6 ms·

This offers an interesting insight into the Internet of Things. Take something simple, and tack on a bunch of new technology that is subject to failure. Don't b

This offers an interesting insight into the Internet of Things. Take something simple, and tack on a bunch of new technology that is subject to failure. Don't be surprised if it fails. To be fair, the issue isn't that there was software involved; digital thermostats have that. I think more telling was the level of complexity, which allows for a lot more to go wrong. Combined with a poor process behind update validation, this is what happens. Keep it Simple. If you're going to use new technology, keep the old stuff around, ideally working in parallel as a failsafe. And generally, anything you can do to distribute the failure points and allow for overlap (such as backup spaceheaters that don't normally kick on) should be considered if someone's life is on the line, or property is at risk. Stuff can always go wrong, but the more you can do to reduce risk the better. This applies to any system dependency; mechanical thermostat or nest.

kuschku 11y ago
In general, the way IoT is treated is wrong. Startups try to treat IoT devices as a standard arm or x86 processor with some additional GPIO pins where they read info, deal with it, and write info. No real-time guarantee, a level of QA similar to websites or consumer software, etc. And here is what previous companies building smart devices (say, SIEMENS) treated them like: As extended microcontrollers. Put all the critical software on one tiny microcontroller, do static verification that that microcontroller is always acting correct. If you need Internet access, add a second chip that does that, and provides the results of that data to the first microcontroller — but don't depend on it. There is a large divide between the Engineers designing these devices in the past, and today's crowd.
- TeMPOraL 11y ago
  Another angle to that problem is that a lot of IoT today is basically making something equivalent to a weekend hardware hack, tweaking it a bit, and then calling for investors and marketing the living shit out of the device. Popular IoT products are toys, not tools.
  - jerf 11y ago
    Call me a curmudgeon, but I'm yet to see a coherent explanation of what IoT is supposed to do for me as a homeowner. I think "all the household IoT stuff is toys" is rather directly connected to the fact that I just don't have massive problems I need solved with the Internet in my house. Of course the resulting market is nothing but toys; there doesn't appear to be any space for non-toys. (I see the commercial and civil IoT possibilities, but an awful lot of them involve being able to monitor and somewhat control things that are physically very dispersed out in the real world. I don't have a house that big.)
    - jarek 11y ago
      Eh... there's some value, it's just really badly implemented currently. A selling idea of the Nest is that it keeps your house warm when you're there, and not-frozen when you're not, and it will do so efficiently, so that the house is warm when you arrive home but not long before. Now of course you can just turn up the heating as you walk in the door wait a few minutes to heat up, it won't hurt. But if technology can do it reliably, there's clearly some value. Compare, for instance, lighting a candle when you get home with just flicking the switch for those newfangled electric lights. The candle works but the technology is more convenient. Of course, electricity doesn't give you failwhales these days, and there's a bit more engineering involved than "move fast and break things."
      - kuschku 11y ago
        Another thing that you can do is tracking which products you have in your fridges, freezers and cupboards – and when you want to know where something is, you can just do "Jarvis, do we have another bottle of ketchup?" – "Yes, it’s in Kitchen Cupboard Number 2 Left" In fact, I’ve actually designed something similar to that, but a lot simpler. A simple barcode scanner where you check in new products you buy, and check out products you throw away.
        jerf 11y ago
        "A simple barcode scanner where you check in new products you buy, and check out products you throw away." Your homework, before you start too far down this road, is to start doing that right now if you haven't already. You don't even need a real scanner; just pretend with a stapler or something vaguely the right shape and weight. You should also go through your cupboards and just scan everything once to simulate the initial load. Then, I don't know whether you mean to design this for yourself or for selling to others, but if it's the latter, consider whether your customers will actually do this for any period of time. I'd also recommend faceoffs between two people, one pretending realistically to ask your system where something is (either literally typing it in somewhere or doing real-enough voice recognition that you can see the correct query came out), and one just looking. Even if I don't know your kitchen, if it's at all sensibly laid out and you give me a quick look around first, decent odds I still beat your system finding something on average, even if it's not my kitchen. If you completely eliminate 100% of the time in a year that I spend looking for kitchen goods that I don't know where they are, you've brought me maybe $20/year in value total; if you make me scan everything going in and out, I'd pay at least $200/year not to do that. (Probably more if I really faced that problem.) (If this is just a personal project, go nuts. Very educational, lots of valuable skill building involved, and it's always good to scratch an itch with it. But if you have any ideas of making money with it, then I suggest taking it seriously and doing some heavy market research and proofing before the tech.)
- Joeri 11y ago
  To be fair, you can still buy traditional building/home automation products. If you go through the list of vendors on the KNX website you see that pretty much every hip IoT device has a standards-based equivalent from those traditional vendors. Learning thermostats, IP-based video doorbells, smart lights, etc... The difference is one of market strategy. The IoT devices are deliberately low-threshold purchases (cheap, easy to install) because they want to build a customer relationship that leads to extra sales. Nest cam, ring subscription, etc... Traditional home automation products instead are targeted towards a one-time sale of a larger system, so they focus more on ability to design the system how you want, with more standards support and more ability to integrate with other products, but also a much higher purchase threshold and a much more involved installation process. In the end, you have to choose what you want. Do you want a single gadget or do you want a whole smart home system?
  - TeMPOraL 11y ago
    > In the end, you have to choose what you want. Do you want a single gadget or do you want a whole smart home system? I want a whole smart home system built piece by piece through buying various smart devices of different vendors and having them talk to each other. This is how it's supposed to work for the "personal use" market, and this is what we need to push for.
- Aleman360 11y ago
  That's how a lot of actual medical devices work. I worked on an anesthesia machine where the control systems were implemented with micro controllers and RTOS. The UI was Windows Embedded running on a separate x86 board and the whole thing would still safely operate even if the UI subsystem crashed. There were completely mechanical backup controls too.
- mavhc 11y ago
  On the other hand, then you get people who've no idea about network security doing things, so a lose-lose situation.

