(Last modified Thu Apr 17 22:42 2008)
Why review?
- Can review sooner than can test.
- Immediate benefits: find problems in current work.
- Indirect benefits: improve people's skills and consolidate the team.
How to review
- Assemble a team of appropriate people to do the review.
- peers at the same level, or
- senior experienced pros, or
- outside experts (for expertise, authority, and scapegoating)
- Except for peer inspections within a team,
include people from other areas of expertise
as well as experts in the area of the artifact being reviewed.
- Schedule it soon.
- Assign some roles.
- The presenter who presents the material
(not necessarily the material's author)
- The scribe who records
action items and important points of the discussion.
- Someone must be in charge (usually the presenter)
- Keep the review meeting on track.
- Make an agenda (items to review and tasks to accomplish) and follow it.
- Start on time and end on time.
- Use the time budget to decide when to move debate offline,
and to enforce the cutoff.
- Produce action items and plans for each identified problem.
- Follow up (after the meeting) on each action item.
Pressman's "Golden Guidelines"
- Accumulate and use checklists for each kind of artifact being reviewed.
- Train senior people to review well.
- Analyze effectiveness of past reviews and make improvements.
- Schedule reviews and allocate resources for them
as part of the development plan.
- 3-5 person review team is optimal.
- Be constructive and not personal.
- Keep to the agenda.
- Focus on defects, not on solutions.
- Can't agree? Make a note, plan to discuss it later, and move on.
- Document the discussion.
- Two hours max.
- Write up a report.
- Follow up.
