"Smart Home" stuff even since the "x10" standards has been a solution looking for a problem to solve. A hobby for the homeowner who wants to ride the high side of the Laffer curve in effort and expense. It could be argued that quiet, flawless functioning of these products would be exactly what the customer doesn't want in this product space.

Compare it to model trains: is that hobby focused around having a lovely little table set up and watching the trains run? Or is the actual focus the torn up mess that's under construction and "we can't play with it right now but i can get it assembled to show the kids by Tuesday".

But what's important is that it's all rock solid and 100% reliable, because none of it leaves my house. I don't have any "cloud" crap. Everything exists on my local network only. Well, that's not totally true. I have a Nest thermostat. It's the only device that ever gives me trouble, and it took a whole day to get automated. Some day I'll replace it.

I'd argue that your plant timer is reliable not because it's old and not networked, but because it doesn't have dependencies on servers thousands of miles away and isn't reliant on the continuing goodwill, solvency and work ethic of the company that sold it to you.

Because just that watering setup is the opposite of rock solid. If the electricity is on, and these several devices in your house didn’t fail, and you have internet, and the cloud service works, then you water your lawn unless the weather forecast says otherwise. Meanwhile a simple mechanical watering timer is basically if you have water pressure and the divide didn’t fail then it all just works.

Don’t get me wrong it can be fun to setup, but setting up and maintaining these “smart” systems end up sucking up more time than those home automation projects in theory save. At least compared to a simple mechanical timer.

> sucking up more time than those home automation projects in theory save

I guess? But it's fun. And I'd rather spend 10 hours automating something than 1 hour doing it. I'm not doing this to create free time, I'm doing it in my free time.

> And I'd rather spend 10 hours automating something than 1 hour doing it. I'm not doing this to create free time, I'm doing it in my free time.

That’s exactly the point the original commenter was making. It’s a hobby, you do this because you enjoy it, and it’s not a general off-the-shelf solution that anyone can pick and plug it in. Nothing wrong with that, I enjoy the same stuff so I know where you’re coming from. But I’d be lying if I said it’s simple and robust like the mechanical timer thing.

The trick is, before you buy anything, look it up on Home Assistant. If its "class" is "local push", you're golden:

https://www.home-assistant.io/blog/2016/02/12/classifying-th...

The other big problem with smart home is that every company wants to own the ecosystem. The technical term for that is “greedy bastards”. It’s not cloud; it’s cloud lock-in that’s the problem as I see it. This is the source not only of reliability problems but also of compatibility and interoperability problems.

When I started on my IoT journey I adopted a hard requirement that everything must “fail normal” - I don’t want a call in the middle of the night for tech support while I’m traveling on business, because my family can’t turn on the lights. Everything in my house fails normal- if the cloud or the internet goes away, the house becomes a dumb house and you can still turn on/off anything with a switch.

And I have about 20 ZWave devices. They've been super reliable, but, it's a small house, so I can't speak for what would happen if I quadrupled. My Caseta stuff has been just as reliable.

Nah, standards assist everyone. Thus Matter[0] is becoming a thing. It will be supported by pretty much everyone (Amazon, Apple, Google, Samsung SmartThings and the Zigbee Alliance).

"Fail normal" is a good thing to keep in mind nevertheless.

https://xkcd.com/927/

I had everything on HA, but at the time they were changing _everything_ way too quickly for me. After the last time I'd lost my entire Zwave network setup, I figured it was a great time to try something now.

So, now it's a single ~1k line javascript file that handles all my automation, built atop zwave2mqtt (hoping to move to zwavejs soon - maybe over the holiday). It's simple, solid, and very easy to maintain.

I could use a UI, and occasionally dream of building one, but honestly I generally forget the place is automated. It all just sorta works.

No idea when I'll find the time, but if you want to email me (check my profile), I'll reach out when I find the chance to do so.

It has some warts, but my automation would not be viable without it. I specially like the NodeRED integration - even if I think that NodeRED is quite quirky and %#@!% sometimes, it's still neat to be able to write automations without programming much. I do plenty of that on my day job. Also a great visual aid to explain to non-technical people.

I use Sonoff plugs, the first thing I do when I get one is solder headers on them and reflash their firmware - hey presto, no cloud

HomeKit stuff can be controlled remotely, but it doesn't "talk to the cloud" as much as IFTTT or Samsung SmartThings based stuff.

After you hit a wall with the Hue native automations + HomeKit, Home Assistant is pretty much the only relevant shop in town, maybe enhanced with NodeRED.

The setup is very cool. I say "Hey Siri, good morning" before I get out of bed, and the lights turn on. Siri also understands more complicated requests like "set the center light to cyan" or "set all the lights to 50%." I also save configurations ("Scenes") and Siri can apply them by name.

Importantly the fixture has a physical switch and the bulbs reset to ordinary white after powering off, so the bulbs fail like an escalator if the WiFi is down.

No relation to any of these products, just a happy customer.

Plus all the time invested into building and maintaining these setups can be measured by tons of fun/adventures missed in ones live. But that's my perspective, I incline more to OP' mindset of needing at most "plant timer".

If you make smart stuff a purely additive layer over your home rather than replacing existing tech with it, keep it out of the cloud, and have everything responding to predictable rules rather than apps/AI assistants, home automation is really wonderful.

And I don't know where on the scale from "old man shouts at cloud" to "living consciously and connected to, and aware of, the world around you" this reaction sits. Even whether that scale has both end points being co-existent.

I find that pursuits intended to give myself more time for the things I enjoy eats into the time otherwise spent on things I enjoy - both in the setup and the maintenance (you already said it's reliable though, so maintenance may not be an issue).

To what end? is my question. Living in the perfect temperature works to narrow the temperature range in which you can comfortably operate. TV on, lights off, blinds closed... it just feels like the next step is automatic fecal extraction and progress towards the body types on the space ship in Wall-E.

To what end? To see if I can. And if you can, which in this case you could, it feels as if you're retro justifying the effort. Are you happy no longer having to do these mundane tasks? Living in algorithmic rhythms.

Getting up from my chair and pressing a switch when I feel it is getting too dark is not really a chore. And it is probably good for my back.

I do have a couple of X10 switches but I mostly use them where I once was too cheap to put an extra cable. One died this year after only 20 years of service. Luckily I had a spare.

I also have a couple of electronic thermostats. I really hate having to replace batteries every few years. Now I set the day and night temperatures to a about the same. But I like that if I feel cold, I can manually turn up the heat and not having to remember to turn it back down. This should also be possible to implement mechanically.

