I wish to categorise my large collection of electronic documents. I need a foolproof set of categories to put each document in. I want to use a classification system similar to how university schools are organized:


Some of my documents are productivity tips (similar to Dave Allen's "Getting Things Done" book). I can't find an academic category under which productivity comes. The term "productivity science" doesn't seem to be a widely accepted term.

What is the most natural category to put the study of "Productivity" in? Psychological Sciences (or other Social Sciences)? Applied Sciences/Professions? I'm tempted to put it in Psychology because I have other documents based on Learning, Memory and Performance. The four of those seem best suited to Psychological Sciences.

  • What software are you using to manage these electronic documents?
    – StrongBad
    Jun 5, 2012 at 14:03
  • Good old file systems (NFS, NTFS, HDFS etc). My resources include, but are not limited to, html web pages saved on my computer, post it notes, Google Docs and printed reference cards. Jun 5, 2012 at 17:30

4 Answers 4


My guess would be Business -> Industrial and labor relations -> Organizational

  • Thank you. Yes, that was another one I noticed but forgot to mention. It's reassuring to hear someone else suggest that. Jun 5, 2012 at 17:30

There is no such thing as a

a foolproof set of categories to put each document in

With paper documents the document had to "live" in a single locations and therefore needed a primary category/label/tag. You then had lots of different indexes that would allow you to find the document from multiple places. You often had to walk to a different library to collect the document. This system was/is inefficient and problematic.

Electronic documents can "live" in multiple places. The concept of assigning a single primary category to an electronic document is dated and silly. You should either modify the file directly to add helpful metadata or create a database that lets you externally create searchable metadata. This then allows you to easily search for the document, and related documents.

  • There is no industry standard tagging system that I am willing to invest effort in, out of fear of future-proofness. The closest thing I know of is XMP metadata tagging. It may be 'silly' to you, but single-parent hierarchies are still the way file systems work. My electronic resources are not academic papers with tag metadata in their header. They are html web pages I've saved, post it notes, Google Documents, printed reference cards among others. You cannot assign physical documents to multiple categories. Jun 5, 2012 at 17:28
  • Sorry, I should have been more clear that electronic documents were just an example of a resource I needed categorization with. Jun 5, 2012 at 17:42
  • @Sridhar-Sarnobat sure tagging the documents directly can be problematic, but what is wrong with creating a database to link the documents and the tags. I would go with a .bib file for use with BibTeX, but I could imagine an Endnote database or even creating your own database app.
    – StrongBad
    Jun 5, 2012 at 18:19
  • 1) I don't want some of my tag links to be in .bib files, others to be in picasa.ini files, others in file header metadata. I want a lowest common denominator that works in all environments/operating systems. 2) Creating an index (that tells you to look somewhere else) would increase maintenance effort. I've done this with symbolic links in my operating system. Links get outdated and broken. 3) Sometimes I create spider diagrams which are tree structures. If you allow a node to have 2 parents, the diagram becomes very messy (I've used yEd and Graphviz for this). Jun 5, 2012 at 21:16
  • @Sridhar-Sarnobat So put all the links in a bib file. There is no reason you cannot create a new IMAGE entry type. A reasonable front end of the database (e.g., JabRef for bib files) will retrieve and open the files directly no matter where the files are. As for symbolic links, maybe you should try hardlinks. They save the disk space but cannot get broken.
    – StrongBad
    Jun 6, 2012 at 8:11

Any of the four categories you suggest would make sense.

On a related note, for exactly this reason, I would recommend against a category-based classification system, as many fields of research are significantly cross-disciplinary. Depending on your field, you'll end up having to cross-reference your article a lot.

I would rather suggest a tag-based method, where you devise an intelligent tagging scheme and simply tag each paper as necessary. This way, you get the benefits of categorization, combined with the ability to cross-categorize by simply adding another tag. You can add as many tags make sense and ensure that you can find the paper again later, which, at the end of the day, is the whole point of the system. Additionally, many bibliography programs support tags references out of the box.

  • Thank you for the reply, but the same response as the one above applies to this. Jun 5, 2012 at 17:26
  • So the "tags" could be something as simple as a plain text file that lives in the same directory as your heterogeneous pile documents. It just has to be searchable. I don't think categorization is bad, but @eykanal is right: You need an effective cross-reference system.
    – Philip
    Apr 25, 2015 at 17:00

The metaproblem here is that the OP has clearly invested a lot of intellectual energy trying to find the “right place” to store documents by some objective standard. But the only “right place” is the one that maximizes your probability of finding stuff when you need it. (This could be an early warning sign of excess rigidity in problem solving more generally. Always think of the big picture.)

If you created an imaginary new academic discipline called zigamazoo and stored all the GTD and similar productivity materials there, would that be a ridiculous name? Of course. Would it be effective? I’ll bet it would.

I’d actually call it personal productivity and make it a peer of the other academic disciplines. I’d probably remember that when I went to look for it. If cross-references would be helpful under topics like psychology, I’d add them, but they’re probably superfluous. You’ll remember you created a (rare!) top-level heading for this.

Why on earth would you care whether it's under the right academic discipline as long as you can find it?

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .