Hacker News new | past | comments | ask | show | jobs | submit login
The “Myths List” is a communication antipattern (cachestocaches.com)
64 points by gjstein on Aug 8, 2019 | hide | past | favorite | 35 comments

I find it slightly ironic that an article about the evils of "Myths Lists" uses the "X is an antipattern" form. Calling something an antipattern by saying that it is bad is just a "myth list" comprising one entry :-)

The original meaning of antipattern is more than a common practice that's a bad idea. It's something that appears to be very good in a certain situation, but is actually bad. In order to document it as an antipattern, you have to describe why people think it is a good idea in the context in which they use it. Then you have to explain why those good ideas don't work out. This article fails to do this, IMHO. It's just a rant about something the author doesn't like (exactly the same as this post ;-) ).

As the author of "Myths about /dev/urandom" (https://www.2uo.de/myths-about-urandom) I gingerly agree with the thrust of this article (although it has not really swayed me, other articles in that vein have).

The essay is still this way, because (a) the title is well-known in certain circles (call it branding) and (b) it's real work to find a new structure.

But your article is about cryptography. And like every article on cryptography aimed at non-cryptographers it has this upshot: "You don't have a chance at knowing what you're doing or talking about, so please just ask a cryptographer for help and get on with your busy day."

If readers of your article come away more confused about dev/random vs. dev/urandom and the veracity of manpages, that is a positive result which will lead them to the aforementioned promised land. So I don't think your article fits the (anti-)pattern.

Can you write an article on how to do attacks on low entropy next? Lol I can't find any good guides.

Newsflash: listicles are garbage. It doesn't need to be a list of myths; any article following the pattern "10 anythings about anything" is clickbait trash.

An article comprised of a list is a kind of promise to the reader that you are not going to have a complicated or nuanced argument. It's a list. The items will be clearly delineated and will probably only superficially refer back to themselves.

I think this is why these lists are popular. They are easy to read and easy to digest. I don't think there is anything wrong, per se, with a list. It has the advantages I point out. The problem comes where you have a complicated and nuanced topic and distill it to a list. Because you have simplified the argument, it is usually one sided and deficient. That's why it is usually trash. It doesn't have to be, though, IMHO. Some things lend themselves well to being expressed as a list (for example, a list of potential alternative solutions to a problem).

Titles with a number generate more clicks so everybody does it.

Numbers in title are a good indicator that the article is 1. Opinion based and that it 2. Lags detail.

At least it's enough for me to know to avoid it when I'm searching for a solution to a problem.

This is one of those things I try to google every once in a while but the results are so clogged with junk that I still don't really know the answer. Why are our brains drawn to click on headlines/titles with a number?

I don't know exactly and I wasn't able to find a reference either but I thought the increase was something like 30%. Maybe hubspot or buffer has a blog on it.

It’s true. Odd numbers do better than even, as well.

You'll be shocked by these eleven ways that lists about things are terrible!

It's generally much more helpful to say what's correct than what's wrong, if only because there are so many more ways to be wrong. Often when I see lists of "myths", the "myths" are true or close enough most of the time, and are only harmfully wrong in rare circumstances. Getting rid of a flawed but workable model without having a better one to replace it with is not an improvement.

Or posing the statements as questions. Like instead of

> Myth: People have unique names

You write

> Do people have unique names?

> No, because XYZ. This varies by region because ABC. And in this case, this assumption caused a major production outage: [link]

Turn it into a detailed discussion, as opposed to a debate. Also, you can share things that people _can_ assume to be true, or can _usually_ assume in specific contexts.

Asking a question whose answer is no is almost as bad as a myth. Just state the true fact you want people to remember.

Sometimes the answer is yes, though!

In addition to myths being mostly true I've also seen articles with the opposite problem, where the myths are over-the-top wrong. One example I just saw was something like, "Myth #123, I don't have to worry about making my website accessible because none of my users are disabled".

> It's generally much more helpful to say what's correct than what's wrong, if only because there are so many more ways to be wrong.

So you should know what is correct before pointing out what is wrong? That doesn't sound right.

> Getting rid of a flawed but workable model without having a better one to replace it with is not an improvement.

This ignores the opportunity gained from not doing something. If "Myth Lists" aren't worth doing, not doing them is an improvement. You don't need to have a "better alternative", because "doing nothing" is already better.

Having said that, the article doesn't exactly show that Myth Lists don't work. It's just an opinion of someone who read one of these lately and didn't like it.

Does the author include "Falsehoods Programers Believe about $FOO" style articles? I generally find them to be clear about the fact that these are lists of false statements, especially since there's generally no content to each falsehood outside of stating it (sometimes they provide counterexamples).

Or I was reading this just the other day from Mozilla: https://support.mozilla.org/en-US/kb/common-myths-about-priv...

