Please forgive me if what I’m asking doesn’t make any sense; I can understand that there may be good reasons for things being as they are, but, being a noob, I don’t see them.
So, every time I look at a new core release, major or minor, I see things that even to me ( 
) look like they belong in the core, but also a lot of things that seem to me like they are specific to things, devices, brands that not everybody has (Victron, Chess, Hisense are just some examples).
So I’m wondering if all these couldn’t be integrations or such, and not part of the core.
I feel like the core, therefore installations, backups, updates, etc. could be smaller and faster for most users.
Thank you all for your patience
1 Like
They are integrations. Core integrations. If you don’t use them then they aren’t loaded.
Those are integrations. If you don’t have those integrations installed there’s nothing to be updated for them and they don’t occupy space in your backups or installation footprint. But they need to be part of the updates in case you do have them installed.
Also your profile suggests that you’ve been here longer than I have, you’re hardly a noob 
, you just may not understand how this works.
Edit: Dam… tom_l just beat me to it
Thank youse 
.
I think noobiness also depends on… density (now that I used this word it feels very fitting in multiple ways 
) not just time.
Maybe I get it… although I thought integrations are outside the core, and need installing. If so, I still don’t know why they are mentioned in the core updates.
Maybe there could be a separate list for integrations, if it wasn’t too much trouble and not undesirable for other reasons?
Like the wiring loom in most cars, there are connectors for added options. If you haven’t ordered them, the options aren’t connected but the wires are there so the manufacturer only has to keep stock of and fit one item. Think heated seats, remote controlled mirrors, etc.
There actually is a list of available HomeAssistant integrations. See Settings/ Devices and Services and click the + Add Intregration. HACS also has many third party developed integrations that are not core components but may solve your requirements.
Reading a little further in the vendot documentation will improve your understanding and moderate your temptation to install every possible option to see what it does.
Yes, being a newcomer can be exciting. The knowledge curve is steep but rewarding.
That does make some sense; although in general there are other ways.
It was like that for cars, and to some extent it still is, although many things have been moved on a bus that itself does not require updates so often. Upgrade lights? Same bus.
I don’t know if a similar single API (I don’t know if it’s the right word) could allow the core to talk to any/all integrations, but if it was possible I feel it could be a more stable core.
I’m talking here philosophically, partly out of curiosity, not expecting much; I suspect there may be reasons for not doing this, even if just large rework.
And being a newbie, while exciting, can also be overwhelming.
Eat your chocolate elephant one byte at a time.
Stand on the shoulders of giants, but be aware of how far off the ground you are if you fall.
It does. That is by design, and its’ strength. Universal interoperativity is very powerful.
It is, considering what is expected of it.
Home automation is a fast paced field. Even AI finds it hard to keep up. Some of the functionality approaches magic for some users. Behind the scenes there needs to be constant tinkering to fine tune new functionality while maintaining well proven solutions that still soldier on, some many years after their original makers and vendors have passed on into oblivion.
The balance between stability and feature creep/rampage is a hard path to balance well. See how vendors like MicroSoft have handled Windoze over the last half century for something to compare and contrast.
I realize now I may have implied it’s not stable (enough), and in fact I don’t think that; I only meant that the smaller a thing, in principle the more stable it would be.
I was going to mention how windows used to need restarts for most updates in the past, and now much less so. Then I realized many HA updates also don’t need restarts, so I didn’t 
Then: If the core had like a universal bus, it would only rarely need updates to accommodate new features of existing integrations.
Now feature creep… that’s another thing.
Much new stuff IS amazing, and I’m glad it’s there, but some things don’t work so well for me; and it’s very likely that some of it is because I don’t know what I’m doing, but maybe not all. For example I have an ESP32… voice satellite that can talk to cloud AI, and it mostly works, but voice responses often stutter or truncate. Also, it seems there are multiple ways to make it turn lights on/off by voice, but so far I felt I had to list all possible ways I can ask, and others sometimes work, sometimes not, so I installed an add-on that, when I first say “command” it knows it’s for HA rather than just generic LLM question, and sometimes it seems like it works, but I’m not sure… It feels like a LOT of work to get things going and then the way often changes, or a better way comes along (can’t complain about that) but often it’s not plug’n play.
I don’t know how I got to this ranting from my original question; maybe because I would prefer less speed for new things, but that they work better from the beginning.
Now it’s all about CI, CD, sprints… I’m not sure that’s the reason.
I used to develop FW, and did not have to release a new version until it was fully working/tested. No monthly or other scheduled releases. I liked that better. Have something new? Release it when it’s finished. If beta testers are beneficial, release it in the beta channel (again, I’m not sure my terminology is right) until it’s ready for release.