Comment by lolinder
7 months ago
I don't think of this as being cargo cult development. Cargo culting has more to do with mimicking practices that have worked before without understanding that they only worked within a broader context that is now missing. It's about going through motions or rituals that are actually ineffective on their own in the hopes that you'll get the results that other companies got who also happened to perform those same motions or rituals.
What OP is describing isn't like this because the thing being copied—the code—actually is effectual in its own right. You can test it and decide whether it works or not.
The distinction matters because the symptoms of what OP calls the Makefile effect are different than the symptoms of cargo culting, so treating them as the same thing will make diagnosis harder. With cargo culting you're wasting time doing things that actually don't work out of superstition. With the Makefile effect things will work, provably so, but the code will become gradually harder and harder to maintain as vestigial bits get copied.
I would almost call this the "boilerplate effect".
Where people copy the giant boilerplate projects for React, K8, Terraform, etc. and go from there. Those boilerplates are ideal for mid to large scale projects. And it's likely you'll need them someday. But in the early stages of development it's going to impart a lot of architecture decisions that really aren't necessary.
That's a great phrase. A perfect example of what you're talking about is actually built-in to the `helm` tool for creating re-usable kubernetes configs: `helm create myapp` creates the most convoluted junk I've ever seen in my life. As a new `helm` user I was extremely confused, as I had been expecting a minimal example that I could start from. Instead, I got a maximal example. Thankfully a more experienced engineer on the infra team confirmed that it was mostly unnecessary boilerplate and I could remove it.
Something to consider for anyone else building tools — boilerplate has costs!
Seeing this exact effect where I am currently working. Main available CI/CD tool is a customised and centrally managed Jenkins fleet. It's pretty much impossible to avoid using and seldom needs changed - until it does. Some attempts have been made at centralised libraries and patterns - but even that requires knowledge and study that most won't know is available or be given time to acquire.
So when the inevitable tweak or change is made it's made in the easiest, cheapest way - which is usually copying an existing example, which itself was copied from somewhere else.
I see exactly the same in other teams repositories. Easiest path taken to patch what already exists as the cost/benefit just isn't perceived to be there to worth prioritising.
I don't understand how the second disproves the former.
In fact, a cargo cult works because there's the appearance of a casual linkage. It appears things work. But as we know with code, just because it compiles and runs doesn't mean "it works". It's not a binary thing. Personal I find that belief is at the root of a lot of cargo cult development. Where many programmers glue things together say "it works" because they passed some test cases but in reality code shouldn't be a Lovecraftian monster made of spaghetti and duct tape. Just because your wooden plane glides doesn't mean it's AC an actual plane
> a cargo cult works
But...it doesn't? That's the whole definitional point of it. If action A _does_ lead to outcome B, then "if we do A, then B will happen" is not a cargo cult perspective, it's just fact.
Sorry, the quotes around "work" were implicit. I thought there was enough context to make that clear. That just because some things actually work doesn't mean it works for the reason it actually works.
This is what I meant by cargo cults working. Where there is a _belief_ in a causal connection where there is none. The Melanesia really did believe there was a causal connection between their actions would cause cargo to return. It's not just about appearance. It is about not understanding the actual causal chain and misinterpreting the causal variables.
Measurement is HARD. It is very common that if you do A then B will happen YET A is not the cause of B. Here's a trivial example: a mouse presses a button and it gets fed. In reality, the mouse presses a button and a human feeds the mouse. It is not the button. It may even be impossible for the mouse to know this. But if the human leaves, the mouse can press the button all they want and they will not get fed. I hope you can see how this can happen in far more complex ways. As the chain of events gets longer and more complex you an probably see why it is actually a really common issue. Because you have literal evidence that your actions are causal while they are not.
For actual cargo cults, yes. Cargo Cult Development just used the name to invoke a comparison..when CCD is being practiced, devs are doing mystical steps because it's part of the incantation. They wouldn't keep doing them if the project then never worked.
Your definition is extremely unlikely to ever be practiced, because those developers would be fired for never getting anything working, and so it's not really a helpful one imo.
4 replies →
> Just because your wooden plane glides doesn't mean it's AC an actual plane
But if your wooden plane can somehow make it to Europe, collect cargo, and bring it back to your island, what you're doing is definitely not cargo culting.
It might not be actual engineering, maybe you don't understand aerodynamics or how the engine works, and maybe the plane falls apart when it hits the runway on the return flight, but if you got the cargo back you are doing something very different from cargo culting.
That's why copypasta doesn't count as cargo culting. It accomplishes the same task once copied as it did before. It may do so less reliably and less legibly, but it does do what it used to do in its original context.
Sure, but these are categorically different and not related to my point.
Let me quote wiki[0]
Even in the example it gives the code will "work." You can collect garbage when the language already does that, you'll get performance hits, but your code won't break.
It "it doesn't _work_" disqualifies something from not being cargo cult programming, then there would be no cargo cult programming. Who is shipping code that doesn't compile or hits runtime errors with any form of execution? You couldn't do that for very long.
Let's take an airplane example. Say you want to copy Boeing[1]. You notice that every 747 has a coffee maker on it. So you also make a coffee maker. After all, it is connected to the electrical system and the engines. Every time you take out the coffee maker the airplane fails. So you just put in a coffee maker.
A cargo cult exists BECAUSE _something_ is "working". BECAUSE they have evidence. But it is about misunderstanding the causality. See also Feynman's "Cargo Cult Science"[2]. As dumb as people are, there's always a reason people do things. It is usually not a good reason and it is often a bad reason, but there is a reason. Even people will explain you "causal" explanations for things like astrology.
[0] https://en.wikipedia.org/wiki/Cargo_cult_programming
[1] Well in the past you might have wanted to lol
[2] https://calteches.library.caltech.edu/51/2/CargoCult.pdf
His explanation explicitly acknowledges the experiment works. In fact, even the math to explain the experiment "works". But it is wrong. Related is Von Neuman's Elephant. Where Freeman Dyson had evidence that a theory explained an experiment, yet it was in fact wrong. Evidence isn't sufficient to determine causality.
2 replies →