I also have a drawer with some Zigbee smart home stuff and a CC2540 debugger to program a USB dongle to allow me turn off all the lamps in my living room. Which I will do whenever I feel like it. Because I have 5 lamps and if I turn them off 3 times a week for 20 years and I save 3 seconds per lamp minus 3 seconds to use some Zigbee switch, I will save 10 hours in total. If I have to find a phone or remote first, it will be a lot less. And I will probably spend a lot more than 10 hours before it all works. And I am not sure that it will work for 20 years, or that I can easily replace broken devices in 20 years.

I'm trying to understand how that could work. My old style thermostat maintains my apartment at a constant temperature based ... what temperature I set it to. If I wanted fancier, might have more sensory, fans or whatever to keep temperature average between floor and ceiling. But I can't see how outside temperature gives you any significant value compared to knowing changing in the inside temperature you're controlled.

Apart from that, I could just set my thermostat to heat-cool between 70-74 at all times. But then if it's 40 degrees out and I'm cooking a roast all day, the thermostat right next to my kitchen could kick the air conditioning on, which could actually damage it. So... I automate it; heat when it's cold, AC when it's hot. :shrug:

There will come soft rains.

How many nines of reliability does your home automation setup need anyway?

To date the only light switch downtime I've had was related to power outages and bulb failures.

It’s worked for…I dunno, a decade maybe. The settings are saved on a small SoC peristent memory chip so even if the battery dies I don’t have to set it back up.

Why mention this. Because a tool like this CANT depend on internet connectivity. Usually when it’s needed most basic power and even internet are either not available or intermittent at best. I live in a hurricane prone area that also by proxy is sometimes succeptible to weak tornadoes. I would NEVER trust that to a “smart device”. It’s just not necessary.

And that’s beyond things like smart bulbs kinda baffle me. I get it, but I would never have one myself. Similar to things like yard services.

I do have some smart things and n my home, like a Honeywell thermostat or even some wifi mouse traps that generally save uneeddd crawl space trips. But other things like bulbs, fridges, even alexas I don’t get. But also understand I’m not a primary market for that either

If there was some logo and brand associated with that guarantee, consumers would prefer it, just like today some only buy organic or freetrade. Then manufacturers would want the extra sales generated by that logo/brand, and have a massive financial incentive to only apply the logo when the product is good enough.

Apple solved this by just never showing error messa...Your computer was restarted because of a problem.

Once the Mac has finished starting up again, the backtrace is easily accessible for those who want it. And if you really want to always see console output, you can turn on verbose boot permanently.

I like these openish interfaces because when it breaks, I can fix it.

Your device stopped working because the server it syncs with is gone because the company that owned the server is gone.

Who gives you your money now?

To do that, There could be a "Will Work Guarantee" organisation. The organisation will administer refunds to all customers with claims, (and combat fraud). Companies who want to be members of the scheme must pay into the scheme to cover payouts for their own products, plus a deposit for future failures of their products if they go bankrupt.

Related: MSI replaced 15 OoW mainboards without proof of purchase, too. They merely asked for serial numbers and shipped them out in a timely manner. Can you guess who's boards I used exclusively when building towers over the last 15 years?

Can't imagine any downsides.

The fault probably lies with the ISP, the router manufacturer, or someone else. Let the gadget manufacturers guarantee pay out, and they can try to chase the ISP themselves. If the issue doesn't get fixed, they can simply say "IoT doorbell (reliability guarantee excludes use with Comcast)"

...and hence the fussiness of home automation. Or to put it another way, clockworks are inherently so complicated that features are literally called "complications." TAANSAFL. YMMV.

Huh? Electric analog clocks have small electric motors to turn their dials and that's it.

https://www.sparkfun.com/products/12708

Following hobbyist-directed instructions that are, also, readily available online

https://create.arduino.cc/projecthub/embeddedlab786/real-tim...

It's not hard to just wire shit directly and use an external plug to control the light.

I worked my way through college with a soldering iron. The wire to my headphones broke, and I thought "I'll just strip the wires and solder them back together!"

The problem was, the wires were about as thick as a hair. My fat fingers just couldn't solder them. I could hardly see them.

The headphones went into the garbage.

I've got both mechanical timers and digital timers. While the digital ones have more programming flexibility (e.g., different hours on different days), I've also got to relearn the button sequences for checking & setting them up every time I use them.

So, what comes out first? The mechanical ones, just set the time, put the pins in the on/off times I want, and plug it in. I only bother with the digital ones if I need to use more than the couple of mechanical ones on hand.

What will I buy next if I need more? Some of these same mechanical ones...

Short, simple, and robust tech stacks really are better than tall, messy and brittle.

If someone could give me a tall, clean, and reliable stack with a clean UI, I'd be all over it, AND happy to pay more. But the trend these days is all about programmer convenience and "productivity". Great, except when the result is mountains of digital crap that results in articles like that one.

Also, I get a kick out of designing the system to withstand the usual challenges of staying running in the usual house: power outages, kids playing with light switches, etc. Its basically got to always work.

Quality may vary between manufacturers, but they are fundamentally more reliable. They are simple devices.

Contrast that to my HomeKit switch. It’s operation requires that a pki based system incorporating WiFi, Bluetooth and some cheap SoC work correctly. It’s controlled by an Apple framework tied to a cloud identity. When it works, it is marvelous. I get power metering, switching, timers, etc.

My lights didn’t turn on this morning. The only telemetry is a red blinking light. I checked my WiFi management app and the switch dropped off WiFi around 3 am. Solution: unplug/replug. Not good for something that is supposed to control security cameras and door locks.

I find the timers are best replaced after say 5-8yr because they're cheap and i dont trust the plastic once it brittles up.If you've been running one over say 50% of its rated load it might be entertaining to crack it open and see how much the conductors have softened.

Good software is simple.

Nothing in the smart home category, as implemented today in retail products, is simple.

The switches aren’t Wi-Fi connected, because Wi-Fi would be an awful choice for 50 devices that you only communicate with a few times a day. They use their own wireless frequency and communicate in a really simple/reliable way to the bridge, which is wired to my router and doesn’t need an internet connection to work.

There’s no always-running cloud service, the light switches work even when the internet is down (and I can still use the home app too in such an event, just no voice control.)

Siri is finicky sometimes, but I’d wager a 97% success rate in saying things like “set the scene Evening Downstairs” and “turn off all the lights except in the bedroom.”

My favorite aspect is that the pico switches can be programmed to switch whole rooms of lights on and off, which makes it so much nicer in the basement where I have 18 different lights to turn on and I can just hit one switch on my way downstairs and it’s fully illuminated (no voice command necessary) and I don’t have to visit 18 light switches to turn them off again.

Home automation tech must be invisible and fall back gracefully, and must always be interact-able in a “dumb” way, or else it’s just playing games IMO.

I've been gone from Coopersburg for 15 years but I still practice the engineering culture I learned there.

Fun fact: I don't know what they do today but when I was there it was a long standing tradition that engineers put in some time on the customer support phones. That way you get direct (and sometimes very unfiltered) feedback from users and their expectations on your product and understanding of the instructions you wrote. So if you call the 800 number it's not going to India and you might get to talk to one of the (lower level) engineers that made the product!

The interview sat across from me and role-played a customer calling to complain about a vague issue happening in their home. I then had to roleplay asking the customer and various internal teams (all played by the interviewer) questions, forming hypotheses, running imaginary tests (the interviewer would tell me the outcome), and so on. I started from "The lights in this room don't work right", then clarified and trouble-shot through to the underlying issue, and finally to a root cause.

At the time it seemed super weird, but looking back on it years later, it was a clear exercise in troubleshooting skill - something that remains invaluable in engineering. It's one of the only times before or since that I can recall being tested on that specifically, even though it comes up all the time in my day-to-day job.

From what you describe, this challenges sounds like that was emblematic of the culture there. :)

Also - typically, no matter where the potential job would be physically located, there would be at least one interview where you had to get on a plane and travel to complete - it weeded-out anyone who wasn't willing to actually travel for a role that required 70%+ time on-the-road.

If you accepted the position, there was a multi-week bootcamp - that would also include a week of paired shadowing telephone support calls.

Before this, the tradition at engineering companies was that the "kids fresh out of school" would start in Manufacturing or Service so they'd have some experience with the difference between design engineering and actually being able to build or repair the thing you designed.

It seems that this doesn't happen much anymore.

I set up all my non-standard lighting fixtures with Lutron Caseta switches, and replaced all the standard bulbs with Philips Hue bulbs. They've both been working flawlessly, together, with my 3 HomePod Minis for about a year and a half now.

It really is a brand thing. Most are bad.

[edit] > My favorite aspect is that the pico switches can be programmed to switch whole rooms of lights on and off...

This is my only gripe with the Lutron hardware. The Picos can only be used to trigger Lutron devices. They don't integrate into HomeKit like the switches do.

This meant I had to go with a bunch of Hue Smart Buttons, which do, but have a tendency of falling out of their homes and onto the ground.

Was looking at the Hue compatible RunLessWire stuff though. [1]

[1] https://runlesswire.com

After a time away, I tried to turn on lights with my phone -- only to be greeted by a message like "There is a mandatory update available for your bridge. You can use your system again after 10 minutes". I found not being able to turn on the lights for 10 minutes because they were installing unprompted updates completely unacceptable, so that was my last day as a Hue user. I now use a mix of Zigbee bulbs controlled with a (stable, offline) zigbee2mqtt gateway.

And then on the other we have the usability pain like this. I'd rather pay a few dollars more for more storage or an A/B firmware system so the upgrade can run async and minimize interruption.

I also have another zigbee dongle with zigbee2mqtt and zigbee2mqtt is what gives me problems particularly if the power goes out over the house.

The hack is worth it for the reliability of Lutron. The Caseta system never fails for me.

They do if you use a “middleware” to tie together the systems like Home Assistant or HomeBridge. I raise and lower zwave blinds using a pico switch.

ArsTechnica has a look inside a similar module in the Hue Tap.

https://arstechnica.com/gadgets/2014/08/gallery-dissecting-t...

RunLessWire looks to be selling two variants of them, one with the same 2.4 GHz module as above for use with Hue, and another version for on/off circuit control, which would be 902 MHz EnOcean in the US.

  > you have to learn how to properly press them (with about a 200ms hold)

Most of them want to proxy all attempts to turn on a light switch via AWS. Which is madness.

With Lutron, beyond the things you mentioned, two more of my unexpectedly favorite features have been:

1) being able to add additional wall switches anywhere and program them to control any set of lights, without any extra electrical wiring (for 3-way or 4-way switches, or controlling multiple lighting circuits with one switch)

2) range-based programming (e.g. automatically turning on lights for rooms from the garage to kitchen when you're arriving home)

Other devices like smart locks, smart garage doors, water leak alarm sensors, and AV equipment have all been useful and generally worked well for me. Some of them have been slightly fussy, but not as a general rule and certainly not to an "unbearable" extent. Overall, they really do make things significantly better, at a surprisingly low cost compared to the rest of home construction.

I do have a few wifi devices for cases where there is no good ZWave option (or the wifi options are just better, like the OpenGarage garage door controller).

I really enjoy being able to double tap the "off" position of the switch in our bedroom when we turn in for the night and have it do all the things (thermostats turned down, doors locked, garage doors closed, lights off, etc). Once that script finishes it triggers a verification script that double checks the status of all other devices and sends a push notification to our phones confirming everything was done.

I previously used a different brand to control my aquarium lights and HATED it. When the app randomly made me enter my password to turn the lights on, I 1-starred it with a very frank, "my old-fashioned light switch doesn't randomly ask me to enter my password" in the review.

One complaint that I have about the Lutron system is that they don't offer relay-based modules. (The old X10 system did.) Aquarium lights run very poorly off of the Lutron modules, so I wired up some 120V relays for my aquariums. I also use a Lutron module to control an outdoor outlet for holiday lights, and had to do the same thing; wire in a high amperage 120V relay into the switch.

Yes they do, here is the cutsheet: https://www.lutron.com/TechnicalDocumentLibrary/369-549b_Eng...

This one can switch 5A using a Pico input device.

Just to be clear, every commercial lighting controls line has relay modules, it’s necessary for occupancy sensing.

At least as far as controlling the outlet for holiday lighting: It is a 15-amp circuit, and I could easily anticipate plugging in a power-hungry tool at some time.

https://www.lutron.com/ExternalTechnicalDocumentLibrary/AL-1...

This seems like a hole in their lineup.

In the analog world we'd just wire those 18 lights on a circuit controlled by a single switch...

zwave doesn’t come close to the reliability of lutron so dumb it shall be

Taking this as joke as one doesn’t often turn on lights when power’s out …

But I used Caseta in a home with a generator since the Lutron light switches and dimmers are physical switches that happily work without any smarthome features.

Unplug your hub and they all work except for the ones that don’t get wired (the ones that are their remote, just stuck on a wall).

Pretty much the top selling point of Caseta for me, was working like normal switches. Gravy is how much more responsive than Hue or similar.

Well I go start up my backup nuclear reactor in the basement of course.

