@User (@User, @User): This is probably a false positive error. There is no "minimum tap area" rule but there is a "minimum psdm area" rule. Since PSDM (P+ implant mask) is automatically generated by magic, it only uses the layer to define regions of p-tap vs. n-tap, and then discards the implant mask geometry. Then it evaluates the DRC rule based on how it would automatically generated the mask in the output. So the original probably had extra PSDM surrounding the tap diffusion area (probably stretching the width of the cell) to satisfy the rule. I am currently working on a method in magic to represent these kind of implant layers in a GDS file that do not match what magic would reproduce in the output, which will get rid of these false positives. I'm not finished with it, yet.