8 ms·

Its a pretty good hill to die on. Everything IOT should be forced to communicate over transparent and self-documenting protocols so a. anyone can write an app t

Its a pretty good hill to die on. Everything IOT should be forced to communicate over transparent and self-documenting protocols so a. anyone can write an app to control said device and b.[0] Alexa/Cortana/Siri can query to set up a voice control interface. [0] Maybe in the future when their capabilities get a bit better, but the gist is every device should respond to a 'hello' ping with a list of commands and NLP'able descriptions such that for an air conditioner 'alexa set temp 67 degrees F' just works.

crooked-v 6y ago
HomeKit is probably the closest thing at the moment as, while it's a proprietary protocol, it's already been mostly-reverse-engineered enough to have open-source HomeKit controllers for most devices, as well as open-source bridges from non-HomeKit devices to HomeKit protocols.
- Mister_Snuggles 6y ago
  Home Assistant[0] is a great tool to act as this bridge. It can act as a HomeKit device, so anything it knows about can be exposed to Apple devices via the Home app. It also has integrations to talk to virtually anything, or you can build your own IoT devices with ESPHome[1]. [0] https://home-assistant.io/ https://home-assistant.io/ [1] https://esphome.io/ https://esphome.io/
  - lostlogin 6y ago
    So much this - ESPHome is incredible. The next amazing bit is how good a system you can cook up with some low cost ESP8266 or ESP32 chips and a few sensors. I’ve been looking for ages for a decent thermometer that connects to Home Assistant. All were expensive and required all sorts of contortions to get into HA. A week or two ago I bought some US$6 chips and a few US$3 sensors and they had me there in short order. I have no electronics background and very little coding experience.
    - Mister_Snuggles 6y ago
      If you want to take this kind of thing to the next level, check out Node-RED. It's a visual programming environment that's well-suited for home automation. One of my most-used automation setups is a dust collection fat near the cat litter boxes. The basic idea is to turn the fan off, if it's on, when the cats approach, then turn it on five minutes after they leave and turn it off ten minutes later. The cats are detected using a PIR sensor hooked to an ESP8266 running ESPHome. I've also got an automation which will turn the fan on and ignore the PIR sensor for ten minutes so that I can clean the litter without breathing in a bunch of dust - this part is kicked off via an input_boolean that's exposed to HomeKit, so I can tell Siri to turn it on and it triggers a whole bunch of work.
    - TeMPOraL 6y ago
      The main thing that prevents me from DIY IoT is worry about fire hazard. Is there a guide out there for picking right components (e.g. are Sonoff devices solid enough) and assembling them in a way they could be safely plugged to mains, or used to switch mains power? Alternatively, are there guides on how to build battery-powered IoT? I'd be fine powering stuff around home from rechargable AAs if I had a clue on how to build IoT devices like that (IANAEE).
      - peckrob 6y ago
        Yeah, I would not suggest DIY'ing anything that's directly connected to mains power unless you are super sure you know what you are doing. That's a really good way to cause you problems with your homeowner's insurance if you have a fire and it could be against your local building codes. That said, there are devices out there that can be used on mains power that are UL certified. I have a ton of Z-Wave [0] switches and relays for various things all talking to Home Assistant. Z-Wave is an open [1] (or soon to be open) mesh networking communications protocol that is pretty widely used and supported, so you're not locked into a single vendor. For low-power stuff, I'm a bit more willing to experiment, but not on mains power. But I am also not an EE. :P [0] https://en.wikipedia.org/wiki/Z-Wave https://en.wikipedia.org/wiki/Z-Wave [1] https://www.theverge.com/2019/12/19/21029661/zwave-open-standard-radios-smart-home-multiple-vendors-silicon-labs https://www.theverge.com/2019/12/19/21029661/zwave-open-stan...
  - ViViDboarder 6y ago
    Home Assistant may be my favorite open source project. The community is great and incredibly active too.
- untog 6y ago
  I have been really happy with a HomeBridge and HomeKit setup. The only downside is the eventual hardware: for various boring reasons I was forced into getting Lutron smart light switches, and they require a hub. That hub has a Telnet connection that Home Assistant uses and it’s super flakey. I just wish there was more truly open hardware out there.
  - LgWoodenBadger 6y ago
    Could you make a regular old float sensor work?
- ex3ndr 6y ago
  HomeKit protocol is open and there are extensive documentation available.
flir 6y ago
Years ago, there was Jini (Apache River now). It never gained much traction. But I always liked the idea of interoprable home devices announcing themselves on my network. "Hi, I'm a mass storage device. If anyone's got data they need storing, send it my way!"
btashton 6y ago
This kind of response is a big part of why I have given up building commercial IoT products for now. It's great to want to have some super abstract high level self discovering protocol, but when you actually start to build on it it really hampers the product. If you want to build a smart light switch you are trying to get the response time very low and worry about things like syncing behavior around the network. Doing these things ends up being very domain specific and you do creative engineering to make it happen. These are very different than the requirements for say a vacuum cleaner. Then we have standards that come out like Bluetooth mesh or HomeKit that say this is exactly how a light switch should work. Great, except your light switch has this cool feature that Philips did not think of in the committee meeting and now you are forcing it in and your product once again suffers. These standards all suck, some small percentage of your customers want custom access (rightfully so), and a large percentage are comparing you on price and experience. The outcome is a closed off product. With maybe a cloud API. Like I said this is why I don't want to work on these products anymore. You cannot win.
- mattmanser 6y ago
  I completely understand and sympathise with this sentiment, as I'm sure many others will. It's one things to want standards, but when it's still an emerging field, with so much different functionality, it's an impossible task. Any poster who advocates standardization at this stage would probably be wise in reading about the early computing days, when you had so many different standards, before it crystallized behind IBM MS-DOS. Or even HTML, where would we be today if MS had listened to "standards" and not released XMLHttpRequest/MSXML library. No ajax, no modern web. This is a perfectly normal, and perhaps desired, period of experimentation where standards will just hold the industry back.
  - moron4hire 6y ago
    > I completely understand and sympathise with this sentiment, as I'm sure many others will. I don't. He's basically admitting his engineers are incapable of reading relatively short, well-written spec documents. This stuff is standardized. BLE is extremely easy to use, both on a device and in an application. BLE makes it possible to create self-documenting devices that any application could use. Device manufacturers are going out of their way to prevent that from happening.
Pxtl 6y ago
Web. A self documenting protocol is it runs a webserver that accepts simple form requests. "But web tech is insecure!" If they can't secure a form post, what hope do they have of securing a proprietary api? Yes, home appliances running webservers would have downsides for management, but that's where routers could add value. It would be trivial for a router to scan port 80 of every device on your network and present a page to pick your devices. That's something I'd even use an app for. "Port 80? But what about encryption you complete monster!" How the hell is that cert going to work? Even if they had some magical way to store a cert on the device securely, surely they're going to screw up keeping it up-to-date, or do you want your "it's MY device! Mine!" to depend on an external provider for security? Just make sure the devices can run on some kind of VPN. Router can keep them on a VPN and run a proxy with proper SSL for outsiders. Then you only need one device in your network secured to modern standards, all your IoT run in the router's IoT sandbox and all access to them is through a proper SSL-encrypted proxy portal. There. It would be possible to make that layman-friendly infrastructure. Router detects you connecting to LAN device over port 80 and says "hey, this looks insecure, do you want to move this device to the secure sandbox? You can access all IoT devices securely through my app! Y/N/OMG this is awful never ask me again!"

