Computers & Graphics 77 (2018) 88–96

Contents lists available at ScienceDirect

Computers & Graphics

journal homepage: www.elsevier.com/locate/cag

Survey Paper

A survey of the contents in introductory Computer Graphics coursesR

Dennis G. Balreira a,∗, Marcelo Walter a, Dieter W. Fellner b,c
a
Institute of Informatics-UFRGS, Porto Alegre, Brazil
b
TU Darmstadt & Fraunhofer IGD, Germany
c
Graz University of Technology, Institute of Computer Graphics and Knowledge Visualization, Austria

a r t i c l e i n f o a b s t r a c t

Article history: Computer Graphics is a very active field, with new knowledge being published every day at a high rate.
Received 2 May 2018 There is, therefore, the pressure to regularly review our teaching contents and adjust accordingly. Among
Revised 24 September 2018
the courses on a standard curriculum, the introductory Computer Graphics course is very often the door
Accepted 3 October 2018
for students into the exciting area of Computer Graphics. It is also the opportunity to attract and engage
Available online 9 October 2018
the best talent for the field. In this paper, we address the question of content in the introductory Com-
Keywords: puter Graphics course as a community. Our main motivation was to find out what our peers are teaching
Computers and education in this first course and use this knowledge to ease the redesign of our introductory course. We have
Computer Graphics surveyed 28 introductory Computer Graphics undergraduate courses from higher level educational insti-
tutions from around the world. We have asked the instructors of these courses to send us data on their
courses, such as the weekly list of topics, and others such as textbooks. We gathered and processed this
data using a bottom-up approach. The final top-level list of subjects and percentages for the introductory
Computer Graphics courses, following the knowledge units defined in the 2013 ACM/IEEE recommenda-
tion, is as follows: Rendering (71.3%), Geometric Modeling (17.4%), Animation (7.8%), Fundamentals (3.0%),
and Visualization (0.5%). We believe this survey will be helpful for institutions considering designing a
new introductory course from scratch or redesigning an existing one, by providing the current state-of-
practice of top Computer Graphics institutions around the world.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction
As in any dynamic and relatively new field, Computer Graphics
benefits from new knowledge being published every day. Consid-
ering only six well-known journals in the field (ACM TOG, Comp.
Graphics Forum, IEEE TVCG, IEEE CG&A, Computers & Graphics,
and The Visual Computer), we as a community published an av-
erage of 2.73 papers a day in the last two years (in Appendix A we
present this data). This high level of production also means pres-
sure on curricular development to cope appropriately and respond
to continuing advancements in the field. Therefore, as teachers
and professors of Computer Graphics, we are always striving for
the balance between breadth of coverage against depth, between
a more hands-on approach versus a more theoretical perspective,
and other balancing decisions. Within the Bachelor of Computer
Science at UFRGS, we started a discussion on the pros and cons
of introducing modern OpenGL in the undergraduate introductory
course and the best way of doing so. This question motivated
R
This article was recommended for publication by J. Jorge.
∗
Corresponding author.
E-mail address: dgbalreira@inf.ufrgs.br (D.G. Balreira).
us to consider how our peer colleagues around the world were
approaching the same issue. With this initial idea in mind, we
thought further that it would be worth seeing what content was
being taught and use this knowledge to help redesign our course.
We surveyed 28 introductory undergraduate courses in Com-
puter Graphics from universities distributed on three continents
(North America, Europe, and Asia) selected according to their pub-
lication records. We gathered data on the topics being taught ac-
cording to the list of topics publicly available – typically on web
pages. We also collected data on textbooks and whether or not
modern OpenGL is already part of the curriculum. We hope our
study will be helpful for institutions which are or will be redesign-
ing their curricula in the near future.
This paper is an extended version of an earlier conference paper
[1]. Based on feedback from the conference paper and a more thor-
ough analysis of our data, we have extended the original research:
we have updated and expanded related work; we have surveyed
28 institutions (instead of 20 from the previous paper), with con-
firmation about the collected data for 71.14% of the courses; for the
final list of topics and percentages we have included a second sce-
nario including two more learning topics; and finally we have as-
sessed, for all universities, how much coverage each learning topic

