Towards Understandable Software (gracefulliberty.com)

A recent commentary highlights the growing dissatisfaction with traditional programming methods, claiming they are not only complex and hard to maintain but often accessible only to a select few. As large language models (LLMs) gain traction in software development, both programmers and non-programmers are increasingly relying on these tools. However, the article argues that LLMs do not fundamentally solve these issues; instead, they can create more problems, including code quality concerns and environmental impacts. The author advocates for a paradigm shift in software development away from reliance on LLMs, calling for a higher level of abstraction that would eliminate the need to write code entirely. To address these challenges, the piece proposes embracing concepts like "literate programming," where documentation precedes code, ensuring clear human intentions are communicated. Additionally, it suggests using graphical user interfaces (GUIs) to create software, making development more intuitive and accessible to non-coders. The author envisions a future where natural language capabilities bridge the communication gap between humans and machines, allowing for deterministic software creation without ambiguous generative processes. This approach aims to transform not just how we write software but the very foundation of software interaction, making the field more inclusive and comprehensible.