What are the Big Unanswered Questions in CAD?

We’ve heard people say for years that CAD is a mature solution, and we’ve done most of what there is to do with this type of tool. So the implication is that the interesting conversation should move on to other areas that can be utilized in engineering, such as AI, simulation, generative design, IT delivery methods, consumer 3D print. Or is the future of CAD just about interface?

Maybe the role of the product designer is going to be automated? Replaced by artificially intelligent server banks? Relegated to maintaining the servers and power infrastructure. “You are belong to us”. The machines will learn language from the internet. Uggg. I dunno. So much of human experience is cyclical, that I don’t think it would be unreasonable for us to have a resurgence of the agrarian society.

Look at the human growth cycle. We grow bigger and stronger for the first 20 years or so, and then we keep growing, but that growth weakens us. Then we actually start shrinking before we pass on. Is it possible that we design organizations that follow this familiar pattern? Do our organizations keep growing until they begin collapsing under their own weight? Humans are obsessed with growth. Everything must grow! Is this natural or just pushed by financial concerns? Population growth is straining the planet, but the big solution is electric cars? Really? Another typical symptom of old age is apparently blindness.

Can CAD avoid bloating until it stops functioning from the non-functional inertia? Or will it simply be dissolved indistinguishably into the rest of the product design toolset?

Here are some things that might make it into mainstream CAD.

1. History is History. What’s Next?

I really believe that technology in this century has outgrown the need for incremental build of geometrical models. There is probably another method out there for building and editing product geometry. I think the whole process needs to become less technical (less about the tools) and more related to the end product. Functional Modeling, perhaps, where we specify what the product does, how it is used, and AI “designs” the product.

One way this could work would be a “body order of operations” which would still be kind of history-based, but based on boolean body operations. “Add this then subtract this”, not for every feature, but maybe for macro features, or several features combined.

Another way is obviously the Synchronous Technology method. It’s direct edit with brains. Intelligence in the software/interface rather than in the rebuild method for the features. It offers a lot of flexibility, and removes parent/child tangles and failures.

Or… Functional Modeling, anyone? Instead of building everything from geometrical sketches, you use functional building blocks, like hinge, thin wall, pin/hole, snap/catch. You start to see some of these kinds of features in CAD now, but if this takes over, designs would be faster and more generic.

This could take the form where the human establishes the location for the interfaces between parts or systems, and establishes the requirements, like “can be dropped from 2 meters without collapsing on itself”, and the AI figures it out from there. 10 hour work week.

2. Should all design be more… modular?

Do we really need CAD tools that are flexible enough to model literally anything, or do we just need a tool that through AI can link technical requirements to materials and thus to manufacturing processes, and  can create geometry for just a couple dozen functional behaviors, such as slide, hinge, clamp, pivot, human interface, etc? This would greatly simplify design, and allow AI to be programmed that could handle it.

3. Combination of Data Types

In a recent interview, we heard Matt Sederberg talk about U-Splines, and how simulation data needs to be more closely related to design data to truly make them interchangeable. Not that this is low hanging fruit exactly, but it is one place where a lot of time can be taken out of the product development cycle. We spend a lot of time working on parallel models – a design model, and a simulation model. Always with this one-way conversion step between them. I think Matt has a bead on how to eliminate this bottle neck in the process.

We can extend that to include the mesh type data. Matt said in the interview,

there has been a strong trend in automotive design for several years to move away from traditional NURBS modeling and towards more flexible, form-finding technologies like subds, T-splines, and polygons. This cuts a significant amount of the modeling time during the initial form exploration stage. However, as the shape is refined to class A, it always requires being completely and tediously remodeled

 Mesh data is more important now than ever before in product development. We need tools that enable us to work back and forth between data types. We need the unification of CAD, subdivision, tessellation, STL, OBJ, 3dsMax, modo, mudbox, MRI, CT scan data, radar, you name it. If it makes a point cloud, it needs to be integrated with NURBS data somehow. This is already happening in places, but the tools are needed across the mainstream/mid level CAD/product development toolset.

4. Yeah, but… Cloud?

I think we all get by now that the Cloud is not a CAD innovation. It’s just an IT delivery method. There are some cool things happening in CAD in the cloud, but I’m not convinced that you need to be in the cloud to do them.

Anything that you can do in the cloud, you can also do locally, aside from possibly the scale, unless of course you have a huge CAD department. In the vein that “all things must centralize”, I get where the cultural drive for this is coming from, but as an individual guy who likes to develop products, I don’t think it benefits me.

They say “kids” (which I take to mean “under 30”) are the ones who really “get the cloud”. That may be true in some ways, I guess. They are the ones who have a hard time with the concept of ownership, and certainly the cloud challenges that concept. Possibly they don’t watch the news that much, because you can’t watch the news for a week without learning of some big security breach where your tightly guarded personal info, or a map of your home created by your robotic vacuum cleaner, or your web browsing habits collected by a spying browser, or any number of things has been stolen and sold to be used against you. Google and Amazon and Microsoft and the ever decreasing number of ever growing monopolistic concerns want us to believe this is ok. In 30 years, those of us who don’t believe it’s ok will be gone, and the kids whose birthright is directed advertising and corporatocracy will just succumb. Wall-E. There is nothing left for humans to do, because the robots do it all for us. We’ve got all these billions of people, but nothing for them to do except fix robots, and we’ve already got robots for that. We consume. But there no longer much value for us all to add to the system. Nations cease to exist, everything is run and owned by corporations.

Maybe I’ll just go back to farming, and mind my own business.

