THEME ARTICLE: COMPUTER AIDED DESIGN I: FORMING

THE CAD SOFTWARE INDUSTRY

Industry-Developed CAD/CAM Software

David J. Kasik , retired, The Boeing Company, Sammamish, WA 98075 USA

Between the late 1950s and mid-1980s, some major automotive and aerospace
companies developed their own CAD/CAM programs. They wanted to leverage the
programs to replace manual drafting and design practices to improve productivity,
produce better designs, and accrue both technical and economic advantages.
Frequently, these internal systems featured proprietary 3-D surface modeling,
numerical control program generation, and engineering analysis capabilities. While
shipbuilding, architecture, petrochemical, and electronics companies also moved
from manual to CAD/CAM methods, they typically trailed the automotive and
aerospace companies.

S
tarting in the 1950s, automotive and aerospace other scientific and engineering programs in terms of
companies purchased significant amounts of performance, scale, and integration. The article then
computer hardware. A number of the companies provides an overview of some of the program func-
developed their own CAD/CAM programs to support tions needed and why most of them were not available
the complexities and scale of their product develop- on a commercial basis.
ment process. Not only were there few commercial This general picture is then followed by a more
CAD/CAM companies, but industrial companies also detailed discussion of CAD/CAM program examples
wanted to protect their intellectual property. from the two industries up through the mid-1980s. The
CAD/CAM programs supported drafting, 3-D sur- automotive industry is covered first, with detailed ex-
face modeling, numerical control (NC) programming, amples from General Motors (GM), Ford, and Renault/
and/or engineering analysis. Drafting let users pro- Citroën. A similar discussion of the aerospace industry
duce engineering drawings that documented designs follows, with a focus on Lockheed, Northrop, McDon-
and contained fabrication and assembly instructions. nell Douglas, Dassault Aviation, and Matra Datavision.
Some industrial companies, especially in the automo- The article ends with a discussion of why and how
tive and aerospace sectors, pushed the CAD envelope these companies led the way in high-performance,
into 3-D surface modeling because surfaces define the large-scale, complex 3-D surface and NC programs. By
external skins that drive automotive style and aero- contrast, early commercial CAD/CAM software com-
space aerodynamics. Using the geometry CAD pro- panies focused on building programs that produced
duced, CAM programs generated NC instructions for a engineering drawings. In some cases, industrial com-
new class of highly accurate machine tools. Finally, the panies purchased commercial programs to produce
geometry was essential input to complex engineering engineering drawings but relied on internal develop-
analysis programs (such as stress and aerodynamics). ment for surface design and NC programming.
This article begins with a general background and
overview of the drafting, engineering, and manufac- BACKGROUND
turing requirements in the automotive and aerospace
Like most forms of computing technology, CAD sys-
industries. It then describes some of the technical
tems have evolved significantly. Some advances have
differences between interactive CAD programs and
been driven by computing technology itself, such as
graphics processing units, personal computers, and
cloud computing. Other have been driven by brilliant
1058-6180 © 2024 IEEE
Digital Object Identifier 10.1109/MAHC.2024.3470710
people developing and improving algorithms (such as
Date of publication 10 October 2024; date of current version finite elements for 3D stress analysis and nonuniform
10 December 2024. rational b-splines). Importantly, industrial companies

October-December
Authorized 2024
licensed use limited Published
to: Amrita School Ofby the IEEE Computer
Engineering Society on IEEE
- Kollam. Downloaded Annals
January of the
01,2025 History of
at 17:07:01 UTCComputing 27apply.
from IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

realized that productivity improvements over manual Mainframe performance was especially necessary for
techniques were possible using interactive graphics. surface design and engineering analysis.
Automotive and aerospace companies have found
benefits in developing and using highly interactive, OVERVIEW
computer-graphics-based CAD/CAM programs since CAD/CAM programs produce two types of basic data.
the late 1950s. Computing helped automotive and First, both automotive and aerospace require 3-D ge-
aerospace companies move into the world of auto- ometry to define their products. Second, they require
mated milling and machining with NC systems (CAM), text and 2-D/3-D geometry as input for engineering
analyzing smooth surfaces to define aerodynamically analysis (CAE) and instructions (such as finish, toler-
efficient and aesthetically pleasing external surfaces ances, and dimensions) for manufacturing and assem-
[computer-aided engineering (CAE)], and producing bly. (Engineering analysis and CAE systems are beyond
engineering drawings (CAD). Starting in the 1980s, oth- the scope of this article.)
er industries, such as shipbuilding, architecture, petro- Because the documentation medium is something
chemical plants, and manufacturing/assembly plants, flat (on paper, a computer screen, or microfilm), com-
adopted CAD/CAM methods more slowly. panies have long used 2-D engineering drawing tech-
Production-level automotive and aerospace CAD/ niques to represent 3-D geometry. The drawings repre-
CAM programs had features commercial companies
sent 3-D objects as a collection of views (see Figure 1).
introduced later. Early commercial offerings, as doc-
Even if the CAD/CAM program defines geometry using
umented in David Weisberg’s excellent book [28], fo-
3-D coordinates, rendering techniques (such as shad-
cused on generating engineering drawings. A few early
ing, perspective, and dynamic rotation) are required
industrial systems, such as Lockheed’s CADAM sys-
to help the user understand the 3-D geometry on flat
tem, which became successful commercially [28, pp.
screens (see Figure 2).
13-1–13-7], addressed engineering drawing, while other
In short, CAD/CAM programs implement the nec-
companies (such as Boeing and Ford) used commer-
essary techniques to define, modify, and communicate
cial drafting systems.
the 2-D/3-D geometry and text needed to build com-
Systems developed by industrial companies in-
plex products.
cluded not only 2-D engineering drawings but also
My Boeing job gave me a broad view of both commer-
CAM, engineering analysis, and 3-D surface design. By
cial and industrial systems. As chief technical architect
contrast, early commercial systems concentrated on
for Boeing’s internally developed CAD system [16], I was
producing 2-D engineering drawings. Daniel Cardoso
invited to numerous presentations from vendors and
Llach’s article [9] in this issue discusses how the 1950s
competitors and became acquainted with their internal
CAM push to improve input definition for numerically
details. Boeing’s CAD/CAM research and development
controlled milling machines influenced some of the
work started in the late 1970s and ended in the late 1980s.
earliest CAD developments. Engineering analysis and
Academic systems are not included in this article
surface-definition capabilities are discussed later in
because the most significant production program devel-
this article and the article by Kasik et al. [17].
opment work was being done by commercial CAD soft-
Industrial and commercial systems differed for
ware companies and industrial companies. A number of
multiple reasons. First, CAD/CAM programs produce
academic research projects inspired CAD/CAM develop-
the complex, digital geometric representations and
ment nonetheless. The Massachusetts Institute of Tech-
annotations needed to design, analyze, manufacture,
nology [28, pp. 3-1–3-25] provided excellent late-1950s
and assemble products. Industrial companies wrote
and early-1960s results focused on interactively gener-
their own programs to protect their proprietary meth-
ating 2-D geometry [24], 3-D geometry [15], and NC ma-
ods. Second, industrial companies chose to directly
chines and programming [23]. Although there were some
hire mathematicians, engineers, and programmers to
academic contributions to solid modeling [27], [25], solids
build customized programs for 3-D surface design and
did not play a significant modeling role until Boeing used
engineering analysis. The programs reflected internal
CATIA V3 and CATIA V4 to define its 777 with solids [21].
company practices and did not need to be as general
When assessing automotive and aerospace CAD
as commercial offerings. A significant amount of the
programs, it is necessary to understand not only the
computer graphics techniques and mathematics im-
data but also the user community:
plemented in industrial CAD/CAM programs still exist
in today’s commercial offerings. Third, industrial com- ❯ Those with technical expertise in one or more
panies were able to purchase mainframe computing scientific, engineering, or manufacturing fields
capabilities that small companies could not afford. often have advanced degrees. Many have

