Change Manager: The Gateway between Parametric and Direct Modeling

Several years ago, the concept of a mechanical CAD program that let you go back and forth between parametric and direct modeling environments seemed like a fantasy. In February, Autodesk announced its development team was on this quest. The Holy Grail of the coveted bidirecional workflow was to be Inventor Fusion. In Release 1, or Technology Preview 1, Fusion offered a way into direct modeling, without a way back to parametric modeling (for more, read “Autodesk Inventor Fusion, Ready for Inspection,” June 24, 2009). But the most recent release, Technology Preview 2, fixed that.

Enter Change Manager, a plug-in for Autodesk Inventor 2010. The latest download of Fusion 2 contains two installation packets: a new version of Fusion and a plug-in for Inventor. (The later requires updates made available to subscription customers, so if you’re not on maintenance contract, it might not work.)

I’ve seen a demonstration of Fusion 2 and Change Manager, seemingly showing how you might make a series of direct edits on a parametric part in Fusion and convert them into parametric edits upon import (that is, when you open the part again in regular Inventor), but I wanted to put this workflow to test on my own.

This week, I did just that. You can see the result in the video clips below:

Change Manager is, indeed, capable of translating direct edits into parametric features. But there are some limitations. Whereas it can easily recognize and convert direct edits that have correlating parametric counterparts (for example, dragging a face forward in direct editing is the same as adding extrusion distance in parametric editing), it has some trouble digesting direct edits involving rounded corners and blended edges.

Overtime, Fusion’s direct-to-parametric translation algorithm is bound to improve, but I think it also illustrates the inherent differences in the two modeling methods. There are some direct edits that just don’t have parametric equivalents. In a direct editing program, you might, for instance, rotate an extruded feature built with a complex profile into an odd angle, but it’s highly unlikely such an edit can be translated into a series of parametric steps. I’m not suggesting that it’s the shortcoming of parametric modelers, merely acknowledging that parametric programs aren’t designed to accommodate such edits.

Eventually, Autodesk plans to incorporate the functions found in Fusion into standard issue Inventor, so parametric editing, direct editing, and Change Manager may all reside in the same environment.

If you enjoyed this post, make sure you subscribe to my RSS feed!

4 Responses to Change Manager: The Gateway between Parametric and Direct Modeling

  • Solid DNA says:

    Hello Kenneth

    Has i comment in our previous article, i don’t believe we are in front of a pure fusion between linear and non linear modeling.

    I stil think the fusion stand more for autodesk merging all their technologies than making the fusion between two modeling world.

    I know this fusion technology is young and it is not yet ready to be roll in production but…. Inventor use IPT and Fusion DWG. What happen to the DWG file of fusion once open in Inventor?

    Also what we need to mention at this point, Change manager seem to compare/evaluate two file body. Look at the top of the feature list in inventor and fusion you will see a body entry. For me this is the raw information of the kernel use by change manager to compare linear and non linear modeling

    So even if everything is in the same environement, from the CAD point of view, if we compare/evaluate/recongnize then translate do we still talk about fusion?

  • Kenneth says:

    Solid DNA: Thanks for the feedback! I just read your post. You raised some points about the difference between linear modeling (as represented in Autodesk Inventor) and nonlinear modeling (as represented in Autodesk Inventor Fusion). I’m duly impressed by your breakdown of how the geometry exchange might have taken place between Inventor and Fusion.

    But I feel your primary objection (“I don’t believe we are in front of a pure fusion between linear and non linear modeling”) is a semantic dispute.

    Are you objecting to the use of the term “fusion” because the way Change Manager handles the information exchange between Autodesk Inventor and Inventor Fusion doesn’t satisfy your definition of how a fused linear-nonlinear CAD program should work?

    By contrast to how Change Manager currently handles the workflow, how would you expect such a hybrid “fusion” CAD program to handle the mixture of linear and nonlinear edits?

  • Solid DNA says:

    Hi Kenneth

    Yes, I found easier to use linear and non-linear to talk about the new and old stuff in CAD. Without being considered this has a reference, I wrote a small lexis to help. I should find some time to make some correction/precision, but the global portrait is there.

    Has for the semantic we could probably have a long discussion about it. The point was more about the idea that is behind the key message.

    « …provides direct modeling for rapid design changes without limitations, and unites direct and parametric workflows within a single digital model created in Autodesk® Inventor® software… »

    From the semantic point of view, change manager does what it say « unite » has it make the bridge between the two to allow informations to travel.

    About the use of «Fusion», I do not think I can object; it is up to each CAD to market their technology the way they want.

    Once all of this will be under the same interface, from the point of view of the user this will be transparent. However, I am not in the god secret this is only a point of view from what I can observe; technically, I do not have the feeling of seeing direct editing. An interim body is manipulate inside Fusion. This body is then compare/evaluate/recognize to the original file. Finally, it is translate into an inventor files.

    Has you may suspect, my answer for the last question will be somehow bias. However, I would give it a try.

    Direct editing implies the direct manipulation of information, so first a single raw 3D database should be use and modifications should be live.

    Second modify the recipe of the model (reordering the feature tree) should not affect the morphology of the model. Ideally, feature tree could be deleting and the model will stay intact

    Derivate from the second one, the model should not rely on plane /2D sketch to stay in a healthy state. However, there is a paradox where 2D elements could be use has a reference elements to position 3D elements.

    So the bias section of my comment would be to behave like Synchronous Technology who is build around the Parasolid Kernel , modifications affect the 3D database directly in real time.

    You can manipulate the feature tree to reorder face collection (unfortunately, we do not have the capability to delete the features (face collection) tree except if export then import from a neutral file format). Sketches are there only to help create shape and face collection could exist without being attach to a plane/2D sketch.

    Geometric constraint can be recognize by the system or create by the user to make them persistent. Equations and conditions can be set from the variable table to set face behavior etc…

    Could probably mention more, but the general idea is there I think. Hope I did not exceed acceptable limit for a reply :-)

  • I just have to say, I enjoy reading your article. Maybe you could let me know how I can bookmark it ? Also just thought I would tell you I found your page through Bing.

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>