pubmed/finder.php?http://medicine.plosjournals.org/test.pdf
The last method is certainly better as it can be used within a scripting environment; note that my domain may not retrieve a PDFs that you can see at your local computer. PDFs also need to have some inline text (run an OCR on scans before submitting them).

Here are some usage instructions: After selecting any article on the left (1.), clicking on relevant keywords (2., 3.) will lead to an immediate Pubmed search (4). Finish by clicking on “rename” or “Pubmed/Zotero” import of the correct reference. If your search doesn’t retrieve any useful hits, keywords will be automatically reset.
What I would like to know all from the AJAX experts who can inspect the sourcecode: I ran into deep problems described earlier with the prototype oncomplete function used in the links of the div2 container. They are checking how the div4 container was being updated. As you see it works, but I failed to clean up the code by having only one Javascript function in the beginning of the page with the div2 links calling this function. Your help would be greatly appreciated.