28 licensed use limited

Authorized IEEE
to:Annals of the Of
Amrita School History of Computing
Engineering - Kollam. Downloaded on January 01,2025 at 17:07:01 UTC fromOctober-December 2024apply.
IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

FIGURE 1. Typical engineering drawing. (Source: https://pixabay.com/vectors/car-vehicle-draw-automobile-motot-34762/; used

with permission.)

tenance staff. Making the staff downstream more

productive was a prime motivator for the development of
early CAD programs. CAM programs started in the mid-
1950s because NC machines were driven by programs that
required part geometry and manufacturing instructions
to fabricate individual parts [2]. Generating the geometry
for NC programs led to the development of tools to make
defining the geometry easier. Engineering programs (such
as computational fluid dynamics and finite-element anal-
ysis) also relied on geometry that defined external surfac-
es for aerodynamic analysis, more detailed part forms for
structural analysis, and many others.

CAD/CAM PROGRAM
FIGURE 2. Annotated 3-D object. (Source: D. Kasik; used with CHARACTERISTICS
permission.) On a technical level, interactive CAD/CAM programs
differ from other scientific/engineering programs and
programming skills and are willing to write their transaction-oriented business systems because of the
own software to solve problems not addressed greater need for performance, scale, and integration.
to their satisfaction in commercial software. However, CAD/CAM programs and their users did not
❯ Specialists with interactive CAD systems build initially levy specific demands on processor speed,
2-D engineering drawings or 3-D models based network speed, memory size, and data storage capac-
on specifications from technical experts. ity. Instead, users tended to start with whatever tech-
nical facilities they could access and then later de-
The models and text guide the activities of down- manded more processor power, network bandwidth,
stream engineering, fabrication, assembly, and main- memory, and data storage.

October-December
Authorized 2024
licensed use limited to: Amrita School Of Engineering - Kollam. Downloaded on IEEE Annals
January of the
01,2025 History of
at 17:07:01 UTCComputing 29apply.
from IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

Performance Requirements Other forms of scientific computing generate or

CAD/CAM interactive drafting and design performance measure vast amounts of data, as in computational flu-
must be close to real time to allow users to manipulate id dynamics or astronomy. When a person produces a
geometry (either 2-D or 3-D) efficiently and comfort- CAD drawing or model, it is most often part of a larger
ably. Immediate response (measured as 0.5 seconds collection of parts, subassemblies, and assemblies that
or less) [28, pp. 13-1–13-7] for simple operations makes ultimately define the entire product. A complex product,
the CAD/CAM program feel like it is responding in real such as a commercial airplane or a building, requires
time. Simple operations include sketching a line and thousands of drawings, hundreds of thousands of unique
rotating, moving, and zooming 3-D models. parts, and millions of individual parts. A configuration
By contrast, many other scientific/engineering pro- management system rather than a CAD system defines
grams are heavily compute bound and can generally and controls interpart relationships and versions. (Con-
be run as batch programs. Even when able to be run figuration management systems are beyond the scope
interactively, users understand how complex the algo- of this article.) The system must be able to handle all of
rithms are and do not expect immediate results. Hence, the original data as versions evolve in addition to the
the necessity for real-time interaction is relaxed. data generated by CAE and CAM processes. All versions
Most interactive, transaction-oriented business are stored to document design decisions and evolution.
systems do not require near-real-time interactive The thousands of people involved in designing, an-
performance. They often feature form interfaces that alyzing, building, and maintaining a complex product
require a person to fill out multiple fields prior to pro- put significant stress on the supporting software and
cessing. Interaction must be fast enough to allow quick hardware. It is critical for the software to keep track
navigation from one text field to another. Once input is and organize all of the parts, drawings, analyses, and
completed, the user starts transactions processed by manufacturing plans. Tracking and organizing gen-
a reliable database system and expects some delay. erally require centralized computing resources (yes-
The real-time interaction aspect of CAD/CAM terday’s mainframes and today’s cloud). Tracking and
programs meant that their implementation differed organizing CAD data on centralized mainframes was
significantly from other types of online programs. Get- difficult enough. The problem got worse as personal
ting acceptable performance for CAD stressed inter- computers started having enough computing power
active devices, operating systems and programming and networking resources to move design to a distrib-
languages; data storage methods; and computing/ uted computing environment. Although tracking and
network hardware. organizing mainframe-based data were difficult, and
distributed work relied on detailed centralized track-
ing and organizing, making sure that a user was work-
Scale: Product Complexity and ing on the latest version added complexity.
Longevity Scale is also measured in calendar time. CAD pro-
The problem of scale stresses computer systems grams generate geometry and documentation data
across both size and time. Then, as computer per- that represent products that could be in use for de-
formance improves, users tend to push the limits by cades (such as cars and military aircraft) or more (such
attacking more complex problems, producing more as power generators). CAD/CAM programs tend to
design and simulation iterations, generating more nu- have a shorter half-life than the product definition data
merous and more detailed models, and so on. For ex- they produce. This puts significant stress on data com-
ample, when Boeing developed the 777 during the late patibility across vendors or across software versions
1980s and early 1990s, each airplane was represented from the same vendor. Different vendors’ implemen-
by a collection of models that contained about 300 tations of the same type of entity could all too easily
million polygons. The fourth version of the Dassault result in translation errors. New versions of a single
Systèmes CATIA CAD system (CATIA V4) was the pri- vendor’s product could also result in translation errors.
mary modeling tool. When the 787 started in 2004, the
geometric models developed using CATIA V5 required
more than 1 billion polygons. Although not necessarily Data Integration
as large in terms of absolute amounts of storage con- CAD/CAM program integration has different variations
sumed as business systems, geometry data are struc- [18]. Effective, active data integration allows different
turally complex (with both intrapart and interpart rela- programs to read and potentially write geometry data di-
tionships) and contain mostly floating-point values (for rectly without translation. For example, a finite-element
example, results of algorithms only come close to zero). analysis program requires geometry from which it builds

30 licensed use limited

Authorized IEEE
to:Annals of the Of
Amrita School History of Computing
Engineering - Kollam. Downloaded on January 01,2025 at 17:07:01 UTC fromOctober-December 2024apply.
IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

a mesh of elements. Many analysis programs (such as interactive performance. In addition, aerospace and
NASTRAN) have been in existence for decades and still automotive companies hired the mathematical and
do not have direct access to CAD geometry models. programming talent needed to build CAD/CAM pro-
Having full data integration across all CAD/CAM/ grams. The programs were tuned to internal corporate
CAE programs is a complex and fragile endeavor that drafting standards, manufacturing, and surface-mod-
remains a challenge for multiple reasons. Different eling techniques.
groups developed the programs and use different in- Commercial CAD software systems were able to
ternal representations that require translation. For ex- penetrate a few large companies in the early days. For
ample, CAD-generated geometry must be translated example, Boeing used them for 757 and 767 engineering
into the nodes and elements that finite-element codes drawings. However, it was more common for large aero-
can process. Similarly, different organizations use dif- space and automotive companies to develop their own
ferent brands of CAD/CAM/CAE programs that also systems to give themselves a competitive advantage
require translation. For example, Boeing used two dif- in surface modeling and NC programming. A few other
ferent CAD systems (Computervision for the 757 and large design and build companies in the shipbuilding,
Gerber for the 767) that forced the company to devel- architecture, industrial design, process plants, and fac-
op its own translators. tory design industries also developed or used early CAD
The translation of geometric data has proven to be systems, like Fluor [20] and GE [2], but they were the ex-
nearly as challenging as translating natural language. ceptions. Automotive and aerospace led the way, but,
Programs often have unique data entities, different in many cases, surface modeling and NC programming
algorithms for the same function, and even different were the prime focus. Engineering drawing programs
hardware floating-point representations. The differ- were developed primarily to save documentation labor.
ences mean that 100% accurate and precise transla- Both commercial software companies and indus-
tion among systems has yet to be realized. trial companies developed dozens of CAD/CAM pro-
grams that had significant functional overlap. As is
the case with other product classes, many competi-
INTERNAL AUTOMOTIVE AND tors initially emerged. However, market evolution saw
AEROSPACE PROGRAM the many gradually coalesce into a few large players.
DEVELOPMENT The CAD/CAM business was no different. Today, a few
Three factors drove CAD/CAM adoption in the aero- large players (Autodesk, Dassault Systèmes, Paramet-
space and automotive industries. First, companies ric Technology, and Siemens) acquired competitors or
observed that engineering drawing preparation was forced them into bankruptcy and now dominate the
time-consuming for both an initial release and subse- market [28, pp. 8-1–8-51, 13-1–13-7, and 16-1–16-48, and
quent modifications. Interactive graphics obviated the 19-1–19-38].
need for drafting tables, drafting tools, and erasers. The internal industrial programs stayed in produc-
Drafters could generate and modify engineering draw- tion through the mid- to late 1980s. Commercial soft-
ings more quickly. Large plotters let companies pro- ware companies started adding functions for 3-D solid
duce the engineering drawings on paper or mylar for and surface modeling and advanced NC programming.
certification agencies, such as the U.S. Federal Aviation The commercial companies were able to spread devel-
Authority, for approval. Second, engineering analysis opment and maintenance costs over multiple clients,
showed real promise in terms of virtually analyzing en- and industrial companies realized that commercial
gineering characteristics, such as aerodynamics, struc- systems could provide cost savings.
tural integrity, and weight. Accurate geometry, espe- The power of personal computers based on raster
cially external surface definitions, was required. Third, graphics devices also started matching and even ex-
NC machines gained popularity and required efficient ceeding minicomputer and workstation performance.
methods to define the geometry of individual parts. Personal computers, which were much cheaper and
Many automotive and aerospace companies devel- offered another cost-savings opportunity, contributed
oped their own programs. Unlike the early commercial to the demise of mainframe-based systems.
CAD/CAM companies, which often relied on minicom-
puters, automotive and aerospace companies had
enough mainframe resources to support a large user AUTOMOTIVE INDUSTRY
community and large amounts of data. A single main- Companies like General Motors [19] and Renault [4] had
frame could be upgraded to support tens and even strong research and development organizations and
hundreds of CAD/CAM users and provide acceptable started recognizing CAD’s benefits in the late 1950s.

October-December
Authorized 2024
licensed use limited to: Amrita School Of Engineering - Kollam. Downloaded on IEEE Annals
January of the
01,2025 History of
at 17:07:01 UTCComputing 31apply.
from IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

Not only did DAC-1 provide body-styling assistance,

but it also forced the IBM–GM team to develop an ear-
ly time-sharing strategy (the Trap Control System) in
1961. Time sharing itself was in its infancy in the 1960s
and generally supported alphanumeric character ter-
minals connected at low speeds (110 or 300 bits per
second). Supporting interactive performance required
higher bit rates and put more pressure on the operat-
ing system. Earlier computers, such as Whirlwind, that
supported graphics and light-pen interactive devices
were dedicated to a single user.
IBM and GM formed a joint development project to
FIGURE 3. Coordinate measuring machine. (Source: https:// develop a light-pen-driven interactive device to meet
www.foxvalleymetrology.com/products/metrology-systems/ GM’s DAC-1 requirements. Even the choice of program-
coordinate-measuring-machines/wenzel-r-series-horizontal- ming language was scrutinized. The Fortran compiler
proved to be too slow, so DAC-1 moved to NOMAD,
arm-coordinate-measuring-machines/wenzel-raplus-
a customized version of the University of Michigan’s
horizontal-arm-coordinate-measuring-machine/; used with
Michigan Algorithm Decoder compiler in 1961−1962 [19].
permission.)
Patrick Hanratty and Dean Beck worked on the CAM
systems that dealt with stamping the designs produced
Automotive surfaces are often defined using by DAC-1 between 1961 and 1964. Hanratty left GM in
full-scale clay models (see Figure 3). While manual- 1965 and went to a West Coast company, where he de-
ly “sculpting” new car body designs in clay was hard veloped his design software. He later took his work and
enough, manually entering computer-processable formed an independent company [2], [38, pp. 15-1–15-20].
surfaces to support design, engineering, and manu- DAC-1 was formally moved from the GMR to GM op-
facturing was even harder. Companies still use full-size erating division in 1967, but that did not stop GM CAD
coordinate measuring machines and numerical sur- system development. Two different surface-modeling
face-fitting algorithms to do so. packages, Fisher Body and CADANCE, appeared in the
The automotive industry especially cares about 1970s. Each ran on IBM 360/370 machines using the PL/I
how a vehicle looks to a potential buyer. Mathemati- programming language. Most users had access to IBM
cians like Steve Coons (Massachusetts Institute of 2250/3250 graphics terminals [17]. Some GM divisions
Technology, Syracuse, and Ford), Bill Gordon (GM and replaced between 50 and 100 IBM devices with DEC
Syracuse), and Pierre Bézier (Renault) solved complex GT40 vector graphics terminals hooked to a PDP 11/05.
computational geometry problems both as academics The 11/05 handled communication to and from the
and as employees. Their solutions became the basis mainframe. In the late 1970s, Fisher Body and CADANCE
for substantial improvements in surface modeling. The were merged into GM’s Corporate Graphics System
methods for defining surfaces, true 3-D objects, var- (CGS). The systems were based on GM proprietary sur-
ied from company to company. For example, General face geometry algorithms. Gordon surfaces [13] were
Motors has full-scale coordinate measuring machines particularly useful when fitting surfaces to scans of data
that capture height along the width and length of a collected from scanning clay auto body models.
full-scale clay model of a proposed automobile. Bill GM developed its own mainframe-based sol-
Gordon’s surface algorithms accounted for height dif- id-modeling system, GMSolid [7], in the early 1980s
ferences in the width and length measurements. that was eventually integrated into CGS. GMSolid used
both constructive (i.e., users used solid primitives, like
GM spheres, cylinders, and cones) and boundary represen-
GM started its CAD developments in the late 1950s tations (i.e., solid faces contained arbitrary surfaces).
[19]. The staff at GM Research (GMR) worked with IBM
to develop time-sharing and graphics capabilities that Ford
were responsive enough to support interactive design. Ford developed a minicomputer-based 3-D system for
The original computer used was an IBM 704 (upgraded multiple programs in the mid- to late 1970s. The Ford
to a 7090 and then a 7094) running a Fortran language Computer Graphics System [6] used a Lundy Hyper-
compiler. The program itself was called Design Aug- Graf refresh graphics terminal connected to a Con-
mented by Computers (DAC-1). trol Data 18-10 M minicomputer. Ford modified the

32 licensed use limited

Authorized IEEE
to:Annals of the Of
Amrita School History of Computing
Engineering - Kollam. Downloaded on January 01,2025 at 17:07:01 UTC fromOctober-December 2024apply.
IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

operating system to maximize performance. There

was one terminal per minicomputer.
The programs supported product design with the
Product Design Graphics System to define an auto
body using Coons [12] or Overhauser [8] surfaces. Other
functions included a printed circuit board design, plant
layout, finite-element modeling, and NC. Ford used com-
mercial CAD systems, such as Computervision and Ger-
ber IDS, for drafting and design functions throughout
its powertrain (engine, axle, transmission, and chassis).
Ford used different minicomputer brands and graphics
terminals for different programs. Computervision ran
on its own proprietary minicomputer and a Tektronix di-
rect-view storage tube; Gerber IDS ran on an HP 21MXE
and Tektronix terminal; and the printed circuit board de-
sign program ran on a Prime 400 minicomputer and Vec-
tor General refresh graphics terminals.
Even though Ford worked in a distributed minicom-
puter-based (rather than mainframe-based) environ-
ment, the company used centralized servers to store,
retrieve, and distribute its design files worldwide.

