Bin
The paper balls in the bin: lessons from things that went wrong, each from a post on the blog, as plain text.
The tidiest debt
Future-proofing
We built a docking point for assessments before product had decided what they were, carried it for a year, then pulled it out. Technical clarity is not the same thing as product clarity. The code was perfect. It just had no business being there.
Too clever by half
A clever fetch wrapper
It shouldn't take 2 days to figure out how to upload a file using React. My own wrapper set the JSON header and stringified every body. How did I make these two mistakes? By being clever!
The marathon I never ran
Just see how far I can run
I ran 20 km because I felt fine. My ankle gave out the next day. I never made the marathon - going too hard too soon killed that goal. Same with coding until 2am. Don't keep doing the thing you think you can until you can't.
from: What Running Taught Me About Persistence and Programming · See it in the bin
My spec is the bottleneck
It's the model
The model isn't the bottleneck as often as I initially assumed. My spec is. A vague brief produces vague results, and then I'd spend more time fixing and redirecting than I would have spent just writing the thing myself.
Be better, I guess
setState is asynchronous
A modal that wouldn't close: it sat inside the thing that opened it. It took me and another experienced developer far too long to spot the problem. What did we learn? Not sure. Be better, I guess.
from: setState wasn't working - annoying bug postmortem! · See it in the bin
Stepping on rakes
You're prompting it wrong!
I've slapped poorly constructed prompts together and picked a random model. Then I did it properly, and I'm still only 85% of the way there.