In No Silver Bullet, Fred Brooks argues that achieving an order of magnitude gain in software development productivity will only occur if the essential complexity of software engineering is addressed. For Brooks, the essential complexity of software engineering is conceptualizing software's "interlocking pieces". This is in contrast to the relatively trivial task of representing the chosen concept in an implementation.
Today's widely adopted AI-enabled tools for software development, like Copilot, aider, and cline, readily produce representations when given a natural language description of a concept. However, if Brooks was correct, these tools address only the accidental complexity of software engineering, not the essential complexity of "specifying, designing, and testing [the conceptual construct]".
In the full version of this article, the authors share their experiences and insights on how LLMs helped them uncover and enhance the conceptual constructs behind software. They discuss how these approaches address the inherent complexity of software engineering and improve the likelihood of success in large, complex software modernization projects.
The authors state that combining rigorous, systematic approaches like static analysis with AI summarization allows for new, customized approaches to modernization. The ability to automate and inspect the outputs of static analysis helps human experts perform their jobs with confidence, while LLM summarization significantly reduces the toil of preparing lengthy documentation.
These advantages bring collaborators from different backgrounds into the process of modernizing software. Only after software teams understand the concepts behind legacy software can they effectively rewrite it, with or without AI tools.
This content is an excerpt from a recent InfoQ article by Michael Wytock, Ken Judy, and Aaron Foster Breilyn, "AI Interventions to Reduce Cycle Time in Legacy Modernization".
To get notifications when InfoQ publishes content on these topics, follow "AI, ML & Data Engineering", "Machine Learning", and "Large Language Models (LLMs)" on InfoQ.
Missed a newsletter? You can find all of the previous issues on InfoQ.
Sponsored
| | New Akka SDK deployment options let you build agentic systems, APIs, and data services and deploy them on any infrastructure: bare metal, Kubernetes, VMs, edge, or on-prem. With self-managed nodes and Akka Platform regions, you get infinite scale and guaranteed resilience, with no dead ends—your code deploys everywhere.
See deployment options, sponsored by Akka |
|
Upcoming Events
InfoQ and QCon: For practitioners, by practitioners
InfoQ Dev Summit Munich 2025 (October 15–16): Master modern software development.
In a landscape of rapid AI advancements and increasing system complexity, senior developers need practical strategies. Join your peers in Munich for two days of deep dives into scalable architectures, resilient systems, and effective platform engineering. Learn from practitioners on the front lines, without the marketing fluff. See the full schedule and register today.
QCon San Francisco 2025 (November 17–21): Lead the future of software.
As AI continues to reshape the development lifecycle and distributed systems grow more intricate, staying ahead is critical. QCon San Francisco offers 12 curated tracks and over 60 practitioner-led sessions designed for senior software engineers and architects. Gain actionable insights into scaling your systems, optimizing your architecture, and making impactful technical decisions. Explore the tracks and secure your spot.
QCon AI New York 2025 (December 16–17): Accelerate the entire software lifecycle with AI.
Move your AI initiatives from experimental to essential. This practitioner-driven conference is for engineering leaders, architects, and senior developers focused on the practical application of AI in the software development lifecycle. Learn from those who are successfully building, deploying, and scaling reliable and secure AI systems in enterprise environments. Register now to lead your team’s AI journey.
