67 comments

[ 3.3 ms ] story [ 127 ms ] thread
I've never seen a compelling use of smart home technology, and I really don't want the continuing functionality my home to be dependent on the whims of some corporation.
Do you know what Home Assistant even is?
A home automation stack that alleges to be open source but can only deploy it the way they want.
What does the deployment have to do with whether or not the project is open source?
Stop hitting yourself, you're mad that someone who doesn't know what they're talking about, in fact, does not know what theyre talking about. Just keep walking.
The fundamental principles of FOSS are around freedom to use, redistribute and modify code. Many different things can weaken the freedom to use and modify. For example, if somebody gives you only binaries, or obfuscated sources, or only minified javascript.

Or projects with very difficult build process and no documentation, or unnecessarily complex deployment or too many functional dependencies. Or making the project not very modular, requiring people to deploy a whole VM or container with tenths or hundreds of dependencies bundled in.

The list goes on and on. For example a lot of commercial "open-core" products deploy a lot of tricks to make forks difficult to maintain in the long term.

Many are open source only strictly speaking aka legally, but they discourage REdistribution. Android is an extreme example of this and you can see how HA goes in that direction.

EDIT: There's a company funding HA development and profiting from it using a subscription model...

A fantastic drive-by comment that misses the point of both Home Assistant and Matter.
Which is exactly why Home Assistant and Matter exist... If you're not going to read the post why even bother commenting.

My Home Assistant instance has only a few devices that are in anyway linked to a third party but even then they have locally accessible API's. Zigbee, Z-Wave, Thread and Matter are all great solutions for this.

Does your household get mad when you forget to get up and stoke the fire?
One day - not too long from now - you're going to sprain your back. Or twist your ankle. Or get a nasty cold. Or, maybe, just get old.

As you painfully shuffle towards your TV in order to increase the volume, you'll mutter to yourself "I've never seen a compelling use of remote control technology."

Once done, you notice it's getting late. Do you continue sitting in the dark or heave yourself up to reach the lightswitch? You smirk. "I'm glad that I'm not dependent on the whims of some corporation."

"Progress" and "better" aren't a straight drive down the road of convenience for a lot of us.
One of the things that helps prevent your body failing is using it. If you are injured, they tell you not to favor your injured limb. As we automate these little movements out of our life, are we doing ourselves a disservice by increasing our own body's atrophy?

To be clear, I like technology and automation. I want to automate dangerous tasks and things that cause repetitive stain. Maybe things that need improvement, home air filters for example are changed every 3 months. Is that ideal? Or by measuring the current draw from the blower fan you can see when the filter is getting clogged and change it when needed.

Having recovered from several injuries: No, we're not doing ourselves a disservice by increasing our body's atrophy, at least not in this particular case. We're getting a better chance at giving the injured limb a chance to rest.

You shouldn't favor your injured limb, but neither should you expose it to unnecessary stress. So if recovery requires parking yourself on the couch, those automations are nice. (Or, alternative, loving friends & family whom you can use to do these things for you :)

And as for the uninjured case: If light switches and TVs are the thing that keeps you from atrophying, you have a problem :)

I expect to die at approximately 94, all things being average. I also expect that I’ll be bedridden for the last three months, and that I’ll spend most of those three months fighting with technology and cursing its frustrating interfaces, not entirely missing the irony of the fact I spent most of my life avoiding anything IoT.

My last words before I die of frostbite will likely be ‘hey google please please switch the heating back on’, to which the reply will be, for the fifty-eighth time, an apology and an offer to give me web search results on ‘swift the heath in baton’

After I buy a smart light switch that connects to my Home Assistance instance, how am I thereafter dependent on the whims of some corporation?
It’s admirable that you just leave this uninformed comment up instead of taking the easy way out and deleting it.
You would like home assistant then. It’s all local and open source. No whims.
Good thing then that those people are working in a way for you to not be dependent on their whims.

(I still don't like the details, but it's absolutely an improvement.)

IMO the only viable IoT is the one that can run completely isolated from the Internet, on a local network.

Home Assistant and Matter are a step toward this.

If only there were some open standard you could use...
> Matter is the reliable IoT compatibility standard that drives category growth by simplifying the purchasing process, expanding the smart home category to more retailers, and lowering operational costs. [1]

In case anyone was wondering what Matter is.

[1] https://csa-iot.org/all-solutions/matter/

The pitch for consumers:

> Matter is the seal of approval that says smart devices work reliably together—taking the guesswork out of the purchasing process. That trust allows you to choose from a wider range of the brands you love, and brings you the comfort of a secure and seamless connected home.

Which is indeed a major pain point with IoT so far: purchasing things that work together (instead of only cooperating in their own ecosystem and with Alexa). Of course that's more or less what HomeAssistant was already trying to do, but Matter might make this more appealing for "casual" users.

I thought zwave certification pretty much solved it? There are of course issues with higher cost, etc. but I thought zwave was the approved cooperating option and zigbee is the yolo option.
The ZigBee alliance next gen protocol is what is nowadays branded as Matter. So it is certification but also improvements on the stack. Which make sense to simplify things.
With zwave, you can't actually implement the spec yourself, you have to buy little bits of hardware from the only company selling them. That, of course, is pretty good for compatibility. Zigbee is just a spec, but no one actually follows it properly. Now the next version of Zigbee is rebranding as matter and hoping that this time everyone follows the spec. I'm assuming through some sort of certification program, which, as cynical as I am, may actually work.
Well, there's also the fact that Apple, Amazon, And Alphabet have said they'll support Matter.

So following that spec suddenly (pardon the pun) matters a lot. I'm sure certification plays a role, but don't underestimate "works with what a huge group of people have ready access to".

Suddenly, compliance becomes self-interest.

Yeah, but it's not _that_ useful to me if only giant mega-corps implement it. I want to buy stuff from Zooz and Inovelli, et all.
Yeah, but if the big three support it then that gives it a lot of gravity. There are a lot of people who will be much more likely to buy smart products if they can be controlled by their Alexa/Nest/Echo. Anyone who doesn't support it could be sacrificing a lot of potential customers.
Is it backwards-compatible with Zigbee, or do I have to buy everything all over again?
on top of what other commenters have said, Matter is IP based and supports Wi-Fi in addition to Thread as well as Bluetooth for initial pairing. With Matter devices should have more flexibility with how they connect physically and the network should be able to mesh with the existing networking in the house.
Do any of these standards give confidence on how much functionality will work when on an isolated network?

I'm never sure what will break if i don't give it internet access :/

Matter is supposed to be a local IP protocol. Explicitly does not specify a cloud protocol. But it is an IP protocol and many devices that interact with a network may expect internet access and not work without (eg hubs and voice assistants). YMMV on controlling matter without such devices.

Speaking for myself not my employer.

Zigbee and zwave products basically guarantee this, but I really hope the matter logo makes it clear what required a cloud.
Is matter actually an open standard? When I last looked into the CSA stuff you had to pay a few thousand dollars a year to become a member to access any specifications
> When I last looked into the CSA stuff you had to pay a few thousand dollars a year to become a member to access any specifications

Don't know if it's what you're looking for but downloadable pdf docs for (Zigbee) Rf4ce specifications are at the CSA Resources page:

https://csa-iot.org/resources/developer-resources/

EDIT reply to: >Matter is not ZigBee.

Setting aside the radio incompatibilities, various web articles says Matter was Zigbee Alliance "Project CHIP" and later rebranded to "Matter". And "CSA Connectivity Standards Alliance" is itself a rebranding of "Zigbee Alliance".

If the pdf reference docs are not relevant, I guess I don't understand why they're hosted at the domain "csa-iot.org".

Matter is not ZigBee. I believe it needs a whole new radio.
As I understand it, Matter uses Thread to communicate with low power devices.

Thread uses the IEEE 802.15.4 PHY and MAC.

Zigbee also uses the IEEE 802.15.4 PHY and MAC, using the same frequency bands as Thread as far as I understand it.

Maybe a new radio is needed if you want BLE commissioning (is this required?), but for operation any chip with a IEEE 802.15.4 PHY should support both Thread and Zigbee.

I think it can run on wifi or Thread (possibly zigbee too if I’m reading the wiki right) but I may be wrong. It also uses BLE but I think only for initial setup.
AFAICT yes, the standard itself is open but certification costs money.
My understanding is that the standard will be open when it's final, but the working draft is not. The reference implementation is open during development: https://github.com/project-chip/connectedhomeip

Disclaimer: not speaking for my employer or CSA/Matter (I have done paid work on the reference implementation).

TLDR: Home Assistant and their community are experimenting today (June 15) with a Matter Workshop, where they will connect a ESP32 dev board to HA via matter.
(comment deleted)
I'm not sure what's in the workshop, but the list of supported devices in the Google Matter sample code can provide you some ideas for ESP32 device compatibility.

