lostapathy wrote > The "trick", really, is to design all your parts around a convenient > origin and then translate them to the assembled position.  It seems most > people's intuition is to model the part in the assembled position and > then try to move it back to a convenient origin for printing (or > whatever else). This also has the added benefit of being more modular. It's easier to reuse components if they have a sensible origin. Say, for example, that I model a bolt. I then put it at the origin, wigh the XY plane just below the head, oriented along the Z. This makes a lot more sense when I want to use that bolt in some other design. -- Sent from: http://forum.openscad.org/

Yes make the origin the most useful point but if there are other points of interest, for example screw holes, than I make a module that moves its children there and use it to make the holes both in the object, and what it mounts to and place the screws. So I never need to calculate positions. It is simply a combination of modules. E.g. motor_position() { motor(); motor_screw_positions() screw(motor_screw(), 10); } Similarly I have functions to get the important dimensions of each object, so I know its bounding box allowing me to place it in an assembly and design other things round it. The rule is never specify anything twice. Always factor it into a function, module, or variable and define it once and re-use it. On Fri, 12 Jul 2019 at 07:45, Troberg <troberg.anders@gmail.com> wrote: > lostapathy wrote > > The "trick", really, is to design all your parts around a convenient > > origin and then translate them to the assembled position. It seems most > > people's intuition is to model the part in the assembled position and > > then try to move it back to a convenient origin for printing (or > > whatever else). > > This also has the added benefit of being more modular. It's easier to reuse > components if they have a sensible origin. Say, for example, that I model a > bolt. I then put it at the origin, wigh the XY plane just below the head, > oriented along the Z. This makes a lot more sense when I want to use that > bolt in some other design. > > > > -- > Sent from: http://forum.openscad.org/ > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

On 2019-07-11 20:00, lar3ry wrote: > I see that many folks have asked for a way to find and use a bounding > box in > OpenSCAD, and I wonder if it will ever be implemented. It would be very > handy, and its use could be implemented as just information in the > console, > or used directly in something that might be called > "absolute_translate(). It is a handy feature, standard in AngelCAD, where any shape can return a bounding box https://arnholm.github.io/angelcad-docs/docs/classshape.html#a9511e310496d8d70c26cf0c4b59b0200 https://arnholm.github.io/angelcad-docs/docs/classboundingbox.html Carsten Arnholm

Bounding box isn't implemented in OpenSCAD because geometry is not calculated for the preview, so it isn't available until you do F6. Not only that but all expressions are evaluated before the geometry is created. That is why you can never query the geometry, you can only specify it. On Fri, 12 Jul 2019 at 09:47, <arnholm@arnholm.org> wrote: > On 2019-07-11 20:00, lar3ry wrote: > > I see that many folks have asked for a way to find and use a bounding > > box in > > OpenSCAD, and I wonder if it will ever be implemented. It would be very > > handy, and its use could be implemented as just information in the > > console, > > or used directly in something that might be called > > "absolute_translate(). > > It is a handy feature, standard in AngelCAD, where any shape can return > a bounding box > > https://arnholm.github.io/angelcad-docs/docs/classshape.html#a9511e310496d8d70c26cf0c4b59b0200 > https://arnholm.github.io/angelcad-docs/docs/classboundingbox.html > > Carsten Arnholm > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >