Gone off the rails have I?

Well, even if you think I am crazy, I still think its a fair question to ponder. Even if only for a moment. Hear me out and you might even agree.

Three Webinars, Three Conclusions

Over the past three months here at ENGINEERING.com, I've hosted three separate webinars on design topics. Back on March 19th, we kicked things off with Clean Sheet Design where we looked at how CAD did or didn't support new design efforts. On April 16th, we looked at how CAD supports Design Reuse. And then finally, last week on May 21st, we investigated how CAD dealt with imported geometry.

In each of these scenarios, I looked a the pros and cons of Feature-History Modeling and Direct Modeling approaches. In each, I came to a conclusion on which ones really supported each of the scenarios. I hate to include a spoiler, but in each case I wasn't in favor of using Feature-History approaches.

So where, if anywhere, is Feature-History Modeling a better fit than Direct Modeling?

Well, that's the catch. I'm not sure there is.

Rules-Based Design and Design Intent

"That's it," you say. Feature-History Modeling is the best approach if you have rules-based design or if you want to drive intelligent changes that represent an engineer's design intent. And I'll admit, until recently, I would have been with you. But hear me out.

Direct Modeling approaches no longer just utilize push, pull and drag manipulations. Many CAD applications that leverage Direct Modeling now allow for parametric dimensions to be created and then used to drive the geometry changes. Furthermore, such systems also allow users to define constraints that are enforced on the model's geometry. And unlike Feature-History approaches, all that intelligence is captured at the model level, not within features.

So at worst case, the difference in the ability of driving programmatic changes between Feature-History Modeling and Direct Modeling is at worst a gray area and not simply black and white.

Another objection is probably the level of complexity that such programmatic approaches that can be captured. Feature-History approaches should be able to capture more intelligence and should be able to drive far more complex changes. However, the reality is that as more features are slapped into a model, the more tenuous its hold on stable modification. You are more likely to encounter a model failure as you get more complex.

Summary and Questions

Is this the death knell for Feature-History Modeling? Watch the webinars. Form an opinion. You tell me.

Obviously, there is a large contingent out there that will hold on to Feature-History Modeling until their careers end. It's incredibly widespread. Many are so highly trained that they are now considered the foremost experts in their organizations. Such a drastic change in the status quo is threatening.

But what about the next generation of engineers? Why would they want to carry on with such a painful approach? I mean seriously: who in their right mind would want to essentially plan and program their design models and then fix them when they break?

So sound off. Let me have it. Let me know if I'm crazy. But maybe, just maybe, I'm right.

Take care. Talk soon. And thanks for reading.