DressUp features (Fillet etc.) references

[tsadowski]

Hi,

I had some trouble with fillets on a complex part. A small change of the base part would wreak havoc with the edge assignments. The edge references appear to depend on topological naming. I did not yet go to the bottom of this. Topological names are very difficult to keep constant.

I am quite certain that commercial packages support topological names with geometry information. SolidWorks for instance displays missing edges. The information does not have to be persistent, it can be recreated during document loading.

Would it be useful and feasible to add geometry information to PropertyLinkSub? Would the right place be somewhere else? The topological naming together with the geometry information could return references with different levels of confidence:
- Topological name and geometry match
- Geometry matches exactly
- Geometry matches mostly (e.g. colinear straight edge with different length)
- Geometry can't be found.

If you think this is a useful approach, I would try to put some effort into it. All ideas for the best place where to implement this are appreciated.

Cheers,
Torsten

Bildschirmfoto von 2020-02-18 23-15-20.png (864.2 KiB) Viewed 1800 times

[vocx]

...Topological names are very difficult to keep constant.
...

Topological naming problem

Topological Naming, My Take

[tsadowski]

Yes, that is it. I also found an approach using TNaming from the OCC application framework:
https://github.com/ezzieyguywuf/freecadTopoTesting

I will give it a try.

Cheers,
Torsten

[vocx]

Yes, that is it. I also found an approach using TNaming from the OCC application framework:...

Topo Naming Solution in OCC 7.4.0 very important

Here is a topic about OCC's OCAF framework :
FreeCAD OCAF and reinventing the wheel

From the beginning on FreeCAD was not designed as a pure CAD system but as a flexible system providing frameworks for different purposes. If it was a pure CAD system then the OCAF framework might have been sufficient but since it's not and we already know that we hit a dead end with OCAF why should we do the same mistake twice?

Reinventing something is not per se a bad thing. At the beginning of a project you can evolve very fast if you use an existing framework but the problems will appear in long-term. If a framework is not designed flexible enough (and I have my doubts with OCAF) then at some point you hit a dead end and either you live with its limitations or you put huge effort into implementing something new -- this effort is a multiple times higher than the effort you had if you implemented your own solution from the beginning on.

[tsadowski]

I just wanted to re-iterate this by Werner.

Coin3D - FreeCAD Delegation to help move Coin3D forward

I find odd that in all those topics about the scene graph issue, the possibility of going with OCCT visualization is never considered. ...

In the early days of FreeCAD we were using the OCCT viewer and even its document framework OCAF. But at that time the viewer was extremely slow for larger objects and lacked of many features that we got with Coin3d for free. The only big advantage of the OCCT viewer over Coin3d was the available highlighting/selection mechanism but by writing our own nodes we could achieve something similar.

And the whole OCAF was a pain in the ass. We got it working for built-in object types but for all custom object types you had to register them with a very weird plugin mechanism which we never got working reliably. And documentation was either non-existent or very poor.

After spending 2-3 years on this we gave up and started to implement something that we have control over and that we understand.

I am not sure if OCAF is not flexible enough but I agree the plugin system is really weird. And to extend the OCAF to new object types it was (still is?) necessary to use the equally weird OCC build system.

I will have a go at realthunders solution and see how far topological naming solves the problem. I still think that geometry information would be useful to aid the user in resolving conflicts.

Cheers,
Torsten

[tsadowski]

Hi,

I did make some tests with the stock FreeCAD and with the topological naming development:
Topological Naming, My Take

I simply made a brick with a fillet on 2 corners:

Bildschirmfoto vom 2020-03-09 21-29-39.png (45.42 KiB) Viewed 1626 times

The first test was to insert an additional fillet. This fails with the stock FreeCAD but worked with the topological naming development:

Bildschirmfoto vom 2020-03-09 21-35-38.png (51.05 KiB) Viewed 1626 times

Topological naming starts as soon as the solid is created. If the sketch is altered, the topological naming fails. The picture shows one wrongly identified edge from the failed fillet:

Bildschirmfoto vom 2020-03-09 21-41-01.png (61.21 KiB) Viewed 1626 times

Topological naming is a big step forward but I think there are some advantages for additional geometry information. If the edge/face etc. does not exist anymore or if the topological naming changes, it is still possible to give the user a hint, where the selected edge/face had been.

Cheers,
Torsten

[galou_breizh]

I also already proposed something stronger that hints: filters for face, edge, etc. selection: https://forum.freecadweb.org/viewtopic. ... 81#p199781.

Hints could be also used in addition to filters. This could be used in the future for example with some filters based on artificial intelligence or based on some kind of heuristics.

The issue is not the lack of good ideas but the lack of workforce to implement them on their free time.

Gaël

[tsadowski]

Geometry search turned up in Realthunders FreeCAD. Now editing the base sketch works as well. I am really impressed.

Cheers,
Torsten