Does Agile/XP frequently fail in heterogenous teams? And if so, is the reason due to white male privilege?
I hesitate to give serious space to this argument, but there are folks who have put some time an energy into it. So perhaps it deserves a response.
I hesitate to give serious space to this argument, but there are folks who have put some time an energy into it. So perhaps it deserves a response.
You’ll find the original argument here: retrosight.github.io/learning/why-a…
The gist of the argument is that since white “dudes” created XP and the Agile Manifesto, they were unaware of the power dynamics that might be encountered by other races and genders.
The author specifically talks about 100% pair programming and 100% TDD. I should point out that neither of these are practices espoused by Agile or XP. Pairing is certainly encouraged; but we long ago realized that 100% was impractical.
My current recommendation for pairing is ~50%. TDD is similar. Yes, it is strongly encouraged; but all serious Agile practitioners understand that there are situations where it simply doesn’t work.
I talk a lot about these limitations of TDD and Pairing in my blogs, books, and videos. Other Agile proponents do as well. So an initial premise of the article in question is a bit skewed.
But now let’s talk about the power dynamic argument. In a pairing situation, the author argues, people of different genders and races will be on the “under” side of the power dynamic. They will feel less able to contribute and collaborate.
The author contends that a female, or a person of color, could be silenced or cowed by the “power” of a white male pair partner.
The power argument is a good one. Power differentials in pairs do, in fact, evoke the issues the author describes. What’s more the author offers several appropriate remedies to resolve such disparities of power.
What the author misses is that these power differentials exist in every team regardless of its heterogeneity. In virtually every pair one member is more powerful than the other. This is independent of race or gender.
One might argue that every time a non-white non-male pairs with a white male, the white male will be higher in power. I have not observed this to be the case. I have paired with women, and people of color, who have assumed the greater power position.
In any case, the fact that power differentials exist in pairs has nothing whatever to do with the fact that XP and the manifesto was created by a bunch of “white dudes”.
The author then moves on to TDD and suggests that women generally prefer a different ordering to test than men do. In particular women like to cover the error cases first and men prefer to go straight to the heart of the technical problem.
I think this argument would make James Damore smile, since it echoes his suggestion that men and women think differently. The author goes on to argue that the power differential means that the woman’s view would be ignored, or diminished by a white dude.
This is essentially the same argument the author employed for pairing. The person on the lower side of the power dynamic gets cowed or ignored and the code gets written the way the more powerful programmer desires.
Once again, the power argument is reasonable. It is true that the more powerful person in the pair will tend to dominate; and must guard against that if the pairing is to be successful.
However, the authors suggestion about the ordering of the tests having a gender component is a bit puzzling to me. I have, for years, taught that TDD should approach problems from the outside in by attacking error conditions and peripheral concerns first.
Indeed, in my blogs and videos you will find this practice under the name: “Going for the Gold”. I advise programmers to delay going for the deep technical issues until after they have surrounded it with peripheral tests.
My conclusion is that the author has a couple of good points about power disparities. The author also offers some reasonable suggestions about how to resolve them.
However, by making race and gender the point from which all the author’s otherwise reasonable arguments derive, the author taints those arguments with a flavor of oppression and privilege which have nothing to do with the actual issue.
The author seems convinced that white males always have the power, and suggests that the white male authors of Agile/XP preserved that power disparity within the values, principles, and practices they authored.
I frankly find that premise absurd, somewhat insulting, and extremely il-considered.
