Software Factories · The ideas, revisited
From faster coding
to better enterprise engineering.
AI opens new possibilities for how we build and evolve systems. This commentary examines whether that capability can be connected with reusable knowledge, business objectives and enterprise architecture.
The question for today
What if AI could help us engineer better enterprises?
Software product line research makes shared capabilities and variation explicit. AI offers new ways to put that knowledge to work.
The aim: reuse proven decisions, adapt systems with less repeated effort, and trace development choices back to business goals.
The commentary explores how this could work—and what evidence would demonstrate its value.
Explore some new ideas
- A 2004 vision, newly relevant. Build families of systems from reusable knowledge.
- Reuse the decisions, too. Carry requirements, architectural choices and tests into the next system.
- Test AI against explicit engineering constraints. Assess whether product line models help it respect choices and dependencies.
- Start with business purpose. Connect development to enterprise architecture and measurable outcomes.
- Judge the whole economics. Consider the cost of delivering, operating and changing systems.
Read the commentary → About 4 minutes
Explore the original thinking
The historical collection preserves the ideas in their original setting. Read it alongside the new commentary to see both the continuity and the questions that remain.
The original magazine article · PDF
OOPSLA 2005 papers and workshop materials
Original biographies, presented in their historical context