four months later

One of the activities from last semester's class was to document our research process, which has four main stages: foraging, synthesizing, externalization, and external feedback. The idea our professor encouraged was that we might aim for reliable heuristics or repeatable methods, instead of waiting for those rare flashes of inspiration.

In the foraging stage, we search for existing work. I use Google, Google Scholar, the ACM Digital Library, journal websites, conference programs, and AI tools to see what's already been published. These systems are great at returning papers and books.

But I also lean a lot on scrolling LinkedIn and other social media, where I stumble on new digital humanities projects, online platforms, and digital tools. I still thought this was a bad habit, because it felt so random and unreliable. I tried to put the feeling into words in class, but I couldn't.

Well, it started to make sense to me a few days after I read one of my advisor's recent papers and scrolled through LinkedIn again. She writes that digital scholarship doesn't fit neatly into the conventional genres that our systems are built around — and that this shows up at three points: discovery systems aren't designed to help people find these projects, there's no shared vocabulary to describe them well, and preservation systems aren't built for their evolving, living nature.

Perhaps neither creators nor users genuinely love social media, but the infrastructure makes it the default. Creators announce and share their work there, and the rest of us stumble on it by chance. This is what scholarly communication looks like here; and maybe there are ways we could make it better, right?

Current AI tools, or foundation models, also mean different things to humanities and CS researchers. For CS, they're interesting because they're generalizable, a statistical achievement. For the humanities, they're a solution to all kinds of data problems. It's about capability and knowledge. So lately we're seeing a huge surge in DH projects, with humanities scholars vibe-coding their own. This brings several problems. Vibe-coded projects aren't reviewed the way they used to be, so they may be fragile and hard to sustain. And on top of that, they continue the older problem we started with — all of them are harder to catalog, track, and discover.