Well this explains why, when I first started to learn to program, it took me hours to find one missing semicolon in 15 lines of C code, but after 20 years I can basically glance at a screen full of code and find them almost immediately.
So I rearranged the code slightly and it was immediately obvious he'd mismatched his closing parentheses.
It wasn't conscious. I just knew it was wrong.
When I worked in a computer lab, I'd often walk past kids writing code in notepad, peer over their shoulder and spot imbalaced parentheses or a missing semicolon. I'd keep it to myself unless they asked for help, because it's a learning process... and then they'd be amazed as I immediately found the spot.
And sure, they probably save you time. But for me, they add visual clutter and only spot trivial issues, and rarely save me a few seconds per day.
It amazes me mostly because this is not the sort of tasks our brains were specialized in doing, by estimating our development path in nature as mammals.
We discover and/or create new models and yet, we are able to adapt to them with some effort. Our new environments will probably affect the DNA that have the brain design, and we will pass all of this as a legacy to our kind.
The brain is such a masterpiece of nature.
concievably this might be similar to spatial navigation, which would be something we'd evolve for as hunter/gatherers
But theres another catch, the tools we use change the shape of our mindset, leading us to design other tools that now map to this new perception of reality we have with the new mindset we built by using the former tools.
The hunter-gatherer example given in the first comment for instance, i bet with enough evolutionary steps, we will not be able to trace back to that mindset, because we have "patched" our own mindset so many times with new tools(over tools) that we will have very little "mindset identity" to share with our former "hunters-gatherers" primordial cousins.
You are brave to handle that much code, imo.
Of course theres a erosion process if you dont touch that piece of code for more than 1 year, and it happens with a codebase of that side, but the 'partial data' is is still there, and with a couple of minutes dealing with it, you recover the information.
Its better when you shape the code yourself.. when you lay down the objects, methods, etc.. But this also works great with code i've happen to just use.
For the architecture and where things are.. be it files, methods, or where i could find something i want.. there were no lost.
I feel i need this partial recovery process mostly for codebases with 2 or more years, and the lost is minimal and can be recovered quickly.
Now being more specific over your question, for me it fells more like a map, smaller maps or trees connected to top-level maps... But yes, it feels a little like rooms as in a quadtree, where you need to specialize and ignore the other parts of the tree for a bit, much like the occlusion/visibility process that algorithm gives.
By the way is great that you have used this analogy, because this is also part of an observation i've had about laying down the information for a specific goal i have in mind.
I cant wait to experiment with this as soon as i finish this thing im working on, to see if its really cool as my intuition is telling me..
After a decade of experience I happily dump all the boring stuff to type checkers and linters.
One day in the morning I was being in a state of half dreaming - half awake, and among other closed-eyes visuals I observed a series of white screens (sort of 80x25), filled with the same digit, like all zeroes or all threes.
I don't know if those were the layers of my brain's NN responsible for recognizing texts. But I generally suck at visual navigation: I don't rely on code-minimaps, scroll-bars, I don't remember if that particular function is above or below. I use tools like quick-jump, fuzzy search, I just need to run a thread of interest through my mind and recall instantly what to search.