To test an aspect of Adobe Acrobat (PDF to text conversion) I installed the program. Not that good an idea. I downloaded the installation program from the Adobe homepage, did a sudo ./AdbeRdr9.4.7-1_i486linux_enu.bin and now my Firefox was entangled with its plugin.“sudo rm -r /opt/Adobe” did not help. It apparently installed a .so file in ~/.mozilla/plugins. nppdf.so, that is apparently not a part of Debian/Ubuntu. I erased that. If you go to Firefox menu Tools
“Manage Content Plug-ins” I still see four lines of the Adobe issue. One should think that right clicking or deleting would work. No. “$ rm -r ~/.adobe/” didn’t help. With “$ egrep -r “Acrobat” ~/.mozilla/*” I tracked down “pluginreg.dat” and erased appropriate lines from “[PLUGINS]” section. Put that didn’t help. “$ locate nppdf.so” showed a lot of issues, and “$ sudo rm `locate nppdf.so`” apparently helped, as now PDFs open in evince. Next time I should really try to see if there is a package, so “aptitude purge” or reconfiguration are possible. BTW: The text conversion in Adobe Acrobat was not that good compared to other programs I tried. It found “fi” and “fl” ligature problem, two-column problem, greek character problem. Python library pyPdf might be promising as you might redefine the ‘extractText()’ method attempting an attack on the ligatures and other strange characters. pdftotext and PDFbox was also interesting. Any comments on this?