"Intellectuals solve problems. Geniuses prevent them." -- Albert Einstein
Posted on Friday, September 29, 2006 9:16 AM

By Randall S. Newton
Editor-in-Chief

Bloomberg News is reporting this morning that the use of two different versions of CATIA 3D CAD software—versions that are incompatible at the file level—is largely to blame for significant delays on the Airbus A380 project.

“Airbus Vows Computers Will Speak Same Language After A380 Delay,” says the headline.

Software used to manage the design and manufacture of the 555-seat A380 at Airbus's Hamburg engineering center isn't fully compatible with that used at company headquarters in Toulouse, France, say current and former Airbus executives, including Charles Champion, who headed the A380 program until September. That's why hundreds of small changes to electrical wiring in the A380 snowballed into at least a year’s delay in delivering the world's biggest passenger aircraft and $2.5 billion in lost profit. Airbus Chief Executive Officer Christian Streiff may announce additional costs or longer delays for the A380 in coming days. The board of Airbus parent European Aeronautic, Defence & Space Co. meets today in Amsterdam.

The incompatible software products are CATIA V4 and CATIA V5, 3D CAD software from Dassault Systèmes. It is no secret that these two versions of CATIA are incompatible at the file format level, and it is nothing new for a CAD software company to update a file format in such a way that compatibility between versions breaks down. We know this well in AEC. Autodesk changes AutoCAD’s DWG format with every two or three annual releases. Bentley is still guiding customers through the transition between MicroStation V7 and V8 formats, even though MicroStation V8 first came out in 2001.

It would be easy to blame Dassault for this mess by criticizing their decision to change the file format in CATIA 5. Easy, but wrong. Doing so sends a really dumb message to software vendors: Don’t Innovate. How much productivity and new functionality would be lost from CATIA V5 if Dassault had stuck with the internal file format of the previous version? Quite a bit. Sometimes you just have to jump the chasm.

The real problem looks to be a lack of leadership at the top levels at Airbus. Somebody inside Airbus knew that engineers in Germany and Spain were using CATIA V4 while CATIA V5 was in use in the UK and France. Somebody made a decision not to push for a single compatible standard, or to at least invest in a translator system. As a result, Airbus has already announced earnings losses of $2.54 billion over the next three years, according to Bloomberg. I am sure the Airbus board of directors, meeting today in Amsterdam, would like to know Somebody’s name.  

What can CAD users in general learn from this? First and foremost, either commit 100% to the use of your software, or find a different CAD standard that gives you the comfort you seek. Don’t be half-hearted. Aggressively demand the best from your vendor and then aggressively work internally to make your purchase work. Allowing different versions to co-exist in a design environment to avoid personnel issues may be penny-wise, but it is also certainly pound-foolish. Whining about lousy software without a will to invest in a solution is pathetic.

The second lesson is to plan ahead. In the case of Airbus, if an analysis showed there was good reason to continue using two different version of CATIA, there should have been a similar decision—made and implemented in advance—to guide how designs from the two versions would come together. All indications at the moment are that none of this happened. I say this knowing that the design work on the A380 started before the release of CATIA V5. It didn't take a crystal ball to know Dassault's plans for the future of the product line.

The third lesson is to directly address cultural issues regarding the use of design software. Time and time again I have heard engineering managers and CAD directors say the most difficult part of implementing new software is the resistance among the old guard. Why should engineers in Germany and Spain have been allowed to continue using CATIA V4 when engineers in France and the UK were already on CATIA V5? It only makes sense if appeasement is a core corporate value.

I can’t help but think of Foster and Partners in the UK, famed both for its striking designs and its aggressive use of design technology. They work closely with their CAD vendor of choice, Bentley, but it certainly is not a case of Fosters being spoon-fed everything Bentley has to offer. They push Bentley for certain tools, and Bentley wisely listens. Passivity in such a relationship would be deadly for both parties. I can’t imagine that either Lord Norman Foster or Iain Godwin, Foster’s executive in charge of technology, would allow cultural issues to prevent them from implementing new software.

CAD vendors all share a passion to create software that makes its users more productive, more efficient, and—they hope—more creative by removing the drudgery from design and engineering. They get rich only if their customers are successful. The purchase and use of CAD software needs to be clearly understood as a strategic issue; top management in design and engineering firms need to make the tough decisions. If Airbus loses $2.54 billion because senior management allowed disagreement in the user ranks about what CAD software should be in use to sway a business decision, then it deserves its losses.

 

Feedback

# re: Lessons For All CAD Users From the Airbus CATIA Debacle

9/30/2006 4:32 AM by CME
Perfect; hit with their own ball of #@?!

# re: Lessons For All CAD Users From the Airbus CATIA Debacle

10/5/2006 5:39 AM by Scott
According to Bloomberg, the delays were caused by incompatible Catia4/5 models. From that, you conjecture that the reason "looks to be" lack of leadership at AB, and speculate that different versions were allowed to "appease" the "old guard" engineers resistent to change. Was that from a crystal ball? Tea leaves? Or just a vivid imagination?

