The Martian principles for successful enterprise systems : 20 lessons learned from NASA's Mars Exploration Rover Mission

Includes index.

About the Author. Preface. Acknowledgments. Introduction. Part 1 The Martian Principles. Principle 1 Don't reinvent the wheel. Someone Else Has Already Solved Your Problem. Understand What Your Added Value Is. Use Commercial Software Whenever Practicable. Principle 2 You won't do better than what's already been done. Adhere to Industry Standards and Best Practices. Seek User Groups, Chat Forums, Online Documentation, Books, and the Like. Do Not Gum Up the Plumbing. Principle 3 Your customers don't know what they want. Do Not Push Too Hard on the Requirements. Do Rapid Prototyping and Lots of User Testing. Customers Are the Best Testers. Principle 4 Get something working as soon as possible. The First End-to-End Thread Is Critical. Working Code Builds Confidence. Always Build on Top of Working Code. Principle 5 Use sound software engineering practices. Use a Component-Based Architecture. Use Design Patterns. Get All the Development Team Members to Agree. Principle 6 Don't trust the client applications. Be Very Paranoid-They Are Out to Get You. If You Do Not Break It, They Will. Principle 7 Plan to make changes. Do Not Hard-Code Values. Use External Editable Parameter Files. Implement a "Read Parameters" Method. Maintain Client Parameter Files Centrally on the Server. Identify the Parameter Values. Principle 8 You can't predict the future. Make Each Service Dynamically Reconfigurable. Create Field-Replaceable, Plug-and-Play Services. Hot Redeployment Allows Reconfiguring without Rebooting. Make Each Service Loosely Coupled. Take a Peek into the Future. Principle 9 Don't tie your services into knots. Keep Your Services Independent of Each Other. Services Should Treat Other Services As Equals. Principle 10 Build early, build often! The Major Challenge Is Not Code Development but Code Integration. Use a Source Code Repository. Maintain a Separate Environment in Which to Build and Deploy. Principle 11 "What middleware?" should be your greatest compliment. The Middleware Should Be Invisible to Users. Good Middleware Creates Ideal Virtual Worlds for the End Users. Principle 12 Expose the invisible. Put Hooks in the Middleware. Do Runtime, Real-Time Monitoring. Principle 13 Log everything. Do Not Turn Off Logging in Your Production Code. "Log Mining" Reveals Usage Patterns. Principle 14 Know the data. Learn the Data Usage Patterns. Create Appropriate Application and Middleware Data Models. Map to Practical Physical Data Models. Adapt to Third-Party and Legacy Data. Principle 15 Know when it will break. Do Lots of Stress Testing. If You Do Not Find Out What the Limits Are, Your Users Surely Will. Principle 16 Don't fail due to unexpected success. Missions May Last Longer Than Expected. Data Repositories May Grow Larger Than Planned. Part 2 Project Management and Software Engineering. Principle 17 Strong leadership drives a project to success. A Good Architect Must Also Be a Good Leader. Any Architecture Is Only As Good As Its Implementation. Strong Project Management Is Necessary for Success. Project Milestones Are Opportunities for Demos and Rebalancing. The Project Milestones Near the End Allow You to Get Your Project Done on Schedule. Principle 18 Don't ignore people issues. Software Projects Are Not Democracies. Agree to Disagree, but Then Move On. Scale the Project According to the Team Members' Abilities and Experience. Do Not Be a Slave to the Latest Project Management Methodology. Foster Good Communication Plus Teamwork, Teamwork, Teamwork. Remove Team Members Who Cannot or Will Not Perform. Principle 19 Software engineering is all about the D's. Discovery. Diplomacy. Definition. Design. Development. Debugging. Documentation. Deployment. Dmaintenance. Principle 20 The formulas for success aren't complicated. Successful Architect = Good Designer + Good Developer + Good Leader. Successful System = Good Architecture + Good Software Engineering. Keep It Simple! Index.

Survival data mining is the adaptation of survival analysis techniques for mining customer databases. Customer history data can be used for building predictive models of time-dependent outcomes such as churn and product upgrade.