Managing wicked projects - a survival guide
cc018927-c596-4d64-b07f-9ceabef3a805
|
Presentation |
This talk will focus on a special kind of wicked software projects - enhancement projects. They are very common and generally overlooked. Wicked enhancment projects exhibit many of the following characteristics:
- The project is delivering something like version 5.7 or 6.0 of a successful software product
- There is a large, buggy, legacy code base with little documentation
- There is a large number of product variations (configurations)
- The customer and user base is large and inhomogenous
- The original development team is long gone
- Sponsor/general management has lost interest (until something goes wrong)
- The current development team is doing all of support, O&M, and fresh development
- Activities other than developent consume a significant portion of the effort
- The developer motivation is low (churn is high)
- There is a high number of external dependencies
- Requirements are volatile
- Firefighting is the norm – defined processes are missing or frequently circumvented
Enhancement projects are especially hard because their origin is a successful product. The risk is generally high (chance of jeopardizing the whole customer base), but not treated as such. Success is expected, but many of the mechanisms for ensuring success are not in place.
Outline:
The presentation will give practical advice on how to manage such projects, and will cover:
- What wicked projects are
- The life (and death) of a software product
- Handling support and O&M in parallell with development work
- Improving quality
- Handling bugs, requirements and releases
- Handling estimation, planning and monitoring of progress
- Handling expectation management and communications in general
- Handling external dependencies
We will focus on presenting practices that have proved beneficial in projects of this kind, and that should be part of any project manager's toolbox. Some practices we'll cover are:
- Bug-Driven Development: Using bugs to improve overall code quality
- Don't Estimate - Timebox. When not to estimate
- Don't Judge - Count. Improving estimates by freeing developers from time estimates
- Pay your debt: When and how to work off technical debt
- Two-page Architecture: Sufficient technical documentation
- Information Gatekeepers: Preventing information overload and thrashing
- Prioritized Feature List: Controlling volatile requirements
- Planning Horizons: How much to plan
Language:
English
Level:
Intermediate
Required experience:
Project management (software development)
Participation in software development projects
Expected audience:
Project managers, project owners, other project participants
-
Jørn Ola BirkelandProject manager with Bekk Consulting As,PMP, ScumMaster In depth experience with long-running (large) software projects, in very volatile business contexts Interests include: Software estimation, project management methods, software development best practices



Intermediate
Methodology and Business