Journey of a hackathon project to a Summer of Code Project
Travel Mate started off as a hackathon idea submission in October, 2015. For any college students who know how to code, a free T-shirt and food giving hackathon is always a must-attend event. TiE Delhi-NCR Hackathon 2015 was a similar hackathon where you had to submit a 500 word abstract and if you’re shortlisted you were called onsite to hack on that idea for 24 hours. We sent the first abstract of what would become Travel Mate in years to come. We were called to the Taj Palace Hotel in New Delhi to build a working prototype of our idea. We spent (wasted — since it was “just” a hackathon) most our time deciding on the contracts between the client (Android app) and the server. Begins the presentations, we got questions for some of the Travel industry leaders and they didn’t like half-cooked very basic looking Android app with a vanilla PHP server. We ended up losing that hackathon but we got a lot of positive feedback on how we could improve on our raw idea.
In coming months, came a bunch of online hackathons. We sent out the same writeup and app we had created during the first hackathon and ended up getting decent positions in two other hackathons (10th in IndiaHacks 2015 and Top 56 in @WalmartLabs Hack Fest 2015). This gave us confidence that there was some substance in what we want to build. We started working part-time on the project, hosted some APIs on our college’s shared hosting server, asked some friends to give feedback, (most importantly) added a section on how to setup everything locally in our GitHub project’s Readme and then moved on with our lives.
During January 2017, we started getting emails now and then with some queries and we had no clue what triggered this after almost a year of being a dormant GitHub repository. Nonetheless, in next couple of months, frequency increased and we started getting some minor contributions on the code too. During August-Septemeber, we decided to open some minor bug issues for people to take up during Hacktoberfest. Proved to be a success, we were not able to keep up with the contributions.
In 2018, we were contacted by one of the organisers of GirlScript Summer of Code (which is very similar to GSOC) to register Travel Mate as one of projects. At that point we decided to move away from unscalable PHP to a much more stable Django Python server and also stop using college servers. On 18th June 2018, we launched v2 of Travel Mate which could handle a decent amount of traffic and had relatively modular and easy-to-follow code with following final commit (Nothing fancy in the commit).
Next couple of months saw a lot of engagement from many students participating in the summer of code. We successfully merged a total of 172 pull requests from 13 different participants fixing a total of 205 issues. We tried to enforce best practices and community-accepted benchmarks like UTs, CI, checkstyle, automated build checks, etc. to both the platforms (Android app and Django server).
Cut to present, we have contributions from over 80 people on both platforms, over 900 stars and definitely more commits and PRs coming than we could handle (that’s why there is always a PR backlog). In the end we would say to every open-source project owner, no matter how good/bad code you start with, it can improve over time as long as others can understand what you want to do and how can they jump in. All big open source projects like DuckDuckGo, Mozilla, etc. have one thing in common is their obsession with developer documentation. They don’t want people to have interest in their projects and not contribute just because the documentation was confusing or project setup was difficult.
Hope to see some contribution from everyone who reads this. Download the Android app from Play Store annd if you can, do support us buying us a coffee.
