Ever since February 22 when I entered the Hackster Hardware Weekend in Seattle, I’ve had a growing passion for the open source side of home automation. What started as a simple idea to automate the closing and opening of windows became something bigger than I ever imagined.
The Hackster.io Hardware Weekend was how the Open Smart Hub was born. I started with a hacked together hub that could run on the Intel Edison and automate a servo to act as the window opening mechanism based on WeatherUnderground API information or light/motion from a Spark.io Core (now named Particle.io). Once the event finished I realized that my implementation couldn’t scale and was horribly confusing to recreate.
I began to research the implementations that were available to the public. What were the open source options? What were the professional products? How did they succeed or fail to solve the problem? My conclusion was that the home automation space was cluttered with all the different companies, organizations, products, and applications. What we as consumers and I as a programmer needed was a simple platform to expand, integrate, and customize my personalized home automation experience. IFTTT is a great alternative but it is impossible to add your own devices, actions, functions, etc. There is no communal collaboration! If you added a device and someone else wanted to use the same sort of device, they would have to recreate it themselves.
That is when I began to reimplement the Open Smart Hub with a modular design. I chose Node.js as my platform because of it’s low barrier to entry for programmers, abundant tutorials, and abundant library of open source modules. The core of the new implementation is the configuration file that declares the available device types (think WeMo switches, Hue light bulbs, Nest, Weather Underground data, etc.) as well as a user’s stored scenarios and devices. I chose an implementation where you could fully own and have the ability to control everything. After all it’s your home!
The implementation is split into two parts, a local hub run on a Raspberry Pi 2 within your home network which handles all the interaction between your devices and an online hub that gives you an accessible UI from anywhere.
I had the chance to participate this past weekend in a Hackster.io Hardware Weekend in Seattle and was blown away by the setup. Like most hackathons, they had big sponsors like Intel, Microsoft, Spark, AT&T, etc. However, unlike most of the software hackathons I have been to, they provided some hardware for people to use including Intel Edison boards with Seeed Studio Starter Kits and more. They also provided some cool and useful swag like a portable charger (which I used to power a Spark Core for my demo) and a small portable Leatherman pocket tool that is perfect for my recent maker lifestyle. The energy they provided was just spectacular and even though it was their first time hosting a hackathon, I think it went smoothly.
The food wasn’t just your normal pizza and salad hackathon meals. It also included a legitimate breakfast with bacon, eggs, bagels, cheese, etc. Lunches and dinners were comprised of delicious sandwiches (kind of like Banh Mi), mexican food, and one meal of pizza. On the side they have a whole bunch of candy, popcorn, and more snacks along with the steady supply of coffee, soda, juice, and water.
Unfortunately, like with most hackathons, the crowd of participants thinned out by day 2 with most of the remaining people being interested in learning more or deeply involved in the hacking process.
I came to the event without a clue as to what I was going to build and not really sure if I wanted to join a team, make one, or run it solo. After hearing about some of the prizes for using certain APIs (Weather Underground and WebRTC) I decided to focus my time on the Weather Underground APIs. Even after deciding what I wanted to use, I didn’t really have a clear understanding of what my final product would be and how it could change the world. I just decided to start hacking something together that I thought would be cool to own and ended up going down the path alone.
Stages of my Creation:
Started with the idea to read the forecast for the day and display it to you through an small LCD screen so that I wouldn’t need to pull up an app to view the forecast. Decided to use the Intel Edison, Cylon.js, and the Weather Underground APIs to do this.
Added functionality that would open your windows using a servo if your indoor temperature was past your comfortable zone and the outdoor temperature was colder. I also added functionality to change these settings through buttons and rotary angle sensors on the board.
Added functionality to push the data to the cloud
Realized that I could also connect to a Spark Core and communicate with it via WiFi and the Cloud from the Intel Edison, so I integrated a lighting scenario with a wireless connection.
Created a prototype case for the now deemed “Hub” in Autodesk Fusion 360.
