...
earthdatadev@conferenceedsc@conference.im.nasa.gov
. Developers use Messages or Screen Hero for remote screen sharing....
We have short daily stand-ups near the start of the work day (8:45 AM) to give status and note any problems.
...
We estimate stories on an exponential scaleusing the Fibonacci sequence, 1, 2, 3, 45, or 8 points. These units do not correspond directly to developer-hours.
...
Our estimation may differ from typical agile projects in the following ways:
We only assign points to stories. Epics are merely an aggregation of stories. Tasks provide no user-visible value. Bugs and improvements are necessary changes that should have been done as part of the original story. Assigning points to bugs or improvements is essentially double-dipping and ruins velocity metrics.
To see this, imagine a backlog with 100 points worth of stories. Let's say we're assigning points to bugs and have an average of 1 point of bugs for every 2 points of stories. Our velocity is 15 points per sprint. We look at the backlog and think we'll be done in about (100 / 15) = 7 sprints. That's wrong, though. 5 points of our velocity is due to fixing bugs, and the backlog doesn't contain bugs. We are assuming that we will write bug-free code for the next 7 sprints, despite evidence to the contrary.
Now assume we don't assign points to bugs. Our velocity is lower, only 10, because we're spending time fixing bugs we've introduced. It also reflects our defect rate. We are finishing 10 story points in a sprint and however many bugs we've introduced. We look at our backlog and say we'll be done in about (100 / 10) = 10 sprints, because our velocity accurately reflects the time spent fixing bugs we're introducing along the way.
Bugs are tackled immediately. In order for the above justifications to work, we Bugs are tackled immediately. We cannot have a huge backlog of bugs. If we've marked something done and it is defective, we fix it. Quality is non-negotiable.
We do not re-estimate stories. The justification for this is similar to the reasons above. Re Re-estimating stories changes the inputs to the backlog and throws off the numbers. Consistency is far more important than accuracy. Velocity will sort it all out.
...