Kernel Wars

A graphics kernel is the unsung hero of CAD systems, managing the rendering and manipulation of graphical elements. This chapter explains the DNA of MCAD applications.

The origins of modern CAD technology come from the laboratories of computer science at the University of Cambridge. Let's explore the story.

Development of graphics kernels is hard, so not everyone could afford to build their own. This is the story of kernels that stayed proprietary and those that became the industry's shared infrastructure.

At the advent of Microsoft Windows stood Jim Meadlock of Intergraph and Jon Hirschtick of Winchester Design. Two teams, the same kernel, very different fates.

Sometimes vendors didn't see a technological shift, or were too late to fix their bugs. This is the story of three dead ends in the history of MCAD.

At the beginning there was only direct modeling. PTC changed the game with parametric modeling. Then CATIA V5 achieved the best of both worlds.

All this MCAD history is directly relevant to AI today — the problems it solved were analogous to those required for advancing artificial intelligence. The full mathematical story.

The battle for controlling surfaces — and ultimately the automotive and Hollywood special effects industries — is a tale of mathematical innovation and brilliant engineers.

Graphics APIs served as the critical bridge between surfacing algorithms and visual output, translating NURBS and Bézier surfaces into renderable forms that powered CAD and eventually AI.

Before wrapping up the Kernel Wars, a look at the hardware side — Silicon Graphics, graphics adapters, and how they became AI's unlikely beneficiary in the 21st century.

From General Motors' drafting tables to the geometry engine wars — the battles between the major CAD vendors that shaped the industry's consolidation.

The story of CAM is fundamentally about translation — converting perfect mathematical surfaces into the messy reality of cutting forces, tool deflection, and heat management.

From Alexander Hrennikoff's 1941 finite element paper to MSC Nastran, Ansys, and the modern simulation cloud — how CAE ate the physical world.

The boundaries between CAD, CAM, and CAE are dissolving as products become more complex. The future lies not in kernel dominance but in seamless integration.

Today's CAD landscape — from Parasolid and ACIS to nTop, Cascade, and implicit geometry — and what the Kernel Wars mean for the next generation of engineering software.