Skip to main content

On Estimating

Estimating is the bête noire of software development: simple to conceptualise, but a mine-field for the unwary. There are multiple levels of problems involved with estimating: technical, theoretical, political, social, economic, which we will touch on in our tour.

Goals​

The main take aways from this track are:

  1. An appreciation as to why estimating is so hard in Software Development. We take a step-by-step approach to understanding the nature of the problem, and why it isn't as easy as estimating other, simpler things.

  2. How difficulty estimating impacts our work. Estimating is practically impossible, yet it is a cornerstone of Scrum. Alternatively, the "No Estimates" movement tries to go the other way and dismiss estimating completely. Here we try to reframe the problem as focusing on the wrong thing.

  3. How we can get faster at development, despite the complexity. Some practical tools and advice for planning in the absence of reliable estimates.

Fill-The-Bucket

Part of the 'Estimating' Risk-First Track, in which we look at straightforward extrapolation.

Kitchen Cabinets

Part of the 'Estimating' Risk-First Track, where we look at exponential distributions in estimates.

Journeys

Part of the 'Estimating' Risk-First Track, where we look at estimates with dependent, connected parts.

Fractals

Part of the 'Estimating' Risk-First Track, where we look at estimating with uncertain boundaries.

Analogies

Part of the 'Estimating' Risk-First Track. Analogies for the problems of estimating.

Fixing Scrum

Part of the 'Estimating' Risk-First Track, looking at the essential flaws in Scrums' time-boxing of work.

Risk-First Analysis

Part of the 'Estimating' Risk-First Track, looking at an alternative to Scrum's Planning Poker

On Story Points

Part of the 'Estimating' Risk-First Track, about improving estimates using risk checklists.