Inventor Coverage Provokes Big Questions

Recent coverage of Autodesk Inventor's 'functional design' touched of a flurry of responses. Despite Autodesk's claim to the contrary, users and employees of competing MCAD programs wrote in to say that their software also did 'functional design,' albeit by other names.

Examples gleaned from comments:

"Pro/E has had this functionality for years. It's called the Behavioral Modeling Extension (BMX)."

"I saw functional design with CATIA v5"

Also:

• SolidWorks has demonstrated a belt/chain assembly tool that -- like Inventor's example -- shows a lot of smarts. In addition SolidWorks Intelligent Feature Technology (SWIFT) also aims to free the user from the constraints of CAD.
• Solid Edge doesn't give its technology a name, but promises "intelligent, functionally realistic" models

In light of all the competing claims, how can the reader not be confused?

"Functional Design" is just one feature of Inventor but its coverage and ensuing discussion has stirred up broader issues. Vendor claims often generate counter claims. Who is telling the truth? Perhaps each believes they are telling the truth. Today's CAD products are complex and deciding between them is no small task. Which is the best for a buyer? Each vendor can drown you with descriptions of features and successes. Who will sort it out? The press? The analysts. They buyers themselves? 

Stay tuned as CAD Insider starts to examine these issues. For example:

• How can a buyer make an educated decision?
• What are the deciding factors when a user buys CAD software?
• What role does the industry press play?
• Has the industry press failed in educating the buyer? 
• Does the industry need CAD companies to compete head to head -- like a CAD Olympics?

With most of the MCAD software still to be bought, these are issues that are important to many. So jump in! Your input is most valuable.

Comments

[Jason]

SWIFT is indeed a buzzword. Its not about creating bad geometry however....more an automated fix for trying to create geometry that fails for whatever reason. It simply gives you an option let the software find a combination that works to give you what you want.

And all parametric modelers have feature failures...especially when things start getting complex.

The functional stuff in Inventor looks interesting....I can see the value in it as more an more engineering related info is getting stuffed into 3d modelers.

First wireframe, then surfaces, then solids, then parametric solids,.....progress and evolution..... can't wait to see where it will all be in 10 years.

[Andrew Anagnost]

It’s been very interesting reading all the interpretations of what Autodesk means when it says Functional Design, and I’d like to offer my view of some of these interpretations.

1.) KBE (Knowledge Based Engineering): KBE tools (essentially tools for capturing engineering knowledge as rules that drive the creation of models, drawings, and simulations) are available from many vendors – including Autodesk. Although KBE tools can be powerful when successfully used, they require expertise and effort in order to capture the rules and turn them into something useful for day to day engineering. They do not address the general problem that CAD tools are geometric modelers – not design systems. KBE and rules are an enabling technology, but they are what Autodesk would consider Functional Design. More on that later.

2.) Behavioral Modeling: This was PTC’s attempt to couple Mechanica to their geometric modeling. The focus is multi-parameter optimization. That is, once you have a “preliminary” model (which you have to create by the way) you then identify one or more parameters as objectives (weight, torque, etc.) and then identify degrees of freedom (that is, geometric dimensions) that are allowed to change. In theory, the system then automatically discovers the optimal value for those dimensions through a series of iterative simulations. This is definitely NOT Autodesk’s definition of Functional Design and requires quite a bit of expertise beyond the simplest of problems. In addition, PTC stopped investing in this capability many years ago.

3.) SWIFT: This capability is neither revolutionary nor innovative; it’s simply automated feature re-order tools that help users repair bad geometry – a Band-Aid for a bleeding paradigm, and similar to tools that have existed in many systems (including Inventor) for years. It’s our contention that users should never create bad geometry, and that is a key component of our vision for Functional Design. I always find it interesting that the same people that decry “marketing buzzwords” let SWIFT easily slip off their tongues. Is this bias or ignorance? Hopefully the latter since biased pundits only hurt the industry.

4.) Belt and Chain Wizards: Belts, chains, shafts, bolts, etc. are simple examples of things that can be generated “functionally”. Often these are used as easy to understand examples (like we did at our press event). In addition, certain vendors added a belt/chain wizards after Autodesk added its functional tools. However, real Functional Design is a paradigm shift – not geometry wizards bolted onto the product.

So where does this leave Autodesk’s vision for Functional Design? What we’re focused on is removing the burden of geometric modeling (parameters, constraints, etc.) from the day to day work of the engineer. We want schematic representations and functional parameters to be the main focus of designer or engineer while the geometric model is, whenever possible, generated for them. In addition, this model should always be a working digital prototype without the user having to modify it to become one. This is not an attempt to have the computer do engineering, but have the engineer be more effective at doing engineering. We envision many parameters and all geometric constraints, in the future, to be completely hidden from the user. This would be a powerful paradigm change and no one else is championing it.

Beyond that, we intend to do this in a way that is easy to understand and use so that is accessible to all designers and engineers, and that is indisputably unique.

Ultimately, the opinions that matter are those of the people that use our software everyday to do real design and engineering. I look forward to hearing those opinions in the years to come.

Andrew Anagnost
Autodesk, Inc.

[Adam]

Is anybody really surprised by this? Each of the CAD vendors are just trying to sell their product and try to differentiate themselves from their competitors. We now have far more access to reseach our decisions on what software to buy that is going to be the best fit for our use. We must remember that many companies will only ever use 10-15% of the functionality of any CAD software as that is all they need for their specific requirements. Many of the features that are often brought to the attention via the media are functions that many people will never use but ultimately help them in their decision to buy. My advice is to do your research and speak to colleagues in the same industry and get their opinions. The fact that people respond in such a way surprises me as you don't have to be Einstein to see that they are just trying to push their product. Geez, if you listened to everything that the CAD vendors say there are 5 different products that are all market leaders and the No.1 selling product in the world.

[James Shipley]

This is where Joe Greco provided great insight. I miss his articles.

[Jason]

Autodesk did this once before that I remember when Inventor was first released. "Adaptive modeling" was the buzzword and it was something other cad programs already did...usually its called "Top-Down Modeling".

Because most of us end users can not afford to spend the time/energy to sift through competing company's marketing claims, it is no wonder that buzzwords du jour (like Functional Design) get recycled year after year after year. Vendors rely on this fact, and occasionally mix things up with new language to describe the same thing. A great example is the sudden explosion in the use of "SOA" in marketing material. One major vendor is claiming breakthroughs in this area, while another is quietly pointing back to a very early (years old) release of their meteorologically named software. As far as functional design goes, I'd rather continue to have humans look after the form and function, and let my CAD/CAE system look after the fit.