CAD Trends | MCAD Advances

What's next for MCAD? The quest for intuitive software

Mark Huxley

Much confusion is circulating in the MCAD marketplace about PLM (product lifecycle management). What does it mean, and how is it changing product development? Though CAD remains a keystone of PLM, both are changing rapidly. Our design processes evolve continually, and challenges to "the way things have always been done" abound. They range from wholesale changes in overall business strategies to critical thinking about reducing the number of mouse clicks needed to accomplish specific tasks.

Innovation and development follow several patterns of varying complexity, regardless of what industry you look at. Incremental advances take place, following Moore's Law, and those are interspersed with major jumps in technology (that is, totally new ways of achieving similar goals). If you believe that few to no truly significant developments have taken place recently, you are not alone. It's been fifteen years since solid modeling made its commercial debut, and more than six years since virtual reality caught our eye. Though many great improvements have taken place in that time, engineers and designers still incessantly click, drag, and bang out products on our keyboards. Meanwhile, we continue to wait for "the next big thing."

Vendor Perspective

Of course, we are all to blame for the apparent lack of radically new technology. Any of us could dream up a viable permutation at any moment. Rest assured that many vendors are working toward that goal. Marcelo Lemos of Dassault Systemes notes that, "For the first time last year, the market did not grow globally." He believes that customers still are not aware of all the benefits PLM can offer. Among other things, PLM has the potential to "minimize new product development cycles, reduce time to market, and lower development costs. The market will soon understand the feasibility and return on investment of knowledge-based PLM," he says.

David Primrose from EDS PLM Solutions points out that, through the use of KBE (knowledge-based engineering), models are becoming smarter. Rather than simply claiming space and defining what the geometry looks like, they interact with other components and assemblies to help define why they look that way. Increasingly impressive implementations of KBE are continually coming to fruition.

Unfortunately, KBE tools are available only to a relatively small niche of users, and they ultimately fulfill the needs of an even smaller group. Part of the reason for this is that most users aren't granted the time in their schedules to exploit KBE.

Dave Primrose says that EDS PLM Solutions is working to expand model intelligence by providing "support for design, development, analysis, and manufacture of electromechanical products."

In addition to 3D topology, the model file "needs to contain additional knowledge about its functional requirements, mechanical behavior, manufacturing considerations, and, increasingly, the software needed to control it," he says.

By incorporating all of this data, you can achieve full product lifecycle support from concept to obsolescence. Interest in KBE should grow as more examples and metrics become available (figures 1-3).

[FIGURES 1-3 OMITTED]

Joe Gavaghan from PTC says, "Product designers and engineers need to focus on delivering great products without having to stop and learn software." Once users adapt to Pro/ENGINEER Wildfire's new interface, "the distinction between CAD, collaboration, and data management is so blurred that one moment, an engineer may be in the midst of a design task, the next, he's using the Pro/E Wildfire embedded browser to access PDM vaults and project workspaces located across the ocean."

Development of "The Next Big Thing"

MCAD development has been sluggish over the past year. Certainly, the major software releases from each vendor have substantial merit, but nothing screams, "You must buy this now!" Several major vendors recently made substantial investments in revamping their user interfaces to make them more intuitive, less obtrusive, and easier to learn. Taking this as a cue, perhaps the next quantum leap in our technology will be a somewhat revolutionary interface. The medical industry, and we as its end customers, are already reaping many benefits that MCAD vendors are still striving for (figure 4, p. 14).

[FIGURE 4 OMITTED]

More than seven years ago, I witnessed an extremely effective VRML demonstration created using SmartScene software from Multigen (now Multigen-Paradigm; www.multi gen.com). The presenter, tasked with landscaping a front yard, dropped a picket fence into place (I didn't say high-quality landscaping). The fence was simply stretched to length and snapped in place to match the edge of the lawn. With two flicks of his hand, the presenter, like a maestro, made the entire "assembly" appear and grow, with the number of pickets adjusting accordingly.

Certainly if any of the MCAD vendors make it that simple, many customers will be calling on them instead of vice versa.

Automation reminiscent of this exists today in MCAD through KBE, but even the current improved interfaces make it challenging to readily accomplish even this simple example.

Industrial design applications such as Maya and Studio Tools from Alias (formerly Alias|Wavefront) have far more accommodating user interfaces for manipulating points, curves, and surfaces than do MCAD applications (figure 5). Though control of all six degrees of freedom is not readily natural to a new user, the controls are always within easy reach. Still, users struggle until they master idiosyncrasies such as using the keyboard and mouse combinations.