https://doi.org/10.1016/j.cag.2018.10.001
0097-8493/© 2018 Elsevier Ltd. All rights reserved.
 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96
is receiving. More details about these extensions are given in ap-
propriate sections below.
2. Related work
The importance of teaching in the field has been acknowledged
as early as 1983 when a panel chaired by James Foley at SIG-
GRAPH discussed curriculum practices in the yet incipient area
of Computer Graphics [2]. In 1986, Ohlson advocated a stronger
position for Computer Graphics within Computer Science [3],
considering the lack of Computer Graphics content in the recom-
mendations at the time. Another panel at SIGGRAPH 1994 [4] dis-
cussed several viewpoints on how to teach an introductory Com-
puter Graphics course, bringing together professors from arts, en-
gineering, and Computer Science to present their views. Jensen
and Nieuwenhuizen [5] presented how Computer Graphics courses
were offered at Delft University of Technology. In 1997 Hans-
mann presented a survey of Computer Graphics teaching in Ger-
man universities, including technical universities [6]. The survey
asked questions about Computer Graphics education in general, in-
cluding the topics being taught.
In the late 90s and early 20 0 0s, mainly due to the availability of
OpenGL, introductory Computer Graphics teaching saw an increase
in research activity. In a panel in 1999 during SIGCSE [7], the mem-
bers proposed a philosophy for the beginning Computer Graphics
course, expressed as a series of recommendations. Among these
the recognition that Computer Graphics deals with intrinsically vi-
sual content, and the importance of interactivity. In [8] Wolfe pre-
sented a small survey conducted during SIGGRAPH 1998 among 20
educators from various institutions around the US. The results are
a clear picture of the state-of-the-art of teaching undergraduate
Computer Graphics courses at the time. Among other conclusions,
the number of identified unique topics was 38, and the most cited
topic was viewing transformations, 95% of educators mentioned it.
Considering the interdisciplinary potential of Computer Graph-
ics, Steve Cunningham presented in 20 0 0 two papers that argued
for a wider audience for the beginning Computer Graphics course
[9,10], for instance, majors in other fields such as engineering and
mathematics. Hitchner and Sowizral [11] proposed a new method
of teaching Computer Graphics with a focus on intermediate and
high-level principles, algorithms, and tools, as opposed to the low
and intermediate levels from the past. A year later Santos [12] pre-
sented a general proposal for incorporating Computer Graphics ed-
ucation for an Informatics Engineering degree, one of the first of its
kind in Portugal. Bouvier advanced a proposal for the introductory
course blending what he called “old” (2D) and “new” (3D) topics
[13].
In 2004, a report from the working group on Computer Graph-
ics in Computer Science from SIGGRAPH and Eurographics put for-
ward the concept of Computer Graphics for all [14], considering,
among other aspects, the growing availability of cheaper graph-
ics hardware in computers for domestic use. In the same year,
Sung and Shirley defended that more mature students would ben-
efit from a top-down approach for the introductory course [15]. In
2006, the SIGGRAPH Education Committee‘s Curriculum Working
Group presented a knowledge base for the Computer Graphics dis-
cipline [16]. This proposal contains seventeen core areas and a pre-
occupation with combining both technical and artistic skills. Tori
and colleagues reported on an experience of combining Java3D and
games for a successful approach for the first course [17].
The Bologna Process affected teaching all over the European
Union and it was not different for Computer Graphics. In 2006 a
Computer Graphics Education Workshop was held just before Eu-
rographics [18]. The goal of this workshop was to provide guide-
lines for Computer Graphics teaching in European universities, con-
sidering the Bologna requirements. The main recommendation was
89
to urge universities to have at least one Computer Graphics course
required for graduation. Another important recommendation was
for the universities to use the term visual communication instead,
since it emphasizes a broader aim for the introductory course in
graphics, helping the recognition of its importance.
Cunningham advocated the use of contexts in the beginning
course as a strong motivational tool to attract and retain students
[19]. This idea was refined and discussed by a group of experts in
the Computer Graphics Education 2009 Workshop [20]. The main
concept is that considering how Computer Graphics has reached all
fields of science, it is important to recognize and adapt the teach-
ing according to a particular context.
Low-level APIs such as OpenGL require considerable program-
ming until compelling visual results are possible. This drawback
motivated the use of the Processing language in an introduc-
tory course at Polytechnical University of Valencia [21] and at the
United States Air Force Academy [22]. Bürgisser et al. [23] explored
further this difficulty by presenting an alternative to OpenGL called
bRenderer. Very recently Chen and colleagues reported an even
more radical API-free approach for teaching Computer Graphics,
with positive results [24].
The availability of GPU programming and pervasiveness of
graphics cards motivated proposals for teaching shader-based in-
troductory Computer Graphics course pioneered by Angel and
Shreiner [25], and later expanded by Papagiannakis et al. [26], and
Toisoul et al. [27].
Considering the need for modern OpenGL programming skills,
Fink, Weber, and Wimmer presented a framework for teaching an
introductory Computer Graphics course at the Vienna University
of Technology [28]. The course has only programming assignments
which guide the student through the concepts and supporting lec-
tures given in a previous semester. Shesh reported the develop-
ment of an undergraduate course at the same institution over a
five semesters span [29]. The paper provides guidelines that can
be used by small and medium-sized institutions where typically
only one graphics course is offered. Ackermann and Bach [30] pre-
sented their experience redesigning a Computer Graphics intro-
ductory course. The trigger factor was how to incorporate modern
OpenGL in the curriculum, the same trigger as for our institution.
Dodgson and Chalmers [31] specifically looked into how to design
a first course for students with elementary knowledge of mathe-
matics and only one programming course. Finally, a modern look
on how Computer Graphics teaching can be improved with inno-
vative approaches was presented by Santos and colleagues [32], as
a collective effort resulting from the Education track from Euro-
graphics 2016.
Our work is similar in principle to the surveys from Hansmann
for German universities [6] and from Wolfe for US educators col-
lected during SIGGRAPH 1998 [8]. Ours revisits the issue of content
for an introductory Computer Graphics course, but with a broader
scope, including courses from all over the world.
2.1. Curriculum recommendations
Together with the contributions above, our field has also seen
the effort of IFIP, ACM, and IEEE in providing recommendations to
help and ease curriculum design. The first proposal was presented
in 1968 [33] and already included reference to Computer Graphics
inside one of three major divisions called “Methodologies”. In 1976
a working group from IFIP discussed issues related to methodol-
ogy and recommendations [34]. New updates from ACM were pub-
lished in 1979, but it was only in 1991 that a major new pro-
posal was put forward and which became known as the ACM/IEEE
Curriculum 91 [35]. A major difference from previous recommen-
dations was that this was the first proposal jointly developed by
both ACM and IEEE. In this proposal, Computer Graphics appears
90 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96

inside a cluster of two courses called “Human-Computer Commu- Table 1

nication”, together with “User Interfaces”. Ten years later a major
revision of the Curriculum 91 was undertaken [36]. In this review,
the body of knowledge of Computer Science was divided into 14
areas, with “Graphics and Visual Computing” one of them. A re-
view of the 2001 proposal was carried out in 2008 [37] and a ma-
Higher ranked 30 institutions according to the number of published pa-
pers in the 2015/16 editions of SIGGRAPH, SIGGRAPH Asia, and Euro-
graphics (only higher level educational institutions). The ∗ marks insti-
tutions for which there was no online course information available. The
† marks institutions for which we did not receive confirmation from the
instructors.

jor comprehensive new version presented in 2013 [38]. This is the Institution Number of papers
most updated version that expanded the 14 knowledge areas into
1. Stanford 18.90
18. In this last proposal the area “Graphics and Visual Computing” 2. MIT † 16.32
was renamed to “Graphics and Visualization”. Each knowledge area 3. Zhejiang University † 14.87
is subdivided into Knowledge Units (KUs). For the “Graphics and 4. Univ. British Columbia † 12.21
5. ETH Zurich 11.55
Visualization” KA, there are six KUs. Later, on Sec. 3 we detail this
6. University College London 7.82
structure and explain how we used this recommendation in our 7. Ecole Polytechnique Federale de Lausanne 7.64
work. 8. Columbia University † 7.51
9. The Chinese Univ. of Hong Kong 7.03
10. Tsinghua University † 6.03
3. Methodology 11. Princeton 5.60
12. Carnegie Mellon 5.52
13. RWTH Aachen Univ. 5.37
In this section, we describe our methodology to collect the data.
14. Univ. Southern California 5.33
All our collected data is available at http://wiki.inf.ufrgs.br/Survey_ 15. Tel Aviv Univ. 4.68
of_the_Contents_in_Introductory_Computer_Graphics_Courses. 16. Vienna Univ. of Technology 4.67
17. The Univ. Hong Kong † 4.65
18. Cornell 4.48
3.1. Selection of universities and courses 19. Tokyo University 4.45
20. University of Wisconsin (Madison) † 4.32
Which universities have the best teaching in Computer Graph- 21. Korea Adv. Inst. of Science and Technology 4.27
22. Simon Fraser University∗ 4.21
ics? This question does not have an easy answer. Global rankings 23. Sapienza University of Rome 4.08
– such as Times Higher Education – usually look the overall insti- 24. Ohio State University † 4.00
tution and not a particular field or, in our case, a particular branch 25. Texas A&M University 3.92
of this field, i.e., Computer Graphics. Besides, their criteria take into 26. Univ. of Science and Technology of China 3.88
27. Universidad de Zaragoza∗ 3.85
account many aspects, not only Education. To answer this question,
28. New York Univ. 3.83
we considered the relationship between the quality of research 29. University of California, San Diego 3.83
versus the quality of teaching. 30. Czech Tech University in Prague 3.74
There are different views on this complex and multidimensional
issue. Galbraith and Merrill [39] tackled the problem of confronting
views on this topic. They showed that faculty research productiv-
ity is positively related to teaching effectiveness. They assessed the received a share of the total. For instance, a paper with three au-
performance of undergraduate students measured by a standard- thors from three different institutions would receive 1/3 author-
ized student learning outcome measure. Other researchers chal- ship each. When the same author listed two or more affiliations
lenge this view. Palali and colleagues found a positive correlation we have two possible situations: if both are with educational in-
only for Masters students [40]. Others argue that there is no cor- stitutions, we used the first listed; if at least one was with an edu-
relation [41,42]. It seems clear, however, that there must be effec- cational institution, we used his/her first affiliation with the educa-
tive instruments in place to promote the synergy between teach- tional institution. We understand that this introduces only a small
ing and research [43,44]. In light of the current research on this bias due to the low number of such cases. Around 18% of authors
issue, we assumed that the universities with the best teaching are are affiliated only with noneducational institutions, as illustrated in
the ones with the best research. To define the institutions which Table B1. These were not included in our datasets. Also, when two
are doing the high impact research in Computer Graphics, we have departments of the same university are listed, we grouped them as
surveyed the 2015/2016 editions of SIGGRAPH, SIGGRAPH Asia, and being the same. The final list has unique 223 institutions.
Eurographics, easily recognizable as venues where high-quality re- In Table 1 we list the higher ranked 30 institutions together
search is published. with the overall number of papers they published. We decided
But are the high productivity professors also the instructors of to use 30 for our survey considering a balance between amount
the courses? To answer this question, we checked the number of of data and representation among all institutions. Fig. 1 illustrates
papers published by the instructors from the selected universities how these institutions are geographically located.
(data in Appendix D). Out of 27 instructors, only seven did not It is important to mention that some institutions with a high
have any full paper published in the above mentioned venues, pro- record of publications do not appear here because they are re-
viding a high ratio of 74%. Further, 14 out of 20 instructors pub- search institutes granting only masters or Ph.D. degrees. Our fi-
lished four or more papers, which is a remarkable number consid- nal study used data from 28 institutions since two universities on
ering the high-level conferences we gathered information from. the list did not have online information available and did not reply
For each full paper of the selected conferences, we collected the to our requests for information. These are Simon Fraser University
names of institutions affiliated with that paper, but only for higher and University of Zaragoza.
educational institutions (HEIs) that grant undergraduate degrees. For each university on this list, we visited their web pages for
Although we realize that these conferences have many tracks for the Computer Science courses and searched for the introductory
submission (short papers, tutorials, etc.), we opted to gather in- Computer Graphics course available for majors in Computer Sci-
formation only from full papers for two reasons: first, only full ence. This was our primary source of data from which we col-
papers are published in journals, and second, full papers repre- lected the following: (i) learning topics; (ii) textbook; (iii) informa-
sent, in principle, more permanent contributions to the field as a tion about modern OpenGL. In only one case we used an automatic
whole. If a given paper had more than one affiliation, each author translation of the content.
 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96 91

Table 3
Top 10 topics in the level 0 set of topics.

Topic Number of occurrences

1 Ray Tracing 16.25

2 Rasterization 13
3 Transformations 11.5
4 Global Illumination 10.5
5 Texture Mapping 10.25
6 Color 9.67
7 OpenGL 9
8 Lighting and Shading 8.5
9 Graphics Pipeline 8
10 Textures 8

