Going Paperless: File Naming Methodology
Breaking my pile of paper down into smaller piles by topic and selecting examples of typical documents in every pile was an excellent idea. I now have a structured overview of what I actually need to scan, file and manage on a regular basis. With that knowledge I can now think about a file naming convention.
People have different recommendations for naming files and I see they make sense each in their way. The key is finding out what would work for me. So here’s my first take.
I want every file name to contain certain information which
- informs me about the type of document, the content, the date or the version
- I can utilise to have a bunch of files sorted
- allows automation of the file management process (i.e. through TextExpander or Hazel)
File Name Information to Include
- TYPE of document
Important: I don’t mean “type of file” such as .txt or .pdf – no – I mean “type of document” such as reference material, correspondence, order, contract, picture, identity document, etc.This is going to be very hard. On the one hand I want as few categories as possible because I like simplicity. On the other hand I may find I’ll need more because reality is so diverse…
Why do I need this at all? For easy searching, sorting, grouping. Simple example: My phone bills, water bills and gas bills, all other bills will most certainly end up in different folders on my harddrive. When I want to quickly get all the bills I paid in October 2012, I don’t want to work through all these folders. I want to tell Spotlight something like “Bill 2012″ and within a blink get a list of all these in a Finder window. That’s why the Finder is called a “finder” ;-) I think.
So you see – I need as few document types as possible to keep things lean and mean. And I must accept as many types as necessary. Here’s my initial list:
DOCUMENT = official administrational documents such as passport or birth certificate
REFERENCE = manuals or articles
CORRESPONDENCE = letters, both private and formal
ORDER = could be a regular order as well as changes to an order or even a dismissal
CONTRACT = core contract, additions, changes
INVOICE = obvious
RECIEPT = obvious
ART = the artistic makings of my childrenI guess I’ll identify more types over time. And hopefully I will also get some ideas how I can ideally merge some of the ones I already have in order to shorten the list again… My main suspects are the order, contract, and invoice types. Maybe I can merge these into a TRADE type or so…
- DESCRIPTION of the contentThis is the heart of the file name. If the type is “INVOICE” the description could be “Phone Bill (mobile)“. It’s the part that actually must tell me most about the file so it should be chosen carefully.I’ve noticed that at least 2 variants could serve as standard guidelines for this part:a) “Sender to Recipient What (explanatory word)” = for correspondence
b) “Institution What (explanatory word)” = for refernence material as well as for contracts, invoices, reciepts, even for my kid’s artwork (in a special way…)
- DATE Putting the date like 07-10-2012 in a file name makes no sense. Because it does not help the sorting of files in a folder. If you sort, the character furthest to the left will define the sorting first, then the second left character and so on.Hence dates in file names should always be YYYYMMDD. Where YYYY represents a 4-digit year such as 2012, MM represents a 2-digit month such as 10 for October and DD of course represents a 2-digit day as in 07. Note the 0 ahead of the 7 – again to ensure the sorting is done correctly. I like to put the date exactly like this because it’s shortest. You may prefer to put dashes or blanks in between the YYYY and MM and DD – all work fine.I often hear that people recommend putting the date at the very beginning of the file name because they like to sort files by date. I understand the point and I even like it – I just apply it a bit differently.I can sort files by date created, date last opened, date last changed even without having the date as part of the file name at all. This information is available as file tags in OS-X anyway. However I do want to put the date in the filename (mainly because the date that I’ll choose by hand may need to be different from the system date on which the file is actually created). I’ll just not include the date as the very first element. Instead I put the TYPE first, then the DATE. I’ll explain the advantage that I see in this later.
Before I joined BT Global Services, I’ve worked for Logica for some years, a large IT Services and Consulting company. They taught me the following methodology of versioning files:A00 = Initial notes
A01 = My first draft version
A02 = my second draft version
A… = my …
A= the first version I gave out for review or negotiation or maybe even as my first attempt to publish
B00 = the first revised version
B01 = my first draft on the revised version
B02 = my second draft on the revised version
B… = my …
B = the second version that I “handed out for review” or brought to the customer negotiation, etc.
C00 = …let’s not hope for that ;-)
C01 = …
I like this methodology very much. It’s very simple, very clear, hard to break and you could go a long way from A to Z… Also if I sort the folder with these versions of the file by name, the A version (without 00, 01, 02) always stands a bit out because it’s shorter. Very easy to keep a quick overview of all iterations and easy to find the right file.
Now maybe for the purpose of my paperless project file versioning is not really relevant. But speaking of a file naming convention, I thought I should close the loop and take this in.
My File Naming Convention
So here’s what I choose for my system:
TYPE - YYYYMMDD - Description (explanatory word) – VERSION.extension
INVOICE - 20121007 - Phone Bill (mobile).pdf (no version required here)
The advantage that I see in this naming has to do with the fact that I want a single file naming convention for all kinds of files. My phone bills I’ll collect in a folder that’s dedicated to the relation with that phone company. This bucket will hold dozens of phone bills and just very few administrational correspondence, orders, changes, etc. In contrast all folders that are dedicated to a project likely will contain very different types of files in a single place. I don’t want these project folders primarily sorted by date but firstly by type of file and only then by date. Nicely made possible by my file naming convention.
Again this is probably beyond my little paperless project, but when I come up with such a ground laying methodology, it should work out nicely not only in that single paperless purpose but should function generally.
I don’t know if I’ll use this exact naming convention forever – probably not. I might find inefficiencies in my processes, probably find out better ways of sorting things, maybe want to apply new tools, switch from system A to system B – whatever. But then I’ll have a systematic in place on which I hopefully can apply an automator script or so and simply change file names to my enhanced preferences then… The beauty of systematics.
However. File naming solved for now – next up will be getting files from my action folder into my apps and services of choice.