[FIGURE 5 OMITTED]

Virtual environments are today's version of virtual reality. Using relatively low-precision sensors, a computer can detect your body's movements. Tracking these movements in combination with a Cyberglove (a sophisticated glove capable of translating about 90% of sign language) or simple switches to change tools, you can perform actions in mid-air. Stereoscopes can then project what you've done in 3D, and magically your work is alive in front of you, ready to be modified.

Computer science doctoral student Steven Schkolne and Dr. Peter Schroder from the California Institute of Technology conducted a study comparing Maya and a CAVE (Cave Automatic Virtual Environment; figure 6). The subjects assembled and manipulated 3D mannequin models.

[FIGURE 6 OMITTED]

One user commented that the CAVE "was much more user friendly. I instantaneously had infinite directions of movement and infinite rotational axes, and it was all done with simple motions of my hands. It was very intuitive how to use the program because it was almost exactly like putting together tangible objects. Visual determinations were a breeze because I could change the angle just by moving my head, and at the same time I could manipulate a shape." Do any MCAD companies you know of get reviews like this from first-time users?

What does the future hold?

Discussing attempts by vendors to create a single interface that is capable of performing a wide range of tasks (sketching, building, assembling, measuring, etc.), Mike Evans, an IT columnist, said, "Can the interface adapt to the task? Or, heretical thinking, do we need different tools?" Applying this train of thought to CAVEs, what if many 3D tools could be displayed in your hand (pencils, tape measures, hammers, and sandpaper would likely be popular requests)? Combining a virtual "dream toolbox" with CAD accuracy and a Cyberglove to create a system that at the same time guides and accomplishes what you are attempting to do could provide an even more immersive experience without prohibitive extra costs. Incorporating haptic feedback could make the process almost realistic.

Schkolne has discovered better code is needed for interacting with virtual models, not necessarily more accurate sensors. In his talks with several larger companies, he finds that they "tend to move more incrementally" rather than by leaps and bounds. He believes that "a killer application needs to appear, and then larger businesses will take hold of it."

Surely KBE will continue to proliferate, and its tighter integration into our CAD programs will foster more automated geometry creation. Perhaps we will eventually be able to create knurl patterns, involute gears, and Class A surfaces with ease and without large file size penalties.

We appreciate today's technologies, but yearn for the next big development that will be as profound an improvement as solid modeling was to t-squares and slide rules. If we are allowed to focus on creativity and problem solving instead of wrestling with our software, perhaps the field of engineering could entice a wider group of prospective students.

Acknowledgements

Mike Evans, Cambashi Ltd., www.cambashi.com

Daniel Keefe, Brown University

Joel Orr, Cyon Research, www.cyon research.com

Steven Schkolne, Caltech doctoral student

Steven Weisberg, industry analyst

Terry Wohlers, Wohlers Associates, Inc., www.wohlersassociates.com

Additional reading ...

For more information on research mentioned in this feature: www.cs.brown.edu/research/graphics/research/ www.cs.caltech.edu/csresearch/groups.html

In this article ...

Alias (Maya, StudioTools) www.alias.com

Dassault Systemes (CATIA) www.3ds.com

EDS PLM Solutions (Unigraphics NX) www.eds.com

Immersion Corp. (Cyberglove) www.immersion.com

PTC (Pro/ENGINEER Wildfire) www.ptc.com

Helping or Hurting? Top cad Interface Annoyance

Certainly layers, windows, and the like help us stay organized and manage data, but on the whole, reacting with them slows the creative process. Among the top user interface concerns with today's MCAD applications are:

* Assigning items to layer(s) with repeated duplicate clicks.

* Manually creating FEA or assembly animations frame by frame instead of simply exporting what is shown on the screen to an animated GIF file.

* Correcting flaws in imported geometry.

* Sketching tangent entities only to have them assume another, improper solution.

* Toying with blends and rounds that are not quite what you envisioned. Ninety percent or more of the feature is acceptable, but that last portion is a definite deal-breaker.

* Cursing when recording macros with all-or-nothing key sequences. Perhaps MCAD vendors could offer better graphical macro creation and editing tools (as in Photoshop) with plain English error messages complete with Continue and Stop options instead of "crashes."

* Resizing windows that pop up at sizes reminiscent of postage stamps. If you have a 190 LCD running at 128031024, why would a file navigation window default to 4903274? Is anyone really running an MCAD/PLM application at 6403480 screen resolution? How many versions of the software need to be built before it actually remembers window sizes, locations, last-used directories, and so forth. At the very least, you should have the option to save the current state so you don't have to change it every time you restart the application.

Mark Huxley is a mechanical engineering consultant who also serves on Cadalyst's editorial advisory board. E-mail him at This e-mail address is being protected from spam bots, you need JavaScript enabled to view it