since they contained expressions like “introduction” or “comments

on assignment #2 and #3” for instance. We labeled the remain-
ing 626 topics as the raw set, since semantically similar entries
Fig. 1. Distribution of the 28 institutions according to the continents. are treated as distinct. For instance, raytracing and ray tracing are
two separate entries. We processed the raw set for lexically similar
topics as in the ray tracing example, generating our level 0 set of
Table 2 topics. We collected an associated counter for each topic express-
Names of courses and occurrences.
ing how many times a given topic appeared. If a given course has
Name Number of occurrences three classes on “ray tracing”, for instance, we counted as three
Basic Techniques in Computer Graphics 1 occurrences instead of one. For individual classes with more than
Computer Graphics 15 one topic, we divided accordingly, and therefore some topics ap-
Introduction to Computer Graphics 6 pear with a non-integer count. We think this expresses the weight
Fundamentals of Computer Graphics 1 of the topic on the course as a whole. As expected, we collected
Fundamentals of Computer Graphics, 1
a large number of different topics, 450 to be exact. In Table 3 we
Image Processing, and Vision
Interactive Computer Graphics 1 list the 10 most cited topics from the level 0 set. Fig. 2 presents
Introduction to Computer Graphics and 1 a word cloud with the level 0 set of topics that have a counter
Imaging greater than 2.
Real-Time Rendering 1
Although the level 0 set is a very rich source of information,
Visual Computing 1
we also need a higher level view of the topics. The next step was
therefore to process the level 0 set bottom-up to create a hierarchy
of topics, by clustering lexically and semantically similar topics.
We think that the use of the online list of learning topics for Due to the high variability of topics, and of the associated seman-
each class, maintained by the instructors of the courses is a reli- tics for clustering, we decided to use a manual clustering method,
able source of information since it is easier to update and edit, and even more considering the difficulties in establishing good distance
likely reflect the current state-of-practice. Nevertheless, we asked metrics for clustering in this context. Therefore, we manually pro-
the instructors to confirm our findings. Of the 28 institutions, we cessed the entries in the level 0 set to compute the clusterings,
received confirmation from the instructors about the data being considering our 20 years experience of teaching the introductory
gathered for 20 of them, exactly 71.14%. We were gladly surprised Computer Graphics course. Some topics were natural candidates to
by this high response rate, considering the usually busy schedule cluster since they appeared many times in the level 0 data set,
of instructors in top-ranked higher education institutions. This is such as transformations. Since we only have access to the topics
valuable information never gathered before at such global scale. instead of the full class material, eventually some topics were dif-
ficult to manually classify. In these cases we checked the previous
4. Results and next topics in the syllabus to help in defining its classifica-
tion. Table 4 shows the level 1 set of 24 clusters as a result of our
In this section we present our findings analyzing the wealth of clustering on the level 0 set. We rounded some of the results to
information provided by the online resources. improve the legibility.
Perhaps surprisingly, “Geometric Modeling” appears as the main
4.1. Names of courses cluster followed by “Graphics Pipeline”. This cluster includes ba-
sically the topics viewing transformations, clipping, and visibility.
We included in Table 2 the titles of courses for reference. The Some topics which are also taught as part of the graphics pipeline
most common one is simply Computer Graphics with a little bit were listed separated, due to their high number of occurrences.
more than half of the titles followed by Introduction to Computer This is the case for clusters “Transformations”, “Lighting and Shad-
Graphics, although other possibilities appear. ing”, and “Rasterization”. Also as expected, clusters with lower
occurrence are typically present in more advanced courses, not
4.2. Main topics being taught necessarily the introductory one. Among these, we mention “com-
puter vision” and “visualization”. Curiously, GPU appears with less
From the 28 institutions we collected all the learning topics ac- than 1%, considering its importance in modern Computer Graphics.
cording to their public data, which generated a total of 637 top- Maybe this topic is given in laboratory classes, not necessarily as
ics. We call them the original data set, since it contains all kinds lectures. The “Miscellaneous” cluster gathers all topics which are
of expressions which were originally available in the syllabus for not clearly part of any of the other clusters. Examples of topics
each institution. From these 637, 13 were considered unclassifiable, in the miscellaneous cluster are history of Computer Graphics (only
92 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96

Fig. 2. Word cloud with the level 0 topics for topics with a count greater than 2.

Table 4 Table 5
Level 1 clusterization of topics. Number of universities teaching each of level 1
topics.
Topic Number of occurrences Percent
Topic Universities
1 Geometric Modeling 95 15.20%
2 Graphics Pipeline 85 13.60% 1 Graphics Pipeline 26
3 Transformations 45 7.20% 2 Lighting and Shading 25
4 Animation 42 6.70% 3 Texture Mapping 25
5 Global Illumination 40 6.40% 4 Geometric Modeling 24
6 Lighting and Shading 37 5.90% 5 Ray Tracing 23
7 Ray Tracing 35 5.60% 6 Rasterization 23
8 Texture Mapping 32 5.10% 7 Transformations 22
9 Rasterization 30 4.80% 8 Global Illumination 17
10 Image Processing 24 3.80% 9 Animation 15
11 OpenGL and WebGL 24 3.80% 10 Color 15
12 Misc 21 3.40% 11 Shaders 14
13 Shaders 16 2.60% 12 Sampling 14
14 Sampling 16 2.60% 13 OpenGL and WebGL 13
15 Color 15 2.40% 14 Misc 12
16 Hierarchical Modeling 11 1.70% 15 Hierarchical Modeling 11
17 Shadows 10 1.60% 16 Image Processing 10
18 Interaction 10 1.60% 17 Shadows 8
19 Applications 9 1.40% 18 Interaction 7
20 Tools 9 1.40% 19 Applications 6
21 Math Review 7 1.10% 20 Tools 6
22 Computer Vision 5 0.80% 21 Math Review 6
23 GPU 5 0.80% 22 GPU 5
24 Visualization 3 0.50% 23 Visualization 3
24 Computer Vision 2

one occurrence), deep learning (2 occurrences), and image-based

rendering (two occurrences). In the “Applications” cluster we in-
cluded topics such as 3D printing, digital fabrication, and non-
photorealistic rendering. The “Tools” cluster lists helpful tools such
as Matlab, Maya, and SolidWorks.
Now that we have the level 1 clusters, we can map back to the
universities and see what the “coverage” of a given topic among
all institutions is. In Table 5 we list the level 1 clusters and the
number of universities that have these topics explicitly in their
web listings. From this table, we can see for instance that “Graph-
ics Pipeline” and “Lighting and Shading” are very popular topics
whereas “Computer Vision” appears in only 2 of the surveyed uni-
versities. When designing a new course or redesign an existing
one, the information in this table indicates the importance of top-
ics as expressed by these universities in their introductory Com-
puter Graphics course.
Clustering yet again the contents of the level 0 set we achieve
the level 2 set of clusters. As a semantic guide for this last clus-
tering, we used the knowledge units (KUs) listed in the most re-
cent curriculum proposed by ACM and IEEE [38] as clusters. The
knowledge area (KA) “Graphics and Visualization” has six KUs:
Fundamentals, Rendering, Modeling, Animation, Visualization, and
Computational Geometry. Computational Geometry is a peculiar
topic since it is clearly a more advanced topic for an introductory
course. It is not mentioned in any of the university list of topics.
Therefore we left it out of our level 2 clustering.
This six KUs do not represent the full range of topics in the
field. Two more KAs from the ACM/IEEE recommendation are re-
lated to Graphics and Visualization. The “Intelligent Systems” KA
has one KU named Machine Vision and Image Processing. Topics in
Virtual Reality are found in the “Human-Computer Interaction” KA.
Therefore we will use either 5 (Fundamentals, Rendering, Model-
ing, Animation, Visualization) or 7 (the previous 5 plus Machine
Vision and Image Processing and Virtual Reality) KUs as labels
for clusters in the level 2. Some of the topics could not fit in
the proposed KUs, leading to a new cluster labeled Non-classified
(NC).
 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96 93

Fig. 3. Percentage of KUs covered by each university according to the ACM/IEEE recommendation. The last line shows the average for all institutions. Data available in
Appendix C.