WiFi is the enemy.

How do they compare with IKEAs "TRÅDFRI" offerings?

While they've always sold dimmer switches at the store for the average homeowner, they're better "known" for commercial installations. Think the lighting controls and software to manage the lighting for an entire office building, corporate campus, museum, etc. The expectation in terms of reliability and scalability is a bit different.

We know from experience that a nebulous interface that can change without notice and provides no clear contract is a asking for trouble and will break your house "build", so to speak, or worse, it will bork at the worst possible moment, when you have "clients" over.

In the early years of home electricity, there was a much higher chance of getting electrocuted or burning to a crisp, outlets had no grounding, insulation and conductors were of a bad quality etc. Similarly, the whole field of home automation is in its infancy, in a few decades most major mistakes will be made and early adopters will work out the kinks out of the technology (and find out why it's a bad idea to have an internet connected microphone listening in your home at all times).

Have you ever talked to another SE? The majority I know don't even care about privacy and have all kinds of invasive auto update devices in their homes (Alexa, Nest, Ring, etc).

I have an aversion to all of the above. But when it comes to Android, it feels like the options are very limited. Either use iOS and only be able to use the device the way they want you to (for a steep price), or use Android and be tracked like crazy. As for Google products, YouTube is lacking competitors, I use YouTube vanced and NewPipe, but those aren't available everywhere.

It seems like the options really suck right now if you want to participate in the modern world with even some of the comforts that are available.

With Android if you know what you are doing, you can stop a lot of it much better compared to iOS. With Apple the defaults are better than Android but many of the options for blocking trackers are not available.

There is a huge disconnect between what you read about online and the real world. Even my conspiracy friends that talk endlessly about Big Tech and their harvesting of data use all the Big Tech things without putting in even the minimal amount of effort to protect their data.

The functionality I want from a light switch is simple enough that it should be able to be built into a finished product. More complexity means more failure modes.

Like you allude to, I'm sure there will come a time where the effort is appropriately rewarded for a grump like me, and I'm grateful there are folks out there working and struggling to improve this.

But I'm enjoying messing with Home Assistant anyways, so maybe that's not enough to convince you (and that's okay!). I did need to two two different zigbee integrations, but now it's working and I can buy the ikea stuff and I know it will work.

I'm curious what goes on in one's mind when you write something like that. Surely you're aware that people other than yourself exist? People with their own thoughts and lives that are actually different from yours? Or you genuinely believe that there are no people out there who are smarter, more experienced and wiser than you that still install smart bulbs or whatever.

This seems like a lot of effort, and I've tried to do it. But even for some products that claim to never phone home, or transmit data outside the network, sales people have lied to me and then I have to go through the hassle of returning the product.

To me, it seems like there is either 0 effort for 0 smart home automation, or lots of effort for any smart home automation, privacy being equal.

it sounds like you very much are concerned about privacy and security

If the Internet connection dies, or the network goes down, the house (largely) reverts to dumb mode and we're no worse off than if we'd not installed any of the smarts. The Hue switches will still control their respective lights as long as there's power to the hub, and the Honeywell controller similarly only requires power for its smarts.

Pretty much all smart devices I own work just like their dumb counterparts if you need them to: Switches, bulbs, etc. I think this fear is overblown.

There's another comment about Hue lights not coming on for 10 minutes because of a bridge update. This makes no sense to me. If you disconnect the actual power to the bulb (via a switch) and then restore the power, they'll turn on at full brightness.

"Computer, lay in a course to intercept the Romulan cruiser." "Intercept course calculated. Did you know, I can automatically reorder toilet paper for the bridge head?"

If any one from Amazon is hanging nearby, please pass along how annoying a chatty home assistant is.

And people just triggering their half-dozen routines they made 3 years ago isn't enough engagement. Thus: a ham-fisted attempt to get people to use it more.

That said, it's a really interesting problem figuring out how to teach people what Alexa can/can't do. How do you tell your users "we built a feature so now you can set subscription orders!"

Personally, I don't get why they don't run ads. Like, instead of a smiley face delivery person, just say "here's a new thing your Alexa can do"

They will. Thing is if you asked Alexa to turn the lights on and it replied with an ad, people would lose their minds.

However, if you slowly erode the experience with these "did you know" quips and other time wasters, eventually people will accept the extra annoyance. Now you have the ability to slip in ads. Why? If you replace the 5-10 seconds useless quips with an ad that could save some one money, suddenly you invert the negative aspects of ads and the outcome is positive. No linger are you subject to needless did you knows and instead "useful" ads that tell you nike is on sale.

Did you know there is a setting to turn the did you know stuff off?

Did you know that this setting even used to work at one point in time?

Its frustrating that companies can't get it into their head that users may not want to know.

Also they have my email, and could send that kind of shit there so no excuses for updating the system to ignore the "follow up recommendations" setting.

Well, for starters, it needs to interpret queries better :) I have Echos across the house, and one Google device I got for free. Google's implementation "understands" way, way more. Amazon's offering is more well supported and they are less likely to pull a Reader, otherwise I'd have switched by now.

A monthly email seems more than adequate.

(...yet?)

Why is it so hard to configure VLANs for my security cameras using the fancy Unifi switches I have? Why isn't there just some documented JSON file somewhere that I can edit? (The key word in the previous sentence is documented -- I'm sure there is an XML or JSON file somewhere they just don't tell you where it is). Why should I have to do configuration of VLANs on my iPhone? And why are you hiding the good stuff with a pretty web-based interface. I'd like to be able to manage my network even if their "cloud" goes down.

Pfsense firewall, no problem. Local ZFS file servers, no problem. Getting my upstairs thermostat (required by the the high end Lenox AC unit) to work, big problems. Again, why does it require me to have a connection to "the cloud" as the AC technicians say. Now I need to remember some dumb password to adjust the air upstairs on my phone.

Lightning hit a power transformer down the road. Everything in my home survived, except the fancy video security night-light motion detecting "cloud" connected doorbell. So now I have two wires sticking out of the wall by my front door with a big note on the door that says "Doorbell broken, please knock". It turns out that works pretty well.

Thinking it was maybe time to declare technology bankruptcy and start over, my spouse and I looked at a few homes for sale. In one, we couldn't hear one another because the "background" music was so loud. We wandered around looking for a way to turn it down; It emanated from speakers around the house and neither the agent nor us could figure out how to silence the infernal music. Did I have to shout out the magic word, like "Alexa" or "Hey Google" or was there a Sonos remote hidden under a cushion--who knows.

I've installed garage door openers myself, but that was before cell phones. Now, I can't get the openers coordinated with my family's cars. The cars have their own distinct button programming protocols quite tricky to manage. Its a bit like learning how to be a courtroom stenographer that types on the funny corded keyboard. A neighbor with an engineering degree claims to have figured out how to program the street's gate opener to his car, but my MB and Subaru have thwarted my attempts to keep up with him.

I'm on my third sprinkler controller. The first had a big rotating switch with ambiguous buttons scattered around and a cryptic page of instructions on the inside of its cover: virtually impossible to figure out. Where were the man pages when I needed them. I replaced it with a "modern" sprinkler controller. It was, naturally, "cloud" connected, so the first step was to climb around in the 120 degree attic dragging cat6 and then install a POE wifi access point in the garage. I foolishly said to myself, "Now I can look forward to using an iPhone app to control my sprinkler." It lasted about a year before I noticed a patches of dead grass.

I figured that maybe I just needed professional help with the sprinkler controller (and I was busy inside because my spouse was complaining about all the wires everywhere--she didn't understand that it takes time to repair a doorbell from the future and run new cat6). The professional irrigation guy was great, but he said I needed a new sprinkler controller. Now, he informed me, I would be able to control my sprinklers with my iPhone, cool, I just need to figure out which app I'm supposed to use! He needed my wifi password of course so he could work on the controller when I wasn't home. No problem, I thought. I'll set up a separate VLAN for IOT (internet of things) and configure it via my Pfsense firewall to be isolated from my home's higher security subnets. At this point, I'm back at paragraph two.

They do have some learning curve, but the interface is rather intuitive once you get the hang of it, even over SSH.

And honestly, that's terrible. Because let's be honest, Home Assistant still kind of sucks. Not because of all the hard work the developers put into it, but because all of the hacks they need to get devices to work with the damn thing. The IKEA smart lamp bridge is unreliable, wake on LAN is unreliable, smart switches refuse to operate without a cloud connection, transmitted data formats are proprietary and unreliable, it all just sucks so bad.

There's work being done by FAANG (or would that be MAANG now?) to make interoperability between devices better, but knowing these companies that probably means the clouds get more integrated with each other.

These terrible internet smart devices should pick a protocol like MQTT or CoAP or whatever IoT protocol they fancy, and stick it proudly on the box. They can have their stupid internet uplink, bht they should Just Work if they're only connected locally. These things run fully fledged ESP32 chips with more power than you could ever need for a plant humidity sensor, just put a basic web interface on them that takes a password, a WiFi network and an endpoint to talk to. As for alternative protocols, companies like Phillips and IKEA should get their shit together already and just expose a good, documented API.

Home Automation is a hobby mostly practiced by the technically inclined. The more bridges and products that work with your smart fridge, the more value your fridge has to consumers.

When it comes to devices that do cooperate I’ve found Home Assistant to be extremely reliable. Every update I’ve applied has gone through flawlessly, but the truth is that once something is set up you don’t really need to update. I’ve set one up for friends with only local devices (WIFI switches and relays) and it’s been rock-solid for a couple years now. It’s probably severely outdated now but it works fine and being behind a firewall means outside security threats aren’t an issue.

It’s been working perfectly. They don’t even know nor care how it works behind the scenes, from their perspective light switches will toggle their corresponding light, some basic automations like turning off lights after 10 minutes in common areas such as the hallway and the heating will pause in any room that has windows open.

Home Assistant does its best to compensate for the bullshit and they're doing an amazing job. Autodiscovery is a godsent and when it works, it's easily the best way to get any home automation system set up.

But if autodiscovery fails, honestly, the process is unnecessarily painful. You need to enter a lot of information and mess with settings to get things working reliably because the vendor decided to change the API between hardware revision or send invalid JSON at random.

But, yeah, it does suck as a consumer product.

I used to have a meticulously managed 'smart home' with HA, but after the system spontaneously self-destructed the fourth time, I ran out of motivation to spend ANOTHER weekend redoing everything.

The unbearable demands of advertising.

EDIT (ADD)

I've installed a couple of ESP32's with humidity / temperature gauges, in two of my abodes, that feed into a prometheus -> cortex store, so I can then map them on Grafana.

This is considered 'excessive' by the family, but for me it's deeply interesting and compelling, and I'll be pushing more of them out over the next few months.

But these are very basic devices, with rudimentary access to the local network, and no way (or interest) in phoning out to anything else beyond my router. This is the ideal scenario for these kinds of devices.

As for smart homes, my toe dipping into them so far has been largely negative. Biggest mistake I've made is probably mixing brands.

If I did it all from scratch I'd:

1) find a single brand that does everything reliably and stick with it. This saves having to have a dozen different apps on my phone and somehow tie them all together, not to mention keep track of half a dozen different passwords, as the apps tend to randomly log me out every so often.

2) choose a decent brand. I know pretty much everything is made in China these days but my experiences with my Yeelink light has been much more painful than my experience with Kasa-related appliances that seem to just work (most of the time) and don't need messing with periodically.

3) ask tech people I trust on what to buy and how to set it up. I don't trust advertising/marketing and many review sites have a vested interest in peddling and recommending shit products.

and possibly

4) wait until it's a more mature market

The Node-Red / Home Assistant (hassio) stuff I need to really dive into in anger at some point. So far my 'smart' stuff is some google gear + Philips Hue (lighting), and these little ESP microcontrollers.

When looking at what vendor to adopt, I've been checking what has the best support by the hassio community - on the assumption they've been down this path already.

I've never really looked in that menu previously, but, yes, definitely have uBO installed everywhere.

I had to Google because I've never heard of any of those. Did I find the right ones?

https://prometheus.io/

https://cortexmetrics.io/

https://grafana.com/

Mine is much more primitive. My indoor temperature monitor is an ESP8266 that uploads the temperature to a simple PHP page that saves it in an sqlite DB. A cron job runs a Perl script every few minutes that extract the data for the last hour, 3 hours, 12 hours, 48 hours, and since the beginning of time and uses gnuplot to produce PNG graphs. There's a static page on my server that displays those graphs.

My outdoor temperature monitor uses a cheap AcuRite 433 MHz indoor/outdoor thermometer I bought. I have an RPi with an RTL-SDR attached spying on the communications between the AcuRite sensor outside and the AcuRite display inside using rtl_433. A script looks at the rtl_433 and finds the AcuRite sensor data and puts it in an sqlite DB. I haven't yet gotten around to making something to graph it.

The nice thing about that approach is that it was also easy to add support for other 433 MHz wireless sensors near me, such as the AcuRite fridge/freezer thermometer I have. I can also see a few assorted sensors of neighbors (temperature, humidity, soil moisture, tire pressure, wind speed, wind direction, rain, and a few other random things). If I wanted to it would be easy to add them to the DB.

