In software development, what is coding

Software development with AI-assisted coding

In his 1987 paper "No Silver Bullet - Essence and Accident in Software Engineering", Fred Brooks postulated that there is no technology that, by itself, will increase productivity, reliability or simplicity tenfold in ten years. Nor could he see a doubling every two years.

A lot has happened since then in terms of developer tools and productivity advances; there are modern IDEs; higher level languages; faster compilers; sophisticated runtime environments; Intellisense; Test tools; Refactoring tools; Code analyzer; Code navigation tools; Code visualization tools; Source management and DevOps tools ...

Still, none of this could be considered a Fred Brooks Silver Bullet. Patrick Smacchia, developer of the analysis tool NDpend, discusses in great detail in his English-language article "Is Artificial Intelligence Assisted Coding the Next Developer Productivity Silver Bullet?" Whether this is still just around the corner.

His conclusion: Visual Studio Intellicode is a good start to improve developer productivity with an AI-based system. No ambitious scenarios are outlined right now, but it can certainly help developers save some time each day which is a win.

Smacchia does not currently see any significant obstacles to most of the scenarios described in its article: there is a lot of open source code available for training AIs.

Code is a purely logical, discrete and predictable material like a game of chess (in contrast to a road that engineers work with who want to develop systems for autonomous driving).

The research and development coding supported by AI will surely benefit from the demand for better tools. There is a demand for more coding productivity. The demand to simplify the development process can also be seen, as confirmed by the growing success of low-code systems.