The Inventor Fusion preview is now available. I am anxious to see this thing. Inventor Fusion “Unites direct and parametric workflows”. This is one of the title statements found on the Autodesk Labs web site. I hate to burst the bubble but “direct” and “parametric” workflows have already been “United”. It first happened about 10 years ago when Hewlett Packard embedded the D-Cubed 3D DCM parametric solver into their SolidDesigner (Now PTC CoCreate) product. Siemens has followed by using the same D-Cubed solver inside their Synchronous Technology. There are certainly differences in how these technologies are “united” and implemented in these systems, but the “unite” has already happened.
The first bullet under this title goes something like this: “With the introduction of Inventor Fusion technology, both direct, history-free and parametric, history-based workflows are united”. There could be some confusion in the usage of terms here, but again direct and parametric technologies have already been united. This statement however takes it a bit farther. It claims that history-based and history-free have also been united. If they are referring to a synchronous uniting of this technology, they are wrong. Either a system is recording the modeling steps, or it is not. It is a “1” or a “0”. And if it didn’t record it … assumptions must be made if a record is to be created after the fact.
It goes on to say, however that “Changes can update into the model's parametric feature history, ensuring critical design intent is maintained”. Now this could be cool, but I’m not sure. Are they simply referring to capturing the direct edit into the history tree, as any other (including Inventor) history-based system does with direct edits (nothing new)? -OR- Could they be referring to creating or modifying sketches, features, parameters, and structure of a history tree based on the B-Rep geometry of a history-free model? This however would be nothing new as well. Many companies, universities and creative individuals have been working for many years on this type of technology. It involves analyzing unintelligent 3D B-Rep geometry in an attempt to recognize characteristics in the model such that a fully functional history tree can be created (or perhaps edited) including sketches, features, parameters and structure/relationships. If successful the resulting history-based model could be edited using the standard history-based editing techniques. The system must make many assumptions to do it. Autodesk has been working on it with their Feature Finder, maybe they have made some major breakthroughs. SolidWorks has been working on it with their FeatureWorks technology. PTC has similar technology in their Feature Recognition Tool (FRT). All of this technology attempts to make sense/intelligence of a non-intelligent B-Rep model – the kind that comes from direct modeling. None of it works very well.
There could be another option I guess. Perhaps Inventor Fusion is tagging the B-Rep history-free model with information about a specific direct edit such that the information can be accessed back in the history-based environment to drive identical changes to the history-based model through the manipulation of a sketch, parameter or structure. This could be real messy in many ways, and would only work if the history tree is structured in such a way that the edit would actually work and not attempt to invalidate the model. It would also mean that inventor Fusion would be an "editing" tool only, not a modeling tool.
And then it is interesting to consider how you would rationalize having a history-based model and a history-free model of the same part. Which one is the master document? How would the PDM system handle it?
In a history-free system the 3D model is the master. In a history-based system the history-tree is the master and the 3D model is just a result. Two very fundamentally different technologies.
Well, speculation is fun but I wonder if Autodesk would let me download the Inventor Fusion Preview. Since they have a link to my blog on the inventorfusion.com site, I think they should. We’ll see. I do like some of the interaction and use model I see in the videos, but I am most interested in the round trip between history-free and history-based, if such a thing exists. As you can tell, I am a bit skeptical.