In these ‘Behind the Magic’ series we will give in-depth information about Homey and the technology behind it. Not only because we like geeky people and hopefully provide them some […]
Posted on April 6, 2018 in Behind the Magic
In these Behind the Magic series, we give an overview of what we worked on the month before. Please be warned that the contents might get a little technical now and then.
The new Homey Smartphone App is still under development, and is —as usual— taking longer than we hoped. Most features have been implemented, and the design definitely needs some polishing over the coming time.
Additionally, we have started working on Homey v2.0.0, which will be an API-focused version of Homey. The web interface will be removed, and the Web API will be more consistent. For example, all input will be validated more strictly, so 3rd party services and apps (e.g. HomeyScript) cannot mess something up by accident.
The core will receive many stability and performance improvements as well, as well as preparation work to make it easier and thus faster to add new features in the future.
In Homey v2.0.0 we will also make some changes in speech and music. We have explained the details in this blog post. Summarised, speech recognition using ‘OK Homey’ will be disabled by default for new users (it will still work) and we will replace Homey Music with music controls in devices, and playlists & search in Flow.
Homey is a very complex product to build, and because of the freedom users can have with it, virtually an unlimited amount of situations can happen. We are very happy with our active community reporting their issues both using our Issue tracker and Athom Support, so we can improve Homey further.
The issues that are deserving most of our attention are the following:
We are unique in our approach to support the entire 433 MHz spectrum, whereas other products only support a very narrow band (e.g. 433.92 MHz). Our approach comes with the advantage of support all devices, but with the disadvantage of a shorter range in some households.
We are currently working on an updated version of our Microcontroller software. It will make it easier to adjust the frequency-parameters on the fly, and get more insight about Homey’s performance in different environments. Users will also be able to tweak these settings with the Developer Tools.
This update is planned for an upcoming release in the v1.5.x range.
Larger Z-Wave networks (more than 15 devices), or networks with faulty devices not adhering the Z-Wave specification, are known to slow down the network and other devices. This is something we want to invest more time in to make better.
Another issue seems to be that some devices stop updating their values to Homey (e.g. a motion sensor). Why this happens is not always clear, especially with so many different brands and bugs in the devices themselves. We are looking into this problem as well.
Zigbee is known to delay or not send commands to some devices in larger Zigbee networks (more than 20 devices). Initially we started with Zigbee using an open source implementation, however the project itself isn’t actively maintained anymore and the source code of the project isn’t up to par with our standards. We will therefore start working on our own implementation after the Microcontroller changes discussed above.
This gives us more control over the entire protocol, and will make it much easier to eventually support additional protocols such as TouchLink (used by Philips Hue, for example) and generally upgrade the Zigbee experience.
We have released the Neato BotVac app last month. Most of our work has been in upgrading our existing Homey Apps to SDK2 for improved stability & maintainability. Major projects we’re still working on are KlikAanKlikUit and Fibaro. We have heavily invested in developing tools to make development easier both for us and 3rd party developers.
Last but not least, we have released Homey for MS-DOS® on April 1st. It is very important for us to support as many platforms possible, so go check it out! It actually works! And don’t worry about our time management — we only spent one Hacky Friday on this project.