The biographer of Robert Moses, Robert A. Caro, refers to the bridges and underpasses of the famed New York State parkways. These bridges and underpasses are quite low, intentionally specified by Moses to allow only private cars to pass. All those who traveled by bus because they were poor or black or both were barred from the use and enjoyment of the parkland and its "public amenities" by the technical design of the bridges. Even at the time of Robert Moses, a political statement of the form "We don't want them blacks in our parks" would have been unacceptable in New York State. But a technological expression of the same prejudice appeared to be all right. Of course, to the public the intent of the design became evident only after it was executed, and then the bridges were there.
In our use of digital and analogue filing tools, we classify information through folders. An article about railway construction gets filed under ‘infrastructure’ or ‘transport’. In Evernote we tag it with ‘rail’ or ‘construction’. This is thinking like a librarian and not like a writer. We are classifying the information as an input. The reason you take notes as a writer is to produce content. It makes sense, then, to take notes in line with this goal.
Traditional filing like this tends to fail when you attempt to write your content. You are stuck trying to figure out which categories will be relevant for your proposal, paper or blog post. Interesting writing often comes from connecting separate fields through a common idea. By revealing the common denominator. By unifying two seemingly-contradictory ideas. How can you possibly achieve this if you’re looking in the same category for your information? The categories simply do not fulfil the function required by the writer.
The notes you take and indeed, the way you process information, should be with a specific project or idea in mind. You must classify information in terms of its outputs. When you take notes on a book, think about how this could apply to a specific idea you had or how it argues against a paper you read last week. The premise is that you should be organising by context and always trying to connect the dots between the content you're consuming.