Week 13 - Machining

This week was machining week! My group’s project can be found on the Harvard section’s machine site.

My Contributions

Involunarily Becoming a Half Section Lead: Brainstorming & Coordination

The first two days, the Harvard section spent brainstorming what we were going to do. When I arrived on Wednesday after 5pm, which was when we were supposed to start brainstorming, nothing was going on. So I took it upon myself to assemble the group, clean the whiteboard, and start the brainstorming discussion.

I did not know that with that, I had set myself up to become the section lead.

The entire group seemed really quiet/shy and tired and nobody else was leading. In situations like this, I just feel that I have to fill the gap and move things foward. So I start the conversation. It doesn’t always feel comfortable because I don’t want to take up too much space and be the center of attention without being given the space for that. However, once I expressed that, the group gave me feedback that they were really appreciating my efforts and they were helping, which I was incredibly grateful to hear.

Unfortunately, I had another project to lead this week so I was really worried about becoming the section lead. We left our first meeting not deciding on one.

During the recitation on Thursday, Neil asked about updates from each section and I replied for Harvard. When I said we didn’t have a section lead, he basically said “I think we’ve just found one.” My mind just went: “No! Not you too!” half-laughing, half-crying. When I refused with the explanation that I had another project, he offered me the “compromise” that I would be the lead until I found somebody to replace me.

At the team meeting that same day, we brought up the topic. “I think it’s you,” Leo also said. “-_-” expresses my reaction well. I guess I couldn’t escape that role.

BUT there was hope! Micheal was interested in leading - and asked me point-blank: “Do you want to co-lead?” YES! And thus we began.

We continued the brainstorming from the previous day and our individual research. It was fun. We had way too many ideas but they were good.

Here is something really important I learned: When making decisions where all options seem interesting, a good mechanism to reach group consensus is downvoting or elimination.

We mapped out the key to do areas and formed task forces around them.

Then we already got started planning the machine.

Helping out here and there

Of course, I was also helping out with here and there. Those contributions are best documented in the group page.

Technical: UI & Code Integration

With my background in web development, I found myself most useful in helping to build photobleaches UI. As we built the project with Modular Things, it was easiest to implement the backend there as well.

I wanted to code up the UI with the vanilla JavaScript but Henghao had already started working on a Gradio UI that directly integrated Catherine’s image processing code.

First, I tried creating a Flask server to serve the code from the Jupyter notebook. However, doing that on Replit took so much time as I was trying to install the required libraries. And on top of that, I would’ve needed to code the frontend with DOM API while we had a perfectly beautiful Gradio frontend.

So instead, Haoheng and I worked to import the Gradio frontend into Modular things as a canvas element. We managed, however, that made the integration really messy and uni-directional. I spent hours trying to trigger our backend code from the Gradio canvas but due to browser regulations (and I guess just how the modern web is built), that wasn’t possible. But anyway, it looks nice and the machine runs.

Shooting the video

Props to Lanson and Keunwook for shooting the video. We did this at like 2am.

Lanson then took off to edit the video. He didn’t sleep all night but the video came out beautifully - we think Neil really liked it.

Demo

Part of being section lead is you’re leading the demo. I was glad our machine was working and people loved the video!