Over-processing is work or complexity beyond what current customer needs require. In software, it can look like premature optimization, an extensible design without observed variability, or extra validation that slows every change. The work may be used and well intentioned; the waste is the gap between its cost and its present value. Unlike overproduction, it is not necessarily work that goes unused. section 2.4.6
Listing 2.6 Overprocessing waste
// Simple, sufficient solution
boolean isValidUsername(String username) {
return username.length() >= 10;
}
// Overprocessed version
class ValidationContext {
private final Map<String, Integer> rules;
ValidationContext(Map<String, Integer> rules) {
this.rules = rules;
}
boolean validate(String key, String value) {
return value.length() >= rules.getOrDefault(key, 0);
}
}
A username-length rule needs one comparison. The alternative adds a validation context, configuration, and another path to understand and test, without evidence that the rule varies. section 2.4.6
The same mismatch appears beyond application design. An AI-generated order service may add caching, validation, and business rules that nobody requested; a polished pull request can hide which decisions were invented. A CI/CD pipeline can grow from 12 to 42 minutes as checks accumulate, leaving commits queued and developers waiting. A production API that returns 47 fields when a client uses six spends resources processing unused data. None of these examples makes abstraction, testing, or rich responses inherently wrong: the question is whether their benefit justifies their cost here. section 3.3.2 section 6.5.2 section 12.5.3
The cost of understanding a long method
To change three lines of tax logic in a 500-line order-processing method, a developer must first navigate payment, inventory, shipping, and other responsibilities. If that adds 30 minutes across 40 changes, the unnecessary reading costs 20 hours. Separating focused responsibilities can make the tax change easier to locate and verify. section 7.2
Before accepting more structure or processing, ask:
- What variability or customer need are we actually observing, rather than anticipating? section 2.4.6
- What does this choice add to review, testing, maintenance, or feedback time? section 2.4.6 section 6.5.2
- For AI-assisted changes, which behavior came from the request or contract, and which was assumed? section 3.3.2 section 3.5
