1. Updated from: Critical Decisions in Software Development: Updating the
Updated from: Critical Decisions in Software Development: Updating the State of the Practice, IEEE SW Sept p84-87

2. Background
Global survey of completed software projects that attempts to shed light on international differences in the adoption of different development practices.
Sample isn’t a true random sample. To initially solicit responses:
ed a worldwide list of industry specialists and researchers and asked them to forward our request for participation to their colleagues and industry contacts.
approached software associations and research organizations (such as the Fraunhofer Institute) and software-focused trade journals (such as InformationWeek) to promote participation among their members and readers.

3. Background Participating companies
Japan
Motorola India Electronics, Infosys, Tata Consulting, and Patni Computing from India; Hitachi, NEC, IBM Japan, NTT Data, SRA, Matsushita Electronics, Omron, Fuji Xerox, and Olympus US
IBM, Hewlett-Packard, Sun Microsystems, Microsoft, Siebel Systems, AT&T, Fidelity Investments, Merrill Lynch, Lockheed Martin, TRW, and Micron Technology;
Europe
Siemens, Business Objects, and Nokia from Europe.
Survey conducted using a secure web server at MIT
104 projects

4. Background Results might not generalize to these broader populations.
Survey required a significant amount of detailed project data, projects under greater managerial control
Better managed projects are more likely to be included in the sample.
Results might represent better, rather than average, project performance.

5. Project Descriptions

6. Practices Used

7. Performance Data Output per programmer-month of effort in new LOC
Number of defects reported by customers per 1000 LOC in the 12 months after delivery to customers

8. Update No one “best practice” fits all projects and situations.
Outsourcing Issues
For every 1% increase in the dispersion of tasks to outsourced locations, productivity decreased by about 0.7% and quality by about 1.5%.
Investments in structured processes, and the learning that comes from their adoption, can mitigate declines in productivity and quality.
Must weigh the savings from distributing work to lower-cost locations against the possible loss of productivity and quality
Updated from: Critical Decisions in Software Development: Updating the State of the Practice, IEEE SW Sept p84-87