Thank you for sharing. I have a system that is kind of like this. Except it's not meant for memory as much as analysis. And I don't consider the entries "articles" but rather refer to them as "models" as in mental models. Like a real life model, say a toy car, I try to cover enough bases that it--metaphorically--at least _resembles_ a car at first. Then as I identify needed leverage points in the model, I refine my analysis and expand the model. So maybe at that point it has a hood you can open, and an engine you can see, so to speak.
I think I'm at or near around 800 of these, and many are very short. But no matter how short they are, they are all there because they provided me or continue to provide me with needed leverage.
I keep the bulk of the information in markdown in a Dropbox folder and also occasionally try new methods. For example, for topics that will quickly benefit from hyperlinks, I developed a LibreOffice web template and a companion PHP script that indexes these files and inserts additional CSS, variables, JS, etc. when they are served up. For searching I like Regexxer a lot, but I also use grep quite a bit :-)
On my XFCE desktop are buttons for opening a random one of these files, and for opening a random journal entry to try to harvest new models, so to speak.
And there's some paper involved here too... Can't get away from it, because paper has its own special leverage points...
I'm not concerned about memory, knowing it's a special weakness of mine. If I'm working in a context where memory is super important, my energy is best spent moving to a more sustainable context. :-) Memory is a hobby for me, a side gig for memorizing pi, that kind of simple and fun thing.
