discuss@lists.openscad.org

OpenSCAD general discussion Mailing-list

View all threads

I discovered CadQuery

J
jon
Mon, Jan 15, 2024 3:05 PM

William:

I purchased MeshCAM specifically so that I could take STLs generated by
OpenSCAD and create them on my Shapeoko.  That works to a large extent,
but MeshCAM is a bit quirky and is not updated that often.  Generating
the gCode directly from OpenSCAD would be interesting.  MeshCAM includes
multiple passes (rough and finishing) with different cutting tools. 
Perhaps too much to expect from your experiment.  In any event, nice
seeing you thinking outside of the box

Jon

On 1/15/2024 9:56 AM, William F. Adams via Discuss wrote:

On Monday, January 15, 2024 at 08:39:53 AM EST, jon via Discuss
discuss@lists.openscad.org wrote:

(not so) Privately.

It happens.

Would your code eventually be able to replace MeshCAM?

Maybe, for a much better programmer than me.

The idea is to model how a tool cuts, so one can create G-code w/o
needing a separate CAM program. This is a weird way to work in
OpenSCAD, but it works for me.

I suppose an absolutely brilliant programmer could use this to analyze
an STL and create G-code --- but such a person could probably just
write their own CAM tool w/o such a crutch.

I am currently concentrating on adding DXF support --- because the CAM
stuff is hard --- this will allow writing out DXF files, importing
them into a 3rd party CAM tool such as MeshCAM, then writing out
G-code using that.

The direct G-code generation will currently make anything which can be
done w/ straight lines (working to add arcs now) in a single pass
(there's a radial flattening example file on Github) --- but it needs
to be generalized for different shapes, and to have support added for
multiple passes.

Hopefully I'll have an announcement about arcs working presently.

If someone has a CNC and has a simple project as a test case which
they want cut out, I'd be glad to look into this with them.

William

--
Sphinx of black quartz, judge my vow.
https://designinto3d.com/


OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

William: I purchased MeshCAM specifically so that I could take STLs generated by OpenSCAD and create them on my Shapeoko.  That works to a large extent, but MeshCAM is a bit quirky and is not updated that often.  Generating the gCode directly from OpenSCAD would be interesting.  MeshCAM includes multiple passes (rough and finishing) with different cutting tools.  Perhaps too much to expect from your experiment.  In any event, nice seeing you thinking outside of the box Jon On 1/15/2024 9:56 AM, William F. Adams via Discuss wrote: > On Monday, January 15, 2024 at 08:39:53 AM EST, jon via Discuss > <discuss@lists.openscad.org> wrote: > > >(not so) Privately. > > It happens. > > >Would your code eventually be able to replace MeshCAM? > > Maybe, for a much better programmer than me. > > The idea is to model how a tool cuts, so one can create G-code w/o > needing a separate CAM program. This is a weird way to work in > OpenSCAD, but it works for me. > > I suppose an absolutely brilliant programmer could use this to analyze > an STL and create G-code --- but such a person could probably just > write their own CAM tool w/o such a crutch. > > I am currently concentrating on adding DXF support --- because the CAM > stuff is _hard_ --- this will allow writing out DXF files, importing > them into a 3rd party CAM tool such as MeshCAM, then writing out > G-code using that. > > The direct G-code generation will currently make anything which can be > done w/ straight lines (working to add arcs now) in a single pass > (there's a radial flattening example file on Github) --- but it needs > to be generalized for different shapes, and to have support added for > multiple passes. > > Hopefully I'll have an announcement about arcs working presently. > > If someone has a CNC and has a simple project as a test case which > they want cut out, I'd be glad to look into this with them. > > William > > -- > Sphinx of black quartz, judge my vow. > https://designinto3d.com/ > > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email todiscuss-leave@lists.openscad.org
BB
Bruno Boettcher
Mon, Jan 15, 2024 6:17 PM

Hello!
excuse the noob question, but it seems to me that writing a slicer is not a
trivial task? I am switching between cura and icesl (both have advantages
and disadvantages)  and both are in heavy development for some time?
So if you absolutely don't want to switch programs,probably a named pipe
towards e.g. icesl (which is open source too) would be better than
reinventing the wheel?
excuse me if i am offtopic :D but the problem seemed weird to me....

ciao
bboett

Am Mo., 15. Jan. 2024 um 16:06 Uhr schrieb jon via Discuss <
discuss@lists.openscad.org>:

William:

I purchased MeshCAM specifically so that I could take STLs generated by
OpenSCAD and create them on my Shapeoko.  That works to a large extent, but
MeshCAM is a bit quirky and is not updated that often.  Generating the
gCode directly from OpenSCAD would be interesting.  MeshCAM includes
multiple passes (rough and finishing) with different cutting tools.
Perhaps too much to expect from your experiment.  In any event, nice seeing
you thinking outside of the box

Jon
On 1/15/2024 9:56 AM, William F. Adams via Discuss wrote:

On Monday, January 15, 2024 at 08:39:53 AM EST, jon via Discuss
discuss@lists.openscad.org discuss@lists.openscad.org wrote:

(not so) Privately.

It happens.

Would your code eventually be able to replace MeshCAM?

Maybe, for a much better programmer than me.

The idea is to model how a tool cuts, so one can create G-code w/o needing
a separate CAM program. This is a weird way to work in OpenSCAD, but it
works for me.

I suppose an absolutely brilliant programmer could use this to analyze an
STL and create G-code --- but such a person could probably just write their
own CAM tool w/o such a crutch.

I am currently concentrating on adding DXF support --- because the CAM
stuff is hard --- this will allow writing out DXF files, importing them
into a 3rd party CAM tool such as MeshCAM, then writing out G-code using
that.

The direct G-code generation will currently make anything which can be
done w/ straight lines (working to add arcs now) in a single pass (there's
a radial flattening example file on Github) --- but it needs to be
generalized for different shapes, and to have support added for multiple
passes.

Hopefully I'll have an announcement about arcs working presently.

If someone has a CNC and has a simple project as a test case which they
want cut out, I'd be glad to look into this with them.

William

--
Sphinx of black quartz, judge my vow.
https://designinto3d.com/


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

Hello! excuse the noob question, but it seems to me that writing a slicer is not a trivial task? I am switching between cura and icesl (both have advantages and disadvantages) and both are in heavy development for some time? So if you absolutely don't want to switch programs,probably a named pipe towards e.g. icesl (which is open source too) would be better than reinventing the wheel? excuse me if i am offtopic :D but the problem seemed weird to me.... ciao bboett Am Mo., 15. Jan. 2024 um 16:06 Uhr schrieb jon via Discuss < discuss@lists.openscad.org>: > William: > > I purchased MeshCAM specifically so that I could take STLs generated by > OpenSCAD and create them on my Shapeoko. That works to a large extent, but > MeshCAM is a bit quirky and is not updated that often. Generating the > gCode directly from OpenSCAD would be interesting. MeshCAM includes > multiple passes (rough and finishing) with different cutting tools. > Perhaps too much to expect from your experiment. In any event, nice seeing > you thinking outside of the box > > Jon > On 1/15/2024 9:56 AM, William F. Adams via Discuss wrote: > > On Monday, January 15, 2024 at 08:39:53 AM EST, jon via Discuss > <discuss@lists.openscad.org> <discuss@lists.openscad.org> wrote: > > >(not so) Privately. > > It happens. > > >Would your code eventually be able to replace MeshCAM? > > Maybe, for a much better programmer than me. > > The idea is to model how a tool cuts, so one can create G-code w/o needing > a separate CAM program. This is a weird way to work in OpenSCAD, but it > works for me. > > I suppose an absolutely brilliant programmer could use this to analyze an > STL and create G-code --- but such a person could probably just write their > own CAM tool w/o such a crutch. > > I am currently concentrating on adding DXF support --- because the CAM > stuff is _hard_ --- this will allow writing out DXF files, importing them > into a 3rd party CAM tool such as MeshCAM, then writing out G-code using > that. > > The direct G-code generation will currently make anything which can be > done w/ straight lines (working to add arcs now) in a single pass (there's > a radial flattening example file on Github) --- but it needs to be > generalized for different shapes, and to have support added for multiple > passes. > > Hopefully I'll have an announcement about arcs working presently. > > If someone has a CNC and has a simple project as a test case which they > want cut out, I'd be glad to look into this with them. > > William > > -- > Sphinx of black quartz, judge my vow. > https://designinto3d.com/ > > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org > -- ciao Bruno =========================================== http://nohkumado.eu/, <http://bboett.free.fr>http://aikido.nohkumado.eu/, <http://bboett.free.fr> http://aikido.zorn.free.fr
TP
Torsten Paul
Mon, Jan 15, 2024 7:43 PM

On 15.01.24 19:17, Bruno Boettcher via Discuss wrote:

towards e.g. icesl (which is open source too) would be better than
reinventing the wheel?

What? Where?

All I can see is a "binary" license which in strict reading would not
even permit normal private use, it only mentions "research use".

Where's the source code?

ciao,
Torsten.

On 15.01.24 19:17, Bruno Boettcher via Discuss wrote: > towards e.g. icesl (which is open source too) would be better than > reinventing the wheel? What? Where? All I can see is a "binary" license which in strict reading would not even permit normal private use, it only mentions "research use". Where's the source code? ciao, Torsten.
WF
William F. Adams
Mon, Jan 15, 2024 8:16 PM

On Monday, January 15, 2024 at 01:18:18 PM EST, Bruno Boettcher via Discuss discuss@lists.openscad.org wrote:

excuse the noob question, but it seems to me that writing a slicer is not a trivial task? 

Decidedly not, hence my weird approach.

I am switching between cura and icesl (both have advantages and disadvantages)  and both are in heavy development for some time?

This G-code is for a subtractive CNC router (or mill).

So if you absolutely don't want to switch programs,probably a named pipe towards e.g. icesl (which is open source too) would be better than reinventing the wheel?

I don't know of any similar projects --- the closest would be CAMotics:
https://camotics.org/

but I don't like Javascript, and since they did a controller for a company which competes with the one I work for, that's not an option.

excuse me if i am offtopic :D but the problem seemed weird to me....

It is weird.
The idea is to avoid the need for CAM software (or to simplify using CAM software) by directly modeling the tool movement, and recording that either as G-code, or in a DXF.
A good example of a problem which this works around is slots which are exactly the same diameter as the tool --- when describing as outlines it is necessary to add 10% or so so that rounding issues won't interfere with the calculations of whether the tool fits or no.
I'm going to post a template file to my Github now, then work on adding support for arcs --- once I've got that done I'll try to work up an example file which explains this better.
I did do one compleat example, a flattening file:
https://github.com/WillAdams/gcodepreview/blob/main/flatten.graph.tres

which when opened in OpenSCAD Graph Editor should show a radial flattening toolpath (some configuration will be necessary, including installing gcodepreview).
William

On Monday, January 15, 2024 at 01:18:18 PM EST, Bruno Boettcher via Discuss <discuss@lists.openscad.org> wrote: >excuse the noob question, but it seems to me that writing a slicer is not a trivial task?  Decidedly not, hence my weird approach. >I am switching between cura and icesl (both have advantages and disadvantages)  and both are in heavy development for some time? This G-code is for a subtractive CNC router (or mill). >So if you absolutely don't want to switch programs,probably a named pipe towards e.g. icesl (which is open source too) would be better than reinventing the wheel? I don't know of any similar projects --- the closest would be CAMotics: https://camotics.org/ but I don't like Javascript, and since they did a controller for a company which competes with the one I work for, that's not an option. >excuse me if i am offtopic :D but the problem seemed weird to me.... It is weird. The idea is to avoid the need for CAM software (or to simplify using CAM software) by directly modeling the tool movement, and recording that either as G-code, or in a DXF. A good example of a problem which this works around is slots which are exactly the same diameter as the tool --- when describing as outlines it is necessary to add 10% or so so that rounding issues won't interfere with the calculations of whether the tool fits or no. I'm going to post a template file to my Github now, then work on adding support for arcs --- once I've got that done I'll try to work up an example file which explains this better. I did do one compleat example, a flattening file: https://github.com/WillAdams/gcodepreview/blob/main/flatten.graph.tres which when opened in OpenSCAD Graph Editor should show a radial flattening toolpath (some configuration will be necessary, including installing gcodepreview). William
BB
Bruno Boettcher
Mon, Jan 15, 2024 9:17 PM

Ah yes, you have to ask them directly...
French universities still have problems with public code repositories....