crooked-v 6y ago

HomeKit is probably the closest thing at the moment as, while it's a proprietary protocol, it's already been mostly-reverse-engineered enough to have open-source HomeKit controllers for most devices, as well as open-source bridges from non-HomeKit devices to HomeKit protocols.

Mister_Snuggles 6y ago
Home Assistant[0] is a great tool to act as this bridge. It can act as a HomeKit device, so anything it knows about can be exposed to Apple devices via the Home app. It also has integrations to talk to virtually anything, or you can build your own IoT devices with ESPHome[1]. [0] https://home-assistant.io/ https://home-assistant.io/ [1] https://esphome.io/ https://esphome.io/
- lostlogin 6y ago
  So much this - ESPHome is incredible. The next amazing bit is how good a system you can cook up with some low cost ESP8266 or ESP32 chips and a few sensors. I’ve been looking for ages for a decent thermometer that connects to Home Assistant. All were expensive and required all sorts of contortions to get into HA. A week or two ago I bought some US$6 chips and a few US$3 sensors and they had me there in short order. I have no electronics background and very little coding experience.
  - Mister_Snuggles 6y ago
    If you want to take this kind of thing to the next level, check out Node-RED. It's a visual programming environment that's well-suited for home automation. One of my most-used automation setups is a dust collection fat near the cat litter boxes. The basic idea is to turn the fan off, if it's on, when the cats approach, then turn it on five minutes after they leave and turn it off ten minutes later. The cats are detected using a PIR sensor hooked to an ESP8266 running ESPHome. I've also got an automation which will turn the fan on and ignore the PIR sensor for ten minutes so that I can clean the litter without breathing in a bunch of dust - this part is kicked off via an input_boolean that's exposed to HomeKit, so I can tell Siri to turn it on and it triggers a whole bunch of work.
  - TeMPOraL 6y ago
    The main thing that prevents me from DIY IoT is worry about fire hazard. Is there a guide out there for picking right components (e.g. are Sonoff devices solid enough) and assembling them in a way they could be safely plugged to mains, or used to switch mains power? Alternatively, are there guides on how to build battery-powered IoT? I'd be fine powering stuff around home from rechargable AAs if I had a clue on how to build IoT devices like that (IANAEE).
    - peckrob 6y ago
      Yeah, I would not suggest DIY'ing anything that's directly connected to mains power unless you are super sure you know what you are doing. That's a really good way to cause you problems with your homeowner's insurance if you have a fire and it could be against your local building codes. That said, there are devices out there that can be used on mains power that are UL certified. I have a ton of Z-Wave [0] switches and relays for various things all talking to Home Assistant. Z-Wave is an open [1] (or soon to be open) mesh networking communications protocol that is pretty widely used and supported, so you're not locked into a single vendor. For low-power stuff, I'm a bit more willing to experiment, but not on mains power. But I am also not an EE. :P [0] https://en.wikipedia.org/wiki/Z-Wave https://en.wikipedia.org/wiki/Z-Wave [1] https://www.theverge.com/2019/12/19/21029661/zwave-open-standard-radios-smart-home-multiple-vendors-silicon-labs https://www.theverge.com/2019/12/19/21029661/zwave-open-stan...
- ViViDboarder 6y ago
  Home Assistant may be my favorite open source project. The community is great and incredibly active too.
untog 6y ago
I have been really happy with a HomeBridge and HomeKit setup. The only downside is the eventual hardware: for various boring reasons I was forced into getting Lutron smart light switches, and they require a hub. That hub has a Telnet connection that Home Assistant uses and it’s super flakey. I just wish there was more truly open hardware out there.
- LgWoodenBadger 6y ago
  Could you make a regular old float sensor work?
ex3ndr 6y ago
HomeKit protocol is open and there are extensive documentation available.

