In late October we (Michael, Nick and I) saw a post on the GTHackers group on Facebook promoting HackNashville. Almost as a spur of the moment decision, we all signed up and less than a month later we were on a van to Nashville. This is what we did.
The Idea
We came with an concept for a light-hearted web app that facilitated (friendly) competition. It’s simple. Integrate an everyday part of life more seamlessly with web technologies and social media.
-
First take an embarrassing pic(ture). A picture that you don’t want to be shared with your social network or the public as a whole.
-
Then pick your wager. Get a group of your friends and make a bet, a game or some type of contest. There must be one identifiable loser and they will suffer the consequences.
-
Now you are stuck in a Pickl(e). No one can back out of this commitment. once the picture has been uploaded to our servers and the clock ticking, there can be no quitting or flaking. Someone must lose.
Implementation
In the past we had dabbled with Node.js and Express and made one or two projects. This was by far our most ambitious endeavor. Our app called for a full stack, database and interfacing with other API’s (Facebook and Imgur). The three of us all have a bunch of experience programming but we’d soon learn that this project was a bit over our heads.
First up was the question of which framework? Knowing that we wanted to use Node simplified the search quite a bit but there was still quick a selection. In the end, (after an enormous amount of trial and error… mostly error) we settled on Drywall. Based on Node and Express, it implements most of the social media and user management tools that we needed to pull off our app.
Getting to work
Even though we found an excellent foundation for our work, it was not smooth sailing to our final product. We encountered quite a few hiccups, false starts and more than our fair share of backward steps throughout the process. Overall it was definitely a learning experience.
First was installing the actual framework. Drywall is meant to be a starting point for projects like ours but we started off having trouble installing it and getting it running. In the process I learned the hard way how Grunt and Heroku worked. In hindsight the problems we were having were very minimal and trivial but in the moment they were each an abyss of misunderstandings and time consumption.
We were pointed in the direction of Imgur to store our images. While trying to use the Imgur API, we stumbled into another strange and mysterious bug. No matter what we tried and copied from the sample code on Github, we would always have an error. Initially we were trying to send the image as in a GET request. We kept getting errors concerning the length of the URL. After that, we were stuck on a problem with cross-site scripting that I still don’t entirely understand.
Through our own misunderstandings we had the hardest time learning how the Jade templating engine works with Node and Express. That would have been more of a problem if we didn’t also struggle so much with signing in and setting up Facebook integration.
Even though Passport and Drywall makes social media integration really easy, we still had silly implementation issues with getting the nuts and bolts working. For quite a while I wasn’t sure how to actually store the OAuth key on Heroku. I tried over and over to put it in the environment variables section of the Gruntfile only to find that Heroku had an online dashboard where all the variables were stored.
Time is Up
In the end we did not finish the product. But we did get crucial parts working:
users could signup and login with Facebook and our servers could upload and
store pictures on Imgur. Our Node app could even send out confirmation emails
from an @alternativeheroes.me email account! Even though we didn’t finish, our
presentation at expo time still got a lot of positive feedback.
Future Plans
HackNashville was an awesome experience but we couldn’t ship a product because we didn’t have the skill or knowledge to implement the backend. Our time at the hackathon was essentially a 48 hour learning experience. It was an intense, exhausting, no-sleep learning experience but an incredibly transformative one nonetheless. We know have a working knowledge of most of the underpinnings of the Drywall framework and are eager to implement one ourselves. Even though we didn’t finish Pickl at HackNashville, we do plan on rewriting the codebase from the ground up and shipping the product within the school year. Stay tuned for updates! You can find our work on our Github.