Is it true that so many software projects are unsuccessful?
I read this article today: http://forums.theregister.co.uk/forum/1/2015/11/11/most_developers_never_seen_successful_project/ I've started *experimental* personal (toy) projects which were abandoned due to my own lack of ongoing interest, but I can't imagine going into a situation in which the core business software was allowed to fail. That's never happened in my experience. Does this really happen so often in industry?
posted by: Nick 11-Jan-2016/14:30:32-8:00
It is bad, but not that bad. A lot of project do exceed time and cost. Most of it comes from overmanagement and underdocumentating. Bringing extra programmers in and simultaneously also bring up frequency and duration in meetings, asking more reports about working hours and introducing more distractions on the workfloor. I could go on for a while.
posted by: iArnold 11-Jan-2016/15:25:35-8:00
Given that projects such as space exploration flights are often successful, and those involve incredibly precise calculations, and I expect require deep research, as well as technical innovation, broad foresight, and a wide perspective about possible problems and human errors which could hinder the outcome, what other types of software are so difficult to produce, that the success rate would be so low? Do software companies hire totally incapable engineers, managers, and developers? Or is the difficultly which third party developers have communicating with CEOs, managers, etc. about the requirements of software, such a barrier that it would lead to that degree of failure? (Communication is always difficult, and specifications change as software gets developed, but that's expected, right?). I'm really curious. If I tell someone that I can produce a piece of software which does X, Y, and Z, then I already know exactly how I will produce it, as well as basically how much work and time is required to produce it. And I generally know enough to ask questions about the business in which that software will operate, to ensure that the software will do the job actually required, and that future feature requirements are discussed before the project begins, so that the full potential scope of the project is at least generally understood before I agree to do any work. How, and in what sort of environment, is this process able to fail totally?
posted by: Nick 11-Jan-2016/15:37:15-8:00
Space exploration flights are mostly governmental prestige projects with large budgets. Projects in company's are often a follow up on a succesful piece of software. Sometimes gonna build something and not even know if the product will be sold.
posted by: iArnold 11-Jan-2016/16:06:50-8:00
I guess I'm curious to hear more about those types of projects which have those sorts of question marks. When I get involved in projects, there's typically a clearly defined, critically important goal which needs to be achieved.
posted by: Nick 11-Jan-2016/16:12:18-8:00
To me this more relates about the commercial success of the project, not the technical side. Still there are good/bad practices on the technical side of course.
posted by: GregP 25-Jan-2016/15:35:24-8:00
I see it in government, because if the core business fails, you just pour more tax money into it. Many projects become so insanely large and complicated that they are never really finished, and it's considered normal that implementing some software solution will take 10-15 years. It's quite crazy. In Denmark, we have suffered many project losses due to such uncontrolled IT problems, where management consistently derailed projects, by insisting on keeping the scale big and "impressive". Currently our tax revenue service is suffering deeply from this. It costs billions.
posted by: Henrik 16-Feb-2016/7:52:19-8:00
|