Renault and Citroën

Pierre Bézier popularized and implemented the curve
definitions for defining the smooth curves needed
for auto bodies [10] developed by Paul de Casteljau FIGURE 4. Sample NACA airfoils. (Source: Summary of Airfoil
(a Citroën employee) in 1959. Bézier developed the
Data, NACA Report 824, NACA, 1945; used with permission.)
nodes and control handles needed to represent and
interactively manipulate Bézier curves via interactive
graphics. He was responsible for the development of 3-D surface-definition technology that was consistent
Renault’s UNISURF system [5] for auto body and tool with their company surface-lofting practices and could
design. System development began in 1968 and went produce surfaces that could be modified relatively eas-
into production in 1975 on IBM 360 mainframes. ily, represented conics precisely, and exhibited C2 con-
Citroën developed two of its own systems (SPAC tinuity. (C2 means continuous in the second derivative,
and SADUSCA) in parallel with Renault [30]. The sys- an advantage when doing aerodynamic analysis.)
tems were also based on de Casteljau’s work and ran Aerospace companies tried to use automotive sur-
on IBM 360 and 370 series computers and IBM 2250 face-modeling methods, but they did not work partic-
graphics terminals. ularly well. Automobile companies care more about
the attractiveness of smooth surfaces, although aero-
dynamics has become more important as fuel efficien-
AEROSPACE INDUSTRY cy demands have increased. Aerospace is based on
In the aerospace CAD/CAM world, companies start- aerodynamic efficiency and demands C2 continuity for
ed defining aerodynamically friendly surfaces shortly analysis, which were not handled well with automotive
before humans took flight at Kitty Hawk. The National surfaces. The development and implementation of non-
Advisory Committee for Aeronautics (NACA) defined, uniform rational b-spline surfaces became and remain
tested, and published families of airfoils [5] in the late the preferred aerospace surface-modeling method [26].
1920s and 1930s. The idea was to assist aircraft devel-
opment by predefining the aerodynamic characteris- Lockheed
tics of wing cross sections (see Figure 4). Lockheed focused on producing engineering drawings
Aerospace engineers must design surfaces that and NC programming, not surface modeling. The goal
balance aerodynamic performance, structural integrity, was to be able to speed up both processes.
weight, manufacturability, fuel efficiency, and other pa- Lockheed California developed computer-aided
rameters. Industrial aerospace CAD systems adopted drafting software internally to run on IBM mainframes

October-December
Authorized 2024
licensed use limited to: Amrita School Of Engineering - Kollam. Downloaded on IEEE Annals
January of the
01,2025 History of
at 17:07:01 UTCComputing 33apply.
from IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