Ciao

Bruno Böttcher

35 rue de la république, FR-6720 Schwindratzheim
email: bboett@adlp.org, mobile:bboett@gmail.com
Fon:+33 3 88 89 91, Mob:+33 6 76 55 82 68

Dev: Java/Perl/PHP OS:GNU/LINUX, Android
Aïkido: http://aikido.zorn.free.fr, http://www.aikido-club-saverne.fr

Torsten Paul via Discuss discuss@lists.openscad.org schrieb am Mo., 15.
Jan. 2024, 20:44:

On 15.01.24 19:17, Bruno Boettcher via Discuss wrote:

towards e.g. icesl (which is open source too) would be better than
reinventing the wheel?

What? Where?

All I can see is a "binary" license which in strict reading would not
even permit normal private use, it only mentions "research use".

Where's the source code?

ciao,
Torsten.


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

Ah yes, you have to ask them directly... French universities still have problems with public code repositories.... Ciao Bruno Böttcher -- 35 rue de la république, FR-6720 Schwindratzheim email: bboett@adlp.org, mobile:bboett@gmail.com Fon:+33 3 88 89 91, Mob:+33 6 76 55 82 68 ------------------------------------------------- Dev: Java/Perl/PHP OS:GNU/LINUX, Android Aïkido: http://aikido.zorn.free.fr, http://www.aikido-club-saverne.fr Torsten Paul via Discuss <discuss@lists.openscad.org> schrieb am Mo., 15. Jan. 2024, 20:44: > On 15.01.24 19:17, Bruno Boettcher via Discuss wrote: > > towards e.g. icesl (which is open source too) would be better than > > reinventing the wheel? > > What? Where? > > All I can see is a "binary" license which in strict reading would not > even permit normal private use, it only mentions "research use". > > Where's the source code? > > ciao, > Torsten. > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org >
JB
Jordan Brown
Mon, Jan 15, 2024 10:22 PM

I have to wonder whether a good (better?) answer for this mightn't be to
use a conventional programming language to emit gcode, and a gcode viewer.

Admittedly most of the gcode viewers that I know of are for
filament-style 3D printing, not for subtractive technologies.  You would
want one that knows how to display the path differently for different
tool shapes.

I keep wanting to do CNC the same way I do 3DP, OpenSCAD -> STL ->
gcode, but that doesn't seem to work all that well because (a) CNC at my
level is 2.5D, not 3D, (b) tool shape matters, (c) feeds and speeds
matter (more), and (d) cut path matters (more), e.g. concentric vs
raster.  Still, a programmatic CAD that feeds into a CAM toolchain seems
like a big win.  I don't want to control it at the gcode level,
controlling every tool movement, but I do want to programmatically say
"put holes in a 10cm grid".  How exactly that would interact with the
desire to do bulk cutting with a big bit and then switch to a smaller
bit to fill in the details, I don't know.

I have to wonder whether a good (better?) answer for this mightn't be to use a conventional programming language to emit gcode, and a gcode viewer. Admittedly most of the gcode viewers that I know of are for filament-style 3D printing, not for subtractive technologies.  You would want one that knows how to display the path differently for different tool shapes. I keep wanting to do CNC the same way I do 3DP, OpenSCAD -> STL -> gcode, but that doesn't seem to work all that well because (a) CNC at my level is 2.5D, not 3D, (b) tool shape matters, (c) feeds and speeds matter (more), and (d) cut path matters (more), e.g. concentric vs raster.  Still, a programmatic CAD that feeds into a CAM toolchain seems like a big win.  I don't want to control it at the gcode level, controlling every tool movement, but I do want to programmatically say "put holes in a 10cm grid".  How exactly that would interact with the desire to do bulk cutting with a big bit and then switch to a smaller bit to fill in the details, I don't know.
LM
Leonard Martin Struttmann
Tue, Jan 16, 2024 3:15 AM

UPDATE: So I finally got CadQuery/CQ-editor installed and spent about three
hours playing with it today to design a simple part.

Conclusion: I will probably NOT be switching over to CadQuery.

I found the user interface to be clunky, and the editor is very basic with
little in the way of preferences.

CadQuery is very flexible, but flexibility breeds complexity.  The learning
curve is very steep.

In practice, I found it to work just as if you were documenting each step
you would take if you were using a drawing program such as TinkerCAD,
FreeCADD, etc. For simple parts, each line of code (or each layer in the
stack of functions) corresponds with a similar action on a drawing program:
Select a plane, do some 2D drawing, then cut or extrude, select some edges,
apply chamfer/fillets, and then repeat.

The devil is in the details and in understanding what is going on under the
hood (bonnet).  For example, I could extrude a regular polygon, but I could
not extrude a polyline (created from a list of points).  After much
searching, I found that one must use the .close() function between the
polyline and the extrude.

Of course, your mileage may vary and if you are transitioning from a
drawing program you may find it easier than this old guy.

End of report.

On Mon, Jan 15, 2024 at 7:44 AM jon via Discuss discuss@lists.openscad.org
wrote:

Great.  Not so privately.  Sigh.

When Reply does not mean Reply.

On 1/15/2024 8:39 AM, jon via Discuss wrote:

Privately.

Would your code eventually be able to replace MeshCAM?

Jon

On 1/14/2024 10:16 PM, William F. Adams via Discuss wrote:

I tried CadQuery a while back --- it was confusing because they were just
transitioning from being a workbench in FreeCAD (which I was hopeful of) to
being a stand-alone project (which I found off-putting and which lacked the
synergy w/ a graphical program I was hoping for).

Agree that there was a striking absence of example files/code, and the
ones which I could find seemed arcane and confusing in how they were coded.

The nascent effort to put Python into OpenSCAD has been a lot more
approachable to me:

http://pythonscad.org/

https://github.com/openscad/openscad/issues/4880

(but I don't know if anyone would want to do anything the way I do it:
https://github.com/WillAdams/gcodepreview )

William

--
Sphinx of black quartz, judge my vow.
https://designinto3d.com/


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

UPDATE: So I finally got CadQuery/CQ-editor installed and spent about three hours playing with it today to design a simple part. Conclusion: I will probably NOT be switching over to CadQuery. I found the user interface to be clunky, and the editor is very basic with little in the way of preferences. CadQuery is very flexible, but flexibility breeds complexity. The learning curve is very steep. In practice, I found it to work just as if you were documenting each step you would take if you were using a drawing program such as TinkerCAD, FreeCADD, etc. For simple parts, each line of code (or each layer in the stack of functions) corresponds with a similar action on a drawing program: Select a plane, do some 2D drawing, then cut or extrude, select some edges, apply chamfer/fillets, and then repeat. The devil is in the details and in understanding what is going on under the hood (bonnet). For example, I could extrude a regular polygon, but I could not extrude a polyline (created from a list of points). After much searching, I found that one must use the .close() function between the polyline and the extrude. Of course, your mileage may vary and if you are transitioning from a drawing program you may find it easier than this old guy. End of report. On Mon, Jan 15, 2024 at 7:44 AM jon via Discuss <discuss@lists.openscad.org> wrote: > Great. Not so privately. Sigh. > > When Reply does not mean Reply. > > > On 1/15/2024 8:39 AM, jon via Discuss wrote: > > Privately. > > Would your code eventually be able to replace MeshCAM? > > Jon > > > On 1/14/2024 10:16 PM, William F. Adams via Discuss wrote: > > I tried CadQuery a while back --- it was confusing because they were just > transitioning from being a workbench in FreeCAD (which I was hopeful of) to > being a stand-alone project (which I found off-putting and which lacked the > synergy w/ a graphical program I was hoping for). > > Agree that there was a striking absence of example files/code, and the > ones which I could find seemed arcane and confusing in how they were coded. > > The nascent effort to put Python into OpenSCAD has been a lot more > approachable to me: > > http://pythonscad.org/ > > https://github.com/openscad/openscad/issues/4880 > > (but I don't know if anyone would want to do anything the way I do it: > https://github.com/WillAdams/gcodepreview ) > > William > > -- > Sphinx of black quartz, judge my vow. > https://designinto3d.com/ > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org > > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org > > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org >
TP
Torsten Paul
Tue, Jan 16, 2024 4:56 AM

On 15.01.24 22:17, Bruno Boettcher via Discuss wrote:

Ah yes, you have to ask them directly...
French universities still have problems with public code repositories....

That is not how open source works. At all.

CGAL is Universtity/French based too, so I don't really buy that claim
(https://geometryfactory.com/who-we-are/).

Quote "Andreas founded GeometryFactory in 2003 in order to commercialize
the technology accumulated through a series of European research
projects developing the CGAL library."

That is exactly what IceSL did/does.

ciao,
Torsten.

