Making projections is something we all do all...

Making projections is something we all do all the time, from projecting what the weather will be in a few hours to who will win the Super Bowl. In our work as software engineers, though, we are also making a lot of projections for all of our work. " "" "Estimating the time to write code is itself more an art than a science, and getting good at creating these projections is essential to success as a software engineer. There are several elements that I tend to think about when I make estimates:" "What's the best case and worst case amount of time this could take?" "How likely is the best case? Similarly, how likely is the worst case?" "How much time do I think it'll take me to code" "How much time will this take to release and be in production so that I can quote an estimate for time to production?" "How busy am I over the next 3 months, should this land on my plate?" "" "These considerations help form a number of days/weeks, which is typically considerably more than just the raw time to code. Working in a big company with large codebases, this is somewhat expected. It takes more time to figure out where to make a change and plan it out than actually to execute that change. " "" "Estimates like these are all about setting expectations at the end of the day. If you can consistently underpromise and over-deliver, that's the goal. And as was said in Superforcasting, it takes practice in a field of projection to get better. So continuously tune your estimates as you go, review how close you were, and continuously try taking a slightly more aggressive approach until you fail, then start overestimating again and start the cycle over! #bendevtip

---

Original post on LinkedIn