Tuesday, March 27, 2007/
How to keep track of computer files? I’ve worked out a simple naming system that tells me the date, version author and content.
The Microsoft Office default name for documents is Document1. Consequently, my Dad has hundred of documents name Document1, spread through hundreds of folders. Because every document has a meaningless name, when he searches his hard disk, the search has to look inside every document to find the one he is missing. This makes the search dramatically slower.
Of course, my father wants to have unique meaningful document names to make life easier, but he doesn’t have a system and consequently resorts to Document1 as soon as he is a bit rushed.
Now I would argue (with him regularly) that from a document name you would want to know the following things:
- Ownership. Which organisation does this document belong to?
- Content. What is inside the document?
- Author. Who wrote it? (in the case of the Churchill Club, not required).
- Version. What version of the document is it?
To make like easier for my father, I have let him in on my system, which has served me well for a decade.
I split every document name into three sections. The first part is ownership (“tcc” for the Churchill Club and “bml” for me). When documents come in from other people, I don’t change the document name as it makes it obvious it’s not for me.
The second part of the file name is the document title (eg, “event-sponsorship-template” or “letter-to-accountant”). Note the hyphens between the words, rather than spaces. I do this because some computer systems don’t like spaces in file names and corrupt the document.
The third part of the file name I use for version control. I use two different types: a date for one-off documents, or a numbering system for controlled documents.
For dates, I write the year, month and day in reverse (eg, 070326 for March 26, 2007). This means that if I have multiple versions of a letter written on different days, they will all sit nicely in a directory, sorted for me in date order. If I update a document on the same day, I normally add an “a” on to the end.
The numbering system for controlled documents is a little more complex. The first major release of a document, I give the number 1.0, the second major release 2.0, etc. Minor releases (small changes) after 1.0, I number 1.1, 1.2, etc. Minor releases after 2.0, I number 2.1, 2.2, etc. Drafts before the first release are numbered 0.1, 0.2, etc.
Using this system a letter to my accountant from me would be called “bml-letter-to-accountant-070326.doc” and a letter to my accountant from the Churchill Club would be called “tcc-letter-to-accountant-070326.doc”. If I wrote three letters to my accountant on different days, and they were all sitting in my accountant’s directory, it could look like:
Now I think this system is fairly simple to remember but powerful, as it means I can write unique, useful file names every time without having to think about it. What does my Dad think? The answer to that is a whole lot more complicated than this blog.
Adam McWhinney from The Federal Publishing Company writes: But then again, don’t fight it, use the search systems developed for you and try naming your documents with relevant keywords (or tags).
Most folks are more likely to remember a topic or theme-related set of keywords rather than a rigid code/based naming convention.
In the example you give, I can’t easily tell what the content inside the document is going to be unless I open it.
The central theme, idea or message described using keywords (think tagging) is far easier to recall than when a document was created or to whom it was sent. Besides, any OS can tell you when you made or modified something.
This concept works a treat for email subject lines too.
Try it, I *guarantee* you’ll love it – or hate it!
PS I bet your Dad loves it. 😉
Brendan Lewis replies: Thanks Adam, but I disagree.
Firstly, you have to remember to use keywords, but are forced to save a filename. Therefore choosing a filename takes less effort and is something you are much more likely to do.
Putting the document title in the filename means I can pick the right docuemnt, just by glancing in a directory. Much quicker than searching a hard disk (or two).
Secondly, putting the document date in the filename is critical for document control. Using the operating system to date your file means that you can’t keep different versions of a document in the same directory. And in the case where is a variety of people working on a document, lack of version control is a recipe for disaster.
Even Wikis, which do this well, keep an archive of old versions of documents.
Thirdly, I see tagging a word document as adding little value since if you are going to go down the search path, why not just search on the documents contents anyway.
However the best system is always the one that works for you. What I offered here is the one that works for me, and has worked for IT services business I have owned.