5 Replies to “What are the Big Unanswered Questions in CAD?”

  1. Wow…such pessimism…product design will always (imho) be driven by customer demand, therefore human interface will always be necessary to interpret the customer through iterative intercourse. This is not one robot talking to another robot…this is human interaction driven by, not only need, but curiosity…”hmmmm…I wonder what it would do if I did this”…and CAD, as a tool, will always be with us until such time as there is thought driven programs to do what we do now…work with the tools we are given…so, no, I don’t see us devolving into agrarians any time soon!

    1. Dave,
      I’ve heard enough executives say “customers don’t know what they want” that I don’t put a lot of stock in the myth of “customer driven” anything. I mean it happens, but it’s not a rule, at least not a 100% or even 90% rule. Product design is not a democratic process. It’s more about individual instinct and that some instincts are better than others. Democratic design would be a mess. Like design by committee, but worse.

      The pessimism accusation is something that has followed me around. Not because I’m a pessimist, I truly believe I’m not a pessimist. I’m extremely curious about life and ideas, too curious to be a true pessimist. Maybe I’m a bit of an opportunist, although I realize that has some negative connotations, I take that to mean I’m open to a lot of options. Sometimes you make your own opportunities, and sometimes opportunities are pushed in your path. Sometimes when I write, I let the stream-of-consciousness thing take over, and goes in a direction that comes from some connection, although it may be not really obvious at the time. There are a lot of robot movies out there. Star Wars, Star Trek (Data, Lore, Borg), Clash of the Titans (weird owl robot), Short Circuit, Robocop, Robots (animated kids movie), Austin Powers (Fembots), Total Recall, Elysium, Minority Report (spider bots), Sky Captain and the World of Tomorrow, Matrix, Transformers, Hitchhikers Guide to the Galaxy, Pacific Rim, etc… AI movies are also out there, like Bicentennial Man, Resident Evil, Transcendence, Tron, Chappie, Ex Machina, Artificial Intelligence, and the ai and robot movies intersect a lot. And there’s no way to look into the future like using a movie where they’ve already done it. The future of robots and AI, if you watch movies, is generally mixed, with a lot of darkness in there.

      Anyway, the “rise of the robots” can go in several different directions. The iRobot direction, where robots are generally good, but there’s a story about what happens when they go bad. Or there’s the “Terminator” direction, where robots are generally a frightening tool used by evil people. Or there’s the “Wall-E” direction. Where robots are good, but man is alternately evil, stupid, or clueless. Robots work to fix our problems, support our laziness, and generally rescue us from ourselves. In the end, robots make it possible to return to our roots, figuratively and literally. My stream-of-consciousness took this third stream. If you’ve never seen the movie, it’s severely quirky, and starts with the contrast of happy music with a distopian world not ruined by war, but by trash. Humans have gone on a space cruise for several hundred years while the robots clean up the mess. The entire planet appears to be owned by Buy-n-Large, a Sam’s Club sort of consumer paradise. The movie is an indictment of humanity’s laziness, and reliance on technology. In the end, a robot rediscovers plant life, which brings the BnL cruise ship back to earth, where the captain of the ship proudly declares to the children “You’re gonna grow all sorts of plants, vegetable plants, pizza plants…”

      That’s where all of this came from. It’s not that I think that’s where it’s gonna wind up, just that it’s one option. I do think that AI is going to play a role in design eventually. I do think that organic shapes from generative design are going to become more common place. Manufacturing methods are going to mean less for design in the future because somethings will be printed, or built with additive technology. Modular design is an option. It’s less interesting, but as we yield decision making control to AI, it seems a logical progression. Centralized control always ends badly. Just ask the movies.

  2. Call me old fashion but I think AI will never replace true Design and Engineering for some time. And even then I doubt it’ll completely supplant human ingenuity. and drive to invent.

    What worries me more, is NOT the CAD tools but the young kids coming up and their education to use it. The fact that you can design it on CAD or even have CAD help you design it, doesn’t mean it’s worth the effort to make it.

    I only pick on Solid Edge because I’m a long time user, but I suspect the marketplace suffers from the same shortsightedness.

    Here’s something we could all use every day… Truly Smart Holes based real world use, that includes the Material. For example I would love to see Solid Edge warn the user that placing a 6-32 or 3mm Tapped hole in Stainless Steel only needs to be 2X the diameter to insure proper strength, yet in plastic it should be 2.5X to 3X. My point is this would be an everyday tool… not once in a blue moon tool, that still requires enormous Engineering input to make good parts.

  3. CAD developed from mechanical drawings. Mechanical drawings came about from a need to shape the material “real world” to our needs in a consistent repeatable manner so that we could efficiently make the same thing over and over without tedious hand work. That moved us one step away from the materials we work with. Many shops that relied on mechanical drawings also relied on the prototype builders to “check” the drawings with reality. But reality was still totally free form.

    3D CAD took over from the prototype reality check and captured 80% maybe of what reality could be. Heaven help you if you were designing anything flexible or deformable or truly sculpted. 3D CAD is still challenged by F&S* curves. So designers were now 2 steps away from reality and swallowed the limitations. Every try modeling a dope and paper balsa airplane model? What exactly is the shape the paper takes after it is sprayed with water to tension it? How does one get that into a 3D model to do a CFD analysis of the wing?

    Then comes simulation and again we can capture maybe 60% of what is really going on. So we limit our world to what we can model and another layer of separation.

    Perhaps CAD will morph into a giant 3D RPG, we’ll all wear VR goggles and headphones and some guys with oxen outside of our dream world will still grow the 100% real food that we need while we sit strapped to our desks debating to NURB or not to NURB, that is the question. 🙂

    The big unanswered question is how do we get to 100% of reality without breaking the bank?

    *File and Sandpaper

Leave a Reply

%d bloggers like this: