Yes, #publishing failed results are necessary.
But the constraints I find are :
1. lack of #recognition. Journals look for positive results to improve impact factors.
2. Time - You spend lots of time compiling and writing the failed #hypothesis and get rejections. You may utilize that time on another project.
Will it make sense to just create a "Failed projects #archive" without any peer review and protected by #CC BY-SA to make things easier ?
@harikesh @mitras2 @cathal That seems reasonable. And I believe I've heard of something similar in other fields.