chrisb wrote: ↑
Sun May 24, 2020 9:11 pm
I don't fully agree. If the sketches are not attached to faces, then creating fillets as early as possible increases the stability of the model with respect to earlier steps.
If the fillets are created at the end, then every
change corrupts the with some certainty the fillets. If the are placed early in the modeling sequence, then all subsequent steps don't corrupt the fillets.
True, you cannot break an already done fillet with any change later in the tree, but with this qualification: as long as the subsequent steps happen later in the tree. It's when you go back up the tree where the fillets (and chamfers) are vulnerable to topological naming. Fillets, by their very nature, will always be vulnerable to topological naming because there is no getting around the requirement to specify which edge by name to fillet, e.g. Edge23. In this sense we think of making a fillet as equivalent to attaching to a face/edge insofar as topological naming is concerned as both operations will leave the model vulnerable to breaking if changes are made further up the tree. In other words, if we are 100% certain we will not be going up the tree to modify some earlier sketch we can safely ignore topological naming issues and do such things as attaching a sketch to a face at that point.