Table 6 Table 7
Percentage of level 2 clusters for 5 and 7 KUs including the Textbooks used.
non-classified topics.
Name Number of occurrences
Topic Percent Percent
Marschner, Shirley [45] 3
Rendering 62.10% 62.10% Steven Gortler [46] 3
Modeling 15.20% 15.20% Edward Angel [47] 3
Animation 6.80% 6.80% Hearn, Baker, and Carithers [48] 2
Fundamentals 2.60% 2.60% Hughes, van Dam, McGuire, Sklar, Foley, 1
Visualization 0.50% 0.50% Feiner, Akeley [49]
Machine Vision and Image Process – 4.80% Alan Watt [50] 1
Virtual Reality – 0.40%
Non-classified 12.80% 7.60%

4.3. Textbooks

Another information we gathered refers to the use of a text-

We summarize the results into Table 6. We can see that “Ren-
dering” is at the top of the contents. At first, it might appear ex-
cessive, but “Rendering” represents the core of what we do and
therefore, for a typical introductory Computer Graphics course, this
amount seems appropriate to us.
Concerning the percentage given to the other topics, if we recall
that the ACM/IEEE recommendation is for the knowledge area as a
whole, not only for the introductory course, we think that the dis-
tribution is appropriate. In a way, this higher level clustering pre-
sented on Table 6 is the snapshot of current state-of-practice of the
28 institutions we surveyed. Out of the original 450 level 0 top-
ics we collected, roughly 13% of the topics could not be mapped
to the ACM/IEEE recommendation when using 5 KUs. For 7 KUs
this number drops to 7.6%. We observe that the new KU “Machine
Vision and Image Processing” is mainly responsible for decreasing
the amount of unclassified topics.
Another interesting result is presented in Fig. 3. The values of
this chart are available in Appendix C. We show how much cover-
age each university is giving to each of the 7 KUs plus the Non-
classified topics. The last line in the charts shows the average. As
expected, “Rendering” prevails in 26 out of 28 universities. One
university has only “Rendering” classes. It seems clear from this
chart that there is considerable variation among institutions. This
is not intrinsically good or bad. If much, it shows the discretionary
freedom of HEIs following their educational vocations.
book. Of the 28 surveyed institutions, 13 required textbooks, a lit-
tle less then 50%. The used textbooks are listed in Table 7.
4.4. Modern OpenGL
Modern OpenGL was introduced in version 3.3 of OpenGL, in
2008. All fixed functions were deprecated, favoring the use of
shaders and the introduction of GLSL - OpenGL Shading Language.
All teaching using OpenGL was affected by this change, and peda-
gogical and technical considerations have to be taken into account
when switching. The so-called modern OpenGL is more complex
to teach and learn, since it requires a fair knowledge of math and
the basics of the graphics pipeline in order to start programming.
On the other hand, the classic OpenGL provides a set of functions
which greatly decreases the amount of work required to render
a scene on screen. Therefore, many courses still use the legacy
mode version, or introduce the modern OpenGL after the basics
are covered. The decision on when and how to change is still in
discussion by the community [30]. Of the 28 surveyed institutions,
16 effectively use modern OpenGL in their classes or assignments.
Plus, two universities use WebGL, consisting of an OpenGL version
which uses GLSL as well. At any rate, around 64% of the institu-
tions teach modern OpenGL. Considering this scenario, an effort to
ease the transition to modern OpenGL as presented by Reina and
colleagues [51] is welcome.
94 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96

5. Discussion provides easier integration among text, images, 2D and 3D graph-

ics as well as UI elements into rich Web applications.
The only two previously available surveys [6,8] reported a sig-
nificantly lower number of topics. In both, the educators had to
6. Conclusions
choose topics from an existing list, 34 and 38 topics respectively.
Our survey reported a much higher number of topics, and we think
We have presented a survey of Introductory Computer Graph-
this is due to two reasons. First, we collected the data directly from
ics courses from 28 higher learning institutions from around the
online resources; there was no list to choose from. Second, in the
world. From these courses we collected data on which topics are
17 years since the last survey, our field has grown significantly. To-
being taught, the textbook used, and whether these courses are al-
day, it would be difficult to agree on a comprehensive list of topics
ready teaching modern OpenGL or not.
if we were to repeat the methodology of the previous surveys.
The source of our survey was the list of topics publicly avail-
In general, courses balance theory with applications and tech-
able. With the confirmation from 71.14% of the instructors, we
nology and it is interesting to see how this balance evolved over
claim that online repositories are a good source for our purposes.
time. The two previous surveys also offer us a view on what was
The criterion for selection of the surveyed institutions with its
“hot” at the time and maybe it is not so hot anymore. For in-
emphasis on research leaves out of the results many medium
stance, in both surveys Fractals appear as a relatively significant
and small-sized high-quality educational institutions. It would be
topic, whereas in our survey there is no mention to Fractals. Also,
worth pursuing a survey similar to ours but over an even larger
in the survey of German universities, there is a topic called Prod-
sample of schools. This would have to be a collaborative effort con-
uct Data Exchange, a standard for information exchange among CAD
sidering the amount of data it would generate and more automatic
systems. Clearly, today’s undergraduate courses would not teach
methods for clustering would be needed. Nevertheless, we think
such a technical topic.
the collected data and clustering here presented should be useful
With respect to curriculum recommendations, although there is
for institutions looking into redesigning their curricula by provid-
a common body of knowledge and proposed recommendations, it
ing a summary of current teaching contents. As we developed our
is not clear whether these recommendations are being considered
investigation, it became more clear that, although we have a cur-
when a new course is designed. As shown in Fig. 3, there is con-
riculum recommendation from ACM/IEEE [38], it would be worth-
siderable variation among institutions. The high variability among
while to look for inspiration from peers. That is basically how we
institutions on how they use the time to teach the KUs suggests
approached this investigation: to give instructors and educators a
that other factors are prevailing on curriculum decisions. It could
look into how 28 institutions around the world are teaching the
be that the instructor’s experience on a given area bias the cho-
introductory course in Computer Graphics.
sen topics; or the environment where each school is reflects cur-
As presented, our survey is a snapshot of current state-of-
riculum content. Maybe the school has a particular bias related to
practice for 28 HEIs around the world for the introductory course
its past, for instance, a strong rendering group. Also, the ACM/IEEE
in Computer Graphics. For future work, we would like to check
recommendation does not suggest any particular distribution of
how these practices correlate with proposed learning outcomes,
hours among the many KUs. It is clearly the decision of each in-
like the ones suggested in the ACM/IEEE recommendation. We
stitution.
would also like to try automatic or semi-automatic bottom-up hi-
A related question is how many courses or hours would
erarchical clustering algorithm, to compare with our clustering. It
be needed to teach most of the contents currently avail-
would also be interesting to compare and contrast our results us-
able in Computer Graphics? Given the current rate of new
ing the KUs from the ACM/IEEE recommendation with other clas-
knowledge being published, we are already seeing full de-
sification schemes such as the publication categories provided by
grees granted for specialized topics within Computer Sci-
the ACM Computing Classification System. Finally, we believe there
ence, such as a Bachelor of Science in Computer Graphics,
is no research showing how instructors put together their list of
granted for instance by the University of Wellington in Victo-
topics for a given course. Which factors are at play when making
ria, New Zealand (https://www.victoria.ac.nz/explore/study-areas/
course design decisions? An answer to these questions would shed
computer-graphics/overview?international=true), and a Bachelor
light on why we see such variability among institutions.
of Science degree in Computer Graphics Technology from Pur-
due School of Engineering (http://engr.iupui.edu/departments/cigt/
undergrad/bscgt/index.php). Another interesting aspect concerns Acknowledgments
textbooks. The classic Foley et al. book [49] disappeared from cur-
rent teaching whereas a previous edition of Angel’s book [47] is First and foremost we have to thank the 20 instructors who
already mentioned in the US survey. Although our field has a good put aside their routine tasks and replied to our request for confir-
collection of excellent references, it appears that approximately mation of information about their courses. We gratefully acknowl-
half of the instructors still rely mostly on course notes provided to edge the partial financial support from CAPES through grant BEX
the students. Considering the relatively young age of Computer Sci- 5824/15-0. This work was developed while the second author was
ence when compared with other fields such as Physics and Math- a guest at the GRIS group at the Technical University Darmstadt.
ematics, maybe we as a community do not have yet a consensual
view of what an introductory course should offer, and this impacts Appendix A. Number of published papers in 2016 and 2017
the use of textbooks.
Finally, regarding modern OpenGL, our survey suggests that it In order to estimate the strength of the Computer Graphics area
is a hard topic to include in teaching. Although it has been around as a whole, we gathered data on the number of papers published
since 2008, only a little bit more than half of the institutions in the top 6 journals in the area according to their impact factor.
are clearly teaching it. Other institutions are considering alterna- The total number of papers is given in Table A1 for the years 2016
tives. Recently, Ackermann and Bach [30] presented the redesign of and 2017. Data gathered from the DBLP Computer Science Bibliog-
an introductory Computer Graphics course using WebGL instead raphy1 and impact factors from Journal Citation Reports.
of modern OpenGL. They presented a few reasons, among them
WebGL being a multiplatform solution, uses open source web-
based JavaScript libraries, is supported by a new textbook [52] and 1
dblp.org
 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96 95

Table A1 Table D1
Number of papers published in 2016 and Number of papers for each instructor responsible for the
2017 in the top 6 journals of the field. introductory CG courses in our surveyed conferences. We
could not find information regarding the instructor at Zhe-
Journal 2016 2017 jiang University. For all the remaining 7 universities the in-
ACM TOG 250 246 structors have no publications in the selected venues.
IEEE CG&A 55 66
University Number of papers
Comp. Graph. Forum 238 247
IEEE TVCG 233 231 Stanford 1
Visual Computer 129 120 MIT 13
Comp. & Graphics 97 82 UBC 9
Total 1003 994 ETH Zurich 13
Daily average 2.75 2.72 Ecole Polytech. Fed. de Lausanne 5
Columbia University 8
The Chinese Univ. of Hong Kong 3
Table B1
Tsinghua University (Beijing) 5
Percentage of authors with only research
Princeton 4
affiliations for the proposed conferences.
Carnegie Mellon University 2
Conference 2015 2016 RWTH Aachen University 7
University of Southern California 4
SIGGRAPH 23.45% 18.71% Tel Aviv University 16
SIGGRAPH Asia 12.47% 22.39% The University of Hong Kong 4
Eurographics 15.32% 11.17% Cornell University 2
Total 18.13% 18.59% The University of Tokyo 1
University of Wisconsin, Madison 4
Texas A&M University 2
Univ. of Sci. and Tech. of China 4
Appendix B. Percentage of research authors in conferences Univ. of California, San Diego 7

We also counted the number of authors with research only af-

filiations among all papers. This data is presented in Table B1.
Appendix D. Number of total papers published by each
instructor

To reinforce our claim about the link between research and

Appendix C. Percentage of KUs covered by each university
according to the ACM/IEEE recommendation
We provide our original data containing the percentage of KUs
covered by each university according to the ACM/IEEE Curriculum
in Table C1. This is the same data presented in Fig. 3.
teaching, we provide data on the number of papers published
by each instructor for the surveyed introductory CG courses in
Table D1. We considered only full papers. The results show that
the majority of the instructors of first year courses are also very
active in high quality research. This find supports our criteria for
selection of universities.

Table C1
Percentage of KUs covered by each university according to the ACM/IEEE Curriculum: Rendering (R), Modeling (M), An-
imation (A), Fundamentals (F), Visualization (V), Machine Vision and Image Processing (MVIP), Virtual Reality (VR) and
Non-classified (NC).

