When designing a back-end system, what granularity do you normally give your stories and tasks?
Most examples of creating stories and tasks usually center around a GUI application with the story being something the user can do (e.g. search for a book by the ISBN) and each task centered around enabling this GUI feature.
When designing a back-end system, i.e. one that doesn’t have a user interface but is just a bunch of services talking to databases, middleware, etc. how do you come up with tasks and stories?
Basically, I try to keep the size of my User Stories in the area of 1 to 10 man-days to complete. That keeps me from passing what Mike Cohn calls “Epics” or “Themes” as user stories to the developers, and on the other size stopping my user-stories to be so specific as to imply the solution (they should be describing the problem, not how it should be solved).
As far as content goes, I make sure that my stories contain only business value – it never describes how I (should) satisfy the demand, nor does it “require” non-user-domain knowledge to understand.
Good example: As a content manager, I want all users to have to log in before writing a talk-back, in order to deny them the ability to spam.
Bad example: Add captcha to the web site.
Tasks, on the other hand, are steps towards solving the solution – they describe components & functionality that need to be added / modified. This is where an “Add Captcha” solution comes in. As far as size goes, I try to have each task‘s size be between 1/2 a day and 2-3 days of work.
Tasks also include a set of standard tasks that apply to each and every feature / requirement / problem / bug, such as:
etc.
Hope this helps,
Assaf.