Shame crazy ideas about rhino horn continued beyond the middle ages http://www.theatlantic.com/business/archive/2013/05/why-does...
Cargo Cult Science (1974)
11–20 of 42 posts
Re: Cargo Cult Science (1974)
#12Earlier quoted context omitted.
I've always wondered about 'Cargo Cult' with reference to software development. When you write software you are expecting a certain result. Surely if you don't get the result, you try something else. EDIT: ah right. So you 'may' get the result, but it's not due to the thing you have ascribed it to.
Wikipedia actually has an article on "Cargo cult programming". They have it as "the ritual inclusion of code or program structures that serve no real purpose".
Re: Cargo Cult Science (1974)
#13I feel a lot of the things described in the article is happening in fields like information security now. It was interesting listening to the 33C3 talk "In Search of Evidence-Based IT-Security"[1] a couple of days ago and read this Feynman-article now. The article may have been posted before, but it was new to me. [1] https://media.ccc.de/v/33c3-8169-in_search_of_evidence-based...
This article is always relevant. There is always an expert telling people to do something a certain way because reasons, despite zero evidence for those reasons.
Re: Cargo Cult Science (1974)
#14I often use the term 'Cargo Cult Programming'. Not a very large but significant amount of developers fall into that category sadly.
I've always wondered about 'Cargo Cult' with reference to software development. When you write software you are expecting a certain result. Surely if you don't get the result, you try something else. EDIT: ah right. So you 'may' get the result, but it's not due to the thing you have ascribed it to.
It's about confusing correlation with causation.
Look at Cargo Cult Agile for example. Company X is very successful. Company X uses points to estimate tasks. Therefore, if we use points to estimate tasks, we will be successful.
Remember when everybody went crazy for Erlang because WhatsApp was written in Erlang?
Re: Cargo Cult Science (1974)
#15I often use the term 'Cargo Cult Programming'. Not a very large but significant amount of developers fall into that category sadly.
I've always wondered about 'Cargo Cult' with reference to software development. When you write software you are expecting a certain result. Surely if you don't get the result, you try something else. EDIT: ah right. So you 'may' get the result, but it's not due to the thing you have ascribed it to.
> Surely if you don't get the result, you try something else.
Results have many "dimensions" and it is often not clear what THE result is. You will usually get A result though.
Re: Cargo Cult Science (1974)
#16For reference, previously on HN - 8 months ago: https://news.ycombinator.com/item?id=11669004 - 3 years ago: https://news.ycombinator.com/item?id=6543791 - 6 years ago: https://news.ycombinator.com/item?id=1629571 - 7 years ago: https://news.ycombinator.com/item?id=723140 https://hn.algolia.com/?query=Cargo%20Cult%20Science&sort=by...
And I hope it comes up every year in the future too
Re: Cargo Cult Science (1974)
#17> It's a kind of scientific integrity, a principle of scientific thought that corresponds to a kind of utter honesty--a kind of leaning over backwards. For example, if you're doing an experiment, you should report everything that you think might make it invalid--not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you've eliminated by some other experiment, and how they worked -- to make sure the other fellow can tell they have been eliminated.
The problem is that in the current system (based on publishing as much as possible quickly and at prestigious conferences) there are few incentives to do this. Instead people learn to game the system. Academics need to relax a bit, but they can't, since there is so much competition.
Re: Cargo Cult Science (1974)
#18This subject is something the data science/ ML world need to take very seriously. I am constantly passed snippets of google analytics as 'proof' something on the website worked with no analysis of what else could have caused it. With pay per click you get some very sulky looks when you say 'how much of that result do you feel is caused by people clicking the ad that would have clicked the search result if the ad wasn't there?'
Taking a tangent to talk about the style rather than content here.
To me the sign of the true greats in the world of thought are not just those who make new discoveries or have new ideas, but those with the amazing power to communicate those ideas without jargon. Feynman, a Nobel laureate, is often remembered for his teaching as much as his discoveries. His books are very funny too.
I remember, when I was college age, a friend telling me they were writing a paper on Existentialism. They kept trying to explain it to me and I felt like an idiot, I couldn't get the meaning of 'existence precedes essence' or whatever, and became sure that philosophy was for smarter people than me...until I picked up a copy of 'Existentialism and humanism' by Sartre, whose idea it was in the first place, and it is not full of jargon and the ideas are perfectly simple. I then read some Camus novels, which actually changed my life(!) and again the ideas are perfectly simple. Their greatness is in their challenge to the norm, the idea that we have everything the wrong way round.
Now I treat anyone who uses jargon unnecessarily (and it can be a useful shorthand at times) as someone who is trying to appear part of something they are merely a commentator on, or who is trying to keep the uninitiated out. I have so much more respect for people who can communicate in plain English. Feynman was a master, as is Orwell.
Startup culture has some really annoying forms of this. Despite my interest and involvement in technology I am actually foremost a finance professional although I do less of it each each day. When you want to stay solvent you need to forecast your cash flows, not f* around with 'burn rates'. Verbalisation is a special form of rubbish. I heard a sportsman saying he was 'hoping to meddle in the Olympics', when I presume he meant '...win a medal'. I especially hate language that is not specific, 'bi-weekly' means both 'fortnightly' and 'twice-weekly' for instance. I can only conclude that people who use too much jargon are pseudo-intellectuals.
Re: Cargo Cult Science (1974)
#19> It's a kind of scientific integrity, a principle of scientific thought that corresponds to a kind of utter honesty--a kind of leaning over backwards. For example, if you're doing an experiment, you should report everything that you think might make it invalid--not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you've eliminated by some ot…
After Peter Higgs won his Nobel Prize he said some great stuff on this subject, the idea that he wouldn't be allowed the time to do work like his anymore. Please google at your leisure.
Re: Cargo Cult Science (1974)
#20I often use the term 'Cargo Cult Programming'. Not a very large but significant amount of developers fall into that category sadly.
I've always wondered about 'Cargo Cult' with reference to software development. When you write software you are expecting a certain result. Surely if you don't get the result, you try something else. EDIT: ah right. So you 'may' get the result, but it's not due to the thing you have ascribed it to.
"No, the code must be correct: I tested it"