Konubinix' opinionated web of thoughts

#ModernAgileShow 25 | Interview With Vasco Duarte, #NoEstimates - YouTube

Fleeting

  • External reference:

It is advised to avoid the trap of thinking about a big and long solution and go back to the problem space. Then try to make micro experiments to find out whether the hypothesis of the problem hold. That way your client can save a lot of money by figuring out that they don’t actually want X or Y.

Then, perform the micro experiments that take 1 or 2 days, faking most of the product, but giving placeholders to be implemented.

If you have the fake product, you can now make small increments to substitute fake parts of the fake product by real ones.

The fake product is very useful, because you already can gather feedback from the users. For the user’s point of view, it is real. You already provided value to them.

Let’s stop focusing on the end date and do high speed experimentation to gain high speed learning.

In terms of scrum, estimation is an active impediment preventing from adding value.

Why are we trying to estimate some big stuff that we know won’t be the end result (because we have always changed our plans in the past)? Why focusing on some 5 years old spec that is most likely obsolete? Why not instead focus on very smaller time horizons that we are more confident reflect what we want?

Lost art of agile is to find the core hypothesis that may help decide whether we continue on that lead. we want to go towards incremental funding, not incremental delivery.

The main issue with the backlog is that it becomes an end instead of being an instrument (measure-objective inversion).

A heuristic is that if you remember it, it is important. If you forget it, it is not. This is ok because you are always discussing with the clients and therefore the clients and the coworkers becomes the trusted system, not the product backlog. The backlog became a mental disease that prevents us from forgetting the bad ideas.

mental disease that prevents us from forgetting the bad ideas

we want to go towards incremental funding, not incremental delivery

why are we trying to estimate some big stuff that we know won’t be the end result

When someone asks how much time something will take, I can tell how much a similar task took and when I will start working on it. Like a Disney land queue.

Notes linking here