From User Stories Applied: For Agile Software Development (link goes to Safari Online

Overview

The task of disaggregating a User Story in to a set of Tasks for backlog is according to the author one most developers have been doing for years.

The focus on disaggregating the User Story should be in translating the User-centric/focused information from the User Story into a set of actionable taks by potentially several different programmers/programmer teams.

The author gives a set of examples from the User Story "A user can search for a hotel on various fields" which can be broken down into

• code basic search screen
• code advanced search screen
• code results screen
• write and tune SQL to query the database for basic searches
• write and tune SQL to query the database for advanced searches
• document new functionality in help system and user's guide

Most of this example turns from a user focused User Story into an almost purely technical list of requirements for programmers, however it is important to note that there is an element for "document new functionality in help system and user's guide". It is critical to keep the user's help/documentation in mind throughout the development process. In some ways each of the action items for the backlog should have the "document new functionality in help system and user's guide" item attached to end of it. The author removes the neccessity for this step by staying that items from the backlog should be pulled one User Story at a time.

Guidelines

The author also gives several general guidelines as to how to go about partitioning the User Story into seperate items.

• Tasks of a story which are particularly difficult to estimate (for example, a list of supported data formats requires approval from a person who is slow to respond),should be seperated from the rest of the tasks.
• Tasks which could could easily be done by separate developers should be split
• If there is benefit in knowing that a part of the story is done, break that part out as a task. Delays in the completion of one task does not delay completion of the other potentially dependant tasks.