Autodesk Develops Own Kernel? 

by MARTYN DAY, editor, CADserver

(continued)

Ramifications

There are two aspects to this deal, technological and political.

On the technology front, Autodesk is now in control of its own kernel development, so it no longer relies on component technology that anyone can license, the development team can start to produce unique features to differentiate Inventor from the competition. In the cut-throat world of the mid-range modeling market, being out of sync with a release of ACIS could mean falling behind. It seems that all the technology moves here are to further develop Inventor. While other products will benefit, Inventor is the main focus of the MCAD team now. AutoCAD is being developed by the Building Industry/ AEC team and they have no need for advanced solid modeling but I'm sure the Autodesk ACIS modeler will at some point find its way into core AutoCAD. Here, if any additional features were added this might force a change in DWG, a situation, which Kross says, would not be intentional.

The ACIS modeling engine is about 14 years old, with 7 being a major re-write. However, by taking on this responsibility, Autodesk will have to, at some point, develop a new kernel from the ground up as technology doesn't last forever. But why reinvent the wheel? The reason the mid-range and mid-priced modelers existed was because the technology was available in affordable, component form. By taking on kernel development Autodesk has signed up to spend millions of dollars in R&D on its mechanical aspirations.

Politically, Autodesk has freed itself from continually paying and relying on a competitor, Dassault, for its kernel technology. SolidWorks, which, like Spatial, is owned by Dassault, is Inventor's main rival. So, despite emphasizing that that it isn't in the kernel business, Autodesk is now in the kernel development business. As proof of this, while summing up the benefits to Autodesk of the move, Kross pointed out that overnight its new ShapeManager technology will be the most widely use kernel in the market!

Conclusion

To get a reaction for the market I talked to Mike Payne, CEO of Spatial and CTO of Dassault Systems. Prior to his current positions, Payne was renowned in the industry as being key in the development of Pro Engineer and more recently SolidWorks, which uses EDS's Parasolid kernel. He told me, "It costs millions and millions and millions of dollars (or pounds and euros) to develop a solid modeler. Obviously Autodesk has chosen to increase its expenses, or not develop other capabilities that it cannot buy. If you were to ask me, would I do the same thing as Autodesk if I were in their shoes? My answer is that I was in their shoes and did not, and I was right!"

Payne continues, "It makes no sense to make things that you can buy. In geometric modeling, both ACIS and Parasolid have demonstrated that they can deliver a product that successful applications can use with no excuses. We have an application called SolidWorks which uses a kernel supplied by EDS (who has products that compete with CATIA). No one would argue that there are any excuses in SolidWorks. It beats Pro/E in benchmarks, ImpactXoft has made a very interesting product based on ACIS (IX Design), and so on."

There's no doubt about it, this is a major decision by Autodesk. Financially this is a huge undertaking but also indicates just how seriously Autodesk takes its place in the MCAD market. Not satisfied with relying on Spatial's developments, Autodesk wants to develop its offerings more rapidly than the previous arrangement would allow. It moves them away from reliance on Dassault and gives them a industry proven, mature product to build upon. The development will certainly be one to watch, for if Autodesk gets it wrong the repercussions will be extremely serious.

previous page

Related Articles

.