I actually wrote about this exact thing: https://xavd.id/blog/post/my-perfect-task-app/#dependent-tas...

My hope for the perfect task tracker springs eternal.

So to that end, I'd probably start with a project-oriented book, or give them tasks that require programming to solve (e.g. tell me the distribution of words in this Shakespeare play. What are the most common ones?).

Editor wise, VSCode is great, but may be too many bells and whistles. Sublime Text may be a better place to start, since it'll do highlighting, formatting, etc without being overwhelming.

Language wise, Python, JS, and Ruby all have ample beginner resources. Depending on what else they're into, JS is probably the most widely useful (since it's on most web pages and the sandbox is built right into the browser).

Good luck! Let us know how it goes!

That's so interesting- I think I have the opposite response. I figure unless I'm doing a very specific case that doesn't fit well with the stdlib sort (for some specific reason), It's probably good enough. Those language folks usually know their stuff, who am I to try something else?

1. They supported 3rd party watch faces based on a file. There was a great little site that let you use a visual UI to build the file and, if done from your phone, upload it right to the watch. Once, when out at midnight beer launch, I made a custom face for the beer and won a little keychain from the brewery. My apple watch is nowhere near this functionality.

2. Once time in 2014, I ran into some Pebble engineers at an ice cream shop in Palo Alto, they spotted my watch, they bought my scoop! Fun, weird little Silicon Valley moment.

  - What bits of a changelog are most helpful / important? 
  - What's not helpful, but usually included?
  - How should the document balance skimmability / comprehensiveness / readability?
  - What projects have the most helpful changelogs?

Comments URL: https://news.ycombinator.com/item?id=30995647

Points: 5

# Comments: 4

It's basically just text and images, but I'm proud that I did the whole thing (content, design, implementation) and the overall attention to small details.

Also, I love the domain. @xavdid is my username on basically everything that isn't an account I've had for a very long time (eg this one).

I mean, I think there's something to that. Bash has a lot of arcane rules; I personally never write an `if` statement right, what with the `[[` vs `[` and having to surround the condition with whitespace. I find that bash scripts carry a million little gotchas that will cause variables to not evaluate/expand the way you'd expect.

Say what you will about Python, but it's much easier to reason about what a script will do as compared to bash (unless the bash is _extremely_ simplistic).

If you only watch one clip, check out this segment from the training montage: https://www.youtube.com/watch?v=yCk9VAxEpD0&t=156s

Their use of visual boundaries and transitions is just so superb.

It's all about adding layers of prevention that each drop your chances of catching it (and its potential severity) by N% each. Get enough of those, and your chances start to look pretty good.

For a lot of the 20th century, _most_ people could buy a house regardless of their chosen occupation ("most" is doing a lot of work here- there's a long history of racism tied into housing acquisition, but my basic point is the same). You could be a blue collar tradesman and buy a house for your family of 4. So yes, I think that's definitely a (theoretical) possibility.

It's not outlandish either. We already have a much healthier model to copy: cars. There's no one buying up _all_ the cars and then leasing them to you at artificially high prices. There's (normally, chip shortages notwithstanding) plenty of cars and you can pick what best suits you- new / old / cheap / expensive / fancy / plain / etc. Housing could be the same way if the goal wasn't to extract profit out of those who can't afford to buy (or weren't born 60 years ago).

I think the value here _really_ depends on your usage patterns. For you, it sounds like individual discs make a ton of sense. But, many people watch more movies than that (during the height of lockdown, we were watching 1-2 movies _daily_), so streaming services made the most sense. Sure, it's annoying that movies come and go, but the price is certainly right.

Once size won't fit all here, so it's definitely great that discs are still an option. I'm glad it works for you!

I think with most things, the answer here was to have a mix of keys and buttons. With 100% TB, it's slow to change volume/brightness. With 100% keys, I've got all these extra keys I can't use (you can re-bind them with 3rd party software, but it's fairly static).

I'd personally pull like, F3-F9 in favor of a touch bar, and then be able to configure the remaining buttons. I'm a little sad reaction was so strongly against the idea on the whole, it definitely had potential.

FaceID is better for:

  * in-app auth- when an app (e.g. a financial one) has a secondary auth check, FaceID passes without extra effort- it's very slick. For TouchID, I usually had to move my phone to the button (because it wasn't already there)
  * using unmasked- It's fast enough that under normal (unmasked) circumstances, it works great
  * dirty/wet hands - especially since the phones are waterproof, it was always annoying to have to do my passcode if my hands weren't totally dry. I didn't think it would come up a lot, but it did

  * Pandemics. The Apple watch unlock works 70% of the time, but that's not 100 _and_ you have to own a separate, expensive piece of hardware. They're fixing this in the latest iOS beta (FaceID can learn your face w/ a mask on by checking details around the eyes).
  * Bed- my face is usually squished on a pillow and the phone is close to my face; it can basically never be read

So, it seems like there's no silver bullet. Both would be great, but I think I understand why they made the tradeoffs they did.