[0] https://nedbatchelder.com/text/bigo.html
[1] https://nedbatchelder.com/blog/201711/toxic_experts.html
I'll double down on my toxicity by saying I didn't like the page layout. As someone with ADHD (and a declining memory), I need to be led through formatting/sub-headings/bullets/colored sections/etc into each detail or it all blends together into a wall of text. The longer it takes to make a point (visually and conceptually), the more lost I am. I couldn't easily follow it. The Simple Wikipedia page was more straight to the point (https://simple.wikipedia.org/wiki/Big_O_notation), but reading the "full" Wikipedia page thrusts you headlong into a lot of math, which to me signifies that this shit is more complex than it seems and simplifying it is probably a bad idea.
I hate that this declaration is both hilarious and one many might suggest I need to prefix with regularly.
:-D
Ask yourself why. The usual answer is that top experts either can't be bothered to create better content, or they actively gatekeep, believing that their field must remain hard to learn and the riff-raff must be kept out.
I think the first step is to accept that laypeople can have legitimate interest in certain topics and deserve accessible content. The remedy to oversimplified explanations is to write something better - or begrudgingly accept the status quo and not put people down for attempts that don't meet your bar.
It's also good to ponder if the details we get worked up about actually matter. Outside the academia, approximately no one needs a precise, CS-theoretical definition of big-O notation. Practitioners use it in a looser sense.
This oversimplifies things.
It's sometimes a third option: the topic is complex enough it cannot be digested into a ready to consume blogpost without previous work (reading and practicing), which in turn requires previous work, and so on, until you've turned it into an introductory course.
And that's not gatekeeping or "cannot be bothered to simplify" -- it's a falsehood, a kind of internet new age mantra, that everything can be made simple enough to explain in a blog post. Some things can't. There's nothing elitist about it, since everyone has the mind power to take that introductory course.
> It's also good to ponder if the details we get worked up about actually matter. Outside the academia, approximately no one needs a precise, CS-theoretical definition of big-O notation. Practitioners use it in a looser sense.
The article made some genuinely serious mistakes and the author here (graciously, it has to be said) admitted to being wrong about some things like Big O being "the worst case".
In this cases, maybe the damage could be limited by saying "I know this isn't Big O, this is what some engineers call it but it's actually something different. Because practitioners find useful nonetheless, I will explain it (explanation follows)".
I found the visual presentation top-notch by the way, it's clear some effort was put into this and it shows!
Before asking why, ask if. There are good articles about complex topics, they just get drowned out by bad articles.
My experience is the opposite. I hate the colorful books with many little boxes, pictures with captions in floaters, several different font sizes on the page, cute mascots etc, where even the order or reading is unclear.
Instead I found it much easier to learn from old books made before the 70s-80s, sometimes back into the 40s. It's single column, black on white, but the text is written by a real character and is far from dry. I had such a book on probability and it had a chapter on the philosophical interpretations of probability, which took the reader seriously, and not by heaping dry definitions but with an opinionated throughline through the history of it. I like that much better than the mealy mouthed, committee-written monstrosity that masks any passion for the subject. I'd rather take a half page definition or precise statement of a theorem where I have to check every word but I can then trust every word, over vague handwavy explanations. Often these modern books entirely withhold the formal definitions so you're left in a vague uneasy feeling where you kind of get it, have questions but can't find out "is it now this way, or that way, precisely?". And I'm not so sure that these irrelevant-cute-pics-everywhere books are really better for ADHD at the end of the day, as opposed to distraction free black on white paragraphs written by a single author with something to say.
It's a bit like the choice between two doctors. One is very polite, spends an hour by your bedside every day, smiles and holds your hand and encourages you not to give up but has no idea about how to interpret your symptoms, has a horribly confused idea of what's going on but can explain that mess in his head to you in a reassuring tone that feels authoritative.
The other doc is morose, doesn't smile, the meetings are brief like an interrogation but he knows exactly what's up, spends the time on reading the literature, case studies, phones up other doctors who treated similar illnesses, cuts you and sews you back at the right spot and hands you the exact pills that will get you on your feet in a few months, but at the same time was distant and cold throughout the whole thing and talked to you as if you were a car in the repairshop.
Ideally, it would be both. Of the two I'd still choose the second guy.
Don't think the entire internet is repeating inaccuracies. :) I also believe there are readers that attempt to learn further than a blog. A blog post can inspire you to learn more about a topic, speaking from personal experience.
If there were no blog posts, maybe there would be no HN I believe.
There should be a place for non-experts. One could remain skeptical when they read blog posts without hating blog posts about complex topics written by non-expert.
Where it falls short is the usefulness to others.
The more detail that’s included, the harder it becomes to write less. If you ask me to write 500 words on indexing strategies for an RDBMS, I’ll probably manage, but very little will be explained. If you ask me to write 5000 words, there’s no way I’ll hit that target, because I’ll want to explain every tangent, every fundamental, and so on.
The single best blog [series] that manages to pull this off, IMO, is Jeremy Cole’s series on InnoDB [0]. It’s not short taken as a whole, but each article is short enough that it (probably) meets people’s attention spans. I think the thing that makes it so great to me is that it’s written assuming that the reader already understands certain concepts. For example, he doesn’t spend multiple paragraphs explaining what a B+tree is; he links to the Wikipedia page, and then describes how InnoDB uses them.
In contrast, when I write something, I usually find myself assuming that the reader doesn’t have some fundamental knowledge. I usually try to either put it in a section that’s easily skipped, or in a pop-out toggle, but either way, it’s there.
To me, these fundamentals are incredibly important, because a. they’re how I understand things b. I don’t feel that you can operate something to its fullest capability (let alone fix it) if you don’t know how the thing works.
Re: gatekeeping, to me the perfect example is the Linux kernel. Torvalds is notoriously unfriendly, and while you can judge that however you’d like, the man has a massive weight on his shoulders, and letting code in that doesn’t meet his standards threatens that weight. I get it. Then, I also came of age during the RTFM era, so I’m used to it.
If you say, “hash maps are generally O(1), and are a great default choice,” you’re not that wrong (subjectivity aside).
If you say, “storing integers as strings guarantees precision, so you should prefer them over decimal or float types,” you’re technically correct about the first part, but are enormously wrong for a variety of reasons about the conclusion.
The thing I was taught wasn't wrong, but it left out important details. There are very specific steps involved in cleaning parts, applying the right lubricant, not applying lubricant, aligning parts, not forcing things, doing the same change on both sides, torque specs, bedding steps. If you don't do them, the brakes just fail again. The person who taught me didn't go over those important yet intricate details - probably because they were taught by a non-expert too.
If the choice is "quit my job to start writing lots of expert content", or "accept the status quo", I choose neither. I have my own life to live. But if during the course of living my life, a wave of inaccuracy begins to lap at my door, I will attempt to stem the tide, in my own way. The toxic aspect, I think, is really just the way in which these corrections are given, and definitely I and others could do better with our tone. But to just give up and not give the corrections at all, I think would be disastrous.
(fwiw, I think HN's overzealous "guidelines" force people to either be toxicly positive or insidiously critical. that's not to say I'm not an asshole, but it's hard to just be normal on this forum without this bizarre culture coming down on you)