I work on the "garbage in, garbage out" approach. With all of the digital tools, including the linearization algorithms we love, it's still best to have inputs as good as possible. "Post processing" has it's limits and is a lot more work with bad starting files.
With respect to deconvolution sharpening, I'm not convinced we have enough information about all of the causes of the unsharpness for it to be of much help. There are a multitude of different causes of the image defects, and they are not spread evenly through the image. I suspect a good "profile" for a specific lens-camera combination can help in many cases, particularly where someone is looking for a "one-button" solution for processing lots of images. However, I also suspect that careful work with standard tools can do as good a job. Photoshop's Smart Sharpening is said to be a type of deconvolution sharpening, and so far with PS CC I see more unintended negative side effects than benefits. I don't doubt that better sharpening is possible, but I suspect it takes specific profiles that are not readily available. Perhaps some company will come up with a way for us to make our own specific decomvolution sharpening profiles for the specific lens-camera combinations we use.
Paul