kuschku 11y ago

In general, the way IoT is treated is wrong. Startups try to treat IoT devices as a standard arm or x86 processor with some additional GPIO pins where they read info, deal with it, and write info. No real-time guarantee, a level of QA similar to websites or consumer software, etc. And here is what previous companies building smart devices (say, SIEMENS) treated them like: As extended microcontrollers. Put all the critical software on one tiny microcontroller, do static verification that that microcontroller is always acting correct. If you need Internet access, add a second chip that does that, and provides the results of that data to the first microcontroller — but don't depend on it. There is a large divide between the Engineers designing these devices in the past, and today's crowd.

TeMPOraL 11y ago
Another angle to that problem is that a lot of IoT today is basically making something equivalent to a weekend hardware hack, tweaking it a bit, and then calling for investors and marketing the living shit out of the device. Popular IoT products are toys, not tools.
- jerf 11y ago
  Call me a curmudgeon, but I'm yet to see a coherent explanation of what IoT is supposed to do for me as a homeowner. I think "all the household IoT stuff is toys" is rather directly connected to the fact that I just don't have massive problems I need solved with the Internet in my house. Of course the resulting market is nothing but toys; there doesn't appear to be any space for non-toys. (I see the commercial and civil IoT possibilities, but an awful lot of them involve being able to monitor and somewhat control things that are physically very dispersed out in the real world. I don't have a house that big.)
  - jarek 11y ago
    Eh... there's some value, it's just really badly implemented currently. A selling idea of the Nest is that it keeps your house warm when you're there, and not-frozen when you're not, and it will do so efficiently, so that the house is warm when you arrive home but not long before. Now of course you can just turn up the heating as you walk in the door wait a few minutes to heat up, it won't hurt. But if technology can do it reliably, there's clearly some value. Compare, for instance, lighting a candle when you get home with just flicking the switch for those newfangled electric lights. The candle works but the technology is more convenient. Of course, electricity doesn't give you failwhales these days, and there's a bit more engineering involved than "move fast and break things."
    - kuschku 11y ago
      Another thing that you can do is tracking which products you have in your fridges, freezers and cupboards – and when you want to know where something is, you can just do "Jarvis, do we have another bottle of ketchup?" – "Yes, it’s in Kitchen Cupboard Number 2 Left" In fact, I’ve actually designed something similar to that, but a lot simpler. A simple barcode scanner where you check in new products you buy, and check out products you throw away.
      - jerf 11y ago
        "A simple barcode scanner where you check in new products you buy, and check out products you throw away." Your homework, before you start too far down this road, is to start doing that right now if you haven't already. You don't even need a real scanner; just pretend with a stapler or something vaguely the right shape and weight. You should also go through your cupboards and just scan everything once to simulate the initial load. Then, I don't know whether you mean to design this for yourself or for selling to others, but if it's the latter, consider whether your customers will actually do this for any period of time. I'd also recommend faceoffs between two people, one pretending realistically to ask your system where something is (either literally typing it in somewhere or doing real-enough voice recognition that you can see the correct query came out), and one just looking. Even if I don't know your kitchen, if it's at all sensibly laid out and you give me a quick look around first, decent odds I still beat your system finding something on average, even if it's not my kitchen. If you completely eliminate 100% of the time in a year that I spend looking for kitchen goods that I don't know where they are, you've brought me maybe $20/year in value total; if you make me scan everything going in and out, I'd pay at least $200/year not to do that. (Probably more if I really faced that problem.) (If this is just a personal project, go nuts. Very educational, lots of valuable skill building involved, and it's always good to scratch an itch with it. But if you have any ideas of making money with it, then I suggest taking it seriously and doing some heavy market research and proofing before the tech.)
Joeri 11y ago
To be fair, you can still buy traditional building/home automation products. If you go through the list of vendors on the KNX website you see that pretty much every hip IoT device has a standards-based equivalent from those traditional vendors. Learning thermostats, IP-based video doorbells, smart lights, etc... The difference is one of market strategy. The IoT devices are deliberately low-threshold purchases (cheap, easy to install) because they want to build a customer relationship that leads to extra sales. Nest cam, ring subscription, etc... Traditional home automation products instead are targeted towards a one-time sale of a larger system, so they focus more on ability to design the system how you want, with more standards support and more ability to integrate with other products, but also a much higher purchase threshold and a much more involved installation process. In the end, you have to choose what you want. Do you want a single gadget or do you want a whole smart home system?
- TeMPOraL 11y ago
  > In the end, you have to choose what you want. Do you want a single gadget or do you want a whole smart home system? I want a whole smart home system built piece by piece through buying various smart devices of different vendors and having them talk to each other. This is how it's supposed to work for the "personal use" market, and this is what we need to push for.
Aleman360 11y ago
That's how a lot of actual medical devices work. I worked on an anesthesia machine where the control systems were implemented with micro controllers and RTOS. The UI was Windows Embedded running on a separate x86 board and the whole thing would still safely operate even if the UI subsystem crashed. There were completely mechanical backup controls too.
mavhc 11y ago
On the other hand, then you get people who've no idea about network security doing things, so a lose-lose situation.