William,

Are you choosing a particular method eg static binding to ensure the core
of the “optional” embedded python stays the same over time?

Edmund

On Monday, May 12, 2025, William F. Adams via Discuss <
discuss@lists.openscad.org> wrote:

I agree with you, Edmund, (although I'm on windows) plus I am sort of
concerned about the reliability of python's stability, since not so long
ago there was upheaval going from v2 to 3. I would not like to build
serious objects on the back of python. Anyway I checked with ChatGPT,
and there is actually an 'official' body looking after python
development (unless it is guessing again) -

Python has a mature, formal governance system with predictable
releases and strong community involvement.
OpenSCAD is more artisanal — driven by passion, caution, and a small
team. That keeps it stable and focused, but slower and less open to
disruptive change.

I find I can do everything I need using  OpenSCAD, although I use it
mainly as a mental exercise. I've tried various alternatives, which seem
to lever off openscad, but I have always returned to pure OpenSCAD.
Currently, I'm playing with pythonscad from pythonscad.org. Other than
that, I think python should be kept as far away from 'pure' OpenSCAD as
possible,  it will likely pollute its purity... OpenSCAD is not broke,
no need to rush into fixing it.

Best wishes,

Ray

On 12/05/2025 13:38, edmund ronald via Discuss wrote:

On Monday, May 12, 2025 at 10:42:24 AM EDT, Raymond West via Discuss discuss@lists.openscad.org wrote:

Unless your project needs mutable variables or the ability to write out arbitrary files --- arguably:

https://github.com/WillAdams/gcodepreview

is the poster-child for that. I certainly appreciate how tolerant the OpenSCAD devs were of my early efforts on this, and also their work/added features which seemed to attempt to accommodate it w/in the limits of OpenSCAD.

On Monday, May 12, 2025 at 09:52:58 AM EDT, Jon Bondy jon@jonbondy.com wrote:

It's a large system which I find confusing, and which uses an approach which does not suit the direction I am trying to take --- I want to be able to model tool movement so as to create a 3D model and create DXFs and G-code directly in that process --- it's looking like I may presently get to a point where I have to re-implement some aspects of it, so will have to dig into it, but overall, thus far, I've found all the examples and mentionings of it confusing to me --- obviously it works well for other folks and is well-suited to a more typical approach.

That said, the first thing which I need is to re-implement the METAFONT/POST curve algorithm, so I'll probably start with that (and extending gcodepreview to write out .mp files).

On Monday, May 12, 2025 at 10:28:16 AM EDT, edmund ronald edmundronald@gmail.com wrote:

Since I'm on Windows, all I did was install Python per the instructions for PythonSCAD and I trust it to make use of that version. Hopefully, moving forward as Python updates, this process will continue to be simple and straight-forward and my code will continue to run --- I'm not doing anything that complex, and I'm not using any constructs which are version-specific, so I believe it will continue to work moving forward. If not, will address it then (there have been long stretches when I haven't updated PythonSCAD, so worst-case is I just stop updating).

I've also been careful not to take on any projects which involve more recent versions of Python --- I don't know if that's necessary or a concern, but I try to keep things as simple as possible.

Similarly, if OpenSCAD had just gained the ability to write out files I'd've probably figured out some way to make it work, but that's not an option for OpenSCAD itself, so the added capability and flexibility which Python adds via OpenPythonSCAD is esp. welcome to me.

William

On 5/12/2025 12:07 AM, Karl Exler via Discuss wrote:

Tough question.

If you're happy with "no programming", if you think that a "for" loop or
an "if" is exotic, then... no.  But then again lots of advanced OpenSCAD
stuff won't be interesting.

If you're interested in programmatic CAD (writing programs to produce
models) and OpenSCAD's limitations and quirks are a problem for you...
yes, definitely.

If you're somewhere in the middle... shrug, it's something you might
want to be aware of but probably not follow in detail.

Eventually I think we might need to split the mailing list into two or
more pieces, with one focusing on the OpenSCAD language and one on the
Python variant, and maybe another on the UI that's shared between the
two.  I don't know when that would be.

Python has a lot of different libraries which could be used to make very
complex sketches.

I mostly use numpy and scipy.

For quite sometime I have not written much but I have already written a lot
of different functions. For people who are using python can use my library.
It is quite powerful.

On Mon, 12 May, 2025, 10:53 pm Jordan Brown via Discuss, <
discuss@lists.openscad.org> wrote:

@Sanjeev Prabhakar sprabhakar2006@gmail.com, which library?  URL?

On Mon, May 12, 2025 at 9:22 PM Sanjeev Prabhakar via Discuss <
discuss@lists.openscad.org> wrote:

I think that it's at:
https://github.com/sprabhakar2006/openSCAD/tree/main
https://github.com/sprabhakar2006/openSCAD/blob/main/openscad2.py

On Tue, May 13, 2025 at 10:43 AM John David via Discuss <
discuss@lists.openscad.org> wrote:

Thanks Leonard:
That is correct

the url: https://github.com/sprabhakar2006/openSCAD

Following files should be downloaded:

• openscad2.py
• dependencies2.scad
• dependencies.scad (optional, Initially I started writing functions in
  pure openscad and very soon gave up, but still there are a lot of useful
  functions written here with explanation).

for learning about the functions (kind of documentation):

• explanation of approaches\ example of various functions.pdf

In my view:
While sketching / modeling you very soon reach a point where openSCAD
native language falls short of doing things and it is very hard to write
complex functions.

Hats off to BOSL developers who have written such elaborate and complex
things in pure openscad. But despite their extraordinary skills, there are
a lot of things which still would be almost impossible to do effectively,
due to the limitation of structured language (e.g. finding the intersection
between solids, fillets in complex cases, offsets 2d and 3d effectively for
complex cases and so on)

I am not very good at python, but still could write very complex functions
with a short study of python language. Till today, I do not know the
concept of classes in python which probably would be very trivial for many
here.

I am sure a good python programmer can write much better and can be useful
to all here.

Thanks and Regards
Sanjeev

On Tue, 13 May 2025 at 21:41, Leonard Martin Struttmann via Discuss <
discuss@lists.openscad.org> wrote:

I think all points of view discussed so far have merit.

One might freeze openSCAD basically as it stands now, why break something
useful? Just as TeX was frozen and serves as a base for the widely employed
LaTex macro system and environments. And then see whether new work which
allows more powerful scripting should build o what is here or on another of
the CodeCAD variants.

On Wed, May 14, 2025 at 3:39 AM Sanjeev Prabhakar via Discuss <
discuss@lists.openscad.org> wrote:

• Maybe because  some of the folks dont like the intentional limitations
  of OpenSCAD  like Scoping and the  inability to communicate with the
  outside world ??

Please note that the intent is not to break  or alter it, but to extend
while highest priority is to stay backward-compatible.
To ensure that, there exist over thousands of ctests to make sure, that we
don't inadvertently break something ...

On Wed, May 14, 2025 at 1:24 PM edmund ronald via Discuss <
discuss@lists.openscad.org> wrote: