I'm rather tired, but here are a few early ideas to play with, because I've been low-level for much of my life, and as my health broke down too much to work for quite a while, I spend my good days doing my own unfocused private research on topics that interest me - some enduring, some Weekly Specials.
I've also fiddled with text processing, with partial tools, though never with any reliability requirements.
Glancing at this "Big German Word" ...
- It seems your mind needs some level of innate talent to make it work, or it could collapse entirely! I don't think you can both be "surprised at what's in your notes" and at the same time have a comprehensive command of the material.
- Tags: I think I put mine right into the file names, so the tags travel along. I think I've grokked the thread we don't like native Windows search. But something I've done in the past was use file directory readers, and then crank the output text file into something else for a refined search. No scripting needed. If you think you've changed more than a trivial bunch files, both new and revised, just re-run the directory. "Can't find the file" - that will guarantee it's FOUND, it just may have a name you weren't expecting if you don't have that talent for consistency!
- More fun things to do with file reader output : if you put a special character (not above the number key but like zz), after the "regular" part of the file name and before all your tag-y things, then you can import that directory output into Excel, and chop it up into sections and then your notes can reference the fragment of the file name.
- In reverse, if you want to share stuff on the web, and you added new tags to your file names, because now it references Cher Banb Bang for "I was Five" and Harlan Ellison's story Jefty is Five, re-run the directory reader, then (I did one in Excel) it concatenates back the cels, slaps some shell code on either side to make it a legible web page, and off you go, and your text files drove the outbound web copy.
- False Negatives: If you think your Tags are 4 leaf clovers, they might obscure some nice creativity because only so many tags can be sanely added to something. The more complex the partial thought is, it might slip away in tags. So if I got to daydreaming, "hey, what was that book about the MMA fighter, where his raw identity was based on that area code, I forget which one, it's in that midwest somewhere"... how do you tag that? At the time you might have tagged fighter, boy, and a couple other things, and it's competing wiht JAckie Chan movies and Mortal Kombat.
So maybe You just write periodic summaries of what you were up to, and then you might get clever recalling what phase of your life it was in. There's a risk, it seems to me, of system decay.
- DonationCoder! That's us! If you find you need some really strange processing of text files, it might NOT exist! So commission it! In the days before certain modern computing chess trends, and even now it's a hobby so I'm old school, I had a problem. In chess, the "text" version of the moves is called a PGN, which is a straight text file, just with headers in brackets in known sequence for the computers to know who played and when. But those headers were a pain, taking up half the page, and some sources had certain hardcoded line breaks and stuff. And I wanted very basic filtering so if I am looking at Tal's games, no games under say 15 moves (tended to be last rounds etc), and no draws. I didn't want a huge database program - same situation as you. (Chessbase used to be the gold standard.) I just wanted to take my little text file I produced in ten minutes to a restaurant.
So I commissioned it elsewhere, (got Mouser's approval it's not for resale and I own the end to end work for hire rights etc) and it's here.
NANY 2012 Release: Chess PGN File Processor http://www.donationc....msg272545#msg272545
(I think there's 1 file that needs to be accounted for on some systems, but I got it back working a few months ago).
So instead of complex scripting, let's say the academic world has these quirks to its materials, they're not conceptually hard, but annoying, maybe get a Coding Snack!
So someone was talking about parallel files, the snack scans the material, reverse sorts by noun and like my program truncates somewhere, then cranks out a parallel "super tags" file, so if you're trying to remember that program with the miners you spent a week analyzing, and remembered the dev changed DBA names, you crank the utility, it skips over Twitch presenter David Miner, and comes up with Ludum Dare Undermined.
So I hope some of that helps!