I worked at a company with a senior (Ex Amazon old dude with a PhD in com-sci and math) coder who every day would update an internal blog about the code he had written that day.
Although it was occasionally a bit difficult to search, it turned out to be amazing documentation for the code he was written. Just volumes of text about what was a fairly complicated subsystem involving machine learning and a bunch of stuff like that.
It gave an insight into the thinking process of someone with 30 years of experience solving some insanely complicated problems, but he insisted it was mostly for his own benefit to organize and clarify his thoughts about a giant project that was changing rapidly as business pressures piled on (The joys of startups).
I had the exact same thing with an older job. About 15 years ago I got hired to finish a PHP project that the original developer had basically run off screaming from, and oh boy was it a super amateurish mess. No use of the OOP system, and very little in the way of "structured" coding, it just ran from the top of the page to the bottom. (I call this style "long drop programming", your code starts from the top, rapidly proceeds down, and that the bottom is a dead programmer in a noose. Oh and lots and lots of just bad SQL practice. SQL injections, joining tables in memory instead of the DB etc.
So I went through the main pages and rewrote much of the logic as single SQL queries, admittedly rather complex ones, reorganized the code and broke it out into functional modules. The resulting code was about 1/3 of the size of the original.
About a month into the task, I got dragged into management where a very confused project manager asked me why my LOC was negative. I told them "Because I actually know what I am doing". The guy seemed unconvinced until I showed them the original code and mine, and the benchmarks of the two. So I ended up the only guy in a team of 30 coders excempt from their idiotic LOC rules.
And then a month later the 2007 stock market crash happened and that company almost immediately curled its toes and died. Possibly for the best.
Although it was occasionally a bit difficult to search, it turned out to be amazing documentation for the code he was written. Just volumes of text about what was a fairly complicated subsystem involving machine learning and a bunch of stuff like that.
It gave an insight into the thinking process of someone with 30 years of experience solving some insanely complicated problems, but he insisted it was mostly for his own benefit to organize and clarify his thoughts about a giant project that was changing rapidly as business pressures piled on (The joys of startups).
I really ought do the same.