PDLC

SFEIR's 22 July 2026 note places the Product Development Life Cycle as the containing frame: ideation and discovery through to market withdrawal, with the SDLC nested inside it under the development phase. ISO/IEC/IEEE 12207 standardizes the SDLC around building software correctly and reliably; the PDLC asks whether it is the right product. Theodore Levitt's PLC (1965) observes a commercial curve, while the PDLC organizes a work of design.

The distinction has teeth now because of arithmetic. Google and JetBrains measured in May 2026 that roughly 85% of developers use coding agents and roughly 41% of new code is AI-generated, compressing implementation from weeks to hours. Marty Cagan's Four Big Risks expose what that leaves untouched: the SDLC natively addresses technical feasibility alone, one risk out of four. An organization strong on the SDLC and blind to the PDLC becomes John Cutler's feature factory, which measures output rather than outcome. DORA 2025, across about 5,000 professionals, found AI adoption correlated positively with throughput and negatively with stability.

Bill Staples describes a PDLC in which business signal becomes verified software and production results feed back into the decision. Human judgment, in his account, moves up to intention, architecture and exceptions. Dorra Bartaguiz arrives from the other end: equipping the development team alone creates a new bottleneck, so she proposes specialized agents across the whole product cycle. Spec-driven development makes the PDLC/SDLC boundary porous, since the specification itself becomes executable by agents.