Presentation
When Size Does Matter: Beyond Lines of Code
Contributors
Event Type
Technical Session
Languages, API's and Framework
All Audiences
Troubleshoot and Resolve Complex Issues
TimeMonday, August 179:45am - 10:05am EDT
LocationRoom 319
DescriptionAsk a development manager how big a program is and they'll cite lines of code. Ask how complex it is, and the answer falls back on gut instinct, tribal knowledge, or familiarity with whoever wrote it. In today's mainframe environment — stretched teams, unfamiliar applications, original developers long gone — that's not good enough.
This session makes the case that Lines of Code is the wrong unit of measurement, and that two proven metrics give development managers a far more accurate and actionable picture of program complexity.
We start with the evolution from card decks to SLOC — tracing why Physical SLOC gave way to Logical SLOC and why even that falls short. Two programs can have identical line counts and do completely different things. SLOC measures volume, not understanding, not risk, not effort.
Enter Halstead Metrics. Introduced in 1977, they shift the lens from how much code to what the code actually does — counting operators and operands to derive Vocabulary and Length. Vocabulary tells you how many distinct concepts a developer must understand to truly grasp a program. These metrics surface complexity that SLOC cannot see, objectively and automatically.
McCabe's Cyclomatic Complexity takes a complementary angle — measuring the number of independent logical paths through the code. A high score signals a program that is difficult to test, prone to defects, and expensive to maintain.
Together, they give managers an impartial basis for assignment decisions, estimates, risk prioritization, and technical debt reduction — elevating decisions above gut instinct.
This session makes the case that Lines of Code is the wrong unit of measurement, and that two proven metrics give development managers a far more accurate and actionable picture of program complexity.
We start with the evolution from card decks to SLOC — tracing why Physical SLOC gave way to Logical SLOC and why even that falls short. Two programs can have identical line counts and do completely different things. SLOC measures volume, not understanding, not risk, not effort.
Enter Halstead Metrics. Introduced in 1977, they shift the lens from how much code to what the code actually does — counting operators and operands to derive Vocabulary and Length. Vocabulary tells you how many distinct concepts a developer must understand to truly grasp a program. These metrics surface complexity that SLOC cannot see, objectively and automatically.
McCabe's Cyclomatic Complexity takes a complementary angle — measuring the number of independent logical paths through the code. A high score signals a program that is difficult to test, prone to defects, and expensive to maintain.
Together, they give managers an impartial basis for assignment decisions, estimates, risk prioritization, and technical debt reduction — elevating decisions above gut instinct.
Contributors
Lead Product Manager