I think the article hits the nail on the head. DNS itself isn't hard, but learning real-world DNS is, due to the (as described) hidden nature of many parts betw

I think the article hits the nail on the head. DNS itself isn't hard, but learning real-world DNS is, due to the (as described) hidden nature of many parts between intending to do a lookup and getting the expected result. Some of this also stems from the divergence of what used to be a 'default' way to be connected to the internet (one interface, one gateway, one DNS server provider) and what is now far more common like having to figure out how to actually do the lookup in the first place (connected to multiple WANs, sometimes concurrently like LTE and WiFi) and as a user, not actually being able to tell which resolving path was used (was it the browser? the standard interface in the system C lib? was it some local resolver or recursor that sits in between the lookup? do any of them have a local cache? do any of them add special options by default?). Even if everything works, you'd still not be able to blindly trust if the query and response used the same path in one application as it did in another. Three browsers might all use something different, and the OS itself might use something different yet again, only to have mDNS add a fifth option.