On Two Desiderata for Creativity Support Tools
Wai K. Yeap1, Tommi Opas1 and Narges Mahyar2
1Centre for Artificial Intelligence Research,
Auckland University of Technology , New Zealand
2Department of Artificial Intelligence,
Faculty of Computer Science and Information Technology,
University of Malaya, Malaysia
Abstract. This paper discusses two important desiderata for developing
creativity support tools, namely ideation and empowerment. We then use them
to guide us in designing a new individual creativity support tool codenamed
Creative-Pad. Creative-Pad is designed to assist individual advertising creative
to develop creative ideas for advertisements. For ideation, Creative-Pad
searches and filters information automatically from the internet to present to the
user with related words and exemplar sentences. For empowerment, Creative-
Pad is designed in such a way that the user is neither distracted nor burdened to
do any other tasks unrelated to conjuring up a creative idea for a new
advertisement. Creative-Pad is fully implemented and some preliminary results
of its use by advertising creatives are reported.
1 Introduction
Developing a creativity support tool is an exciting and a very challenging problem for
HCI researchers. This is because the interaction between the computer and the human
in this task is almost magical. That is, despite significant past research into the nature
of creativity (for example, see [1, 4, 19, 20]), we do not understand how the mind
works in a creative way. Yet, the challenge here is to develop tools to assist the mind
in its creative endeavour. Shneiderman [17, p. 116] remarked: “Developing software
tools to support creativity is an ambitious, but some would say, vague goal.”.
Developers of such tools thus face some acute problems. For example, when users
of such tools fail to develop a creative solution, it is difficult to know where the
problem lies. There are many other factors, such as one’s lack of attention, skills, and
interests, which could affect the user’s performance. The interplay of these factors
occurs in the mind of the user and is therefore difficult to weed out. Without doing so,
it would be difficult to develop a set of criteria or a framework for developing and
evaluating these tools (although attempts were made, see [16, 18]). Another example
is a general lack of distinction between a creativity support tool and a problemsolving
tool. For instance, if one were to use a sketch pad to help sketch out various
ideas, should that be a creativity support tool or a drawing tool? Distinguishing
between them might not be a straightforward task. This is because a creativity support
tool is often perceived to be very much a part of a problem-solving tool. Yet, without
180
doing so one could complicate the design of such tools or at worse, be confused with
the kind of tool that one is supposed to design.
Attempts to define creativity support tools in the past thus tend to be quite
comprehensive. For example, Lubart [10] considered four categories: computer as
nanny, pen-pal, coach and colleague, while Johnson and Carruthers [7] considered
three classes which range from tools that do not produce creative ideas/artefact to
those that could assist in many different ways. Although these classifications provide
a good scope for discussing work in this area in general, they lack definitive
statements about these tools and in particular those designed to assist individuals to
solve a particular problem.
In this paper, we discuss two desiderata for developing creative support tools,
namely ideation and empowerment. The former emphasizes on generating new ideas
to the user and the latter, empowering the user to be creative. Section 2 discusses
these two desiderata in details. We then show how they guide us in our design and
implementation of a new creativity support tool, codenamed Creative-Pad. Creative-
Pad is designed to assist individual advertising creative. An advertising creative (or,
in short, a creative) is a person working for an advertising agency who is responsible
for developing creative ideas for a new advertisement. For ideation, we emphasize on
developing a process which generates ideas that bear some relations to the problem on
hand. For empowerment, we emphasize on providing the user much time to conjure
his/her idea for a new advertisement while being hinted with some “seed” ideas.
Section 3 discusses the design and implementation of Creative-Pad. Section 4
concludes the paper with a general discussion of future work and the lessons learned
from developing Creative-Pad.
2 Desiderata
The first desideratum, ideation, emphasizes a process which has been well observed
to be an inherent part of creative thinking, namely the ability to generate/discover new
ideas. Much has already been said in the literature regarding the way in which ideas
emerge in a creative process, first as a set of divergent ideas and then as a set of
convergent ideas. However, we argue that the ideation process implemented for any
creativity support tool should focus only on generating a set of divergent ideas. In
particular, we consider the convergent part, for now, to be the responsibility of the
user. Partly, this is because we lack an understanding of how creative thinking arises
and partly this provides a clear goal for designing these tools. If not, the design of
these tools would become too intertwined with the two roles and thus becoming
unnecessary complicated.
Furthermore, the ideation process should be able to generate a new set of ideas
when used repeatedly and the ideas generated must somehow be able to inspire the
user to then work towards a creative solution. If not, the tool itself will be limited in
its ability to support creative thinking, both in terms of quantity and quality of ideas
generated. It is worthwhile distinguishing between creative thinking tools from
creativity support tools as defined here. The former incorporates methods (such as
Osborn’s [13] brainstorming, de Bono’s [5] lateral thinking, and MacCrimmon and
181
Wagner’s [11] techniques for “making connections”) which encourage users to come
up with new ideas themselves whereas the latter automatically generates ideas to
inspire/lead the users to develop a more creative solution for the problem on hand.
