I remember reading a blog post from a certain project management software startup bragging about their genius lead engineer who created this wonderful NIH frontend framework for them, how it worked, et c.

Having used the product (which would not-infrequently manage to make keyboard input lag by full seconds on strong hardware, ate hundreds of megabytes of ram and tons of processor cycles while idle—even worse than your average webshit "app") and putting 2 and 2 together, I'm pretty sure I could have written a list of things they'd failed to take into account, and reasons the results were so poor. It was a fun read thanks to all the details, like "oh, that's why this is so awful, I see exactly what's wrong now".