Nyere forskningsresultater som er viktige for software arkitekten¶
Recent research results important for software architects
Date: June 2014 Slides: slideshare.net/totto
What does software engineering research actually say — and how much of it contradicts what we think we know? This talk reviews recent empirical findings on programmer experience, code coupling, and defect rates, and draws practical implications for software architects.
The four research topics¶
1. Technology choices, experience, and quality. How do the tools and languages a team uses interact with their experience level to produce (or undermine) quality?
2. Skills, experience, and defects. The relationship is less straightforward than expected. More experience does not simply mean fewer defects.
3. The senior programmer paradox. Experienced programmers show higher error rates and lower productivity when working with highly coupled code. Less experienced programmers show better productivity and quality in the same conditions.
4. "Good" vs "bad" systems and experience. The architecture of the system shapes who thrives working on it — with implications for team composition and hiring.
The architectural implication¶
High coupling is often treated as an unconditional negative. The research suggests a more nuanced view: highly coupled codebases may actually be less problematic for projects with frequent programmer turnover or active training of new developers — because the experienced programmers who suffer most in that environment may not be the ones you have.
Language: Norwegian