When I made a wireless tipping range gauge recently. I used a 433 MHz transmitter module [1] and added a decoder [2] to rtl_433 that understands my data stream format. That gets my data into the rtl_433 output. No need to futz around with 433 MHz receiver modules which appear to be a pain in the ass [3]. An ATTiny85 counts the tips and runs the transmitter. The ATTiny85, the transmitter module, a battery holder, an RJ11 socket because the rain gauge has an RJ11 connector, a board to put those things on [4], and a small waterproof case is pretty much the complete parts list.

I think I'm going to standardize on this general approach. For things that do not have WiFi and only need to report data 433 MHz modules and custom decoders fro rtl_433 on the RPi. For things that do have WiFi, such as any future ESP projects I do, they will just use WiFi to talk to the RPi. If anything needs to get sent outside of my LAN the RPi will handle it.

The RPi is also currently controlling a space heater in my living room, getting connection data from my cable modem periodically and recording that in an sqlite DB, and serving a simple web page that lets me quickly change inputs and volume on my Denon receiver and so I'm already pretty much committed to keeping it running all the time.

[1] https://www.sparkfun.com/products/10534

[2] Decoders can be specified in a simple text file. Here's the one for my rain guage as an example:

  decoder {
    name        = TZS Rain Gauge,
    modulation  = OOK_PWM,
    short       = 416,
    long        = 612,
    reset       = 1452,
    gap         = 0,
    tolerance   = 0,
    sync        = 1484,
    bits        = 96,
    match       = {24}0x545a53,
    get         = Tips:@80:{16},
  }

  {"time" : "1637771949", "model" : "TZS Rain Gauge", "count" : 1, "num_rows" : 1, "rows" : [{"len" : 96, "data" : "545a5304436f756e743a03da", "Tips" : 986}], "codes" : ["{96}545a5304436f756e743a03da"]}

[4] https://www.amazon.com/gp/product/B07ZYTZ48N/

Nothing wrong with 'simple'. Although you are kinda reimplementing the wheel in some of the setup.

If you don't have an use for the whole Prometheus shebang (it's a great piece of software, but might be overkill, and you probably don't need Cortex), you can upload stuff to Graphite. Installation is simple. Sending data is just a netcat away, can't really get simpler than that. Bonus points is that you don't have to maintain (or create) the PHP page. Or maintain a database schema for any new metrics you come up with.

Send what amounts to a string with your metrics to Graphite, and then you can connect Grafana to it and quickly build all sorts of graphs, even alerts.

Now that you have built it, it makes no sense to replace. If you ever find you need different metrics, different graphs, queries, alerting, a dashboard or what have you, take a look at that. It's the simplest solution I've found.

Like their examples show:

    echo "foo.bar 1 `date +%s`" | nc localhost 2003

Send that to the listening port and you are done.

I have some 433MHz kit, plus a bundle of various arduinos, but (for me) the complexity of getting them to talk back to base has kept me procrastinating for years.

The ESP8266/ESP32 devices, with WiFi built-in, are effectively the same price as arduinos (here in AU, via ebay) but so much more convenient because of that extra memory + the wifi. I'm going to have some frustration with the 3 vs 5 volt, especially with some of the more esoteric components, but so far it's been a breeze to setup.

As Outworlder observes, my back-end is way more complex than a normal human would need - I'm replicating a stack we use at work, so it's basically taking up a bit of space on my home lab. Cortex is for a serious (enterprisey) amount of long-term storage of time-series metrics. Prometheus is easy enough to set up - it scrapes web end points that contain key / value pairs in plain text, and puts those in its own time series data store. Sqlite will scale just as well, I'm sure.

If you have the bandwidth, I can recommend playing with some of these things, just in case they may make your life easier later. Prometheus (server) will run on a Raspbery Pi easy enough.

https://github.com/G1Tech/youTubeStream/blob/master/promethe...

That's a code fragment to run on an ESP and present a prometheus-compatible end point with a handful of key/value pairs - if you have a spare ESP, run it up, and hit the endpoint to see what I mean. The simplicity is compelling.

If / when I go down the path of custom components plugged into arduinos - and I'd like to one day build something to measure the levels in my rainwater tanks - I think that I'd try to get those data back into an intermediary device (esp or RPi) that could present them in this opentelemetry format, as it would make it easier to swap things around in the future.

Grafana is fantastic, and can produce some gorgeous visualisations from Prometheus (and other) sources. You may even be able to get it doing something with your sqlite DB.

Running up a monitoring agent (Prometheus' node_exporter, or InfluxDB's telegraf - functionally very similar) on your laptop may be a good way to experiment with live data and visualisations with low-effort. (Note that Telegraf will by default try to push into an InfluxDB -- I'm not a huge fan of InfluxDB -- but you can configure it to provide an otel / prometheus-compatible scrapable web endpoint instead.)

The transmit side is pretty easy. Here's the xmit.ino file from my rain gauge project [1].

The receive side using one of those typical 433 MHz receive modules looks like a complete pain in the ass. I spent a while with the transmitter and receiver both on a breadboard hooked up to my oscilloscope, one channel on the transmit data in signal, one channel on the receiver data out signal, and one channel on the receiver receiver linear out signal.

Triggering on transmit and looking at the receiver output it didn't look too bad. But triggering on receiver output it was terrible. It was almost always outputting a mostly random looking stream of 1s and 0s of variable lengths. You have to watch that and look for things you recognize embedded in it.

Triggering on the linear output was more promising. My understanding is that the way these cheap receivers work is that they automatically adjust the gain until they are seeing something, so when you aren't actually receiving a signal they are just amplifying noise. When you actually are receiving a decent transmission the gain turns down. When that transmission ends the gain goes back up. (This, I've read, is one of the reasons most transmissions start with a long intro pulse before going into the actual bits. It gives the receiver a chance to turn down the gain so that it won't be including a lot of noise in the output).

The linear out signal is analog. It is higher when receiving a strong signal, and lower when receiving a weak signal or ambient noise. So it looks like using the linear out signal to decide when it is worth looking at the digital out to check if your device is being received could considerably cut down the processing needed at the receiver. By adjusting the trigger level on the scope I could get it to usually only trigger when an actual device that I can get a decent signal from was transmitting.

The good news is that you can say "to heck with all that" and cheat like I did. Forget the 433 MHz receive modules. Cheap SDR dongle on the receiving computer, rtl_433 [2] to do all the work of finding and decoding the signals among the noise, and you are all set plus as a bonus you can also see other 433 MHz sensors that happen to be in your area.

