Sanjeev's approach is to do all the computations in Python and use OpenSCAD
just as a display engine. I don't feel like it represents a solution to
the problem of filleting objects "in OpenSCAD" in any meaningful sense.
You seem dismissive of fillets---"beloved of cast-iron foundries"---but
fillets (and roundovers) are very important, both to produce nice feeling
parts, but also for strength of parts. Any non-filleted corner creates a
stress concentration which can make the part very significantly weaker.
Early research in materials science struggled with a 10x-100x discrepancy
in theoretical material strength compared to actual observed material
strength and the explanation turned out to be stress concentrations. My
personal aim is that no model should have any corners. And this is
remarkably difficult to accomplish in OpenSCAD.
One could imagine the intersection() command having an option to fillet the
intersection, for example. That would provide easy access to fillets for a
lot of situations.
On Sun, Jul 23, 2023 at 2:40 PM Roger Whiteley via Discuss <
discuss@lists.openscad.org> wrote:
---------- Forwarded message ----------
From: Roger Whiteley roger.whiteley@me.com
To: discuss@lists.openscad.org
Cc:
Bcc:
Date: Sun, 23 Jul 2023 19:40:18 +0100
Subject: [OpenSCAD] Re: python support for openscad
As a recently retired Systems Architect, having served over 35 years in
two global IT services companies, specialising in Networks and Systems
Security, anything that compromises my system's security is a Bad Thing.
I have found that there are very few things that OpenSCAD, as a 3D
modelling tool, does badly, simplicity is good, complexity is a loophole
waiting to happen.
As a 3D generative engineering tool, it can be more challenging to
create objects using OpenSCAD that are structurally sound [no sharp
internal corners] and look 'right' , but the wonderful libraries and
community contributions go a long way to fill in the gaps.
One question, if a pair of objects can be mesh filleted using Python
computations to create the image in Sanjeev's email, why does OpenSCAD
need Python support at all?
Fillets, so beloved of cast-iron foundries, are one of the elements
which OpenSCAD could improve, but I'd like to thank all developers of
OpenSCAD for creating something a mathematical numpty like me can use
productively without resorting to mind bending computations to create
something useful.
---------- Forwarded message ----------
From: Roger Whiteley via Discuss discuss@lists.openscad.org
To: discuss@lists.openscad.org
Cc: Roger Whiteley roger.whiteley@me.com
Bcc:
Date: Sun, 23 Jul 2023 19:40:18 +0100
Subject: [OpenSCAD] Re: python support for openscad
OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org
Sanjeev's approach is to do all the computations in Python and use OpenSCAD
just as a display engine. I don't feel like it represents a solution to
the problem of filleting objects "in OpenSCAD" in any meaningful sense.
You seem dismissive of fillets---"beloved of cast-iron foundries"---but
fillets (and roundovers) are very important, both to produce nice feeling
parts, but also for strength of parts. Any non-filleted corner creates a
stress concentration which can make the part very significantly weaker.
Early research in materials science struggled with a 10x-100x discrepancy
in theoretical material strength compared to actual observed material
strength and the explanation turned out to be stress concentrations. My
personal aim is that no model should have any corners. And this is
remarkably difficult to accomplish in OpenSCAD.
One could imagine the intersection() command having an option to fillet the
intersection, for example. That would provide easy access to fillets for a
lot of situations.
On Sun, Jul 23, 2023 at 2:40 PM Roger Whiteley via Discuss <
discuss@lists.openscad.org> wrote:
>
>
>
> ---------- Forwarded message ----------
> From: Roger Whiteley <roger.whiteley@me.com>
> To: discuss@lists.openscad.org
> Cc:
> Bcc:
> Date: Sun, 23 Jul 2023 19:40:18 +0100
> Subject: [OpenSCAD] Re: python support for openscad
> As a recently retired Systems Architect, having served over 35 years in
> two global IT services companies, specialising in Networks and Systems
> Security, anything that compromises my system's security is a Bad Thing.
>
> I have found that there are very few things that OpenSCAD, as a 3D
> modelling tool, does badly, simplicity is good, complexity is a loophole
> waiting to happen.
>
> As a 3D generative engineering tool, it can be more challenging to
> create objects using OpenSCAD that are structurally sound [no sharp
> internal corners] and look 'right' , but the wonderful libraries and
> community contributions go a long way to fill in the gaps.
>
> One question, if a pair of objects can be mesh filleted using Python
> computations to create the image in Sanjeev's email, why does OpenSCAD
> need Python support at all?
>
> Fillets, so beloved of cast-iron foundries, are one of the elements
> which OpenSCAD could improve, but I'd like to thank all developers of
> OpenSCAD for creating something a mathematical numpty like me can use
> productively without resorting to mind bending computations to create
> something useful.
>
>
>
>
>
>
>
> ---------- Forwarded message ----------
> From: Roger Whiteley via Discuss <discuss@lists.openscad.org>
> To: discuss@lists.openscad.org
> Cc: Roger Whiteley <roger.whiteley@me.com>
> Bcc:
> Date: Sun, 23 Jul 2023 19:40:18 +0100
> Subject: [OpenSCAD] Re: python support for openscad
> _______________________________________________
> OpenSCAD mailing list
> To unsubscribe send an email to discuss-leave@lists.openscad.org
>