name FreeCAD 0.91 instead of 0.20, means version after 0.19 could be 0.91

[Mark Szlazak]

The FreeCAD project needs a project leader with a clear vision of the project goals and of the road to get there [...] In fact it seems to be a sort of darwinian selelection where many developers make different experiments creating different versions of the same thing.

that's the Cathedral and the Bazaar story.

Also, and it's independent of the development model, is the question of the governance: in free/open source projects, there are essentially 2 possibilities:

• the BDFL (Benevolent Dictator For Life), represented by Linux Kernel/Linus Thorvarld, GNU/Richard Stalman, OpenBSD/Theo de Raadt, Python/Guido van Rossum
• the collegial organised organisation, best represented by Debian or Fedora, probably LibreOffice, with clearly written rules and identifiable development process, without any person being put forward.

I don't recognise any of them for FreeCAD: 3 names appear as copyright holders when launching from the shell:

Code: Select all

© Juergen Riegel, Werner Mayer, Yorik van Havre 2001-2018

Who are these people ? How did the become the copyright holders ? How are they making decisions ? Where is it documented ?

(probably through a fork)

going from the obsolete Coin3D to OpenSceneGraph would be a good justification for this fork. Funnily, I've made a Fligh Simulator with OpenSceneGraph (a fork of the Plib-based crrcsm). But I wouldn't know how to handle OCC-type objects with OSG. Or, for that matter, what OCC objects are.

Another possibility is to use the worker-owned cooperative model of management. Workers in our case would be all developers of FreeCAD. They democratically vote for a manager or management team and it’s planning or to remove an existing one. They rate the performance of the management team when that team is in place. Studies have shown these structures are on average superior to top-down enterprises, the size and type of the enterprise does not seem to matter. This FreeCAD enterprise is *not-for-profit*, so this cooperative structure would also bring in planning strictly based on user *needs* giving FreeCAD direction along with democracy in it's work space.

[wsteffe]

going from the obsolete Coin3D to OpenSceneGraph would be a good justification for this fork

But I wouldn't know how to handle OCC-type objects with OSG

I think that, for visualization purpose, it would be easier to use the V3d_Viewer class which is part of the OpenCascade library instead of relying on an external package like Coin3D or OpenSceneGraph. V3d_Viewer (see https://www.opencascade.com/doc/occt-7. ... iewer.html), together with AIS_InteractiveContext, is designed to work on OCC objects providing also the selection mechanism needed in any CAD environment.

I really do not understand why the FreeCAD developers wanted to reinvent the wheel (with an unecessary conversion from OCC objects to a different kind of representation) instead of using the visualization tools available in the Opencscade library, which is already linked to and extensively used by the FreeCAD code.

[Zolko]

I think that, for visualization purpose, it would be easier to use the V3d_Viewer class which is part of the OpenCascade library instead of relying on an external package like Coin3D or OpenSceneGraph [...] I really do not understand why the FreeCAD developers wanted to reinvent the wheel

I believe that some sort of scene-graph is necessary ... but only for assemblies. As long as one only wants to visualize a single 3D object, there is no need for that, raw OpenGL would do, and V3d_Viewer seems to be that, a search for "scene" didn't return any result. But for rendering a scene — meaning a root container with children and children's children — you need a scenegraph. And then, using an existing toolkit/library is the open-source way to go.

And Coin3D was probably used for compatibility reasons (with Inventor).

[ickby]

I really do not understand why the FreeCAD developers wanted to reinvent the wheel

You need to think 15 years or more back, when freecad started. From what I hear as reasoning is that occ visualization was extremely limited back than. And OCC being easier may be true for occ topological shapes, but the moment you go for general visualization, which is used in freecad quite often (Fem, Arch, Mesh...) a full blown scene graph toolkit is way better! just my 2 cents

[easyw-fc]

I don't recognise any of them for FreeCAD: 3 names appear as copyright holders when launching from the shell:

Code: Select all

© Juergen Riegel, Werner Mayer, Yorik van Havre 2001-2018

Who are these people ? How did the become the copyright holders ? How are they making decisions ? Where is it documented ?

IMO this is an other user spending time in whiling and blaming around and not using the time to get information on what he is talking about.

[wsteffe]

I believe that some sort of scene-graph is necessary

Actually it is not. In fact I have developed a code named "EmCAD" (see http://hierarchical-electromagnetics.com) with which I am regularly importing and visualizing assemblies that are exported (as step files) from Catia. The visualization is done using V3d_Viewer inside of a QWidget.
In this code I am relying as much as possible with the OCC infrastructure. The step file is translated using a STEPCAFControl_Reader into the OCAF data structure. This data structure was designed by OpenCascade developers to hold CAD models (parts or assemblies) composed of OCC shapes.

a search for "scene" didn't return any result

I think you should look for TPrsStd_AISPresentation which, probably, is the OCC equivalent of a "Coin3D scene".
I do not know very well Coin3D so I may be wrong with this analogy.
I can just tell you that in OCC a presentation can be assigned (as an attribute) to a label of the OCAF data structure.
The latter is a tree like structure where each label is associated with an object (in example e body/part/subassembly..) and may
have several attributes.

PS The EmCAD code that you may find following the link above is quite old and it doesn't work with the current version (7.3) of OCC.
Recently I have made several improvements and I have ported the code to Qt5 and to OCC 7.3 but I still have to update the site.

[wsteffe]

but the moment you go for general visualization, which is used in freecad quite often (Fem, Arch, Mesh...) a full blown scene graph toolkit is way better!

Ok, that may be a good point. In my code I am using V3d_Viewer only for the visualization of CAD models, while for FEM/meshes (which are also generated/used in my application) I prefer GMSH. I have no direct experience but I suspect that a general scene graph toolkit can not handle those very different data structures as well as it is done by the specialized tools (i.e. V3d_Viewer for CAD objects and GMSH for mesh objects).

[bernd]

@Zolko:
Jürgen is the father of FreeCAD, and as you may know father sometimes leave their childs ...

Werner is the brain of FreeCAD, without him FreeCAD would not be FreeCAD.

Yorik is the heart of FreeCAD.

For me these are very easy to understand positions.

[wsteffe]

Jürgen is the father of FreeCAD, and as you may know father sometimes leave their childs ...

Werner is the brain of FreeCAD, without him FreeCAD would not be FreeCAD.

Yorik is the heart of FreeCAD.

For sure those persons played/are playing a very important role in the starting and development of FreeCAD project.

Nevertheless I think that their contributions are not enough without a clear vision of which form should take the CAD architecture and how it should be developed to that point. And, honestly, I do not see anybody in that group of persons which is providing this direction.

I may understand that it is not easy to find a person which may bring fresh new ideas in the mature field of CAD software. In fact it is a rare event in the CAD industry. But then, in the absence of an original CAD architecture to develop, the developers could, at least, select a reference CAD architecture amonge the existing commercial products and then try to replicate it. This plan would be similar to what has been done by OIpenOffice/LibreOffice versus MS Word. An I think that it would still be a successfull story if it were possible to replicate a high end CAD system (like Catia/NX) in the open source world.

[bernd]

I do not know about other developer, but my aim is for sure not to copy another mechanical CAD system. BTW: I neither know catia nor solid works or any other mechanical cad. Furthermor I have never ever seen any mechanical cad live on a computer. Means even if I would I can not because I can not copy something I have never seen in my live.

I only know CAD from construction industry. For sure I would not like to copy them too. Why should I do that? It woud be much easier just to by a license than instead of writing a copy of one of them.