An anonymous reader writes: A programming buddy and I have cooked up an experimental graphics algorithm (a "2D image processing algorithm" to be precise) that we believe could prove quite useful to people working in a number of different fields if we were to open source it under a GPL license. But there is a small hitch. While we R&D'd every line of the algorithm from zero, putting in 100s of hours of exhausting trial & error and experimentation over a number of months to get it working correctly, we don't know what its "patent compliance status" is. Neither one of us is particularly skilled at reading and understanding what are often highly technical and obscurely worded image processing patents. And there are thousands of them scattered across many different patent databases, filed under all sorts of titkes — i.e. we don't even know WHERE TO START LOOKING to check if our algo overlaps with one or more existing imaging patents.
If we were to open source our graphics algorithm without knowing for certain whether it violates some other person's — or worse — some large, well endowed and potentially "agressively protective" commercial entity's image processing patents, can this bounce back to harm us? i.e. could we get sued for "loss of business" or similar because our "crafted-from-zero" open source graphics algo happens, by chance, to overlap with some 5 or 10 year old patent that we didn't see, didn't understand fully, or didn't realize existed in the first place because its filed under some obscure field (e.g. 'Method for detecting x type edges in y type CMOS images — or — Method for recognizing and removing z type pixel artifacts from x-y-z type medical MR images')?
If you were in our shoes, i.e. strongly wanting to open source a neat 2D graphics algorithm so other people can benefit from it, but uncertain what its patents status is, how would you proceed?