Lyceum — dev log 32
Lyceum

I've been sick for the past couple of days, which has been relatively fun. I enjoy getting sick every now and then. Breaks the monotonocity of lockdown.

We have our halfway hackathon this week. It's the mid-point of the MLH Fellowship and to mark that we have our second hackathon. We formed our team quite early and did come up with an idea that we all liked but we've been running into problems with execution. The project involves the use of websockets. A team member who's acquainted with socket.io suggested that we use it. The idea is a web app and we wanted to keep the backend and frontend decoupled. I really wanted to take this opportunity and get my feet wet with writing go, but as it turn out, go's implemtation of socketio supports a very old version of the socket.io client. The python implemtation has a flask version, so I set it up, but again we ran into probelms. When we tried testing some common routes like connect and disconnect, we noticed that socketio wasn't using /ws for websockets but was using long polling and making http requests instead. We wresteled with the problem for a while, but decided it was probably going to be a good idea to just use the node implemtation, as that's what socketio was written for, but we've been running into problems again. It's looking pretty troublesome right about now.

The other piece of news is that I've switched projects again, this time on my own volition. Most of Kiwi's issues deal with writing test cases. Our conversation with the maintainer suggested that those were the only triaged issues that we should work on, until we run out of them. While yes, the bounty on the issues is enticing and I'll be the first to admit that was what attracted it to me at first, I've been getting pretty bored of writing them. The perfect solution came in the form of BentoML. We were in a meeting with our MLH Mentors when I asked them if any other projects had devops related issues. I've been interested in the field for a while and this would provide the perfect excuse to learn it. They directed me to the BentoML project which had an open issue for Apache Kafka Integration, which I scooped up. While I have a distaste for things Machine Learning, the project has enough of devops to keep me interested and learn new technologies. More updates on how the issue goes in the future.