University R M A F V MVIP VR NC

Stanford 70.59% 11.76% 0.00% 5.88% 0.00% 5.88% 0.00% 5.88%

MIT 51.67% 13.33% 26.67% 3.33% 0.00% 0.00% 0.00% 5.00%
Zhejiang University 77.78% 11.11% 0.00% 11.11% 0.00% 0.00% 0.00% 0.00%
UBC 97.67% 0.00% 0.00% 2.33% 0.00% 0.00% 0.00% 0.00%
ETH Zurich 26.92% 3.85% 15.38% 7.69% 0.00% 46.15% 0.00% 0.00%
University College London 80.77% 13.46% 0.00% 1.92% 0.00% 1.92% 0.00% 1.92%
Ecole Polytech. Fed. de Lausanne 53.85% 23.08% 7.69% 0.00% 0.00% 0.00% 0.00% 15.38%
Columbia University 66.67% 12.50% 12.50% 0.00% 0.00% 0.00% 0.00% 8.33%
The Chinese Univ. of Hong Kong 71.43% 0.00% 2.38% 7.14% 7.14% 0.00% 4.76% 7.14%
Tsinghua University (Beijing) 57.58% 27.27% 0.00% 3.03% 0.00% 3.03% 0.00% 9.09%
Princeton 47.83% 21.74% 13.04% 8.70% 0.00% 4.35% 0.00% 4.35%
Carnegie Mellon University 41.67% 15.00% 23.33% 3.33% 0.00% 0.00% 3.33% 13.33%
RWTH Aachen University 44.00% 44.00% 0.00% 0.00% 0.00% 12.00% 0.00% 0.00%
University of Southern California 52.17% 17.39% 8.70% 0.00% 4.35% 4.35% 4.35% 8.70%
Tel Aviv University 40.24% 4.88% 2.44% 2.44% 0.00% 25.61% 0.00% 24.39%
Vienna University of Technology 80.77% 11.54% 0.00% 7.69% 0.00% 0.00% 0.00% 0.00%
The University of Hong Kong 69.23% 15.38% 7.69% 0.00% 0.00% 0.00% 0.00% 7.69%
Cornell University 75.00% 12.50% 4.17% 0.00% 0.00% 0.00% 0.00% 8.33%
The University of Tokyo 30.00% 40.00% 15.00% 0.00% 0.00% 15.00% 0.00% 0.00%
University of Wisconsin, Madison 66.67% 22.22% 0.00% 0.00% 0.00% 5.56% 0.00% 5.56%
Korea Adv. Inst. of Sci. and Tech. 65.22% 13.04% 4.35% 8.70% 0.00% 0.00% 0.00% 8.70%
Sapienza University of Rome 55.00% 6.67% 26.67% 0.00% 0.00% 0.00% 0.00% 11.67%
Ohio State University 81.67% 16.67% 0.00% 0.00% 0.00% 0.00% 0.00% 1.67%
Texas A&M University 55.56% 33.33% 3.70% 3.70% 0.00% 0.00% 0.00% 3.70%
Univ. of Sci. and Tech. of China 45.83% 16.67% 0.00% 0.00% 0.00% 8.33% 0.00% 29.17%
New York University 10 0.0 0% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00%
Univ. of California, San Diego 73.53% 17.65% 0.00% 2.94% 0.00% 0.00% 0.00% 5.88%
Czech Tech. University in Prague 47.44% 23.08% 0.00% 2.56% 7.69% 0.00% 0.00% 19.23%
96 D.G. Balreira et al. / Computers & Graphics 77 (2018) 88–96

