I joined an open-source team a year ago to develop an extremely complicated signal processing algorithm. Ended up doing everything myself, with limited guidance from one of its admins, call him Tom. Tom followed my work closely for most part, and expressed much appreciation for it. Few others from this 10-person team chipped anything beyond a few comments.
Recently I coauthored a paper, per Tom's invitation, showcasing the algorithm. My role was only to review, but their failure to get “state of the art” in a machine learning task required my intervention (that I lacked time for), which pushed it several points past.
Nothing hinted something's wrong up to that point, but I did feel the others didn't really treat me as an equal part of the team. Tom denied any such thing, until I pushed to be promoted to admin, when he admitted he thinks I “lack technical competence”. It would seem I, a mere Bachelor's with a questionable username, can't keep par with an all-Ph.D team… despite fixing serious implementation flaws, doubling existing code's speed, and beating SOTA against their design philosophy.
Yesterday, as I'm close to wrapping up, two members casually inform me, “we aren't using your code, it's too complicated to review”. The same members who can't use a debugger beyond bare minimum, and admit they “aren't really software people”. Surely Tom didn't know, and doesn't approve? … Lest to say, he struggled to see anything wrong. So I wrote,
I poured my passion into this project, working with a mutual understanding and a verbal promise from you of an eventual merger, did extra work just for sake of merging with [library], and last minute you declare it's a no go? The solution to “too much to review” is to scrap my work and start from scratch, instead of having those who can't understand start their own damn repository if they want something easier?
I advertised this library, promoted their papers, wrote extensive tutorials. Tom, to whom I had to extensively explain his own library that he didn't write, now scraps my work because he can't understand it either. Now he proposes I rebase my code, with tons of features they won't reproduce, onto one they'll make from scratch… This, knowing that I need things for my resume and affiliation helps.
I'm leaving the library, taking my code, and demanding they remove “in [library]” from the paper's name. Been a ride.
As it's pointless otherwise, library name: Kymatio. I do ask, please don't harass anyone there: they don't deserve it. I do want people to know what happened though, in hopes of preventing more me's.