Imports and Exports

By Mike Hudspeth

As any designer knows, interoperability is a big issue poised to get bigger. If you’ve had any experience with collaboration you know that pairing the modeling software you use with another used at a different location can make or break your competitive edge. Say you need to send files to an outside design firm. Will your SolidWorks files translate into your vendor’s Ashlar Vellum? Can your tool maker in Asia accept your Pro/Engineer files? Will your painstakingly crafted models work with the art department’s rendering software? What about FEA? These are thorny questions indeed, born of extensive experience with all manner of applications.

Each year Kubotek USA, makers of KeyCreator modeling software (formerly CadKey), undertakes a survey of as many CAD users as possible on the subject of interoperability. This year’s survey turned up some interesting results.

As expected, 94 percent of those surveyed said they have to work with at least two different CAD file formats regularly. I can testify to the relevance of that number. Loaded on my computer at work are no less than three different 3D modeling or CAD programs, all of which must be accessed at various times to do what I do. And that doesn’t include other software I use to design products and their associated materials. When my company either buys or licenses a product, we have to incorporate yet other file formats into our database. We usually end up remodeling the parts in our chosen software (as do almost 50 percent of respondents); and if they come in a format we can work with we count ourselves lucky. The survey also indicated that 95 percent of those surveyed have issues — such as missing or corrupted data — with transferring files. That’s more than enough to say it’s a major consideration.

Figure 1: According to the survey, a sizable group of respondents rebuild 3D models from scratch to make changes to them.

With everyone going global, interoperability should be in the forefront of the corporate mind. Companies are increasingly taking it very seriously, especially in light of the rise in popularity of product lifecycle management (PLM) products. In 1999, the National Institute of Standards and Technology of the U.S. Department of Commerce stated that interoperability issues cost the automotive industry more than $1 billion that year.

Consider the more recent news from Airbus concerning the multi-billion-dollar delays on its newest monster airliner: The company is blaming it on software incompatibility — and it’s the same software! It’s just a difference in revision. (Ever sent a file to a vendor who uses the same software you do only to find out you upgraded but they didn’t? I can’t be the only one out there.)

Imagine what the dollar figure would be if you factored in the aerospace industry, construction, or even the washing machine industry. Now imagine the savings and increase in profits if the problem were solved.

A total of 2,869 CAD users responded to the Kubotek survey between May and July of 2006. They were given the option to participate when downloading a CAD model viewer on the Kubotek website. Here are some of the numbers:

Fifty percent of respondents worked at small businesses of fewer than 100 employees. The vast majority of them — 69 percent — were located in North America. The largest single group — 35 percent — worked in some kind of manufacturing industry.

Figure 2: The vast majority of the importing issues engineers have are with their translators. In second place is missing or corrupted data.

The survey asked questions related to the kinds of CAD files used. Fifty-seven percent said they used the same software as the files they work with to make necessary changes at least half the time. Sounds good right? Well, that also means that the other 43 percent have to deal with translation. The survey found that tooling vendors tend not to use the originating software of files sent to them. And the 45 percent that use the same software as the supplied files do so only 25 percent of the time. That should be a concern to all.

Any translation carries with its intrinsic risks. Not all software uses the same number of decimal places or standards of accuracy. If your model is accurate to three places, even 16, and your tooler’s is accurate to 32, what’s going to happen when your software says two edges meet and their software says it doesn’t? That’s how little bitty holes show up in models and people have to expend a great deal of effort to find and repair them.

The survey indicated that a large number of users rebuild their models from scratch (see Figure 1, above) because the imported data is not suitable for their purposes or the changes they need to make. There were many things that gave respondents trouble (see Figure 2, above), but the main issue was the translator. This might not be anything new to anyone who has ever tried to translate a file, but what is really interesting is that the survey found that the same problems persisted no matter what segment of industry the respondent worked in (see Figure 3, below). While the magnitude shifted somewhat, the rest of the problems mentioned in the survey generally ran the same across the board. Translators were the most troublesome, followed by missing or corrupt data, geometry errors, issues with the sender, large file sizes, and not enough data.

The survey asked respondents to indicate whether their software was parametric (46 percent) or non-parametric (29 percent). It’s interesting to note that a full 25 percent of respondents said they create or edit their drawings directly (see Figure 4, below). I think it’s significant that 54 percent of respondents don’t use parametrics at all. To hear some people talk, only those who use parametrics are doing real design work. But think about it, there is still a huge number of users out there who haven’t gone to 3D.

Figure 3: Interestingly, respondents had the same main issues regardless of the industry segment they belonged to.

The survey asked the users of parametric modelers how often the part’s feature tree needs major rework. Seventeen percent of the respondents said always, 28 percent said three-quarters of the time, and 21 percent said half the time. I can understand these numbers. Using parametrics can be tricky. Some users neither build their parts nor make changes well. I have seen people who won’t be told how to model and refuse to model with any consideration for those who will need to work with the data down the road. In a parametric modeler, how you do something makes a great difference. This is not so true with a non-parametric modeler.

Figure 4: Almost half of the survey’s respondents use a parametric modeling system.

With a non-parametric modeler it doesn’t matter how or when you do what you do. The geometry’s the thing. It doesn’t matter whether you have access to the original feature tree. You can always make changes. In fact, the survey indicated that 61 percent of the non-parametric designers indicated they had to rebuild their 3D models less than one-quarter of the time.

Of course, it’s important not to read more into the survey results than we ought. It signals neither the demise nor ultimate success of anything. It merely points out that there are some significant issues still troubling 3D design software and its users. Having used both parametric and non-parametric modelers for many years, I can honestly say that neither is head and shoulders above the other. Both have strengths and weaknesses. Both have their place. Both will be used for the foreseeable future. Interoperability is going to be an issue for some time to come.

Mike Hudspeth is a senior designer for a global medical company and has been using a wide range of CAD products for 20 years. He, his wife, two daughters, and their cats live outside of St. Louis, MO. Send your comments about this article through e-mail by clicking here. Please reference “Imports and Exports” in your message.

Kubotek USA
Malrborough, MA