Well, with a single json file I can pull it into my text editor (Sublime) and do find and replace (and it's a little nicer than find in all files) and I can do things like code block folding, and probably others Im forgetting. Also, I'm not sure how to `jq` across 2000 files but how to with one json file. I've also heard that the filesystem is not a great place to store larger amounts of files like, say, 500k records?
My "memex" tries to use well-known abstractions so that I can do certain tasks in other programs until my "memex" can do it. That and because for quite a few operations, I'd like to simply mount the window of a separate program into my "memex" as a sort of window manager (but deeply integrated with the other stuff the "memex" does, like data and tab management) instead of trying to remake such monolithic, decently-built programs
My "memex" tries to use well-known abstractions so that I can do certain tasks in other programs until my "memex" can do it. That and because for quite a few operations, I'd like to simply mount the window of a separate program into my "memex" as a sort of window manager (but deeply integrated with the other stuff the "memex" does, like data and tab management) instead of trying to remake such monolithic, decently-built programs