and 2250/3250 graphics terminals [28, pp. 13-1–13-7]. the McDonnell Aircraft Company. It was initially a 2-D
Development started in 1965 as “Project Design” to cre- design and drafting system that was extended to 3-D
ate 2-D engineering drawings quickly. Project Design in 1973, integrated with NC software in 1976 and sold
was rechristened as CADAM in 1972. An acceptable commercially beginning in 1977.
response time was deemed to be 0.5 seconds or less. McDonnell Douglas Automation (McAuto), the
CADAM operators were often judged by how fast they computer services business unit, purchased Uni-
seemed to be working even if little was actually hap- graphics (UG) from United Computing in 1976. McAuto
pening. Basically, CADAM has a lot of relatively short- rewrote and expanded the United Computing system
duration functions that made operators appear busy. software based on a license to Hanratty’s ADAM soft-
Project Design drawings were used to drive NC ware. The first production use of UG occurred at Mc-
machines as early as 1966. Use of the software spread Donnell Douglas in 1978. Bidirectional data exchange
quickly inside Lockheed, which established a separate between the two was not completed until 1981 even
business to sell CADAM in 1972. The new business though both were in production use.
started sending CADAM source code to others in 1974, The two systems’ implementations differed sub-
including IBM Paris, Lockheed Georgia, and Lockheed stantially. CADD ran on IBM mainframes, and its geom-
Missile and Space in Sunnyvale, California. Eventual- etry was based on parametric cubic polynomials and
ly, IBM started a successful effort to sell CADAM (ac- evaluators. Graphics support was primarily the IBM
quired from Lockheed) to drive mainframe sales. 2250, a 2-D-only device. Evans and Sutherland (E&S)
Additional CAD development occurred at Lock- [11] sold a number of Multi-Picture Systems (MPSs) as
heed Georgia in 1965 [28, pp. 4-3–4-4]. Spearheaded by a 2250 alternative. The MPS featured hardware for 3-D
Sylvan (Chase) Chasen, the software ran on CDC 3300 transformations, which had the potential to offload the
computers and Digigraphics terminals. The purpose mainframe. E&S modified its controller to allow two
was more to assist in NC program path planning than terminals to share a single controller through a device
to create engineering drawings. called a Watkins box (named after the designer and de-
veloper, Gary Watkins). The Watkins box was attached
to a small DEC minicomputer, which handled commu-
Northrop nications to and from the mainframe. This configura-
Military program funding often drove the development of tion provided enough savings over the 2250/3250 to
aerospace company systems. Northrop Computer-Aided justify the purchase of dozens of E&S terminals.
Design and Northrop Computer-Aided Lofting (NCAD/ UG ran on multiple brands of midrange minicom-
NCAL) is an excellent example [1]. Northrop based the puters, including DEC PDP and VAX systems as well as
system design for the mid-1970s B-2 Spirit stealth bomb- the Data General S/250, S/230, and S/200. UG derived
er on NCAD/NCAL. Other Northrop military programs its geometry from the ADAM system. Early versions
and Northrop subcontractors used NCAD/NCAL for 3-D of ADAM relied on canonical forms and special case
surface modeling and CADAM for drafting. geometry algorithms. Interactive graphics for UG was
Northrop used funds from the B-2 program to develop provided on Tektronix storage-tube devices.
NCAD/NCAL [14] rather than use similar systems from
other contractors. NCAD/NCAL ran on IBM mainframes
Dassault Aviation
interconnected with classified networks. Importantly, the
Dassault Aviation started its journey in computer
mainframes and networks crossed multiple corporate
graphics to help smooth curve and surface data in the
boundaries, including Boeing, Hughes Radar, GE Engines,
late 1960s. In 1974, the company became one of the
and Vought. All partners had to use NCAD/NCAL and pro-
first licensees of Lockheed’s CADAM software for
vide their own IBM mainframes. This approach simplified
2-D drafting.
data integration and transfer issues and resulted in the
Designing in 3-D took a different route. In 1976,
first military aircraft fully designed on a CAD system. The
Dassault Aviation acquired the Renault UNISURF pro-
B-2 program started in the early 1980s, and its first flight
gram and its Bézier curve and surface capability to
occurred 17 July 1989. The airplane is still in service today.
complement CADAM.
CATIA itself started in 1978 as the Computer-Aided
McDonnell Douglas Tridimensional Interactive (CATI) system. Francis Ber-
McDonnell Douglas implemented two distinctly dif- nard [3] gets credit for extending CATI to surface mod-
ferent CAD systems [22]. The first, Computer Aided eling to generate geometry that would be easier to
Design and Drafting (CADD), was developed in 1965 by machine, a capability particularly important for wind

34 licensed use limited

Authorized IEEE
to:Annals of the Of
Amrita School History of Computing
Engineering - Kollam. Downloaded on January 01,2025 at 17:07:01 UTC fromOctober-December 2024apply.
IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

tunnel models. CATI became CATIA in 1981 when Ber- firsthand by developing their own CAD/CAM pro-
nard convinced Dassault Aviation to commercialize grams. Because of that experience, industrial com-
the system through the Dassault Systèmes spinout. panies can clearly articulate the problems CAD/CAM
As both an internal and commercial product, programs have with interactive performance, scale,
CATIA ran on IBM mainframes with attached IBM and integration to today’s commercial vendors.
2250/3250 and IBM 5080 graphics terminals. The early As noted earlier, the basic performance and in-
underlying geometry forms included Bézier curves and tegration requirements for CAD/CAM programs are
surfaces and grew to include canonical solid defini- essentially the same today as in the early days. Scale
tions and constructive solid geometry operations. Lat- adds another zero or two to the left of the decimal
er versions ran on IBM RS/6000s and other Unix-based point as CAD/CAM data quantities grow.
workstations. The mainframes and minicomputers of 1960−1985
were supplanted by workstations that cost significant-
ly less. Workstations were overtaken by the ever-in-
Matra Datavision Euclid
creasing compute power and the ability to network
French aerospace company Matra’s Euclid system (not
personal computers in the mid-1990s. Computing to-
to be confused with the Credit Suisse Euklid system
day has a turn-the-clock-back feel as cloud computing
for NC machining) addressed modeling for fluid flow.
has gained momentum, and current CAD systems are
Euclid [29] was a modeler sold by the French start-
being delivered via the cloud. As was the case with
up Datavision in 1979. It was originally developed by
early mainframes, cloud computing centralizes pro-
Brun and Theron in the Computer Science Laborato-
cessing and data resources. Users take advantage of
ry for Mechanics and Engineering Sciences in Orsay,
high-performance networks and access cloud sys-
France. Its initial purpose was fluid flow modeling.
tems remotely via lower cost PCs. When CAD software
The French conglomerate Matra, which had aero-
is executed in the cloud, license sharing becomes fea-
space components, bought the controlling interest
sible and software updates occur remotely.
in Euclid in 1980. Dassault Systèmes purchased the
When applied to CAD, cloud computing faces the
software in 1998.
same scale and performance issues present in the early
days with centralized mainframes and minicomputers.
CONCLUSION Cloud scales well from a raw processing perspective. It
Even though internally developed CAD/CAM pro- is easy to add more processing power, and servers are
grams are unusual today, a number of commercial sys- generally in the same physical location, which decreases
tems had their roots in early industrial programs. In- data transfer costs. What is hard for cloud computing is
ternally developed programs had direct access to user satisfying CAD systems’ requirement for near-real-time
communities and were able to develop math software interactive performance, especially at significant dis-
that matched company practice. The interactive meth- tances. Many cloud services are based in data centers
ods and the mathematics influenced other industries, that are tens, hundreds, and even thousands of miles
such as electronic games and animated films. away. Such distances make achieving near-real-time in-
Early commercial CAD/CAM programs were pack- teractive performance difficult. Interactive performance
aged as turnkey systems. Each turnkey system sup- continues to force many CAD/CAM applications to run in
ported only a few concurrent users at relatively slow a distributed manner. Applications run on a PC near the
speeds. Industrial companies, which had to support user, whose data are stored on a file server that is con-
hundreds and even thousands of users, had the com- figuration managed. When requested, the data are most
puting power (generally large, multiuser mainframes often checked out from the server, downloaded to the
and high-speed vector refresh terminals), the talent PC, processed locally, and checked back in.
(mathematicians and programmers), and the money As AI has become more popular, using it to improve
to build their own proprietary CAD/CAM programs. CAD/CAM user productivity is also being pursued.
By the late 1980s, however, there was not enough of There has been significant research into design opti-
a competitive advantage to continue development mization and automated documentation production,
and support. At that point, commercial companies with limited success to date. Design optimization relies
had developed enough manufacturing, surface design, on one or more engineering analyses to tweak the ge-
and other capabilities that internal development and ometry. Not only are multiple runs needed to optimize
maintenance were no longer cost efficient. the geometry, but the suggested optimization can force
Industrial companies experienced the require- changes to the geometry (such as folds and tears) that
ments for high-performance, interactive systems cannot be computed. Other efforts have attempted

October-December
Authorized 2024
licensed use limited to: Amrita School Of Engineering - Kollam. Downloaded on IEEE Annals
January of the
01,2025 History of
at 17:07:01 UTCComputing 35apply.
from IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

to create useful exploded-view drawings typical in a [11] “Evans and Sutherland multi-picture system.” Evans
parts catalog from the raw geometry. This seems to be and Sutherland. Accessed: Oct. 9, 2024. [Online].
straightforward because assembly components can be Available: https://archive.org/details/evansutherland
easily moved along an x-, y-, or z-axis. The issue is that an [12] A. R. Forrest, “On Coons and other methods for
exploded view in a parts catalog shows a disassembly/ the representation of curved surfaces,” Comput.
reassembly sequence. Automated disassembly is a task Graph. Image Process., vol. 1, no. 4, pp. 341–359,
that has been unsuccessfully researched for decades. 1977 . Accessed: Oct. 9, 2024. [Online]. https://doi.
CAD/CAM programs are still evolving with signifi- org/10.1016/0146-664X(72)90020-2.
cant amounts of work still needed. The field remains as [13] W. J. Gordon, “Spline-blended surface interpolation
pertinent and as challenging as it was in the early years. through curve networks,” J. Math. Mech., vol. 18, no.
10, pp. 931–952, 1969 . Accessed: Oct. 9, 2024. [Online].

BIBLIOGRAPHY https://doi.org/10.1512/iumj.1969.18.18068.
[14] J. M. Griffin and J. E. Kinnu, “B-2 systems engineering
[1] N. S. Argyres, “The impact of information technology
case study,” Air Force Center for Systems Engineering,
on coordination: Evidence from the B-2 ‘Stealth’
Air Force Institute of Technology, 2007. Accessed: Oct.
bomber,” Org. Sci., vol. 10, no. 2, pp. 162–180, 1999.
9, 2024. [Online]. Available: https://apps.dtic.mil/sti/
Accessed: Oct. 9, 2024. [Online]. Available: http://www.
citations/ADA464771
jstor.org/stable/2640310, doi: 10.1287/orsc.10.2.162.
[15] T. E. Johnson, “Sketchpad III: A computer program
[2] C. Baltes, “Patrick J. Hanratty: Pioneer in commercial
for drawing in three dimensions,” in Proc. Spring
CAD/CAM software systems,” IEEE Ann. Hist. Comput.,
Joint Comput. Conf. (AFIPS Spring), 1963, pp.
early access, 2024.
347–353. Accessed: Oct. 9, 2024. [Online]. https://doi.
[3] F. Bernard, “Dassault systèmes: A CAD success story,”
org/10.1145/1461551.1461592.
IEEE Ann. Hist. Comput., early access, 2024, doi:
[16] D. Kasik, “Developing and using CAD/CAM/CAE
10.1109/MAHC.2024.3457165.
systems in Boeing,” IEEE Ann. Hist. Comput., early
[4] P. E. Bézier, “Example of an existing system in the
access, 2024. Accessed: Oct. 9, 2024. [Online]. https://
motor industry: The UNISURF system,” Proc. Soc.
doi.org/10.1109/MAHC.2024.3468340.
London Ser. A, Math. Phys. Sci., vol. 321, no. 1545, pp.
[17] D. Kasik and J. C. Dill, “Interactive graphics in industry:
207–218, Feb. 1971. Accessed: Oct. 9, 2024. [Online]. The early days,” IEEE Comput. Graph. Appl., vol. 40,
https://doi.org/jstor.org/stable/77846. no. 5, pp. 89–99, Sep./Oct. 2020. Accessed: Oct. 9,
[5] R. Bilstein, Orders of Magnitude: A History of the NACA 2024. [Online]. https://doi.org/10.1109/MCG.2020.
and NASA, 1915–1990. Washington DC, USA: NASA, 3012228.
1989, pp. 9–14. [18] D. Kasik et al., “A flexible approach to alliances of
[6] F. W. Bliss, “Interactive computer graphics at Ford complex applications,” in Proc. 21st Int. Conf. Softw.
Motor Company,” in Proc. 7th Ann. Conf. Comput. Eng. (ICSE), 1999, pp. 23–32. Accessed: Oct. 9, 2024.
Graph. Interact. Techn. (SIGGRAPH), 1980, pp. [Online]. https://doi.org/10.1145/302405.302408.
218–224. Accessed: Oct. 9, 2024. [Online] Available: [19] F. N. Krull, “The origin of computer graphics within
https://10.1145/800250.807495. general motors,” IEEE Ann. Hist. Comput., vol. 16, no.
[7] J. W. Boyse and J. E. Gilchrist, “GMSolid: Interactive 3, pp. 40–56, 1994. Accessed: Oct. 9, 2024. [Online].
modeling for design and analysis of solids,” IEEE https://doi.org/10.1109/MAHC.1994.298419.
Comput. Graph. Appl., vol. 2, no. 2, pp. 27–40, Mar. 1982, [20] T. Lazear, “CAD development at Fluor, T&W Systems/
Accessed: Oct. 9, 2024. [Online] Available: https://doi: Versacad, Computervision, and Archway Systems,
10.1109/MCG.1982.1674155. 1959–2024,” IEEE Ann. Hist. Comput., early access,
[8] J. A. Brewer and D. C. Anderson, “Visual interaction 2024. Accessed: Oct. 9, 2024. [Online]. https://doi.
with Overhauser curves and surfaces,” SIGGRAPH org/10.1109/MAHC.2024.3464371.
Comput. Graph., vol. 11, no. 2, pp. 132–137, 1977. [21] O. Memon, “How the Boeing 777 represented a
Accessed: Oct. 9, 2024. [Online]. https://doi. milestone for computer-aided design in aviation,”
org/10.1145/965141.563883. in Simple Flying, Saint-Laurent, QC, Canada:
[9] D. C. Llach, “Computer-aided design in the United Valnet Publishing, 2023. Accessed: Oct. 9, 2024.
States, 1949–1984: Notes for a historiography,” IEEE [Online] Available: https://simpleflying.com/
Ann. Hist. Comput., early access, 2024. boeing-777-computer-aided-design-milestone/
[10] “Engineer, Pierre Bézier, inventor of the Bézier curves.” [22] M. Mills, “A totally integrated systems approach to
Engology. Accessed: Oct. 9, 2024. [Online]. Available: design and manufacturing at McDonnell Douglas
http://archive.is/EhFU Corporation,” in Proc. 18th Des. Automat. Conf.,

