[Posting type: slightly nerdy]

Start putting dates in front of file names and in documents. It helps finding which version of your essay you opened (extra tip: use the "Save copy..." function to create dated backups). Once you've read this post, you'll understand how simple and hair-saving this can be.

Timeless web

In the course of my work I am called to do some searching on the internet... and in the course of my lesiure too, mind you. And there's one thing that annoys me - sometimes short of exasperating, but hey, you can't get totally worked up about everything can you?

It's astonishing how many documents get published without a date. They get archived for posterity, with the title "recent news" with references like "software to be published next June" or "artist playing next Saturday". Sweet. Until you realize that the document is back dated by a year.

Endianness

It gets better: some pages actually do display dates - "Exhibit will end 5/4/08" That's good to know. But is that 5th of April, or 4th of May?

At our office, we never hear the end of it. Switching between the case management system, the email client, the customer's email client, and the document reference system, not to mention all others, we end up looking at a date and saying "Oh that case is okay, it's only been opened since the 12th of March". No, sir. 3rd of November. It's ancient.

I don't see Americans adopting the European format, nor the inverse for that matter (had I my way, we would all be writing yy/mm/dd to follow on with hh:mm:ss)

"Middle-endian": adj. Used of perverse byte orders [...] Non-US hackers use this term to describe the American mm/dd/yy style of writing dates - Europeans write little-endian dd/mm/yy, and Japanese use big-endian yy/mm/dd for Western dates.

See endianness for discussion on endianness in byte ordering. It's can be slightly annoying when you're programming low-level code.

But back to dating documents.

Unmistakeable: format it.

There's a more pragmatic way of dating a document that will please the human, the computer and the US/non-US/Japan divide on how to write out dates:

• The day of the month can be expressed in 2 digits.


• The month can be expressed generally in 2 digits or 3 letters


• The year can be expressed in 4 digits.

Let's use both the 2 digits followed by the 3 letters for the month.

Now tell me you can't read one of the three formats:

USA: 01JAN-04-2008
EUR: 04-01JAN-2008
JAP: 2008-01JAN-04

You could even do it Spanish, where January is "Enero":
04-01ENE-2008

There's no doubt that it must be January, since the month is th one with 2 digits and 3 letters

Get the computer involved

Now why those three formats, for each part of the date, specifically?

Well it makes it easy for the human eye to read, and also for a computer to read. It remains concise, and even if the month is not expressed in the language you're used to, you still can't get it wrong.

In computing terms, you'd be looking for a match to one of the following patterns (regular expressions):

([0-9]{2})-([0-9]{2})[a-zA-Z]{3}-([0-9]{4})
([0-9]{2})[a-zA-Z]{3}-([0-9]{2})-([0-9]{4})
([0-9]{4})-([0-9]{2})[a-zA-Z]{3}-([0-9]{2})

Even when sorting folders on a computer it makes it easier: compare 20080904-birthday_photos with 2008-09SEP-04-birthday_photos which not only gives a clear dating before the folder is opened, but also allows name sorting (uploading photos to a web share for example nearly always forgoes the creatin date and relies on file names)

More nerdy people like me would also be able to reference items by writing crawler scripts to extract dates to check relevance - disregarding languages or date formats. Because all could be recognized (we say "parsed" in the computing world) without any ambiguity.

So there we have it. STANDARDIZE AND TIMESTAMP. So that we may concentrate more on the content of our documents with the assurance we aren't being fed old info.