I guess it depends what you run, and how the projects/containers are configured to handle updates and “breaking changes” in particular.

But also, I’m being a bit broad with the term “breaking changes”. Other kinds of “breaking changes” that aren’t strictly crashing the software, but that still cause work include projects that demand a manual database migration before being operational, a config change, or just a UI change that will confuse a user.

The point is, a lot of projects demand user attention which completely eclipses the effort required to execute a docker update.

Are you updating 1000’s of stacks every week? I update a couple critical things maybe once a month, and the other stuff maybe twice a year.

I don’t recommend auto updates, because updates break things and dealing with that is a lot of work.

Documentation is for onboarding other people. Why on earth would I need to onboard other people to something self-hosted?

How is the average person going to know that? If Joe blow can’t easily get to the distro they “should be using”, Linux ain’t happening for most people.

I enjoyed the depth of this answer. That being said…

4 copies seems like a level of paranoia that is not practical for the average consumer.

3 is what I use, and I consider that an already more advanced use case.

2 is probably most practical for the average person.

Why do I say this? The cost of the backup solution needs to be less than the value of the data itself x the effort to recover the incrementally missing data x the value of your time x the chance of failure.

In my experience, very few people have data that is so valuable that they need such a very thorough backup solution. Honestly, a 2$ thumb drive can contain most of the data the average user would actually miss and can’t easily find again scouring online.

I’m not sure why the author chose this sentence and why you are picking it out. The author provided no evidence of it. Instead, when you read on it seems to be an ownership problem. People rotating in and out for a year on a ten year project. You can be the most competent and skilled worker, if you don’t get the opportunity to become invested in the success of a project, of course you won’t see the project become successful.

I love Linux. I use it wherever I can. I don’t use Linux on my primary gaming workstation, for the simple reason that the display drivers, specifically mixed extended desktop and screen mirroring is just straight up ass.

I think the design philosophy is that each tile represents no more than a single entity at a time, and compact enough that you can arrange the entities in a section to represent a device, room or group.

In your example, perhaps the room also has a humidifier and/or heater. So in reality, the room temperature entity isn’t truly tightly coupled with the fan. The heater and or humidifier are also a part of the whole.

With tiles, you are in complete control to arrange groups of entities that represent a larger whole, in whatever scope you like.

Python is just distancing itself from JS.

I’m using off-the shelf CT-clamps with an ESP. Obviously it’s a fair amount more work, but it’s cheaper than a commercial solution, fully offline and no subscriptions, you know exactly what you are getting, and you can build a solution that is just the right size for your application, and infinitely modifiable if your needs change.

Surely the SVGO package can be compiled into a browser bundle.

I might look into this myself…

Does this support SVG, i.e. SVGOMG/SVGO? If not, that’s a glaring omission.

Tried a couple matter devices. Ended up having to create an account with the manufacturer. Was there truly a local option? Who knows. So far I haven’t been impressed.

I think it’s a good idea, everyone should be automating this anyway.

This is still not possible in all scenarios. For example, wildcard certificates for DNS providers with no API support.

My experience with TP-link matter devices is you need a TP-Link account authentication. Excuse me? Fuck off TP link. I only want local-first devices.

I spent an unhealthy amount of time on Reddit. Getting bored of Lemmy is a feature, not a bug. Embrace it.

Before I understood Docker, I used to have HA installed directly on bare metal side by side with other “desktop” apps.

To be able to access devices, HA needs many different OS-level configurations (users, startup, binding serial ports, and much more I don’t have a clue about). It was a giant mess. The bare OS configuration was polluted with HA configurations. Worse, on updating HA, not only did these configurations change, the installation of HA changed enough that every update would break HA and even the bare OS would break in some ways because of configuration conflicts.

Could this be managed properly through long term migration? Yeah, probably, but this is probably a ton of work, for which a purpose-built solution already exists: Docker. Between that and the extra layer of security afforded by dedicating an OS to HA (bare metal or virtualized), discouraging the installation of HA in a non-dedicated environment was a no brainer.

If it’s such a problem, maybe we just collectively move on to ES or TypeScript nomenclature?

This is what happens when stack overflow is used for training.

The thing holding me back is multi monitor support. It’s just atrocious on Linux (and has been for a long time). I have issues with:

• can’t duplicate screens (“not supported” huh?),
• random screen not detected,
• broken layout when screen changed,
• flickering and constant layout changes

Not sure if anyone has tips? NVIDIA and Intel displays, scarred to even try VR.