flir 6y ago

Years ago, there was Jini (Apache River now). It never gained much traction. But I always liked the idea of interoprable home devices announcing themselves on my network. "Hi, I'm a mass storage device. If anyone's got data they need storing, send it my way!"

btashton 6y ago

This kind of response is a big part of why I have given up building commercial IoT products for now. It's great to want to have some super abstract high level self discovering protocol, but when you actually start to build on it it really hampers the product. If you want to build a smart light switch you are trying to get the response time very low and worry about things like syncing behavior around the network. Doing these things ends up being very domain specific and you do creative engineering to make it happen. These are very different than the requirements for say a vacuum cleaner. Then we have standards that come out like Bluetooth mesh or HomeKit that say this is exactly how a light switch should work. Great, except your light switch has this cool feature that Philips did not think of in the committee meeting and now you are forcing it in and your product once again suffers. These standards all suck, some small percentage of your customers want custom access (rightfully so), and a large percentage are comparing you on price and experience. The outcome is a closed off product. With maybe a cloud API. Like I said this is why I don't want to work on these products anymore. You cannot win.

mattmanser 6y ago
I completely understand and sympathise with this sentiment, as I'm sure many others will. It's one things to want standards, but when it's still an emerging field, with so much different functionality, it's an impossible task. Any poster who advocates standardization at this stage would probably be wise in reading about the early computing days, when you had so many different standards, before it crystallized behind IBM MS-DOS. Or even HTML, where would we be today if MS had listened to "standards" and not released XMLHttpRequest/MSXML library. No ajax, no modern web. This is a perfectly normal, and perhaps desired, period of experimentation where standards will just hold the industry back.
- moron4hire 6y ago
  > I completely understand and sympathise with this sentiment, as I'm sure many others will. I don't. He's basically admitting his engineers are incapable of reading relatively short, well-written spec documents. This stuff is standardized. BLE is extremely easy to use, both on a device and in an application. BLE makes it possible to create self-documenting devices that any application could use. Device manufacturers are going out of their way to prevent that from happening.

Pxtl 6y ago

Web. A self documenting protocol is it runs a webserver that accepts simple form requests. "But web tech is insecure!" If they can't secure a form post, what hope do they have of securing a proprietary api? Yes, home appliances running webservers would have downsides for management, but that's where routers could add value. It would be trivial for a router to scan port 80 of every device on your network and present a page to pick your devices. That's something I'd even use an app for. "Port 80? But what about encryption you complete monster!" How the hell is that cert going to work? Even if they had some magical way to store a cert on the device securely, surely they're going to screw up keeping it up-to-date, or do you want your "it's MY device! Mine!" to depend on an external provider for security? Just make sure the devices can run on some kind of VPN. Router can keep them on a VPN and run a proxy with proper SSL for outsiders. Then you only need one device in your network secured to modern standards, all your IoT run in the router's IoT sandbox and all access to them is through a proper SSL-encrypted proxy portal. There. It would be possible to make that layman-friendly infrastructure. Router detects you connecting to LAN device over port 80 and says "hey, this looks insecure, do you want to move this device to the secure sandbox? You can access all IoT devices securely through my app! Y/N/OMG this is awful never ask me again!"