Monday, July 20, 2009

Direct Modeling - Thinking Outside the Box

Most CAD users are very familiar with history-based parametric modeling. It is likely what they learned in school and actually may be the only 3D CAD technology they have ever used. As a matter of fact, most do not know that there is even an alternative. When the topic of 3D CAD comes up, it is almost always in the context of parametric history-based modeling. “3D CAD” IS parametric history-based modeling for most people.

There is now certainly more attention being given to history-free, direct modeling and as such, many people are now experiencing it for the first time. This first exposure can be a good experience or a bad experience depending on many different factors. Here are a few things to consider before getting too frustrated with it and putting it back on the shelf.

Creating models without a history tree is fundamentally different than creating models through a history tree. And don’t confuse direct modeling with direct editing. Direct editing is available in either technology and as such many CAD users may be familiar with the concepts. Direct, history-free modeling is much more than direct editing. To take full advantage of history-free modeling, it will require that you think outside the familiar history-based box.

For most 3D CAD users the common reaction to history-free direct modeling is that it seems to be too freeform and “unstructured.” A history-based CAD system, to a certain extent, guides the user through a structured modeling process. Many have developed modeling standards and best practices to ensure consistency in the process. A properly structured modeling process may yield a properly structured model. With a properly structured model, flexibility AND inflexibility can be provided where necessary. The investment in learning how to use the structured modeling process to get a structured model that works as intended can be rather high. For those that have made this investment, it may be difficult to accept this “unstructured” history-free environment.

Structured Modeling

For some, when stating that history-free modeling is “too unstructured” they are in fact referring to the modeling process. They find value in a system that, to some extent, guides or imposes a particular modeling process. They find it helpful in that there is only a few ways to get a particular result. Again, the process is guided and structured. They gain some sense of “safety” in this environment. With history-free modeling there is no concept of a structured modeling process. There is complete flexibility in how the model is created and edited. This “unstructured” modeling process may be considered too “unsafe” for them. Perhaps this “unsafe” environment might allow them to do something to the model that they should not do. However, it must be considered that a structured modeling process certainly does not guarantee “safety”. Recreating history-based models is all too common.

Structured Models

Closely related to the “structured modeling process” is the “structured model”. For others, “structured” refers to the structured model, a model that is embedded with design intent to the point that it closely represents the physical. The concept of the structured model (design intent) came into existence with the introduction of history-based modeling. “Capturing design intent” was the justification used for the high investment required in the structured modeling process of history-based CAD. Many have come to accept that a structured modeling process is the only way to develop well structured models and to yield models that behave as intended. In the early years of direct history-free modeling, embedding structure and intelligence into the model was not possible, so there is certainly some validity to these notions. Several history-free systems today still do not provide the capability to add structure and intelligence to the model. PTC CoCreate Modeling is one of the few that does provide a good set of tools for adding structure. Structured history-free models are a reality, but the process of adding this intelligence is not as inherent with the modeling process as it is with history-based CAD. With a history-free tool like CoCreate, you can add structure to the model anytime in the process, even after the model is fully developed and refined or even to an IGES or STEP model. With history-based CAD, the development of model structure starts the moment you begin the process of modeling – and it is not optional.

So if you are one that is concerned about an unstructured modeling process, be sure that you clearly understand the value of the structured process and its support of your design process. Be sure that the significant investment in this structured process provides real value. In many cases it does not, and that is why many have added direct, history-free modeling to their toolset. Mistakes can be made in both technologies and in either case your work must be validated. This validation can happen real-time with or without a structured modeling process.

If you are one that is concerned about the structured model, you actually have two choices now. Both technologies can provide similar results, but the means to the end is very different. In a situation where every part model needs to be well structured and changes are predictable and controlled, history-based modeling may be a good choice. In a situation where it is more important to get quick results, and the investment into the structured model is only needed occasionally, your best choice may be direct, history-free CAD.

Initial reactions to direct history-free modeling are usually based on a few key factors:

  • Previous CAD experience and individual productivity will certainly impact your initial reaction. However, these should not be significant factors when determining the viability of the technology. (Unless you work by yourself, and the resulting CAD data has no lifespan or future value.)
  • The type of work you do, including your process requirements, will also have a big impact on your initial reactions. Process and process requirements should always be key criteria in the analysis of technology.
  • The capabilities of the direct modeling tool may also have an impact on your initial reaction. There are many new and very immature examples of direct modeling out there. In some cases they are very poor representations of what direct modeling can and should be. Be sure to look at a few before making any final judgment.

Be careful in weighing your criteria (reactions) as you experience this new world of “explicit” “history-free” “direct” modeling. Take a step outside your comfort box, keep an open mind and focus on your product development process requirements.

Paul

No comments: