We use Joomla with Remository to store and manage publications (don't ask me why). Files (PDF) are stored in a database and can be accessed via dynamic, rewritten links of the form
Here is an example: some file
Current browsers reliably detect that they get a PDF.
wget uses the
.html filename but after renaming I get a working PDF file.
curl behaves similarly; piping its output into a (suitably named) files gives a working file. All this leads me to believe that -- against all odds, one might say -- the data our system provides is generally valid and understandable for clients.
However, Google does not seem to index PDF files referenced by such links. Our publication list is indexed, but the PDFs linked there are not (they don't show up in web and Scholar searches).
How can we tell search robots to retrieve our files and index them?