On 15.01.24 22:17, Bruno Boettcher via Discuss wrote: > Ah yes, you have to ask them directly... > French universities still have problems with public code repositories.... That is not how open source works. At all. CGAL is Universtity/French based too, so I don't really buy that claim (https://geometryfactory.com/who-we-are/). Quote "Andreas founded GeometryFactory in 2003 in order to commercialize the technology accumulated through a series of European research projects developing the CGAL library." That is *exactly* what IceSL did/does. ciao, Torsten.
BB
Bruno Boettcher
Tue, Jan 16, 2024 5:17 AM

Well as said, french universities have, usually, indeed a problem with open
source (I made my PhD at one...) since in the administrative layers there's
the big illusion of successful software/hardware patents and the big
money....
But also, usually, it's possible to retrieve, under the guise of a
federated project, the stuff under a Eugpl like license (since the people
writing the stuff usually are in the inverse opinion) ...

But no matter, indeed I didn't knew there weren't any open source CNC
"slicers"

Ciao

Bruno Böttcher

35 rue de la république, FR-6720 Schwindratzheim
email: bboett@adlp.org, mobile:bboett@gmail.com
Fon:+33 3 88 89 91, Mob:+33 6 76 55 82 68

Dev: Java/Perl/PHP OS:GNU/LINUX, Android
Aïkido: http://aikido.zorn.free.fr, http://www.aikido-club-saverne.fr

Torsten Paul via Discuss discuss@lists.openscad.org schrieb am Di., 16.
Jan. 2024, 05:56:

On 15.01.24 22:17, Bruno Boettcher via Discuss wrote:

Ah yes, you have to ask them directly...
French universities still have problems with public code repositories....

That is not how open source works. At all.

CGAL is Universtity/French based too, so I don't really buy that claim
(https://geometryfactory.com/who-we-are/).

Quote "Andreas founded GeometryFactory in 2003 in order to commercialize
the technology accumulated through a series of European research
projects developing the CGAL library."

That is exactly what IceSL did/does.

ciao,
Torsten.


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

Well as said, french universities have, usually, indeed a problem with open source (I made my PhD at one...) since in the administrative layers there's the big illusion of successful software/hardware patents and the big money.... But also, usually, it's possible to retrieve, under the guise of a federated project, the stuff under a Eugpl like license (since the people writing the stuff usually are in the inverse opinion) ... But no matter, indeed I didn't knew there weren't any open source CNC "slicers" Ciao Bruno Böttcher -- 35 rue de la république, FR-6720 Schwindratzheim email: bboett@adlp.org, mobile:bboett@gmail.com Fon:+33 3 88 89 91, Mob:+33 6 76 55 82 68 ------------------------------------------------- Dev: Java/Perl/PHP OS:GNU/LINUX, Android Aïkido: http://aikido.zorn.free.fr, http://www.aikido-club-saverne.fr Torsten Paul via Discuss <discuss@lists.openscad.org> schrieb am Di., 16. Jan. 2024, 05:56: > On 15.01.24 22:17, Bruno Boettcher via Discuss wrote: > > Ah yes, you have to ask them directly... > > French universities still have problems with public code repositories.... > > That is not how open source works. At all. > > CGAL is Universtity/French based too, so I don't really buy that claim > (https://geometryfactory.com/who-we-are/). > > Quote "Andreas founded GeometryFactory in 2003 in order to commercialize > the technology accumulated through a series of European research > projects developing the CGAL library." > > That is *exactly* what IceSL did/does. > > ciao, > Torsten. > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org >
N
neri-engineering
Tue, Jan 16, 2024 7:08 AM

CadQuery is very powerful; it's for the power user, much like UNIX can be compared to Windows. A power user who does not fear learning things would always gravitate towards UNIX or a UNIX-like system (even MacOS I suppose).

Yes you're right, the installation of CadQuery and of cq-editor was a bit more difficult than is typical, and it's because these tools live in "Python world" which needs to know how to load libs etc. That's why they have tools such as conda, anaconda, mambaforge, etc. to handle library dependencies and library installations. Even I was not an expert in Python but I was able to install these tools after careful reading and a few trial and error steps. Yes it took a couple of hours in my case as well.

I still do like OpenSCAD, but for me it's crucial to have a tool that generates STEP files for CNC machining at my disposal. This is why I had to look beyond just OpenSCAD. Furthermore I'm very repulsed with the idea of using graphical user interfaces to create parts that are very similar to each other, with minor differences in between. In other words having a language with which to "program parts" is very agreeable with me; this is exactly what drew me to OpenSCAD initially.

