| > In my case, I have found trying to contribute to open source to be a huge waste of time. You mentioned the full extent of your attempt at contributing to open source consisted of a grand total of two meager pull requests. I'm sorry to say but if throughout all those years you only managed to put together two patches then quite obviously you did close to zero contributions, both in absolute and in relative terms. In fact, it seems you complain more about floss than actually contribute stuff. I find that I general it's terribly easy to get your pull requests accepted to random FLOSS projects. Whether your PR is janitorial work, fixing bugs or adding features, you don't need much work to get your contribution in. So, quite frankly it sounds like you are grossly exaggerating and overstating not only the extents of your work but also the challenges you supposedly face. > Then there's the issue of dependencies. My cool tool for analysing satellite photos won't be of much use to anyone who doesn't have API access to a satellite ground station. In that case your so called cool tool to analyse satellite photos sounds very poorly designed. I mean, the whole FLOSS earth observation stack is designed literally from the ground up with extensive separation of concerns in mind (see GDAL and proj4 and orpheo toolbox for example) and apparently somehow you failed to learn the lesson that everyone doing professional, academic, and even amateur work knows by heart. |
Similarly, I've had unpleasant experiences with entitled users more than once, I just picked the Heroku buildpack randomly.
As for the satellite tool, open tools will only take you so far. I do use GDAL, for example. But when you need to take photos on demand, you need very precise real-time trajectory and weather data, for which there is no free equivalent. Plus you actually need to send your photo request somewhere. And that's when you become dependent on an NDA-ed proprietary API.
There's good open source tools if you want to do offline analysis of static archive images. But not everyone has such relaxed requirements for refresh period and processing latency.
Consider using satellite photos to analyze traffic. You schedule photos with a proprietary API, use proprietary modules to download them, followed by proprietary file format converters. And then there's a block of image processing that I could open source, but that won't be helpful without the data acquisition and conversion pipeline. But actually it's a TensorFlow AI trained on proprietary data, so even releasing that is a legal nightmare. The whole project is commercially tainted from start to finish.