Live vs Static Models

The popular surge of parametric history-based starting in the 1990s lulled a lot of engineers into a sense that a pinnacle had been achieved the end of the line in CAD or design philosophy development. It’s hard to argue for change. We have so much invested in what we have already done. Even if we were to agree that what we’ve done in the past isn’t perfect, making a change to how we do things is tough. There isn’t a huge precedent for making change, this industry (CAD tools) has only existed for 50 years, enough for only a couple generations of the tools in question. But I think this deserves another look.

The first important change was agreeing that computers could do CAD at all. Moving away from the drafting board was an important step. Drafting had to move away from the entrenched art form that for many was an end unto itself.

Here we have to take a step back and distinguish the difference between the three functions of representing manufactured goods. It’s tempting to see these as phases of product development, but that’s not what I’m looking at here, just the functions of representations of manufactured goods, whether CAD or non-CAD, electronic or otherwise.

• Concept – Just getting your initial idea out in a form that other people to see. Until we have intra-neural links between people, we have to have some sort of external medium to use for communicating ideas. Often CAD is too structured for this stage. This might be hand drawings, clay models, verbal descriptions, etc.
• Design – Still making ideas, but now we start adding some detail to lay out how something works, how it’s put together, scale, esthetics/ergo as required, depending on the product.
• Detail – Adding more detail, ensuring manufacturability, assemble-ability, function, mechanical standards, etc.
• Documentation – Conveying the details of manufacture to people who make the individual parts. Can also include pre-production marketing information, including photos of prototypes and computer renderings from the digital model.

To some extent, the items up the list are always getting tweaked as you go down the list. The more disciplined your process, the less of that goes on. To me, this says that you only need highly flexible models for the first two stages in this list. In Concept, you certainly need to adjust big ideas, size shape, overall function. In Design, you should be focusing less on the big picture, more on function, analysis and test.

The difference between live models and static models is that in live models, the “code” or the thing that makes changes possible is in the actual part and assembly data itself. One of the implications of this is that if you interpret those instructions differently, you can wind up with different geometry. I’m sure everyone has a story where this “feature” of history-based modeling has bitten them at some point.

My story was that I was doing some work for a large home-wares company, making a product with a lot of non-linear shape to it. I finished my work and sent it to an employee. The employee wasn’t the most sophisticated user, but they had just received a promotion. They opened my file, translated it out, and sent it for 3D printing.

Sounds standard, right? The problem came in that this employee had upgraded to a different version of the software, and when I sent my file, the new software “rebuilt” my live model using new algorithms. Which of course flipped a parameter on a spline. The whole model deformed (rather than simply failing…) without the employee noticing. The model was exported and printed in this new unintentional form. And I lost my contract, even though it wasn’t my fault.

This is what happens with live data. You trust it to stay the same as it was when you last opened it, but is it really the same? The software interprets the rebuild directions differently, everything is lost.

Static models are sometimes called “dumb” models by those who don’t know any better. The model just describes geometry, not directions to rebuild it, just the finished shape itself. We use static formats to transfer data between different CAD programs all the time. We don’t have a way to universally and reliably transfer live history-based data between CAD programs.

Even some Solidworks users sometimes prefer the static models over the unpredictability of live models. I have heard from users who will – in the middle of a project – round trip an assembly out then back in through STEP or Parasolid specifically to get rid of all or some portion of the live model they feel like they can’t control. I don’t really condone this sort of thing, it sounds more like a training issue than a problem with the process. I think as long as you’ve built in the live machine part of the data, you ought to be able to make use of it. But still I get it that there are times when you just want to shut it all off, or completely change directions with the design intent.

Plus, when you get toward the documentation phase of the project, the live data becomes superfluous. The changes must be complete at some point, and when you start documenting the detailed design, you need things to stop moving. This is when the static model is most valuable. At this point the model has to stop “thinking for itself” and just sit where you tell it to sit.

It turns out, though, that software has been built specifically to deal with situations where you want the live behavior sometimes, and the static behavior at other times. And you may have predictably overlooked it already. And no, this is not a paid advertisement.

Solid Edge has the live history-based legacy process built in. But it also has the ability to work with data in the other way – as if the software is the home of the intelligence instead of the CAD data. This enables you to work on a “dumb” model and apply changes to it, or just let it sit where you tell it to.

It also enables you to share your model with someone else, and still be able to work on it. Through the synchronous technology, Solid Edge can edit imported parasolid models, and even reassign features like patterns. With a recent enhancement, Solid Edge can read PMI through STEP AP 242 and edit the model using the imported PMI (dimensions). Jealous yet? At what cost are you clinging to the past with your old history-based modeler?

The question really isn’t Live Vs Static. You shouldn’t limit yourself by needing to choose one or the other. You really should be able to work with both kinds of data. Where the smarts is in the software, not in the CAD data. And you get to pick when and where to use it.