Flexible project performance

Question

I remember reading an article about someone evaluating a project’s performance in Agile (or maybe just TDD?), And found that it implemented projects more efficiently than its alternatives. Does anyone know where I can find this, or a similar article?

Answer 1

Here are some interesting works:

• Implementing Quality Improvement through Test-Based Development: Results and Experience from Four Industrial Groups
• Evaluating Test-Based Development Efficiency: Industry Examples
• Yahoo! Flexible Transformation Research
• SirsiDynix example: Jeff Sutherland on high-performance distributed scream (paper available here )
• Towards accelerating the growth of traditional processes by introducing the Scrum process
• Some data is also contained in The Scrum Papers.

I tend to be a little skeptical when I see terms such as "hyper performance" and not all teams can reach the performance level shown in some of the above studies. But since you asked for some performance measurements, here are some data.

Answer 2

take a look at the statistics on the use of Scrum in MySpace teams - http://www.rapidscrum.com/shock.php

Answer 3

Hoping to improve project efficiency is the wrong reason for Go Agile

If you plan to switch to agile methodologies to make your project development more efficient, you will be disappointed. Flexible transformations require a nontrivial time for the team to learn and tune processes. In addition, you will need a number of sprints before you can realistically evaluate team speed.

Even a team that is developing rapidly may or may not be more effective than a team using more traditional methodologies. More likely benefits of flexible approaches are:

• lower defect rates (as mentioned in @ire_and_curses)
• more predictable delivery planning
• creating something that is closer to the needs of the client.

Answer 4

There are several interesting studies cited in a Wikipedia article. Whether this supports your business depends on what you mean by "more effective." In general, studies show that a couple does better work and can be a little faster than a single developer working alone. However, since there is no work in the second draft, the overall pace of development is lower. In other words, you trade from the initial development time for a more reliable design, lower defect rate, knowledge transfer, time reduction, etc.

Answer 5

I have to be honest with you. I am dying to do TDD ... but I can’t even scratch the surface. And it's not just me, the whole team (50+ developers) do not even discuss it.

Last week I attended the p & p symposium in Redmond. I talked with Scott Hanselman about this. He was fanatically absolutely sure that ANY project could apply TDD. "Uncle Bob" in his speech also fanatically mentioned that TDD is the ONLY way to go.

I like testing. I believe in testing at all levels. But first write a test, and then code ... a green-red switch with loops that last like 2 minutes?

I can not do it. Do you feel the same?