A recent study from Stanford University has sparked controversy in the software development community, claiming that nearly 10% of software engineers are "ghosts" who do "virtually nothing" but collect high salaries. The research, which analyzed data from over 50,000 engineers, suggests that these underperforming developers could be costing companies a staggering $90 billion annually. However, experts in the field are pushing back, arguing that the methodology used in the study is flawed and misleading.

The study's lead researcher, Yegor Denisov-Blanch, claims that by cutting these "ghost" developers, companies could save billions of dollars. But critics argue that the research oversimplifies the complex role of software developers, focusing solely on code commits as a measure of productivity. Honeycomb CTO Charity Majors notes that writing code is not the most important role of a senior engineer, but rather understanding, maintaining, and managing software in production, as well as translating business needs into technical implementation.

Moreover, the study's approach has been criticized for not accounting for bad contributions, such as useless abstractions or vanity rework that can add negative value and confusion. Nvidia Senior Engineering Manager Aaron Erickson points out that the researchers might find "another 10% of engineers who do add code, but it's useless abstractions or vanity rework that adds negative value and confusion." This raises questions about the accuracy of the study's findings and whether they truly reflect the productivity of software developers.

The controversy surrounding the study highlights the challenges of measuring engineering productivity. While the goal of increasing transparency, accountability, and meritocracy in software engineering is laudable, using code commits as the sole metric can be misleading. As the Stack Overflow team notes, the "hardest part of building software is not coding, [it's figuring out] requirements." The best engineers will figure out what to build and reduce that to as little code as possible, making it difficult to quantify their productivity solely based on code commits.

The implications of this study are far-reaching, and experts warn that it could lead to misguided decisions by CEOs and managers. If companies start using code commits as the primary metric for evaluating engineer performance, they may end up removing some of their top engineers, not necessarily their worst ones. As the debate continues, it's clear that a more nuanced approach is needed to accurately measure engineering productivity and identify areas for improvement.

In conclusion, while the Stanford study raises important questions about engineering productivity, its methodology and findings have been widely criticized. As the tech industry continues to grapple with the challenges of measuring and improving software development, it's essential to approach the issue with a more comprehensive understanding of the complex role of software developers.