36 licensed use limited

Authorized IEEE
to:Annals of the Of
Amrita School History of Computing
Engineering - Kollam. Downloaded on January 01,2025 at 17:07:01 UTC fromOctober-December 2024apply.
IEEE Xplore. Restrictions
 COMPUTER AIDED DESIGN I: FORMING THE CAD SOFTWARE INDUSTRY

Nashville, TN, USA, 1981, pp. 160–165. Accessed: [28] D. E. Weisberg. “The History of CAD, Shapr3D.”
Oct. 9, 2024. [Online]. https://doi.org/10.1109/ Accessed: Oct. 9, 2024. [Online]. Available: https://
DAC.1981.1585347. web.archive.org/web/20240423194420/http://
[23] D. T. Ross and J. E. Ward, “Investigations in computer- images.designworldonline.com.s3.amazonaws.com/
aided design for numerically controlled production,” CADhistory/85739614-The-Engineering-Design-
Elect. Syst. Lab., Dept. Elec. Eng., Massachusetts Inst. Revolution-CAD-History.pdf
Technol., Cambridge, MA, USA, 1968. Accessed: Oct. [29] “Euclid.” Wikipedia. Accessed: Oct. 9, 2024.
9, 2024. [Online]. https://doi.org/core.ac.uk/download/ [Online]. Available: https://en.wikipedia.org/wiki/
pdf/4378783.pdf Euclid_(computer_program)
[24] I. E. Sutherland, “Sketchpad, a man-machine graphical [30] “History of CAD software.” Wikipedia. Accessed: Oct. 9,
communication system,” Ph.D. dissertation, Dept. 2024. [Online]. Available: https://en.wikipedia.org/wiki/
Elect. Eng., Massachusetts Inst. Technol., Cambridge, History_of_CAD_software
MA, USA, 1963.
[25] 3D Software History 1980–1985. (2011). Computer
DAVID J. KASIK retired as a senior technical fellow at the
Graphics Museum. Accessed: Oct. 9, 2024. [Online],
Boeing Company and has since stayed active with consulting
Available: http://computergraphmuseum.free.fr/
and professional activities. His research interests include highly
softwarehistory.htm
interactive systems for complex 3-D data, interaction methods,
[26] K. Versprille, “My story of NURBS and computervision,”
lessons to be learned from history, and how to ease technology
IEEE Ann. Hist. Comput., early access, 2024.
[27] H. Voelcker et al., “The PADL-1.0/2 system for defining transfer. Kasik received his master’s degree in computer science

and displaying solid objects,” in Proc. 5th Ann. Conf. from the University of Colorado, Boulder. He is a fellow of the
Comput. Graph. Interact. Techn. (SIGGRAPH), 1978, pp. Association for Computing Machinery fellow, a member of
257–263. Accessed: Oct. 9, 2024. [Online]. https://doi. the SIGGRAPH Academy, and an IEEE Distinguished Scientist.
org/10.1145/800248.807400. Contact him at dave.kasik@gmail.com.

Rejuvenating Binary Executables Visual Privacy Protection Communications Jamming

■ ■
Policing Privacy Dynamic Cloud Certification Security for High-Risk Users
■ ■
Smart TVs Code Obfuscation The Future of Trust
■ ■

IEEE Symposium on
Security and Privacy

January/February 2016 March/April 2016 May/June 2016

Vol. 14, No. 1 Vol. 14, No. 2 Vol. 14, No. 3

IEEE Security & Privacy magazine provides articles

with both a practical and research bent by the top
thinkers in the field.
• stay current on the latest security tools and theories and gain invaluable practical and computer.org/security
research knowledge,
• learn more about the latest techniques and cutting-edge technology, and
• discover case studies, tutorials, columns, and in-depth interviews and podcasts for the
information security industry.

October-December
Authorized 2024
licensed use limited to: Amrita School Of Engineering - Kollam. Downloaded on IEEE Annals
January of the
01,2025 History of
at 17:07:01 UTCComputing 37apply.
from IEEE Xplore. Restrictions