[jancsika]

It's worth reflecting on just how many VCS data (or pain) points Linus had ingested right before writing git. Probably more than anyone in FOSS outside of maybe a few Debian devs. Add to that his experience successfully using Bitkeeper prior to that and you can easily see why git is where it is in 2026.

Given a large enough amount of data/pain, designing/optimizing to attack specific, known pain points always beats trying to solving a more general problem elegantly. I mean, kudos to whoever decided Zoom clients with shoddy connections should buffer then race back to realtime at 1.x - 2x speed (can't remember exactly how fast it goes-- perhaps it's dynamic?). One could come up with 1000 toy examples of where that breaks (music lesson, drama class, etc.), or just implement it and save a gazillion people gazillion hours of repeating themselves in boring meetings.

Edit: clarification