I’ve been committing to OpenCV’s CVAT. This is for a couple of reasons. As I’ve gotten into more management and sales positions, I found that I didn’t have any reasons to keep my technical skills sharp. In fact, what I found with my last roles, is that by trying to commit to projects, I would often cause problems.
My coworkers were not thrilled to find out that my code introduced bugs. They also couldn’t count on me to be technically engaged, because the job pulled me in different directions. Me contributing to our tight turn project ended up being more of a distraction than a help, so I stopped.
So I set out to find a project that was related to the business space that we were in that I could work on while sharpening my technical skills. Specifically, at the time I was working for a custom services shop, and I was trying to push us to have a few computer vision AI projects. Thus we needed to have solid tools and experience to train the AI against. Bringing us back around to OpenCV’s CVAT.
Part of what’s interesting to me is building a community. So now that I’ve got a few commits into CVAT, including some buggy ones, it’s been interesting to me to see what works to help foster building a community.
I’ve already talked about trying to pay for open source contributions. However I’ve yet to see any traction with that method.
One thing I do know, is that when people run into problems or have bugs, they are HIGHLY motivated to contribute. Sometimes challenging them to do so is enough to get them started (it certainly worked on me).
I’ve had some limited success with throwing out some high level ideas, linked to code to get successful code submissions. I think this is a decent practice depending on your familiarity with the code and the technical prowess of your consumers. Diagnosing and providing some ideas of how to fix stuff is often easy. It’s testing and iterating through the solution that becomes difficult.
So if you are running a similar project, that’s my suggestion. Challenge people, and if they run into a specific problem, throw them a couple of breadcrumbs.
You never know who might become a regular contributor to your project this way!