A response to the IEEE article: Yahoo’s Engineers Move to Coding Without a Net

link to the article

First, a disclaimer

Let me start out with the understanding that the article as written is being interpreted through the writer’s own biases and background. I certainly have been misquoted or had a radically different interpretation of my words by an author than was meant.

This is a response to the article as written. I would be happy for someone from Yahoo to let me know if they feel the article was an accurate depiction or not.

Yahoo’s new approach to quality is something many other companies already do

What yahoo learned is what many other companies have learned a long time ago. That segmenting test and development is a bad idea. It leads to developers and testers living in their own silos and playing a blame game on each other. Agile methodologies challenged this idea 16 years ago. Having separate test and development teams is a pretty big red flag for waterfall development.

It wasn’t the fact that there were QA that was the problem

Taking away QA can improve quality to a point because it makes quality everyone’s job, but it won’t get you a high-quality product. You could do that by creating a culture of quality, making everyone feel responsible for quality and keeping the QAs on the teams. You can’t find every bug in the product through test automation. Developers tend to be poor testers of their own code, because they expect it to work and they have a hard time using the product like a non-developer. Having someone on the team with an expertise in quality thinking and a user mindset can do real wonders for product quality.

What Yahoo should try next in their quality evolution

Yahoo could also have increased quality and velocity substantially by keeping the QA, integrating the teams of developers and testers and by creating a culture where everyone feels responsible for the quality of the product. I contend that in their evolution they might yet find another factor increase in these by doing exactly that.