Supplementary material [26] Papagiannakis G, Papanikolaou P, Greassidou E, Trahanias PE. glGA: an openGL
geometric application framework for a modern, shader-based computer graph-
ics curriculum.. In: Eurographics (education papers); 2014. p. 9–16.
Supplementary material associated with this article can be [27] Toisoul A, Rueckert D, Kainz B. Accessible GLSL Shader Programming. In: Pro-
found, in the online version, at doi:10.1016/j.cag.2018.10.001. ceedings of the education papers. The Eurographics Association; 2017. doi:
10.2312/eged.20171024.
References [28] Fink H, Weber T, Wimmer M. Teaching a modern graphics pipeline using a
shader-based software renderer. Comput Graph 2013;37(1–2):12–20.
[29] Shesh A. Toward a singleton undergraduate computer graphics course in small
[1] Balreira DG, Walter M, Fellner DW. What we are teaching in introduction to
and medium-sized colleges. Trans Comput Educ 2013;13(4) 17:1–17:21. doi:10.
computer graphics. In: Proceedings of the education papers. The Eurographics
1145/2522689..
Association; 2017. doi: 10.2312/eged.20171019.
[30] Ackermann P, Bach T. Redesign of an introductory computer graphics course.
[2] Foley JD, Bork A, Brown M, King R, van Dam A, Wozny M. Computer graphics
In: Proceedings of the education papers. Zurich, Switzerland; 2015. p. 9–13.
in higher education. In: Proceedings of the computer graphics (SIGGRAPH);
doi:10.2312/eged.20151021.
1983. p. 31–3.
[31] Dodgson NA, Chalmers A. Designing a computer graphics course for first year
[3] Ohlson MR. The role and position of graphics in computer science education.
undergraduates. In: Proceedings of the education papers. The Eurographics As-
SIGCSE Bull 1986;18(1):232–7. doi:10.1145/953055.5903.
sociation; 2017. doi: 10.2312/eged.20171020.
[4] Bresenham J, Laxer C, Rose-Hulman, Lansdown J, Owen GS. Approaches to
[32] Santos BS, Dischler J, Adzhiev V, Anderson EF, Ferko A, Fryazinov O, Ilčík M,
teaching introductory computer graphics. In: Proceedings of computer graph-
Ilčíková I, Slavik P, Sundstedt V, Svobodova L, Wimmer M, Zara J. Dis-
ics (SIGGRAPH); 1994. p. 479–80.
tinctive approaches to computer graphics education. Comput Graph Forum
[5] Jansen FW, Nieuwenhuizen PRV. Computer graphics education at delft univer-
2018;37(1):403–12. doi:10.1111/cgf.13305.
sity of technology. Comput Graph 1995;19(3):461–5.
[33] Atchison WF, Conte SD, Hamblen JW, Hull TE, Keenan TA, Kehl WB, Mc-
[6] Hansmann W. A survey of computer graphics education at german universities.
Cluskey EJ, Navarro SO, Rheinboldt WC, Schweppe EJ, Viavant W, Young DM
Comput Graph 1997;21(1):113–16.
Jr. Curriculum 68: Recommendations for academic programs in computer sci-
[7] Hitchner L, Cunningham S, Grissom S, Wolfe R. Computer graphics: The intro-
ence: a report of the ACM curriculum committee on computer science. Com-
ductory course grows up. SIGCSE Bull 1999;31(1):341–2. doi:10.1145/384266.
mun ACM 1968;11(3):151–97.
299801.
[34] Guedj R.A., Tucker H.A., editors. IFIP workshop on methodology in computer
[8] Wolfe R. Bringing the introductory computer graphics course into the 21st cen-
graphics. Seillac, France: North-Holland Pub. Co.; 1976.
tury. Comput Graph 20 0 0;24(1):151–5. https://doi.org/10.1016/S0097-8493(99)
[35] Tucker AB. Computing curricula 1991. Commun ACM 1991;34(6):68–84.
00145-4.
[36] ACM/IEEE-CS Joint Task Force on Computing Curricula. Computing curric-
[9] Cunningham S. Powers of 10: the case for changing the first course in com-
ula 2001: computer science. 2001. http://www.acm.org/education/education/
puter graphics. ACM SIGCSE Bull 20 0 0a;32(1):46–9.
education/curric_vols/cc2001.pdf.
[10] Cunningham S. Re-inventing the introductory computer graphics course: pro-
[37] ACM/IEEE-CS Joint Interim Review Task Force. Computer science curriculum
viding tools for a wider audience. Comput Graph 20 0 0b;24(2):293–6.
2008: an interim revision of CS 2001. 2008. http://www.acm.org/education/
[11] Hitchner LE, Sowizral HA. Adapting computer graphics curricula to changes in
curricula/ComputerScience2008.pdf.
graphics. Comput Graph 20 0 0;24(2):283–8.
[38] ACM/IEEE-CS Joint Task Force on Computing Curricula. Computer science cur-
[12] dos Santos MP. Computer graphics in the scope of informatics engineering ed-
ricula 2013. Technical Report; ACM Press and IEEE Computer Society Press,
ucation. Comput Graph 2001;25(5):909–15.
New York, NY, USA; 2013. doi:10.1145/2534860.
[13] Bouvier DJ. From pixels to scene graphs in introductory computer graph-
[39] Galbraith CS, Merrill GB. Faculty research productivity and standardized stu-
ics courses. Comput Graph 2002;26(4):603–8. https://doi.org/10.1016/
dent learning outcomes in a university teaching environment: a Bayesian
S0 097-8493(02)0 0109-7.
analysis of relationships. Stud High Educ 2012;37(4):469–80. doi:10.1080/
[14] Cunningham S, Hansmann W, Laxer C, Shi J. The beginning computer graphics
03075079.2010.523782.
course in computer science. SIGGRAPH Comput Graph 2004;38(4):24–5.
[40] Palali A, van Elk R, Bolhaar J, Rud I. Are good researchers also good teachers?
[15] Sung K, Shirley P. A top-down approach to teaching introductory computer
The relationship between research quality and teaching quality. Econ Educ Rev
graphics. Comput Graph 2004;28(3):383–91. doi:10.1016/j.cag.2004.03.005.
2018;64:40–9. doi:10.1016/j.econedurev.2018.03.011. http://www.sciencedirect.
[16] Orr G, Alley T, Laxer C, Geigel J, Gold S. A knowledge base for the emerg-
com/science/article/pii/S0272775717302315
ing discipline of computer graphics. In: CGEMS - computer graphics ed-
[41] Gibbs G. The relationship between quality in research and quality in teaching.
ucational materials. The Eurographics Association; 9-7-2007. doi: 10.2312/
Qual High Educ 1995;1(2):147–57.
cgems04- 11- 1363.
[42] Hattie J, Marsh HW. The relationship between research and teaching: a meta–
[17] Tori R, Bernardes JL Jr, Nakamura R. Teaching introductory computer graph-
analysis. Rev Educ Res 1996;66(4):507–42.
ics using java 3d, games and customized software: a brazilian experience. In:
[43] Khan MA. Achieving an appropriate balance between teaching and research in
Proceedings of the ACM SIGGRAPH educators program. ACM; 2006. p. 12.
institutions of higher education: an exploratory study. Int J Inf Educ Technol
[18] Bourdin J-J, Cunningham S, Fairn M, Hansmann W. Report of the CGE 06 com-
2017;7(5):341–9.
puter graphics education workshop. In: EG education papers. Vienna, Austria:
[44] Prince MJ, Felder RM, Brent R. Does faculty research improve undergrad-
The Eurographics Association; 2007. doi: 10.2312/eged.20071017.
uate teaching? An analysis of existing and potential synergies. J Eng Educ
[19] Cunningham S. Computer graphics in context: an approach to a first course in
2007;96(4):283–94.
computer graphics. In: Proceedings of the ACM SIGGRAPH ASIA educators pro-
[45] Marschner S, Shirley P. Fundamentals of computer graphics. CRC Press; 2015.
gramme. New York, NY, USA: ACM; 2008. ISBN 978-1-60558-388-4. SIGGRAPH
[46] Gortler SJ. Foundations of 3D computer graphics. Mit University Press; 2012.
Asia 1:1–1:4 doi: 10.1145/1507713.1507715.
[47] Angel E. Interactive computer graphics: a top-down approach using openGL.
[20] Case C, Cunningham S. Teaching computer graphics in context. In: Computer
Addison Wesley; 2008.
graphics education 09 workshop; 2009. online at http://education. siggraph.
[48] Hearn DD, Baker MP, Carithers W. Computer graphics with open GL. Pearson;
org/media/reports/CGE09- Workshop- Report. pdf.
2010.
[21] Linares-Pellicer J, Santonja-Blanes J, Tormos PM, Cuesta-Frau D. Using process-
[49] Hughes JF, van Dam A, Foley JD, McGuire M, Feiner SK, Sklar DF, Akeley K.
ing.org in an introductory computer graphics course.. In: Eurographics (educa-
Computer graphics: principles and practice. The systems programming se-
tion papers); 2009. p. 23–8.
ries. Addison-Wesley; 2014. ISBN 9780321399526. https://books.google.com.br/
[22] Schweitzer D, Boleng J, Graham P. Teaching introductory computer graphics
books?id=OVpsAQAAQBAJ
with the processing language. J Comput Sci Coll 2010;26(2):73–9.
[50] Watt AH. 3D computer graphics. No. v. 1 in 3D Computer Graphics; Addison-
[23] Bürgisser B, Steiner D, Pajarola R. bRenderer: a flexible basis for a modern
Wesley; 20 0 0. ISBN 9780201398557. https://books.google.com.br/books?id=
computer graphics curriculum. In: Proceedings of the education papers. The
tQJEAQAAIAAJ.
Eurographics Association; 2017. doi: 10.2312/eged.20171023.
[51] Reina G, Müller T, Ertl T. Incorporating modern openGL into computer graphics
[24] Chen M, Xu Z, Rippin W. On the pedagogy of teaching introductory com-
education. Comput Graph Appl IEEE 2014;34(4):16–21. doi:10.1109/MCG.2014.
puter graphics without rendering APIs. In: Post F, Åra J, editors. Proceedings of
69.
the education papers. The Eurographics Association; 2018. doi: 10.2312/eged.
[52] Angel E, Shreiner D. Interactive computer graphics with WebGL. Addison-Wes-
20181007.
ley Professional; 2014.
[25] Angel E, Shreiner D. Teaching a shader-based introduction to computer graph-
ics. IEEE Comput Graph Appl 2011;31(2):9–13. doi:10.1109/MCG.2011.27.