Trillions spent and big software projects are still failing (spectrum.ieee.org)

Article summary

The article discusses the high failure rate of large software projects, despite significant spending on IT, with examples such as the Canadian government's Phoenix payroll system and the UK Post Office's Horizon system. These failures are often due to management and organizational issues, rather than technical problems. The article suggests that the use of AI and other new technologies is unlikely to solve these underlying issues. The cost of these failures is estimated to be in the trillions of dollars.

Main themes

• Software project failures
• IT management issues
• Organizational complexity
• Technical debt
• Legacy system maintenance

What commenters say

• The failure of large software projects is often due to poor management and organizational decisions, rather than technical issues.
• Simplifying complex systems, such as payroll rules, is crucial to preventing failures, but this is often difficult due to political and bureaucratic constraints.
• The use of Agile and DevOps methodologies can help improve software development, but their success depends on consistent leadership, organizational discipline, and cultural change.
• Putting technical people in charge of projects is necessary but not sufficient for success, as non-technical factors such as politics and finance also play a crucial role.
• The tech industry's tendency to constantly reinvent and innovate can lead to a lack of learning from past mistakes and a failure to establish best practices.
• The high failure rate of software projects is not solely due to technical issues, but also due to the fact that software is often treated as a cheap and disposable commodity, rather than a critical component of modern infrastructure.
• The constant churn in technology can be seen as a deliberate design choice, allowing companies to hire new graduates at lower salaries and perpetuating a cycle of reinvention and obsolescence.