Estimate
- In preparing for [projects] I have always found that [estimates] are useless, but [estimating] is indispensable.
- It’s better to be roughly right than precisely wrong
We keep on estimating and planning according to those estimates, expecting to meet deadlines and firing the wise fools that question the practice, and report the actual numbers that are not accepted by management. However…
- Insanity: Doing the same thing over and over and expecting different results.
Watch out for:
- Estimates become targets.
Use yesterday’s weather, which implies “results from the past give guarantees for the future”.
Collect metrics about the real system, like average lead time distribution, cumulative flow diagram, average throughput, and predictability.
Thou shalt not treat estimates as commitments.
Setting deadlines and promises based on estimates is dysfunctional.
The problem with estimates isn’t so much the concept of estimates themselves—because they are fine if we know that they are estimates—but that the way we treat them in software is not like an estimate. They drive deadlines and they drive promises and because of this, we get all kinds of problems.
- → metrics drive behavior
- → relative estimation
- http://www.djaa.com/noestimates-beef-and-agiles-trojan-horse
- #noestimates
- http://en.wikipedia.org/wiki/Probabilistic_method
- http://www.infoq.com/articles/software-development-effort-estimation
- http://softwaredevelopmenttoday.blogspot.co.nz/2012/01/story-points-considered-harmful-or-why.html
- http://www.infoq.com/resource/minibooks/emag-agile-estimation/en/pdf/Agile-Project-Estimation-and-Planning-eMag.pdf
- http://focusedobjective.com/forecasting-error-accounting-scope-increase/
See plastic plan.