In a recent post, I stated that some aspects of the #NoEstimates approach is not yet a professional Software Development approach. I believe it could become a professional Software Development approach, but it isn’t one yet. Some people raised an eyebrow to this statement and rightfully so. When you drop a bombshell like that you certainly should provide rationale and criteria as to how you ended up with that opinion.
Remember that this is also just my opinion, so do with it what you will. 🙂
Here we go!
1. The proposal and hypothesis should be clear, concise, and consistent
God help me, I really do have problem with knowing what #NoEstimates is proposing. Sometimes it is not doing any estimates – but do then again do them if you need to. Sometimes it is just working on a prioritized backlog and delivering in an incremental way delivering the highest priority items first. I find this ambiguity hurts the #NoEstimates cause more than it helps. If you believe in something, propose exactly what it is and stick to it. A lot of times we really don’t know what you are proposing that we should do. If you firmly believe no one should do estimates, then say that and stick to it.
2. Don’t just criticize alternative theories
All the great ideas over time were able to stand up without criticizing alternative theories. Einstein never got on a soapbox and criticized Newtonian Physics. He simply proposed a new theory and illustrated why it described reality in a better way. The #NoEstimates proponents are actually hurting the acceptance by calling other methods ‘insane’ and ‘crazy’. It is almost the equivalent of swearing. Swearing is the easy way out of rationally describing your position to convince others of how it is a superior approach. To be clear, I don’t believe that a big-bang waterfall approach is better than Agile. But calling it ‘insane’ is intellectually lazy. I should be able to easily prove that to you. (and I can) Sometimes the discussions seem to resemble political attack ads more than anything else.
3. The facts should be available for peer review
Many times in discussions with #NoEstimates proponents I have seen opponents ask for the established companies that are using the extreme #NoEstimates processes of not estimating at all. Each time I usually get the response that they are soon going to interview them soon and would publish the results. I still haven’t seen any numbers published from companies that don’t estimate at all. So it is very difficult to validate the claims and whether the process would work in other situations.
In addition to these three characteristics, the inherent value of an approach is how wide the applicability is in the real world. If an approach is only viable in a controlled lab environment, then the approach is of not much use in the wild. For my money, an approach that is only applicable in a product company addressing only a very small share of the Software Development industry.
It probably is controversial, but the #NoEstimates movement has not yet demonstrated the widespread applicability in the Software Development Industry.