Rust, the numbers

This is a look at some number of how long the payback period for Rust is. I'm not going to cite sources of these numbers, but they are fairly widely known in the programming community. The goal is a fermi estimation that might help companies navigate a tricky period.

To the maths!

Lets start with a normal software project. 80% of the tasks take 80% of the time. The other 20% takes 80% of the time. This infers that an average software project requires 160% of the resources to complete. Lets add another 40% for "Project Managment" duties giving us a nice round 200%. For guess-timates sake, we are going to say that second 80% is bug hunting. Over a year then, this means that a developer spends about 20 weeks bug hunting (Big and small ones).

The common sentiment is that it takes 6 weeks to be productive in Rust. Lets have a worst case of 8 weeks. You might be dealling with cutting edge stuff that is tricky to do.

Microsoft has said that 70% of their bugs are memory related. Other sources broadly agree with this number. Lets say that most of these (but not all) will be solved by moving to Rust. We will call it 50% of all bugs. Looking at our time estimate above, that is 10 weeks in a year looking for memory bugs.

So, in our estimation, it is possible that transitioning to Rust will pay for itself within a year. Of course your milage may vary and Rust isn't suitable for every project, but give it a go and understand the joy of compiler driven development.