TechCrunch, a top startup-focused Bay Area media company, held its annual hackathon this fall. InterSystems joined four other companies in supporting the efforts of hundreds of hackers over a 3-day period, and giving away over $140,000 in prizes. This post talks about some of the things we learned while introducing the Silicon Valley startup community to our technology.
Day 1: Getting Started
Being new to the Silicon Valley hackathon scene, we didn’t really know what to expect. For most of our company’s history, we’ve focused on serving large institutions. While that emphasis continues, we’re also looking to connect with the broader developer community and introduce a new generation to our unique way of providing a fast, reliable data layer for building the world’s most critical applications.
This year, our team took part in the TechCrunch Hackathon at Disrupt San Francisco 2019. We came armed with a host of resources to attract developers to try out our new data platform, InterSystems IRIS. As part of our participation, we sponsored two prizes. One for the general use of the platform, and one specifically for doing something cool with our health-related features. Since the audience was relatively new to the technology, we brought plenty of engineers and educators to ensure participants had the support and help they’d need during the competition.
This year, InterSystems has focused significantly on language “freedom of choice” at the product level. We now have full support in Java, C#, Python, and Node.js, in addition to ObjectScript. The focus has been on accessing IRIS via SQL, but native global support is also there. Also, this summer we dramatically improved the documentation available for new users, launching Developer.InterSystems.com, which brings together everything you need to start using IRIS from your language of choice in one place.
The audience was a mix of coder ninjas moonlighting from their corporate day jobs, aspiring students looking to make a name for themselves in the startup tech scene, and professional hackers who feed on the energy at these events.
On the first day, I had the pleasure of introducing InterSystems IRIS during a workshop, along with my colleagues, Derek and Nicole. It was a smart and spirited group. These were not junior programmers, but seasoned veterans having experience with multiple programming languages and myriad databases. Everyone was familiar with the usual suspects like MySQL, PostgreSQL and MongoDB. Our job was to introduce IRIS as the platform you choose when you want to take what you’ve learned with those tools and build something faster, more scalable, reliable and with the best support team in the industry standing behind you.
At first blush, a lot of databases look the same, so there wasn’t an immediate rush to use our tech, but people were really impressed by our speed and ability to scale. Our dedication to customer support was also evidenced in the quality of the staff we had onsite. Many teams decided to give us a try, and learn more about getting started.
Day 2: Plugging Away
Day 2 of the hackathon was all about coding, as the hackers got down to the business of building their solutions. What was especially interesting to me was watching the programmers in their working environments. What tools were they using? How did they find support when they got stuck? Which did they choose first: an application framework, a database, or a programming language? Of course, with a group this diverse, the answers were all over the place, but a few patterns did emerge. Most obviously, the IDE of choice was Visual Studio Code. Other than a few outliers, I saw using SublimeText or Notepad++. VS Code was all over the place, whether the coder was on Mac or Windows.
What surprised me was how much more Python was used versus Node.js. Maybe it was because people interested in InterSystems technology tended to be more data science oriented, but most hackers who used InterSystems IRIS were writing SQL in Python using our PyODBC driver. I expected that due to the time constraints of a hackathon, more people would be choosing Node.js for its rapid prototyping and schema-less development benefits. The takeaway was that the SQL/relational paradigm is still #1, despite those that will try to tell you everything is going towards JavaScript and JSON at all levels of the stack.
Given the time constraints of the competition, it was critical for teams to be productive. One thing that helped people achieve that quickly was my colleague, Evgeny’s, ObjectScript Docker template. This is an invaluable tool that automates the process of installing InterSystems IRIS Community Edition from the Docker store, creating a namespace and populating it with scaffolding code, and installing ZPM, the InterSystems package manager to do npm-style code imports.
Finally, one smart developer even learned globals in a few hours and built two different applications using native access.
Shot of the hackathon room and our SE, Rich offering some guidance
Day 3: The Finished Hacks
Friday morning brought a bustle of frenzied activity while over-caffeinated coders brought their projects across the finish line. We got to see a lot of great applications, all of which can be seen here. It was hard to choose the winners. FlyerTV were the only ones to implement globals, showing a deep learning of our platform. But that wasn’t the main reason they won. The basic premise of getting inspired to visit a place while watching a travel show about it, then being able to book the trip on your TV as you watch the show felt like a great use of technology in making our lives simpler and more streamlined by putting fewer screens in front of our faces.
Antibiogram took home the 2nd place prize. Their app involved finding ways to identify the right antibiotic for a patient easier, based on up to date data, and available to a broader range of doctors. They used InterSystems IRIS as a data layer writing SQL in Python.
Our winner for best healthcare application was MEL (mental health evaluator). MEL used real-time health and facial expression data paired with machine learning to analyze the mental state of a patient, then combined that with general mental health data to come up with a mental health score. If MEL decides you’re feeling dangerously blue, it can connect you to a mental health professional. The team used IRIS as a data layer and our FHIR data model for storing patients’ health profiles.
Honorable mention for us went to the Familiar project (which incidentally was done by the same team as FlyerTV). Did you know that about 25% of dogs are returned to shelters each year? Familiar’s goal is to reduce that number by matching pets with owners who look like them, which a study shows is a big factor in return rates. They used “facial” recognition to match pets with owners, which was a big visual hit with the judges.
We’ve always known our tech is a great fit for big enterprises with longer development cycles, but there were a lot of question marks around whether our tech could be learned and put to use in 24 hours. Thanks to the new language support options, this is now possible. Developers can be productive immediately using tech they already know like Python and SQL, and dig deeper into the features that make InterSystems special over time. Coming to this realization while providing talented developers around the world with a data layer that scales with their ideas made the cross-country flight from InterSystems Cambridge, MA headquarters well worth it. As we continue to win the hearts and minds of developers, I look forward to participating in future hackathons, meeting some of the most creative developers and witnessing all of the non-conventional ways InterSystems IRIS can be used to continue drive our digital worlds.