A single LED would be on a different pin, driven by a different protocol, and would likely require a different Matter lighting endpoint sans color and brightness, in addition to the CPU differences (unless you did your own PWM).

What exactly is a "smart home" now other than a voice recognition program that plays music and turns on the lights? Why do I need a "standard" for that?
You want a music player and light switches that actually do the things you wanted them to do? Without being tied to a single vendor?
For me "smart home" is more about opening the window blinds in the morning, and turning on the bedroom lights to wake me up. Easily adding light switches that do what I want, and where I want it, without worrying which cables are installed (want to have two separately controlled lamps instead of one? Add a switch next to your bed? no problem).

Yes, all these things have been possible for decades, "smart home" just gives them a new branding and a different UX.

But once you have a couple of things it becomes handy to have devices communicate. My smart power plug tells me how much energy is being drawn. I could use that to dim the lights when I turn on the projector, but that requires the two devices to live in a common ecosystem of some kind. That's what standards are for.

You need a standard so you can use lights and music players from different vendors at the same time. You also need a standard to ensure future devices that don't yet exist also work with your smart home. A third reason: a new upstart can't easily gain traction in the smart home space without a standard to quickly interoperate with existing smart devices.
Some examples, inspired by my own automations or ones I've seen implemented:

- Warns me if I leave my garage door open, but only if I'm not home. Optionally close it remotely. Provide video feed to see if it's safe to remotely close.

- When I walk out the door, warns me if I have windows or doors open or the stove's on. Automatically arms security. Automatically turn off all lights, music, heat, etc but only if I don't have guests over.

- If burglar alarm goes off, turn on interior lights, flash exterior lights, turn on the radio, trigger my car's alarm, push relevant video feed to phone. If the alarm is disarmed, use the same presence and door sensors to turn lights on and off automatically.

- Triggers notifications (or the alarm as appropriate) if cams detect a human. Turn off certain cams if anyone's home.

- When any of my 3d printers finish, ring my phone. Optionally cut power to the printers if something goes wrong.

- Warn me when the garage door opens and it's below freezing outside.

- If any of my electrical circuits go over-current, warn me and optionally automatically shed some load.

- If it's about to rain, automatically close windows and disable the sprinkler system.

- Open the blinds when my smartphone alarm goes off.

- Defer car charging, running the dryer, etc until solar panels have excess capacity.

- Warn me if my fridge or other critical things stop working.

All of these things use equipment from different manufacturers. It's great to have them all play together - but it's only possible because Home Assistant has an entire army of contributors who reverse engineer all sorts of closed protocols. A standard helps with that. Ideally, staying exclusively on the local network.

Ah, so another thing to sell to rich people at a huge markup. Got it.
You can find contact sensors, motion sensors, cameras, thermometers, and smart plugs for < $15 each. Being rich is definitely not required.

One use case I have, to save significant amounts of money, is that I get a notification if a window is open and the inside vs outside temperature difference is too great. The sensors have paid for themselves many times over already.

Removing any and all monthly subscriptions is also a huge monetary saving - as is being able to convert hardware that was abandoned by some cloudy company. If you're techy, you can buy cheap used stuff from ebay and flash it with open source software.

Plus, I know that if I invest my $15 in a widget flashed with open source software, I know I probably won't have to save up another $15 in two years.

I personally don't use any cloud products, and would strongly recommend against it. I buy local network/Zigbee, and use Apple Home/homeassistant for cloud access.
I admit I experienced some schadenfreude when Insteon went down a short while back. Local-only support or bust.
I've got a total investment of well under $500 to achieve the above. Part of the beauty of this is that I can buy the cheapest, dumbest stuff and get the capabilities of the marked up junk.

Edit: I don't think I'm addressing your point. I find my system is extremely useful, keeps me safer, and saves me money. I feel that's worth my modest investment. Can you clarify your position? I do think that most IOT junk is hugely marked up (and hostile). I just think that Home Assistant is one of the few players in the IoT sphere that is _not_.

Is this just another example of https://xkcd.com/927/ ?
It could be. Or it could be more like how USB came along and obsoleted DB9, DB25, parallel ports, and PS/2 ports.

I hope it's the second one, but time will tell...

I think this is different in that, currently, there are no standards - just a lot of companies doing their own thing.
(comment deleted)