Meanwhile, you let Dessault completey off the hook, saying that providing compatibility between releases would stifle innovation. Bull! No, I don't expect releases to be completely compatible at the file level. Yet there is no reason why Dessault couldn't provide more conversion cross-compatibility -- which would easily have solved this problem for AB, without restricting innovation in the new version.

Second, have you ever tried to manage the deployment of a Catia5 upgrade over several thousand workstations, in the middle of critical design efforts? It isn't pretty. Dessault could spend a little less time advertising and more time making the upgrade process less disruptive and easier to deploy.

I am not here to defend AB (I work for the competition) nor to villify Dessault. But your conclusions that the fault is AB management's attempt to appease the "old guard", or that Dessault couldn't provide better version transition features, is pure fantasy -- and flat wrong.

# re: Lessons For All CAD Users From the Airbus CATIA Debacle

10/5/2006 7:46 AM by Barb
While Airbus leadership shouldn't be let off the hook, nor should that of Dassault. Dassault's "take-it-or-leave-it" arrogance about their product and processes is legendary. It's the attitude of a monopoly: We make it; you buy it; our job is done; you (the customer) figure out how to make it work and keep it working. Backward compatibility between releases? Not our problem. Minimizing downtime in the customer environment? Customer, you're on your own.

Dassault customers could hope for market forces to change the scene in one of two ways:

Some worthy competitor offers a viable alternative to Dassault's product along with a solid migration plan, or;

Dassault figures out that, yes, software IS a service industry and there's revenue to be generated by actually offering services.



# re: Lessons For All CAD Users From the Airbus CATIA Debacle

10/5/2006 10:17 AM by aircraft engineer
Gee - Micro$shaft does it routinely (make versions that aren't compatible with older versions) - only they call it a FEATURE. The major problem here is that they didn't make the versions UNREADABLE from old-new. IF they don't play well together, they need to be separated so they WON'T SHOW TO EACH OTHER. Remember when JPL used metric for the Mars lander and Lockheed used "US costomary" units? Didn't show up until the 5th decimal place supposedly (or whatever) - gee, it crashed. The right hand not knowing what the left hand is doing. THAT'S WHY THEY MADE PROGRAM REVIEWS

# re: Lessons For All CAD Users From the Airbus CATIA Debacle

10/10/2006 8:13 AM by MDS
Do any of you realize that a large fighter jet program (hint: 3 letters, starts with a 'J') is currently being designed by a combination of V4 and V5? Successfully! Because of good teamwork.

# re: Lessons For All CAD Users!!!

10/11/2006 2:51 AM by Skymaster, CATIA 5 experienced user
Well, if they were more flexible, they would allow to work with Russian design-engineers, and such disaster would not happened . We are the best in the engineering, it is not joke. But when we are trying to find a job in Europe, there are always problems with job-permission.

# re: Lessons For All CAD Users From the Airbus CATIA Debacle

10/13/2006 3:01 PM by Steve Bedford
JXX is not finished yet therefore no yet a success

# re: Lessons For All CAD Users From the Airbus CATIA Debacle

10/17/2006 9:11 AM by A smiling ex CV AE
AB should have stuck with CADDS5!!

# Bad management to implement without knowing the consequences

10/17/2006 6:17 PM by CAD Design
First off - to those who don't like Dassault on principle - I understand your sentiment. As a business owner, I more than pay my share to them. But let's make no mistake; when it comes to companies of this size, nobody is without fault. Not Dassault, and certainly not Airbus. In my estimation, their size and respective corporate cultures cancel out all faults on one another from that perspective.

Now, to the real matter at hand. Anyone who has ever successfully implemented a new program of this scope (or even lesser) should know that it does not come easily, and requires a serous plan of action. It requires all operating groups, at some level, to understand what is involved, and form some sort of concensus, even if it be only a loose concensus. The decision to go forth with the data migration and/or PLM implementation seems to have been a shoddy bit of work, at best. As mentioned in the comments, it has been done before by other companies. Not without some bumps, and certainly not without some growing pains. But it HAS been done.

Any indication that there would be a hiccup of this colossal nature should have been realized long ago, before ever reaching a production stage. Any decision to go forward with an implementation based on faulty integration can only be blamed on one of two things: 1) a failure to test adequately, 2) a failure to pull the plug (early) on a bad platform. The blame for that lies squarely with the program decision makers at Airbus.

You cannot just push the start button, cross your fingers, and hope for everything to work out fine. For this reason, it is not enough to blame the business dealings of Dassault, no matter how sinister we may believe them to be. They are a large corporation. Their motives will always be suspect in the eyes of the end user. It doesn't mean that they single-handedly sank the A380.



Comments on this post are closed