Talk:Agile software development

Low Quality Article

This article is too unscientific. It mostly regurgitates non-scientific professional literature, while ignoring relevant research. I'm going to add some research. If any agilisits out there take offence at my changes, please read the research before reverting anything.Paul Ralph (University of Auckland) (talk) 12:49, 5 February 2018 (UTC)

Please note that this content has been moved (rather than deleted); it is not significant enough in its own right to warrant being in the lead / lede paragraph. Also, the link you posted is to content behind a paywall; it's a bit hard if other editors are challenged to read this before they can challenge this edit. As it happens, my masters research topic was whether Agile Transformation worked or not, in which I referenced a paper by Dingsøyr which argued that agile practices do in fact help. I will reach out to you offline to suggest meeting up to discuss (as we know each other). Davidjcmorris  Talk  02:23, 7 February 2018 (UTC)

I never said that agile practices were devoid of benefits. There is, for example, evidence that Scrum increases productivity. I said there's no evidence that they increase agility. There are no validated measures of team agility, and you can't demonstrate a causal relationship on a dependent variable you can't measure. The fact that all these practices are called "agile" without anyone ever demonstrating that they increase agility is possibly the most important thing you could know about agile. It's like pfizer selling "cancer drugs" that don't fight cancer, or "anti-virals" that don't kill viruses. It's a "productivity practice" that doesn't increase productivity. It is very common on Wikipedia to point out the lack of scientific evidence supporting a thing in the introduction. And you don't get to discount research that's published behind a paywall. Paul Ralph (University of Auckland) (talk) 21:11, 7 February 2018 (UTC)

Can you clarify:

* What are you meaning by agility? Common uses included: Hypothesis driven experimentation, trying out ideas on a small scale. Early delivery of value (whether to customers / end-users or to the product owner / development team). Fast feedback to uncover change early and avoid too much investment in scope not required. The ability to absorb change without causing additional cost, by dropping lower priority scope. Etc. There is evidence that this happens.

* What you mean by scientific evidence? As Popper and others have said, it is hard to scientifically prove a hypothesis; so a theory should stand until there is evidence that disproves it. There is, however, an abundance of research in this field, following a wide range of research methodologies. Which methodologies would you regard as scientific and which not?

Incidentally, the research was not questioned, but rather the insistence that people read it before challenging the edit. The offer to discus the article and points raised in person still stands (I have emailed you separately about this). Davidjcmorris  Talk  01:26, 12 February 2018 (UTC)