Sometimes It Is Worth Not To Make Software
Some statistics about software projects:Source: Extreme Chaos. The Standish Group, West Yarmouth, MA, 2001.
| Outcome | Share |
|---|---|
| Projects that deliver no working software | 23% |
| Average budget overrun | 45% |
| Average schedule overrun | 63% |
| Of the planned functionality, actually delivered | 67% |
On top of that, as I argue in The Half-life of Software, nearly half of the software released today will be retired within the next four years.
Given those odds, the question "should we build this at all?" deserves more respect than it usually gets. The cheapest, longest-lived, most reliable software is often the software you decided not to write.