[Cad] Which CAD Kernel Should the FOSS and OSHW Communities Focus On?

[Matt Carney]

I don't have a strong programming background, but I do have 10+ years of
professional experience working with Solidworks.

F-rep is worth looking into from the perspective of simplifying the
inside/outside of a boundary, problem - something helpful for declarative
design.


>
> Although 1 and 2 could be part of the same interface, I don't believe that
> you can force all FOSS users to do scripted CAD. You need a traditional
> point-and-click CAD GUI as well for the majority of users. Even with one on
> one training, I've had a hard time getting traditional CAD users to embrace
> CadQuery. It's too much of a shift in thinking.
>
>
The "shift in thinking" is not correctly empathetic to the goals here.
Scripting is fine for highly repetitive generative design, but fails at
one-off design details of complex machines (ie robots, smartphones, etc).
Scripting isn't as useful in these conditions, because, for example: stock
components are not continuous, there are discrete and discontinuous size
options, and every corner you look there is an entirely unique design
problem, mechanism, etc. - it is difficult to generalize every condition. A
database of design rules needs to be able to be referenced by the scripts.

Further, making machines is more experience, art, and empirically derived
design rules than analytical equation. That means your best machine
designers have spent their years taking apart and building things, rather
than programming. That means they have highly developed 3D visualization
skills, but probably not as much programming - these are discretely
different skills.  That means a pure scripting tool isn't useful to those
with the skills and experience to build robust systems.

As an example/challenge: please find me an example of a scripted machine
that's more complicated than some laser cut 2D parts, or static 3d printed
object.

I think it's important to keep in mind what the point of new tools are; I
would think it's about leveraging the experience and skills of the user and
combining that with computational capacity of code. If you drop the gui you
are severely limiting the scope and scalability of your project.

If on the other hand we come up with an interesting way to tie a coding
interface that can smoothly and natively integrate code with a 3D gui point
and click interface then I think we're really onto something cool and
useful that leverages what each type of user interface is good at doing.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/cad-rg/attachments/20151111/5ab3e6ab/attachment.html