PS: if anyone looks at that transmit code and wonders why there is no checksum that is because I haven't gotten around to adding one. It is a tipping rain gauge and according to models I found for my area at my state's department of transportation site at the highest short term (5 minute) intensity that I would expect to see on average once per 100 years it would be tipping at 10 tips a minute. Thus the tip count should be an increasing sequence that when it changes mostly should change by 1, and rarely by more. It is easy to spot when the data is corrupt and so I've not found much need to add a checksum. I'll do that when I add my wind speed and direction sensors).

(Departments of transportation are good places to look for rainfall models for an area because their engineers need them to figure out how much runoff they have to deal with when designing drainage for road projects).

[1] https://pastebin.com/7Aus3U3R

[2] https://github.com/merbanan/rtl_433

I really do have a drawer full of the 433 gear, and achieved some very rudimentary comms going years ago using a pair of arduinos sitting next to each other on the desk, but quickly abandoned it. I'm impressed that you've achieved what you have using this gear - and equal parts impressed / distressed that you use a CRO to get there.

I also have a tipping rain gauge, but I expect to wire it up to an ESP32 too - as the Hall effect doesn't care about 3v3 or 5v : ) And yes, rainfall models in my (rural) part of the world are poorly documented - but public transport systems are even further distant than official weather stations.

For some of my (few hundred metres from a power point) needs, very low power is a pre-req, but pricing now for LoRa, PVC panels, batteries, etc, means even that becomes orders of magnitude simpler to solve by NOT trying to scrimp on a few hundred mA.

For example, I have my home office in my basement. I have an automation in Home Assistant that when I turn on the Z-Wave light switch at the top of the stairs, it turns on my desk lamp and recessed lighting, and if it is a weekday during work hours, it will send a wake on lan packet to my work computer.

Another example, I have a tilt sensor on my garage door, and a contact sensor on the door into the garage, so that when either of those are opened, the main garage lighting turns on.

There are also odd niches. I have an air compressor in a far corner of my shop (for noise reasons) with pipes running to where I need air. I want to control that without having to walk across the shop. (So far I don't have this because finding a switch that control a multi-horsepower induction motor is hard - most top out an 1/4 horse)

* schedule my non-smart EVSE to charge our car when power is cheaper at night

* monitor (but not control) the energy usage of our electric hot water heater. It's the second biggest electricity user in the house but I have to infer how much power it's using based on the other separately metered components.

Cost is the thing holding me back. Will it have positive ROI in a reasonable time for our PHEV minivan that only has a 30KWh battery? Seems somewhat unlikely. Would monitoring the power usage of the hot water heater have positive ROI ever? Also seems unlikely.

At my first job, I worked for a company that built (among other things) X-10 powerline-control interfaces that would let your PC do the same thing: switch on lights/fans, etc.

As far as I could tell, the only people buying that stuff were hobbyists with a lot of disposable income. Part of my duties was customer support and I can't recall a single instance of anyone doing it for any reason that didn't boil down to "it's cool."

Not making judgements, but I really don't see what the big deal is with getting up to turn on a light, and I live in a pretty big house.

But an optocoupler relay is also about £1, so there's that.

But things that can use it seem to be expensive. I checked, the chips are fairly cheap; but all the PoE stuff I might use aren't "one per light socket" prices. Plus I'd still like a decent mains relay board that is appropriate to put in a wall - the Shellys' are a good form for this, but I cannot see a similar wired solution (although some Sonos devices get close).

But I'd sooner order from a company that is liable e.g. if the IoT device burns down my house.

If your IoT device burns the house down your circuit breaker failed, or you connected it without proper failsafe measures.

Is it burning it down by being on fire itself, or by turning something on that shouldn't be?

If the thing just heats up and catches fire, it might not trip the breaker. I meant overheating itself, not erroneously turning things off/on.

Having one switch control all your table lamps, floor lamps, and ceiling lights is useful. Being able to add new wall switches (really remotes) in new locations for ~$25 also nice (excluding the cost of the rest of the system, obviously).

A few years ago, we had no switches or light sources by the entry way; we had to stumble our way to a switch in the dark. I fixed that with remote controlled outlets at first and later upgraded to Hue bulbs and dimmers. In the bedroom, my wife and I both have remotes for the lights (though she actually prefers voice control, ironically).

We have a few useless wall switches that control outlets instead of lights; turning those into in-wall remotes makes it so we can control multiple lights and use the outlets for devices that need continuous power.

We also have our light bulbs automatically change brightness and color temperature based on the time of day. We get daylight-bright light during daylight hours and dim warm light in the middle of the night. It's really nice.

Another nice smart home feature: checking that the range is off. My wife used to stress about this, it's no longer an issue. Don't worry, you can only turn the thing OFF remotely, not on.

The smart speakers are mostly for music, checking the weather, and time. Occasionally for light control.

Smart/energy monitoring outlets we use mostly for tracking down energy hogs and reducing our energy bill.

The security and privacy stuff is a risk, but as long as you stick to major brands IMO the risk is overblown. Those brands are subject to a ton of scrutiny. Fly-by-night alphabet soup brands off Amazon or Alibaba are another matter.

1) Motion sensors

* Our hallway lights turn on via motion and turn off automatically.

* My work desk lights turn off and on via a motion sensor.

Both apply the correct color temperature and brightness based on ambient light and the time of day, so that I don't get blinded buy 1100 lumen lights at 5000K when I go play a few rounds of $game at my desk.

Just having the bathroom lights turn on and off automatically is magical.

2) Voice control

We have a few Alexa routines programmed for the house:

* "Movie time": All lights off, TV and AppleTV turn on

* "TV Time": TV off, few lights near the TV turn off

* "Good morning": Lights at a reasonable level, Alexa tells us the weather forecast and traffic (well not anymore, we work remotely).

...and a bunch of others.

3) Wireless switches.

All of the above can be controlled via physical switches mounted around the house where practical - not where the wiring happens to be.

I've got a whole home full of these, switches and bulbs and smart speakers. They do some very useful things for me:

* Turning the lights on before entering a dark room with inconvenient switch placement inside

* Ensuring all the lights are off before leaving or going to bed

* Turning on my porch light when it gets dark, and off again in the early morning

* Simulating liveness when away on vacation

* Turning the lights on when I've got my hands full of grocery bags

* Ramping my bedroom lights up slowly in the morning as a gentle wake-up

* With some cheap motion sensors, ensuring that lights that are often accidentally left on are turned off, saving power and money.

I'd rather have a 2-way light switch at either end of the hallway and hit a switch when I walk down it, rather than program in every exception like "turn on the hall light when it detects motion unless it's 3 am and I'm walking to the bathroom to pee, or if it's 3pm and my girlfriend is napping on the sofa next to the hall and I don't want to turn on a light and wake her, also ignore cats triggering the motion detectors." My small apartment with regular old deterministic switches is fine. I had a Nest once and hated it. At least for me, the ideal temperature for a room must be a very subjective and non-rational desire. Sometimes I want to blast the heat, sometimes I want to feel cold, it's probably emotional/psychological on a level that Nest can't model.

I know someone will point out that 100 years ago I would think washing machines sounded ridiculous because I could scrub my clothes by hand easily enough. That might be true, but I would counter that there are diminishing returns on home automation. Indoor lighting was a massive improvement over living by daylight hours. Electricity over gas was a big improvement, but not as big as going from dark to light. Turning lights on with Alexa, to me, sounds like the benefit/cost curve has plateaued.

For example, when I come downstairs and say “computer lights on” it’s easier than flipping two switches and turning on two lamps.

Both android and iPhone have voice control so if you have a smart phone you have capability for voice control, you just don’t use it.

But the phone home stuff is just insane to me. I bought an Insignia garage door opener trigger that works with Homekit natively without an app, figuring that would be the simplest possible setup (and cheaper than RF-based physical remotes). I Wiresharked the traffic it was sending out (in order to block it) and it phones home over a *Wireguard VPN tunnel*. I've seen some shady phone-home behaviour from cheap Chinese devices but never this.

So now it is totally blocked from accessing the internet and it does open the garage door, but I am told "there was an error communicating" every single time.

I will mention one device I absolutely love: CloudFree Smart Plug 2 for energy monitoring. As I mentioned on my Twitter※: [CloudFree Smart Plug 2] are the first IoT product I've ever used with zero friction whatsoever. First time usage, from unboxing to pulling power stats over HTTP API in under 3 minutes. No other products come close.

※ - just copy/pasted my Tweet, no need to hyperlink it

I could open it and connect more stuff to the GPIO, patch out the annoying LED or do something else with it. Install tasmota instead. It's quite liberating :)

I also use Alexa to set reminders, get Roomba to clean specific rooms (or the whole house) and I use it constantly to control Spotify.

You can argue that Alexa is spying on me or whatever but i'm not sure how that's worse than the iPhone I carry in my pocket literally everywhere I go...

Don't know if I would use the adjective "smart" compared to a switch.

On the day we were closing on our 1920s house here in Massachusetts, we noticed that the oven wouldn't turn on. Then we noticed that it was shut off at the breaker.

We asked the owner why, and he said "oh, the mainboard in the oven (a Thermidor) has some issue where it beeps incessantly in the middle of the night and wakes us all up – but not a big deal: you can just turn it off and on at the breaker and it works just fine!" He'd apparently sent the mainboard in multiple times to have it serviced, but the problem persisted.

That wasn't a really satisfying answer at the closing table, so we pressed to get compensated for the oven we were rather sure we'd have to replace.

A few weeks later, as we were pricing out $2000-3000 replacement units, I had an idea to look for a smart relay to control the oven's dedicated circuit. I managed to find a 40-amp smart switch for $90 that I could control with SmartThings.

Now we can use the app or say "Alexa, turn on the oven," the switch is activated in the basement, and the oven turns on. At midnight, SmartThings turns it off if we haven't already, preventing the beeping.

$90 got us where we needed to be, and it's been working great for almost five years now.

And given that we were aiming to have the oven turned off at/near the breaker most of the time, except when in active use (and therefore at least reasonably human-monitored), the risk seems acceptably low.

I try my hardest to stick to Zigbee and Z-Wave devices wherever possible which has lead me to purchase devices like Samsung's SmartThings plug, or Sonoff Minis etc that I can connect straight to a Zigbee/Z-Wave dongle to expose them to HomeAssistant. This setup is actually very reliable and the only downtime we've had over nearly three years is when I've pulled the HA virtual machine down to do updates - I think we're up to about 40 or 50 zigbee devices in our network now that all work flawlessly and locally.

So what about Google and TP-Link? Google Home is becoming chatty and has started to regularly mishear commands, or it will say a device is unavailable but still perform the action successfully. TP-Link pushed a silent update out to the couple of smart plugs we had from them that disabled functionality that exposed them to Home Assistant - rendering them a complete waste of time and money.

It's mind boggling to me that the (m/b/tr)illion dollar companies of the world are the ones fucking up my smart home. It annoys me enough that I would no longer purchase hardware from any company that uses WiFi or relies on the cloud for "smart" home stuff.

Devices are “Updating…” for minutes when I want to switch them on. Once a month something gets stuck as “Not responding” and I have to delete it, re-add it, and redo from scratch all it’s configuration and automations.

I’ve tried Thread that was meant to fix the issues of Wi-Fi devices, but the Thread bulbs are the least reliable ones I have.

It’s all awful, and I’m an idiot sitting in the dark listening to Siri making excuses.

When I arrive home it sometimes takes a few tries before HomeKit realises that I’m back on the network and starts working.

I’m frankly not sure WTF is going on, the bridge is on a static IP so it doesn’t even have to do the whole Bonjour/ZeroConf discovery and can just hit the latest IP it saw the bridge at.

Fortunately my main automation is via Home Assistant and HomeKit is only used as a UI, so if HomeKit craps out I just fire up the Home Assistant web interface or toggle the WiFi a couple times.

But if you were looking at HomeKit for a sign of quality I’m not sure it’s a good idea. The protocol is unnecessarily complex - it’s good in that it’s somewhat documented and open-source gateways such as Home Assistant can talk to it as a client (so instead of reverse-engineering a proprietary API they can just pretend to be a HomeKit client) but using it to connect directly to your control devices (phone, Mac, etc) is a recipe for disaster.

So if you have HomeKit devices, maybe look into using Home Assistant as a gateway. That way HA is talking to all your devices (so its network situation never changes and hopefully there won’t be problems) and then control your devices from HA or have it present itself as a HomeKit bridge as well - at least your client device will only have to figure out the whole network situation with one bridge instead of all of your devices separately.

For lights I do use zigbee ones just because they're cheaper, but my Hue and IKEA mix do have communication issues sometimes (I have them both on a Deconz stick attached to my server).

But all of this relies on Home Assistant. I honestly can't imagine trying to use smart home devices as a "normal" consumer, relying on the software of specific companies. They're all largely terrible walled gardens, and I'm constantly surprised by how bad they actually are.