Hacker News Reader

> Read at least HackerNews and possibly some other tech news site regularly. And then make sure to not get too caught up in the hype for new stuff.

Interesting because I advise the exact opposite: do not waste time reading HN, just find one thing that is interesting and master it. Then repeat.

As much as I like HN there is a lot of noise here. Novice mind can easily fall into a trap looking for signal here.

The advice elsewhere in this thread about focusing on one stack and getting good at it is definitely true.

Broadening your knowledge to be aware of other stuff is also useful though. The most useful people are those who understand how multiple things work and fit together.

> there is a lot of noise here

This is true. However, much of what I know about how software development can work, I've learned from reading HN. I've learned loads about cloud, containers, micro services, project problems, etc. This helps me understand all the discussions around me while I was a junior and allowed me to successfully lead development of a SaaS product. If I just focused on learning a single thing, I wouldn't have been able to grasp all the other stuff that is necessary to actually build a product from scratch.

I've gotten a lot of value from reading HN. I've also wasted a lot of time. It helps to develop a personal filter. Mine is

* Avoid headlines that make me angry, unless the news is really important. The discussion is never worthwhile, don't even bother.

* WSJ, Bloomberg, and WaPo won't let me read articles with w3m, so I skip the article. If the topic is tremendously interesting, I'll check the comments.

* TED and nautil.us are the information verison of empty calories. I've never gained any long term benefit from reading them.

* Javascript framework churn is intense. It's not worth trying to keep up with the details unless you do web development. If, like me, you don't, it's still worth reading the headlines to keep abreast of what frameworks are popular, what paradigms they embody, and who's pushing them. This changes every six months or so.

* It's always worth scanning the comments in threads about programming languages. There are some really interesting people who sometimes jump into those threads and blow your mind.

* Know your prejudices and look for information that challenges them. My prejudices are that I think Windows is lame, that MacOS is infuriating, that Javascript is terrible, that vim is better than emacs, and emacs is better than IDEs, and object orientation is a failed paradigm. So I look out for articles that have a different point of view. I've learned that Microsoft has some really interesting technology, that Apple is infuriating even to its own users but it doesn't matter because they're sitting on the world's biggest pile of dollar bills, that Javascript is usually unnecessary but occasionally useful, that kakoune is even better than vim, and that nobody agrees about what object orientation even is.

"object orientation is a failed paradigm"

uh... wtf?

I know my prejudices. I recognize that a lot of people think OO is just dandy, and I'm on the lookout for a good argument in its favor. I'm open to the possibility that the reason I think OO doesn't work is that I don't understand it well enough. Maybe somebody out there will explain encapsulation in a way that doesn't make hidden mutable state seem like a nightmare. Maybe somebody's got a blog post that explains why inheritance only looks like it makes spaghetti code, but actually I can safely ignore the dragons lurking three levels up the hierarchy. Maybe there's some piece of OO code out there that embodies a beautiful solution to a problem I've been struggling with. If those things are out there, they're going to get posted to HN eventually.

Here's an explanation from Casey Muratori: https://youtu.be/GKYCA3UsmrU?t=4m50s

See also Mike Acton (Engine Director @ Insomniac Games): https://www.youtube.com/watch?v=rX0ItVEVjHc

You aren't going to get a lot of support for OO in places where it isn't really necessary like games, and especially games with single developers. At that point the management of data is on the developer. Essentially one is saying that if they fuck up the data, that's on them.

OO is handy in enterprise software where there are lots of developers who may not understand why things are the way they are, they just need to do something. In big, complex systems where a lot of things are going on, if you give someone a noose, they will hang themselves. OO can protect data from some (of course not all) unintended consequences.

> Avoid headlines that make me angry, unless the news is really important. The discussion is never worthwhile, don't even bother.

> Know your prejudices and look for information that challenges them.

How do you prevent the two of these from coming into conflict?

Headlines that make you angry are designed to make you angry. It's not about information that challenges your prejudices, it's about being intentionally provocative. I'm not interested in being baited.

> WSJ, Bloomberg, and WaPo won't let me read articles with w3m

w3m, nice! I tried lynx, w3m, eww (emacs) lately, and liked w3m the most, but found it terribly slow on many sites. In particular GitHub. Just measured 12 seconds for loading my Profile. lynx takes 5s. Opera takes 2s.

Do you have similar experiences?

Yup, it's slow for me too. I like it anyway. I want to like links, which is way faster, but w3m is just so simple. If I had to guess, I'd say it's because w3m doesn't support gzip compression.

I eschewed QBasic early on because it wasn't "real" or it was derived from Basic. This was a mistake.

I spent a long time trying to become good at C++. This was also a mistake.

I ignored Python for a good 2 years because of indentation. Again, a huge mistake.

Great suggestion on anger inducing articles/comment sections. I was just thinking it would be nice to do sentiment analysis and incorporate the emotional vector into both the dynamic rules and the layout. At least put a badge on the discussion so I can know to stay away.

I'm curious - could you elaborate on your mistake regarding spending too much time on C++?

It was too large and too complex for the kinds of things I needed to do. I over focused on getting better with my tools rather than working on domain problems.

Since it has affordances for speed, it encourages one to make fast code, but this is rarely needed. Making fast code is a trap.

Trying to become good or very good at something also puts you in direct competition with others doing the same thing. I did not have time to compete, so it wasn't a big differentiator for me.

C++ helped me learn when do give up, and when to refocus my attention in areas where I can have greater impact. I wish I was a little more technically and philosophically mature before I undertook C++. Having concrete goals is a huge clarifier and allows one to judge something with a much better light.

> and master it

Mastering something is expensive, one should have multiple good justifications for doing so.

Deep familiarity is what people usually need, beyond exposure, but below mastery/trial by fire.

Along the lines of reading HackerNews, read TheDailyWTF.com on occasion. It's fun and entertaining, but that's not why I'm recommending it. A lot of the bad code, bad error messages, bad designs, etc. were not necessarily done by someone _incompetent_, but rather people who are simply inexperienced in one way or another, or were in a rush to meet a deadline and weren't careful.

(At a former company, working in a team full of very strong engineers, we still found one dialog box of ours end up on that site once, and the issue it demonstrated led us to a bug in the code.)

If you start seeing things on there that look like something you'd have done, don't beat yourself up over it, but learn why those are WTFs and re-evaluate how you're writing code/designing software.

> Read at least HackerNews and possibly some other tech news site regularly.

Just be warned that places like this are the worst kind of habit-forming variable-reward site. Worse than Facebook, really. Every now and then you'll find here something that will tremendously improve your worldview, programming skill, or propel your career forward. You might then find yourself using this as a justification for wasting disproportionate amount of time on reading random things, instead of doing.

(Yes, I'm very much stuck in that trap right now.)

Do you have any recommendations for news sites that focus on embedded software? I've recently started a career focusing on embedded tech but most of the interesting stuff I've come across online seems to be confined to blogs from the 90's/00's

My favorite, along with about everyone else in the office, is https://hackaday.com