Because CadQuery uses a B-rep lib as its backend, things like "shelling", "fillet", or even "chamfer" are built-ins. In other words smoothing things is easy; it's part of the backend library. These are frequently discussed topics on this forum. 'Helix' and other complex operations are also built-ins. Yes there is a learning curve with the extensive API, but once you start working on simple examples it becomes clearer and clearer. STL can be generated by CadQuery.

Well here is an example. Threads for screws are very easy to program. With about 300 lines of code I was able to write a complete thread library in two days, which generates both internal and external threads, with many options. (Now of course I will be maintaining it when I find bugs, which is expected because I'm using this in all of my projects.)

If you are interested in trying the thread generation code, it is here:

https://sourceforge.net/p/nl10/code/HEAD/tree/cq-code/common/metric_threads.py

If you add these few lines of code

show_object(internal_metric_thread(diameter=3.0,
pitch=0.5,
length=2.0,
bottom_chamfer=True,
base_tube_od=3.5))

to the bottom of thread lib linked above, opening in cq-editor will get you this:

[threads.png]

Cheers.

Sent with Proton Mail secure email.

On Monday, January 15th, 2024 at 9:15 PM, Leonard Martin Struttmann via Discuss discuss@lists.openscad.org wrote:

UPDATE: So I finally got CadQuery/CQ-editor installed and spent about three hours playing with it today to design a simple part.

Conclusion: I will probably NOT be switching over to CadQuery.

I found the user interface to be clunky, and the editor is very basic with little in the way of preferences.

CadQuery is very flexible, but flexibility breeds complexity. The learning curve is very steep.

In practice, I found it to work just as if you were documenting each step you would take if you were using a drawing program such as TinkerCAD, FreeCADD, etc. For simple parts, each line of code (or each layer in the stack of functions) corresponds with a similar action on a drawing program: Select a plane, do some 2D drawing, then cut or extrude, select some edges, apply chamfer/fillets, and then repeat.

The devil is in the details and in understanding what is going on under the hood (bonnet). For example, I could extrude a regular polygon, but I could not extrude a polyline (created from a list of points). After much searching, I found that one must use the .close() function between the polyline and the extrude.

Of course, your mileage may vary and if you are transitioning from a drawing program you may find it easier than this old guy.

End of report.

On Mon, Jan 15, 2024 at 7:44 AM jon via Discuss discuss@lists.openscad.org wrote:

Great. Not so privately. Sigh.

When Reply does not mean Reply.

On 1/15/2024 8:39 AM, jon via Discuss wrote:

Privately.

Would your code eventually be able to replace MeshCAM?

Jon

On 1/14/2024 10:16 PM, William F. Adams via Discuss wrote:

I tried CadQuery a while back --- it was confusing because they were just transitioning from being a workbench in FreeCAD (which I was hopeful of) to being a stand-alone project (which I found off-putting and which lacked the synergy w/ a graphical program I was hoping for).

Agree that there was a striking absence of example files/code, and the ones which I could find seemed arcane and confusing in how they were coded.

The nascent effort to put Python into OpenSCAD has been a lot more approachable to me:

http://pythonscad.org/

https://github.com/openscad/openscad/issues/4880

(but I don't know if anyone would want to do anything the way I do it: https://github.com/WillAdams/gcodepreview )

William

--
Sphinx of black quartz, judge my vow.
https://designinto3d.com/


OpenSCAD mailing list
To unsubscribe send an email to
discuss-leave@lists.openscad.org


OpenSCAD mailing list
To unsubscribe send an email to
discuss-leave@lists.openscad.org


OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

CadQuery is very powerful; it's for the power user, much like UNIX can be compared to Windows. A power user who does not fear learning things would always gravitate towards UNIX or a UNIX-like system (even MacOS I suppose). Yes you're right, the installation of CadQuery and of cq-editor was a bit more difficult than is typical, and it's because these tools live in "Python world" which needs to know how to load libs etc. That's why they have tools such as conda, anaconda, mambaforge, etc. to handle library dependencies and library installations. Even I was not an expert in Python but I was able to install these tools after careful reading and a few trial and error steps. Yes it took a couple of hours in my case as well. I still do like OpenSCAD, but for me it's crucial to have a tool that generates STEP files for CNC machining at my disposal. This is why I had to look beyond just OpenSCAD. Furthermore I'm very repulsed with the idea of using graphical user interfaces to create parts that are very similar to each other, with minor differences in between. In other words having a language with which to "program parts" is very agreeable with me; this is exactly what drew me to OpenSCAD initially. Because CadQuery uses a B-rep lib as its backend, things like "shelling", "fillet", or even "chamfer" are built-ins. In other words smoothing things is easy; it's part of the backend library. These are frequently discussed topics on this forum. 'Helix' and other complex operations are also built-ins. Yes there is a learning curve with the extensive API, but once you start working on simple examples it becomes clearer and clearer. STL can be generated by CadQuery. Well here is an example. Threads for screws are very easy to program. With about 300 lines of code I was able to write a complete thread library in two days, which generates both internal and external threads, with many options. (Now of course I will be maintaining it when I find bugs, which is expected because I'm using this in all of my projects.) If you are interested in trying the thread generation code, it is here: https://sourceforge.net/p/nl10/code/HEAD/tree/cq-code/common/metric_threads.py If you add these few lines of code show_object(internal_metric_thread(diameter=3.0, pitch=0.5, length=2.0, bottom_chamfer=True, base_tube_od=3.5)) to the bottom of thread lib linked above, opening in cq-editor will get you this: [threads.png] Cheers. Sent with [Proton Mail](https://proton.me/) secure email. On Monday, January 15th, 2024 at 9:15 PM, Leonard Martin Struttmann via Discuss <discuss@lists.openscad.org> wrote: > UPDATE: So I finally got CadQuery/CQ-editor installed and spent about three hours playing with it today to design a simple part. > > Conclusion: I will probably NOT be switching over to CadQuery. > > I found the user interface to be clunky, and the editor is very basic with little in the way of preferences. > > CadQuery is very flexible, but flexibility breeds complexity. The learning curve is very steep. > > In practice, I found it to work just as if you were documenting each step you would take if you were using a drawing program such as TinkerCAD, FreeCADD, etc. For simple parts, each line of code (or each layer in the stack of functions) corresponds with a similar action on a drawing program: Select a plane, do some 2D drawing, then cut or extrude, select some edges, apply chamfer/fillets, and then repeat. > > The devil is in the details and in understanding what is going on under the hood (bonnet). For example, I could extrude a regular polygon, but I could not extrude a polyline (created from a list of points). After much searching, I found that one must use the .close() function between the polyline and the extrude. > > Of course, your mileage may vary and if you are transitioning from a drawing program you may find it easier than this old guy. > > End of report. > > On Mon, Jan 15, 2024 at 7:44 AM jon via Discuss <discuss@lists.openscad.org> wrote: > >> Great. Not so privately. Sigh. >> >> When Reply does not mean Reply. >> >> On 1/15/2024 8:39 AM, jon via Discuss wrote: >> >>> Privately. >>> >>> Would your code eventually be able to replace MeshCAM? >>> >>> Jon >>> >>> On 1/14/2024 10:16 PM, William F. Adams via Discuss wrote: >>> >>>> I tried CadQuery a while back --- it was confusing because they were just transitioning from being a workbench in FreeCAD (which I was hopeful of) to being a stand-alone project (which I found off-putting and which lacked the synergy w/ a graphical program I was hoping for). >>>> >>>> Agree that there was a striking absence of example files/code, and the ones which I could find seemed arcane and confusing in how they were coded. >>>> >>>> The nascent effort to put Python into OpenSCAD has been a lot more approachable to me: >>>> >>>> http://pythonscad.org/ >>>> >>>> https://github.com/openscad/openscad/issues/4880 >>>> >>>> (but I don't know if anyone would want to do anything the way I do it: https://github.com/WillAdams/gcodepreview ) >>>> >>>> William >>>> >>>> -- >>>> Sphinx of black quartz, judge my vow. >>>> https://designinto3d.com/ >>>> >>>> _______________________________________________ >>>> OpenSCAD mailing list >>>> To unsubscribe send an email to >>>> discuss-leave@lists.openscad.org >>> >>> _______________________________________________ >>> OpenSCAD mailing list >>> To unsubscribe send an email to >>> discuss-leave@lists.openscad.org >> >> _______________________________________________ >> OpenSCAD mailing list >> To unsubscribe send an email to discuss-leave@lists.openscad.org