[dismantlethesun]

Understanding the computer science underpinnings are fairly useful. For instance, what if you are unhappy with the rendering speed of React? Rather than faff about complaining about it, you can just pull out your textbook, re-read that chapter on graph algorithms and you too can implement your own virtual DOM that calculates the least set of changes to apply to the directed graph that is the browser DOM.

That's how we end up with projects like Preact, Inferno, etc.

Now does that mean you need to read the latest papers, and be a proper computer scientist? Nope. I find that reading papers from the 70s and 80s will take you pretty far. Just rip and reapply to new circumstances.

I am a web developer.

A lot of code that I write is 'good', but its hardly clever or original---it's a simple reapplication of old practiced done 100 times in desktop GUI applications, now applied to the web.

I found that a lot of old-school GUI developers never really crossed over to the web, so a lot of that knowledge has been lost to the community at large. You can still find it in text books though :)