2. A Quick Story
After converting ten classes in a day, I thought - this is likely to take about 30 work-days to do. And, sure enough, after about 28 days, I was done. It didn't really matter that I started with the easy classes first, because I got better at resolving the issues in the conversion process as I went along.
The only problem was, having got there, I realised I needed a load of other stuff to get things working in the browser too. So that took about another 10 days. Within the narrow definition of the work "converting everything to Kotlin" I hit the estimate, but the wider task of "getting it to run in the browser" I was late.
Arguably, the failure here was that I didn't set the success criteria properly at the start.