Em qui., 24 de mar. de 2022 às 23:57, Jan Öhman via Discuss < discuss@lists.openscad.org> escreveu: > Is it possible to combine the examples below > I don't see exactly how you intend to combine Jordan's and mine examples for two reasons. 1) the examples aim to build different kinds of models; Jordan's model is a linear extrusion upward of a shell of your wall shape so requiring one input data: the shape; mine is a sweep of the wall shape along a generic path requiring two set of data: a shape and a path; (*) 2) Jordan's model is produced by operating (by hull() and linear_extrude() ) on primitive objects (circles); the principles of my example are very different because essentially only one operator is applied (BOSL2 sweep() ) to two lists of points (section and path); at the very end, sweep() calls an OpenSCAD operator: polyhedron(). Those approaches might be compared to defining a polygon either by operating (with union(), difference(), etc) on circles and squares or by passing a list of points to the primitive polygon(). > But is it possible to put together several formulas? (in some way) > 1) A formula from the first corner (90 degrees) to the second corner > 2) The second corner (a circle sector of 90 gtrader) should not be > difficult. > 3) A line to the third corner. > 4) 3rd corner (circle sector <90 degrees) > 5) Where should the next straight line begin > etc. > > As Jordan showed there are many ways to define a shape by a list of its vertices. I used the concept of BOSL2 region in order to define the wall by a code similar to your original wall() module but it could have used any list of points defined by list comprehension. Other alternatives are possible like unioning and/or differencing models produced by sweeps. (*) it is possible to argue that linear_extrude is a specific kind of sweep operating on polygons where the path is a vertical segment.