SteveTech

I just read some more of your comments and thought I might properly explain VLANs:

VLANs let you create a whole virtual network within your physical network, there can be upto 4096 of them that can be tagged and 1 untagged per port, the VLAN ID defines which one to use.

A tagged VLAN is often used between routers and switches, so the connected device can pick which VLAN to use, but an untagged VLAN dedicates that port to that VLAN making it appear to the connected device as if it's the physical network.

Since it's a whole new network you need some sort of router to route between them.

As a rough example you could have something like: Router --2T--> Switch --2U--> TV, where the T is for tagged and U for untagged. Or replace Router with Pi if you use that, the Pi will access the internet with the (technically untagged) physical network, and route between tagged VLAN 2, meaning you can do everything on the Pi with 1 ethernet port.

Disclaimer: Most of this was learnt from experience so it might not be completely correct.

I believe that's the VLAN for the WAN, basically you only need to enable that if your ISP is using VLANs, but you want to enable VLANs on your LAN. I have the Telstra version of that modem (I recognise the Technicolor UI) and it doesn't allow you to use VLANs like you want to.

You could probably set up some routing stuff on a raspberry pi though, and use a switch.

Or if needed put your modem into bridge mode, and acquire a router that supports VLANs, I don't know of any cheap consumer ones, but I'm in the process of switching to OPNsense with an old computer. Unrelated, but in my experience technicolor has severe bufferbloat anyway.

Plus having to basically rely on a built-in app database/store to easily install apps

Someone else has mentioned that all GUI 'stores' suck, and in my experience that's true. I usually find things either from Google, suggestions, or apt search.

Kinda stinks to me, and not being able to simply download an installer from a website.

I also thought I might mention that standalone Linux executables do exist, I believe ookla speedtest publish one, and additionally AppImages are really cool in that it's a container the app runs in that you can just double click.

and having the program, whatever program, up and running reliably within a minute

I've only had issues with this when I've done something dodgy and I know that I've done it, when it's an app in the middle of development, or snap packages. Anything installed using the distro's repo, Flatpak, or AppImage, has always worked. I still use Windows for study, and I'm pretty sure it takes longer on Windows to go from download to app running.

I just did it like I would for a subdomain (with a CNAME).

Cloudflare does leave this little message though: CNAME records normally can not be on the zone apex. We use CNAME flattening to make it possible. Learn more.

I can confirm it works on root domains.

I guess if you have to use a classical terminal or terminal emulator, but I was more talking about drawing apps directly to the Linux console without X or another sort of windowing system.

For example this is Midori, a web browser:

The title's kinda clickbait, they're removing the 'Full' option and adding a choose your own apps dialogue to the 'Minimal' (and now only) option, and installs the selected apps over the internet. This reduces ISO size since the apps aren't installed by default.

Which is an action I can agree with.

You could probably get it to work in a framebuffer... 😏

Probably not though, although some apps like mpv, (maybe vlc) and mplayer can, plus QT and a GTK fork have support too.

I doubt it'll overheat, Pi's do thermal throttle when hot and from memory the Pi 3 runs fairly under clocked to begin with. If it's swapping you're more likely to burn out the SD card.