Created a webpage using AngularJS on Azure that would showcase the data my back-end was receiving so that I could view information on the go.
Some things I’ve come to realize for my next hackathon:
Work in teams! I worked solo this weekend and although I did a lot of work to combine all the components, I definitely could have gone further with the idea and took it to the next level ending up with a professional product rather than a hacked together demo.
Set up a team before hand and know the expertise of everyone in the team and how best to leverage them. (This also might mean vet out the people who might have less to contribute if you are going hardcore)
Sometime’s it’s more about the presentation, the story, and the idea than the execution during the hackathon (although that might be due to the hardware nature of this hackathon). After all, you only have so much time both to hack and to present your creation.
Network! This is really just a great opportunity to meet other people in a related field and find out their skills and platforms of choice. Who knows? You might find a couple new tools that might be useful for your future endeavors.
Roll with it. A vision is great, but be able to adapt if things don’t work out quite like you expected. Sometimes code breaks and it can be stressful but learn from it and debug better.
(Update 2/18/2015): I received an email from Ben, one of the Hackster.io founders in response to the feedback I had in my blog. In it he mentioned that they had already been working to improve some of the areas I mentioned and were definitely open to suggestions for improvement. I am psyched about their future updates and the possibility of improving the community with a more social experience. Also excited to attend the Hackster Hardware Weekend in Seattle and see what they have in store!
I recently posted up my Internet Connected Spark RGB Strip project to Hackster.io and despite a couple annoyances, found myself being drawn to posting up more content. It has become so common for websites and content aggregators to gamify their content in order to draw people to continue using their system. (Respects,Views, and Rewards in Hackster.io are similar to Likes and Follows in Instagram, or Likes, Views, and Subscribes on Youtube). Something to keep in mind for my future web/mobile projects will be to gamify certain aspects.
What Hackster.io does well:
They provide a broken-down view of projects that might be of interest to Makers and Hackers. (By platform, trending, or latest)
They make it somewhat easy to create a project and it’s pretty straightforward.
They monitor their content for spam and cool projects.
They gameify their content so that as a project creator, you want to make and post more.
They have drawn in a lot of cool projects already!
What they need to work on:
Adding useful content is WAY too tiresome and formatting is a huge pain on their site (it’s easier to link to other sites where your content is already hosted, but then you are less likely to interest people with your project page).
For Github project owners who want to show code or their README has all the content, transferring it over to Hackster.io’s projects should be painless and update when the Github README changes!
It would be nice if they were faster about monitoring and responding / notifying a project owner that their project has been accepted and made public.
I will continue to use Hackster.io for posting my projects and I think that with time they will improve their content creation system and gain some real traction for the Maker counterpart to Instructables.
After a week of continual use with the Spark Core I have an updated view on the Spark ecosystem.
It’s incredible how easy it is to develop and connect your core to sites developed in Node.js (check out the project I made: http://sparkrgb.azurewebsites.net/). The biggest part of that project was making sure the RGB strip worked and the UI for the website interface. Connecting the two together was the simplest part. The only major disadvantage to using the Spark ecosystem is the fact that you are required to use their Spark cloud to communicate to your device. (Although they are open sourcing a spark-server on Github, it uses the same format as the Spark Cloud.)
Here are some of my personal remarks on the ecosystem:
Connecting the Spark to a Mac is pretty easy through USB, but doing it on your Windows machine requires a little more driver interaction. See my blog post on that here.
Resetting your Spark’s preferred WiFi connection with the mobile apps is horrible when you compare it to the simplicity of using the command-line interface on your development machine through USB.
The Spark Core remembers 7 WiFi credentials. If you add an 8th it will remove the 1st.
In order to compile an INO into the firmware, an internet connection is needed on your development machine at least, so there is no way to develop on a self-contained network even if you do not plan to use internet connected capabilities.
Overall, I think that the Spark ecosystem provides the best experience for developing IoT devices currently.