In the right context--where there is widespread misunderstanding (or ignorance) around a number of points--it seems a reasonable format.

Lists and FAQs are arguably somewhat lazy writing but it imposes a structure that's easy to follow for both the writer and the reader and that's not a bad thing. And talking to editors I work with, it does seem to be a format that readers like based on the data.

Maybe people like the listicle format because it promises to be quickly digested.

Actually, I wonder if it even needs to be a list. I think a headline like "Economics explained in 20 words" would be attractive in the same way. If it was 100 words, I doubt it would work as well, even though both headlines would be obviously absurd.

What's going on, there?

Attention spans are definitely down. I know I read a lot more shorter format myself. And there's actually a lot of data on it.

I certainly don't feel compelled to write for SEO or eyeballs. But easy and quick doesn't need to mean bad and is often OK for day-to-day stuff.

Most things I publish are in the 800-1000 word range. These days I'll tend to go multi-part rather than the 3K-5K stuff I used to do fairly routinely.

There's nothing wrong with the listicle format, but as another poster described, the problem with 'X myths about Y' is that they tell you that the myth is wrong... Without telling you what is right.

I'm thinking it's not linear, ordered, etc. So the list works.

"Falsehoods programmers believe" tend to be implicit assumptions people have, while "myths" tend to be deliberately propagated ideas.

Assumptions can simultaneously be commonly held and uncontroversially false. The trick is pointing them out.

Here is one of my favourite "myth list" articles: https://www.seriouseats.com/2013/06/the-food-lab-7-old-wives...

I think the form fits great with the content. So it can work, but the myths listed have to actually be widely held and promulgated false beliefs.

Humans being humans love to make the myths easy to knock down so they end up being very general, and often absurd.

The answer then ends up being just as general ("nuh uh!") and so simplistic that it is useless.

So I guess what I'm saying is that it turns into... Twitter.

This is very context driven. I find that myth lists with regards to health topics are helpful. They usually have links to studies and generally aim to dispel common misperceptions. They usually spur me to read more into each myth, especially if it is something I am new to. They aren't always concretely correct, but they spur investigation. However, as this article is clearly labeled "general computing" on the site, I agree that almost all articles in that category use myth lists as poorly executed opinion argument. It comes across as very juvenile. Myth: Rust is difficult to learn. Truth: Actually, Rust is quite easy to learn. Ugh...

Would you go so far as to say Myths Lists Considered Harmful? :)

It's interesting to see the difference in complexity between an application in theory and examples of the actual engineering requirements. It's no different than reading Glamour magazine for makeup tips or Men's Fitness for health advice, etc. Not everyone is going to be into the same level of depth and complexity here.

In the potential failure to recognize the intended audience of these lists, I wonder if the antipattern accusation is really an antipattern (hehe). But I honestly don't know - Were these lists intended to be used as programming guides?

I agree with this, but I think the bigger problem with "myths lists" is that they're condescending. Even when written dispassionately, they implicitly leap out of the gate with "everyone is wrong, including you, reader, and I am going to drop my wisdom on you."

I can't find it now, but there was a great comment on a previous "falsehoods" post about how the only compelling reason to frame content this way is to get clicks and drive engagement through contention and disagreement, and that the world would simply be a better place if the tone was more in line with "here are some super interesting cases you may not have thought of when you do X". See https://xkcd.com/1053/ for a similar sentiment.

Seems to me that "myths lists" are most useful for, ya know, actual "myths"[1]. If somebody is likely to google for X, they'll find your page, and read your refutation. As a matter of SEO, the closer you get to repeating the commonly-repeated story, the more likely they'll get to your page.

If the myth isn't something common enough to be googled for, then you're just confusing people. Just state the facts, in as concise and clear a fashion as you can.

I think people are drawn to the "myths" structure because it breaks up the page visually. It tells people, "Here is a small unit of information. You don't have to read the whole page to get a story". There are lots of other ways to indicate that: headers, lists, images. It's a correct intuition that a flat page full of paragraphs is intimidating, but there are other ways.

[1] in the sense of "myth" as "commonly-repeated falsehood". Most falsehoods aren't myths, and a better use of "myth" means a fiction that resonates deeply and universally. Fictions aren't mere falsehoods, either.

I use Quora a lot, which has some good stuff but at this point I admit the good stuff is far outnumbered by the bad these days.

One of the most annoying things people do is to write posts (technically answers) that attempt to "surprise" people by telling them something they thought was true, but is in fact, not true.

And so many of them are really convoluted and depend on weird non-standard definitions of things. Something like "technically red things are blue" or the like. (because they absorb blue light maybe? So you are defining things by what they absorb rather than what they reflect? Whuuu?) So, so many answers are like that, but maybe not to that extreme.

Yes it is a clickbaity technique. They are always wanting to blow people's minds.

These myths lists are definitely falsehoods programmers believe in.

but will they click?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact