I have been using Fujitsu ScanSnap double-sided auto-feeding scanner for years. I started with a Mac model (S300M I think? no Windows drivers?) and then moved to a Windows model, the S1500. For the last couple of years I've been running a Windows virtual machine in VMWare Workstation primarily so that I can use my scanner and the included OCR features (turn scans into text). Recently, I've been having a lot of trouble with VMWare Workstation, and I run libvirt/qemu-kvm on my web server, so I decided to try it on my workstation as well. It works pretty well for Windows 7, not a fast as VMWare when it worked properly, but the scanner will not function correctly using it. Last time I looked into using the S1500 on Linux I found almost nothing. I could scan stuff but it wasn't very useful for a paperless office workflow, the whole reason I have such a scanner. I am happy to say that, on Debian Wheezy/Sid, gscan2pdf with the libsane-perl backend and tesseract for OCR seem to work nicely. Some caveats:

• I downloaded the newest version of gscan2pdf and installed it. At the moment that is 1.0.6.
• The 'Page Options' tab of the 'Scan Document' window only works properly when I choose options that don't make much sense. I selected 'ADF Duplex' in the 'Standard' tab, and then in the 'Source document' section of the 'Page Options' tab, I have selected 'Single-sided' and 'Side to scan: Facing'. Otherwise, the page numbers come out strange. I do still get double-sided scanning, though.

I couldn't find a lot about this on the web, so I hope it helps someone else in my position. Happy scanning! EDIT 2012-10-28 Strangely enough, I was able to get the scanner working just fine in my Windows VM using Spice USB redirection. It's good to have the fallback, but I'll try to stick with Linux, it's much more convenient and removes a dependency.