I name my files like this: 2017-04-04 #S4907 #Choir List of names.pdf 2017-04-05 #EXCITE #S=5005 Notes on Data Repositories.pdf 2017-04-06 #ARAG #S=5031

I name my files like this: 2017-04-04 #S4907 #Choir List of names.pdf 2017-04-05 #EXCITE #S=5005 Notes on Data Repositories.pdf 2017-04-06 #ARAG #S=5031 #Amount=14.2 #CUR=EUR Invoice.pdf with the following semantic: * date. Every file name is prefixed with the ISO date to facilitate sorting * tags. The syntax #<tagname> to categorize documents with tags. * key-value pairs. The syntax $<key>=value let's us attach structured information to the document name. and keep them in a single large folder. On top of this, I have written some shell tooling to normalize, view, and shuffle around those documents: https://github.com/HeinrichHartmann/pile https://github.com/HeinrichHartmann/pile E.g. `pile extract EXCITE` will extract all files with tag #EXCITE to a separate folder named #EXCITE. There is also a HTML form that helps with proper naming of new files. File management is still a pain for me, but this at least gives me some confidence that I can retrieve stored documents reasonably well. I hope, that one day I'll be able to auto-generate expense reports and tax filings, from properly tagged up filenames.