As many have noted before, there isn’t a reputable scientific journal where you can report failures. Which is a shame, because failures are an extremely important part of scientific progress.
Today I spent much of the day testing a theory about optics. To do the required experiments, I needed to implement a particular kind of ray tracing program, as well as the math to support ray tracing to various sorts of curved surfaces.
After several hours of coding and testing, I realized that my theory was wrong. But I also understood why it was wrong, and quite a bit else besides.
Not only that, but I now have a handy dandy suite of software tools for testing and visualizing all sorts of optical ideas. Which means that the next time I have some crazy theory about optics, I’ll just be able to jump in and test it, without having to build all the machinery first.
Someday one of those theories is going to be correct, and maybe even useful. Whatever that future discover may be, it’s something I’ll be much more likely to find using my shiny new software laboratory, built on failure.