For many years now, I have been maintaining an ever growing, cross-linked personal memo system.
It is, if you like, my own personal knowledge database. The website you are looking at is, to a large extent, a "view", in database terms, on this material.
I guess you could call my whole memo system a mind map. But it started out purely as a filing system.
From childhood on, I used to collect clippings and notes. First in shoeboxes, then in scrapbooks, then in binders, then in folders, then in filing cabinets, only to find that in the end systematically grouping the material was bound to fail. I was forever drowning in a deluge of old and new information.
I found out later that many of my colleagues had suffered the same problem, and had taken refuge in a chronological system of notebooks and reports in which they found their way by remembering in which month of which year they had worked on a problem. One of them referred to his desk as "my piling system".
Since I barely know what day it is today, this system could never work for me.
Any order other than chronological is prone to redundancy. Most items belong to different classes. A single thought or clipping can simultaneously belong to the categories of aircraft, mathematics, mechanical engineering, design methodology, and career planning.
I started making copies of the same article for different folders, and then could not locate the copy that I had scribbled my latest notes on.
Shortly after graduating from university I followed a course on database normalization.
The principle is that every data item ( someone's first name, their home address, etc. ) occurs only once in the entire database. Each item gets a meaningless but unique identifier, usually a long number, and all other items are not allowed to repeat the data item, only refer to its identifyer. If someone moves house, their home address is exactly the one item in the database that is updated. All the references to their home address stay the same, so the database is always consistent.
This became a clear inspiration to me to maintain only a single memo on every subject, and handle the relations between the subjects purely by cross-referencing between the various memo's, never by repeating the information or adding to it, in a duplicate memo.
At some point, personal computers and word processors were becoming affordable. Around 1996 I decided to organize my material, first work related only and later everything in sight, into a system of Word documents linked back and forth by hyperlinks.
Typically, every memo starts with a list of between 5 and 50 links to various other memo's. There is a central memo called "index.doc", which is an ( inverse ) chronological list of all the others.
The contents of these memo's vary from the yearly memo with doggerel poetry for that year's St. Nicholas presents, to 100+ page treatises on arcane subjects.
The proper granularity of these memo's is an issue. I find that typically, sticking to a clear subject per memo limits the size to about 5 pages, which can be taken in at a glance. It works best for me to start a new memo for every subject which can be given a clear title, whether a visit to a particular lab, or a particular piece of analysis.
I guess this is related to a rule of thumb I once learned from a very good software engineer. The criterion for separating off a piece of code into a subroutine or a class is not size. It is that you can give the new code object a clear and meaningful name. The new object may be anything between two lines of code, or two pages of A4. It is the self explanatory nature of its functionality that matters.
Over the years, my memo system has become an asset and a liability to me. At times, the overhead of starting a new memo for every little subject has become oppressive. At other times, I have been very happy that I could find notes and analyses from 30 years ago with ease, and link new subjects quickly to several other building blocks.
I am not sure that I can recommend a memo system like this to anybody else. But it has worked for me.