What about STEP or IGES export?

Any plans on exporting to standard parametric/solid cad formats? Came up
on the thingiverse list. Translating to the brl-cad lang or using their
file format might be an option too.

Elmo

On Feb 6, 2010, at 22:00 , Elmo wrote:

> Any plans on exporting to standard parametric/solid cad formats?  
> Came up
> on the thingiverse list. Translating to the brl-cad lang or using  
> their
> file format might be an option too.
>

No plans, but it would be an interesting feature to add.
You'll most likely lose (most of) the parametric features of the  
exported models (I don't think STEP or IGES can represent such  
parametrizations as OpenSCAD allows, but I'd love to be proven wrong  
here).
A partial parametric conversion could probably easiest be done using  
Open Cascade.

To my (very limited) knowledge, brl-cad is a bit limited in terms of  
what type of geometry you can bring into a CSG operation, but there  
might exist ways to get around this. Conversion could be tricky though.

~/= Marius

--
We are Elektropeople for a better living.

Do you have anything to contibute to this? Especially about the
capabilities of the formats?

Elmo

Marius Kintel wrote:

On Sun, Feb 7, 2010 at 10:14 AM, Elmo <[hidden email]> wrote:
> Do you have anything to contibute to this? Especially about the capabilities
> of the formats?

Elmo, sorry I haven't replied to Marius yet, but for the record I am,
in fact, on the openscad mailing list :-). I am simply slow.

> Marius Kintel wrote:
>> On Feb 6, 2010, at 22:00 , Elmo wrote:
>>> Any plans on exporting to standard parametric/solid cad formats?  Came up
>>> on the thingiverse list. Translating to the brl-cad lang or using  their
>>> file format might be an option too.
>>
>> No plans, but it would be an interesting feature to add.
>> You'll most likely lose (most of) the parametric features of the exported
>> models (I don't think STEP or IGES can represent such parametrizations as

I don't know about IGES. As for STEP, I was reading a paper recently:

Exchange of CAD Part Models Based on the Macro-Parametric Approach
authors: Guk-Heon Choi, Duhwan Mun and Soonhung Han

"It is not possible to exchange parametric information of CAD models
based on the current version of STEP. The
design intent can be lost during the STEP transfer of CAD models. The
Parametrics Group of ISO/TC184/SC4 has proposed
the SMCH schema, which includes constructs for exchange of parametric
information. This paper proposes a macro-
parametric approach that is intended to provide capabilities to
transfer parametric information including design intents. In
this approach, CAD models are exchanged in the form of macro files.
The macro file contains the history of user commands,
which are used in the modeling phase. To exchange CAD models using the
macro-parametric approach, the modeling
commands of several commercial CAD systems are analyzed. Those
commands are classified and a set of standard modeling
commands has been defined. Mapping relations between the standard
modeling commands and the native modeling
commands of commercial CAD systems are defined. The scope of the
current version is limited to parts modeling, not
assemblies."

I have never seen anything using SMCH, so I don't know how wide spread
that is. When I was examining the output of STEP-compatible programs,
I thought there was support for parametric data. For instance,

#1=DIRECTION('',(0.E0,0.E0,-1.E0));
#2=VECTOR('',#1,1.56E-2);
#3=CARTESIAN_POINT('',(0.E0,0.E0,0.E0));
#4=LINE('',#3,#2);
#5=DIRECTION('',(-1.E0,0.E0,0.E0));
#6=VECTOR('',#5,3.12E-2);
#7=CARTESIAN_POINT('',(0.E0,0.E0,-1.56E-2));
#8=LINE('',#7,#6);

On #8, you see a line referencing a point and a vector, previously
defined in the file. I really dislike how the units are not readily
apparent (-1.56E-2 *what*? meters? kilometers?). Anyway, I thought
that this looked like it is parametric. By that I mean that
influencing variables in certain areas of the file will influence the
size, geometry and shape of the other objects in the design.

At the moment I don't recall anything about constraint representation
(if there is any), but I also don't recall if that's available in
OpenSCAD.

>> OpenSCAD allows, but I'd love to be proven wrong here).
>> A partial parametric conversion could probably easiest be done using  Open
>> Cascade.

Maybe. If you guys get some python bindings to the OpenSCAD parser,
then I'll work on a pythonOCC script to call the OCC StepAPI routines.

>> To my (very limited) knowledge, brl-cad is a bit limited in terms of  what
>> type of geometry you can bring into a CSG operation, but there  might exist
>> ways to get around this. Conversion could be tricky though.

Well, if I was to spend time working on this, BRLCAD would be a
secondary target through IGES or STEP (it already has loaders/savers
for these formats). I'd focus more on OpenCASCADE, but this might be
personal bias on my part shining through. I could be convinced to go
either way.

- Bryan
http://heybryan.org/
1 512 203 0507

On Feb 7, 2010, at 17:30 , Bryan Bishop wrote:
>
> Maybe. If you guys get some python bindings to the OpenSCAD parser,
> then I'll work on a pythonOCC script to call the OCC StepAPI routines.
>
What I interpret from this is that parametric modeling in STEP is  
either an immature or seldomly used feature or it would require a  
language or extension layered on top of it. Simple parametrization,  
like referencing earlier defined objects or variables, appears to be  
available. IGES is still unknown territory.

Exposing the OpenSCAD kernel (really the data structure of the CSG  
tree resulting from compiling the model) to  external programs or  
plugins would be a cool feature to have. This would force us to  
standardize the representation of said format, which is a good thing,  
albeit slightly early with such an immature piece of software as  
OpenSCAD, especially if we want to aim towards maintaining forward  
compatibility of such extensions.

However, what would help immensly in terms of evaluating the viability  
of representing parametric solid objects in STEP or IGES, would be to  
manually remodel a representative selection of the OpenSCAD examples  
in one or both of these formats to get an understanding of the  
inherent limitations of/differences between the OpenSCAD vs. STEP vs.  
IGES file formats/descriptions.

This was a bit fuzzy, but as I see it, OpenSCAD does something right  
but could do more right if we can find suitable wheels for those that  
have been attempted reinvented so far.

~/= Marius

--
We are Elektropeople for a better living.

On Sun, 07 Feb 2010 13:50 +0100, "Marius Kintel" <[hidden email]>
wrote:
Before finding OpenSCAD I started out to create my own CSG IGES files
before eventually being told that the CSG stuff in IGES is all vapor.
However, I was under the impression that the stuff that is implemented
in IGES-supported software so-call BREPs is at lease analogous to what
STL has.

Britton