Hacker News Reader

Comment by akerro | Hacker News Reader

That's how we ended up with 1000s of garbage quality tutorials for android development. I was once looking how to save app specific data and all top 10 tutorials in google and DDG were saving data to userspace by creating new directory called something like 'myapp'. Years later Google had to create a series of tutorials with Udemy (as I recall) to actually tell people how to do things in a not stupid way.

Please don't write tutorials about things you do NOT know. Newcomers will visit your blog and think you're an expert and do it this way.

View on HN

Just be honest with your readers by stating your level of expertise when you're putting material out there, whether it's a blog or a book. Besides, even if we got rid of the 1000s of garbage quality tutorials, there would still be 1000s of expert quality tutorials that are nevertheless worthless because technology has moved on and made them obsolete. Of all the many books and blogs published around technology each year, very few will be relevant the next year, let alone make it to timeless classic status.

this is why (esp w/ js) I always use the google search w/in the past year feature. When looking for src code for ideas on ghub, I always sort by last updated or most starred but make sure the ones I check have been updated in past year... etc... Standards change so fast, and I feel coding is getting 'better' in terms of processes, that doing it like they did in 2015 is a step back sometimes, plus js itself has changed a lot, so has laravel, and other frameworks. Always good to get the latest data on something.

As a side note, I've found some 'gems' from tutorials that I knew were shitty novices or intermediate devs at best just writing to expand their career or do inbound marketing or something, but they made me think about a solution to something differently, or they touched on something I didn't know anything about, having never done it before like some api or auth system.

TLDR: I think with all tutorials it's buyer beware - especially coding tutorials, filter by date, don't expect perfect 10x dev tips unless you're on egghead, laracasts, railscasts, pluralsight (some other coding mooc that claims master level expertise), and if you write posts as a jr/intermediate dev, maybe try coming back to them when you 'level up' and make them better/more relevant.

Definitely agree with your tldr, there's some responsibility on the reader side as well. Even when authors don't declare their expertise, the context of the information, which might just be its age, can often be enough. You should know what you're getting when you read and take examples from the old PHP doc commenters, for instance.

I was setting up emacs + evil mode the other night to give it a real spin for the first time, one blog post from 2014 was pretty helpful but I suspected I'd have to throw out a bunch of things (if they worked at all) as I got further along and learned more about the ecosystem. Sure enough the way people add packages has changed, but this blog author helpfully updated the post in 2016 with the new information (and left the marked-as-old info there for context).

Yeah, part of becoming a better coder, is learning to scan through the 'chaff' to get to the good stuff. I'll usually open 5 different articles at once, scan em, see what applies to my situation the best, which one seems to 'know' their stuff better, etc...

One reason I hate video tutorials -- I don't want to sit through 30 minutes to figure out the 20 second tidbit I need to know, I don't even read 50% of blog posts I just scan it for what I want and go... lol

Exactly. No tutorial claims to teach you EVERYTHING about the subject. A tutorial is a light weight introduction to a topic.

You're using it wrong if you expect it to be prod grade. New comers have a big task of understanding the language & framework. Writing "prod" worthy things is the last thing on their minds

But that still fills search engine results with garbage I have to filter through. It's hard enough searching for solutions to problems when there's five StackOverflow answers with full of bad data and incorrections scattered through the comments, answers, and comments on answers.

Exactly. Bad and/or wrong is worse than nothing at all.

Not the writer or publisher's problem imo.

It's also not my problem when I throw my garbage into the river.

Interesting. I don't have any opinion on the incidence rate of low-quality technical tutorials, but I think this is the first time I've seen Google search results analogized as a fragile public good we all have a more-or-less moral responsibility to be good stewards of.

It's absolutely excellent way to frame the problem.

I've umm'd and ah'd over this as well, thinking it'd be a good idea to have a structure to learning by basing it around teaching others - and then I think about the number of tutorials I've gone through, only for the last paragraph be "obviously, that's not how you'd do it in production" with no next step and wanting to punch the screen.

If you're going to teach things, make it actually work for what the person reading needs. Personally, the chances of me reading a tutorial for interest vs wanting to accomplish a task are minimal.

Tutorials for learning, articles for interest. I'd say you could do the latter without expert knowledge, but the former needs to be pretty tight.

(All of this with the knowledge that imposter syndrome is so prevalent that we don't want to put off technically brilliant people writing tutorials when they think they might not be up to snuff).

Tutorials are for getting quick bootstrap on the platform. If you want to learn how do something in production, there's usually a manual or official user guide in place.

But why would you want to bootstrap your knowledge on something detached from your actual use case?

Lots of people on HN bemoan the lack of formalised education and crappy code stitched together, but that's what half arsed tutorials and impenetrable documentation breed.

>Years later Google had to create a series of tutorials with Udemy (as I recall) to actually tell people how to do things in a not stupid way.

Well, Google should have done that from the start and there wouldn't be a problem...

To be honest, this sounds like an issue with search engines. Why are "garbage" tutorials getting high rankings in search engines? Why aren't good quality tutorials rising above them?

Because search engines are designed to show what’s popular, not what’s correct and thorough. Search engines still cannot weigh an article’s sense, only its “pagerank”. This is an issue with people, not with a search engine.

But that merely moves the question on. Why are garbage tutorials more popular than good quality tutorials?

Probably because people who look for tutorials are not experts, so it's hard for them to tell if a tutorial is good or not.

Because a garbage tutorial isn't necessarily incorrect. The solution in the tutorial might be slow, inefficient, un-optimized, hacky, etc., but if it works when people have been slamming their head on the desk for hours, people will use it. It's a matter of which tutorial captured the audience first.

I think it is just numbers.

There are many more poor quality tutorials written than high quality tutorials because high quality tutorials are much harder to write.

The writer has to know the domain, be up-to-date in the domain, have the communication skills and the time to write and edit.

Because people find them using search engines optimised for popularity (not quality) of content, as perceived by the readers, not a community of experts.

I can speak for "data science" tutorials: there's A LOT of people saying you don't actually need to know the math!

Learning math makes you feel stupid.

>Why are "garbage" tutorials getting high rankings in search engines?

Because John experiments with X for the first time and writes about it on his blog. Janette has the same problem while learning about X so she takes his code, makes a few changes and puts it on her blog. This scenario happens 8 more times and we have 10 low quality tutorials about how to do things, but none of the authors has experience and expertise in X, and they won't admit it. Bad practice is copied to many places. 1000 garbage kill 1 gold.

This explains why 1000 garbage tutorials are created, not why they are ranked high in the search engine.

The same exact mechanism (and the parent already alluded to it in his comment: "Janette has the same problem while learning about X so she takes his code, makes a few changes and puts it on her blog").

Non experts look for tutorials,

They chance upon tutorials by non-experts

They don't know enough to tell they're junk

They use the code, link to, and suggest those tutorials to others

Those tutorials are getting higher ranks, attracting even more non-experts

Because search engines don't rank quality of source code and tutorial, they rank page layout, bolds, italics, font color and page speed. Literally any garbage can get to top3, you dont need to be good Ruby tutorial writer for it.

and they won't admit it.

There's your problem. If you are writing a "here's how I did it" post, you should be clear that you don't have any idea what you're doing.

That doesn't address the indexing problem.

Maybe the best would be to write a tutorial, and then throw it away.

Keep it to yourself, then revisit it a year later and laugh at how little you knew. Maybe then update it and publish it since it will have both a beginners perspective and an experts perspective.

>How Not to Xyz

There is a high possibility that you'd be surprised how many dark corners you visited a year before. Depending on the subject you may deal with quite small surface area on the daily bases while preparing the tutorial you had to dig much deeper.

