I don't understand this widespread sentiment at all. If the dogma-idea is right, it's right. No matter how dogmatic the people who are defending it are.
Take the pizza example. Which is better, linear code or small functions? It's a series of tradeoffs. Once you get above a screen of code, or 10 ifs, functions become hard to read. Once the same logic has been written 3x, abstracting it is usually a win. Even if it is small. And there is a fuzzy area where it isn't obvious which is better, and debating it is probably a loss over writing it and moving on. Doubly so if you're defaulting to the same kind of decisions every time and so the style is consistent.
In a world full of pragmatic tradeoffs, dogmatism is rarely the right choice. (Unless you haven't learned the tradeoffs.)
dogma invites people to stop using their critical thinking skills.
One of my favorite examples of this:
Everyone would agree that having a newborn in a car means safety is paramount. Everyone also agrees that left turns are less safe than right turns. No one would agree that this implies you should never make left turns in a vehicle with a newborn.
^But the above is how both security people and TDD proponents tend to act, as if there can be no risk assessment and critical thinking involved. We've all made right hand turns when we really wanted to go left because there was just too much traffic, even without a newborn in the vehicle.
Because ideas are highly unlikely to be universally correct, no matter how good they are. Even if an idea is supported by all available evidence, it MUST subject itself to scrutiny and possible falsification, all the time.
A dogma flies in the face of that. It is, by definition "any belief held unquestioningly and with undefended certainty" (quote from wikipedia). Once people follow an idea dogmatically, they are very likely to apply that idea no matter if it makes sense or not. They stop following logic and start following scripture.
It is bad enough if this happens in science, where we really do have systems so far supported by all available evidence. But it becomes a lot worse in areas like software development, where we know there isn't "the one" true way of doing things.
And to head off one likely reply to this: Yes, I apply this logic also to the assumption that "dogmas are always bad". If someone could present proof showcasing a dogma that only has good outcomes, with no negatives attached, then I am willing to change my mind on this.