24.83 fatalities in 2019 and 25.95 in 2018 per 100 million miles ridden on motorcycle according to the source of the other article: https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812979.pdf

If we take the value for personal flights:

21.1 fatal accidents per million flight hours

I’m going to assume 125 mph as I’m skeptical of your 145 mph figure.

= 21.1 fatal per 125 million miles

= 16.88 fatal per 100 million miles of personal flights

I’d basically conclude that personal flights and motorbike are in the same ballpark whether you’re looking at hours or miles.

Motorcycle:

16 to 18 fatal motorcycle accidents per million hours ridden

Personal flights:

21.1 fatal accidents per million flight hours

I’m really surprised at that.

Thinking about it though most bikers I know are young, have good reflexes and are mostly using a motorbike for commuting or travelling which is probably lower risk than the cliche teenage fool on a superbike my mind jumped to.

By contrast, the vast majority of light aircraft pilots I know are 60+, many 70+ with extensive health issues, heart problems and likely comparatively poor reflexes.

corporate aviation – the world of bizjets – is relatively safe. This sector only saw 0.48 fatal accidents per million flight hours from 2012 to 2021

General aviation overall:

9.5 fatal crashes per million flight hours

No, I would agree that comparing by hour is reasonable. Most light aircraft use (in my area at least) is generally recreational. People are flying because they enjoy the hobby rather than because they have a destination in mind, therefore they’ll fly for a long duration but the distance travelled might not be far

Plus planes are checked before every use, and very few people are flying light aircraft in a reckless manner.

I’d be very surprised if there were any stats to backup a claim that light aircraft are as dangerous as motorcycles.

So, you don’t know Python at all AND you don’t know Bash, but you feel compelled to talk about how one is so much better than the other?

I have plenty of experience with Bash, hence why I was eager to question the implication that bash was less complicated than other solutions.

You’re correct that I don’t know python, but I do have plenty of familiarity with PHP, JS, C#, and Rust. From my experience with those languages I guessed that python probably has similar libraries for making API calls.

Thanks for providing the actual examples. Looking at them I’m curious if you still think I’m wrong?

In my opinion the bash is much more difficult to understand than the python and therefore it’s more likely for bugs to creep in. For example I think curl_exit_code=$? should be called immediately after the curl command as the way it’s presently written isn’t it capturing the exit code of the tail command?

You’ve explicitly called --connect-timeout and --max-time. imo it only comes from experience that you need to add these options. I had a script that had been funcitoning without issue for months then suddenly started to hang and it was a while before I figured out that the defaults for curl had no timeout so it didn’t gracefully fail like I would expect.

These are the kind of traps that I fall into all the time with bash and it’s painful to debug.

response=$(
  curl \
    --silent \
    --write-out "\n%{http_code}" \
    --connect-timeout "$CONNECT_TIMEOUT" \
    --max-time "$MAX_TIME" \
    "$API_URL"
)

http_body=$(echo "$response" | sed '$d')        
http_code=$(echo "$response" | tail -n1)

curl_exit_code=$?

Show us on the doll where the shell script touched you…

Everywhere.

I believed bash would be enjoyable, I believed we’d have fun. But all it ever gave was pain. I was young and naive.

That’s profoundly untrue. Scripting in bash is an indescribably painful experience.

You have absolutely no idea what version of a binary the user will be running so you’re limited to using only options that have been well established.

I’ve never worked with python but I understand it has at least got some semblance of package management providing assurance that methods you’re calling exist, and I imagine it has some standardised mechanism for handling errors unlike bash.

A simple example is making a GET request to an API and deserializing a JSON response if its successful, handling a timeout if the server can’t be reached or handling the HTTP status code if it’s not a 200 response.

JS, python, Rust, C#, Java etc will all handle that simple scenario with zero effort but in bash it’s a nightmare.

To run virtualization and and WSL in a debloated manner you may as well just dual boot Linux and Windows. It’ll remove so much headache that WSL creates.

Dual booting is well documented and you get the best of both worlds. If you later find it’s not for you then you just erase the Linux partitions and stick with windows

Do people in the US buy caged eggs?! In the UK free range is almost a ubiquitous minimum.

It’s in the article too. I don’t know if that’s a later addition but it’s there.

It really would not surprise me if it were an edit and the journalists’ source was the reddit thread

According to commenters on reddit who listened to the ATC recordings: ATC asked the blackhawk to confirm visual of the passenger aircraft. When the blackhawk confirmed ATC told them to wait and then manoeuvre behind the passenger aircraft.

The military aircraft was therefore responsible for maintaining a visual separation which clearly they failed to do. Possibly they were looking at the wrong aircraft. Therefore it seems likely ATC wasn’t responsible.

Obvs this all based on idiots on reddit claiming to have expertise so trust it as much as you would ChatGPT.

Surely the point of your own domain is to personalise it? Why would you go for something generic?