Yes, this is exactly what artist do when practicing their craft

I do feel your pain with these crappy tutorials, and many of them are glorified blogspam or recycling/reheats of the same thing you could have already found out by reading the docs. There's no depth, no referencing of documentation, very little reasoning about why the thing you're explaining works.

That said...ideally if the author abides by (1) do proper research and (2) "Your tutorial can be reviewed by other brilliant engineers who can point out your mistakes" then there's no reason why you couldn't produce a reasonable quality tutorial in this way.

I think this article is written for developers/engineers who's overall skill level and expertise is a cut above inexperienced folks who're just starting out. i.e. you know how to research, you have a working familiarity with the tech stack you're working with (and broader awareness of other tech stacks), you know how to ask a "good question" on Stack Overflow (or have read and understood Jon Skeet's "Writing the perfect question" article[0]).

It's kinda like taking rubber ducking to the next level and sharing your learning experience.

[0]: https://codeblog.jonskeet.uk/2010/08/29/writing-the-perfect-...

Completely, completely disagree.

First, some people probably will find what you write to be somewhat useful. Second, it's the absolute best way to learn something.

In your example, the problem was Google, not the myriad of people learning and writing what they learned.

Writing is one the best ways to learn something. It doesn't follow that that writing should be published.

The problem with blog tutorials is that it is democratization without meritocracy. There is no curation mechanism to bubble the good stuff to the top. Search engines rank by popularity, not quality. At least with forums and Q&A sites, there are natural feedback/reputation mechanisms to achieve some semblance of quality control. Blogs don't have that (they only have commenting, which depends on someone caring enough to comment).

This results in very low signal to noise, and a tragedy of the commons situation happens. In most systems, there is a cost to high-volume low-quality content because it discourages participants. (I don't know if you remember MySpace...)

You can't stop people from publishing low quality content, but if only there was a way to curate such content based on some standards (not gatekeeping... but curation). Quora tried and was successful in the early days because it started with a high quality contributor base, but that base has now been diluted and the platform no longer offers an enjoyable experience.

>> "to actually tell people now to do things in a stupid way."

You mean "how not to do things in a stupid way", right?

The alternative and positive aspects of this is that new people learning provide their own experiences in the current process of understanding at a basic level. And with that the problem areas or more difficult concepts or just interesting solutions to figuring problems out will have recently discovered routes and tricks on how they overcame those obstacles, versus when experts speak only like they are teaching other experts.

This especially is annoying when there are simpler parts of some of these currently dense technologies and it's just a bunch of jargon, of what is supposedly introductory content.

Newcomers should learn by reading the official documentation first and maybe some "official" books. In my opinion blog posts should be about very custom things or more specific use cases. Like integration with different technologies, hidden features or advanced stuff. If I want to learn React, I won't be looking for blog posts or googling "how to create a react app". I'll go to the official documentation, then grab some classic book endorsed by some community.

If you want to learn the whole thing then yes, but a blog post can be a nice 'easy' tutorial that lets you start using the technology with far less initial effort, and then you can learn the rest modularily according to your needs.

I think you mean Udacity rather than Udemy

I would rather have to search through a mountain of trash to find gems, then have nothing to search through at all.

Well at least you have 1000s of tutorials. Even if the way they are doing it is bad or misguided it get's you on the path.

I disagree... 1000s of tutorials explaining how to do things incorrectly just leads to thousands of crappy productions deployments.

I’d rather say, “no tutorial? Ok, to the docs like the old days,” than be misled by a shitty tutorial.

Another problem is that thousands of turorials get few to none updates. It’s not unusual to google for whatever.js quirk and find yourself in a tutorial that uses method that was deprecated 2-3 versions ago. Of course part of this problem is in whatever.js itself, that bumps versions like hell, but you end up with tons of crap that will be never relevant anymore, but still hit the top because its author still writes popular articles.

Idk about HN-ers, but I find myself in this situation at least few times a week since the beginning of my webdev.

Hey at least you have a home, who cares that roof is leaking, walls are breaking and you sleep on pillow of mould.

Reply on news.ycombinator.com