Developing such an ideation process suggests that its des ign needs to be crafted in
a way that combines both the need to have fresh ideas and ideas that are in creative
ways linked to the problem on hand. Consequently, attention needs to be paid to the
exact nature of the creative aspect of the problem for which the tool is designed and to
where the possible source of inspiration lies. Bonnardel [2, p.158], in analysing the
use of analogies in creative activities, also emphasized the importance of knowing
“the nature of the situations that can be used as sources of inspiration”.
The second desideratum, empowerment, relates to the tool’s usability. Following
from the first desideratum, it becomes clear that this should be about empowering the
users to develop his/her ideas freely. By “freely”, it is meant as little interruption as
possible to the user’s thought process. Again, this is important because we lack an
understanding of how creative thinking arises. The user is best left alone to develop a
creative solution.
Fig. 1. Two different views of a creativity support tool: (a) Creativity support
embedded in a problem-solving environment; (b) Creativity support view as
independent moments in a problem-solving process – after each session, the tool
could be updated (dotted arrow)
One way to ensure that this desideratum is met is to separate one’s initial thinking
process from one’s later “action” process. The latter is when the user begins to
implement his/her initial ideas fully. A creativity support tool when designed to assist
only the former activity would require an interface whereby the user literally has to do
nothing. In contrast, researchers who focus on developing a suitable environment to
support creative thinking (example, [6]) often tend to provide many tools or a tool
with many functions. These environments often allow the users to experiment with
alternative ideas prior to moving on to developing the final idea. Creativity support in
this latter case is very much embedded as part of the problem-solving process (see
Fig. 1a) whereas in our approach, creativity support is very much an independent
process (see Fig. 1b). It captures the moments when users take time out to think about
certain aspects of the problem. There could be more than one such moment and each
might require the use of a different creativity support tool.
(a)
Creativitysupport
Problem-solving
environment
Possibly different
creativity-support
tools
Problem solving process
(b)
182
3 Creative-Pad: Design and Implementation
The process of creating an advertisement could succinctly be described as having
three key elements, namely a message, an idea, and an execution. A creative first
draws out a message from the brief describing the product. Then he/she develops an
idea which in turn is executed to produce an advertisement. It is evidently clear that
much of the process for finding some initial good ideas for developing an
advertisement involves words association. It is interesting to note that this method is
also one of the simplest and most popular methods used for generating ideas in many
of the commercial creative software. For example, IdeaFisher is one such program
and it has an idea bank of more than 700,000 words associations to generate ideas for
its user. However, creatives working in this area have often noted that if the word
associations are generated in a manner unrelated to the problem on hand, there is a
danger that the ideas generated may not be of much use [15]. Furthermore, Poltrack
[15] noted that for advertising, related ideas should come from “all corners of life”
and one need to “stay tune with the world”. These observations suggest the need for a
rich source of contemporary ideas.
Without doubt, one rich source of such contemporary ideas is the World-Wide-
Web, or in short, the web. It also has the added advantage of being a huge resource
which is readily available, constantly updated, and information is literally coming
from “all walks of life”. However, using search tools to retrieve information from the
web often produce an overwhelming amount of information [9]. Consequently,
researchers are constantly designing new ways to help filter the information. Of
particular interests here are Otsubo’s Goromi-Web [14] and Koh et al.’s
combinformation [8]. The former extracts and displays keywords that frequently
appear in the search results and also displays images and blocks of text as floating
images on the screen. The latter extracts text and image clippings from the found
documents. These clippings are then presented to the user in a composition space as a
group of related ideas and with which the user can interact. For instance, he or she
could mouse over it to retrieve further information.
However, unlike Goromi-Web and combinformation, an ideation process for
Creative-Pad must not detract the users unnecessary. Consequently, we design a
process which does not engage the user with its search for ideas. The user simply
enters the keywords from the message and the extraction of ideas is then done
automatically. Our current ideation algorithm is as follows:
1. User enters keywords from message;
2. Send request to search engine (currently using Altavista.com) for related
information on the web;
3. Extract links from the results returned and download Html files from each link;
4. Extract all sentences from the Html files that contain the keywords;
5. Extract “interesting” words from these sentences; for now, a simple algorithm is
used. We extract all adjectives and verbs found in these sentences.
183
Fig. 2. Creative-Pad Interface
Thus, the ideation process produces as ideas, a set of words and sentences. To
provide an environment conducive for the user to develop his/her own creative ideas
for an advertisement, we developed an interface whereby the demand on the user is
again kept to a minimum. Basically, Creative-Pad projects ideas onto the screen and
the user constantly works on his/her ideas with little interruption. Each session using
Creative-Pad consists of the following steps:
1. Words, in randomly assigned colour, are beamed to the user one at a time and up
to a maximum of 150 words per screen (see Fig. 2a). While the words are
presented, music is being played in the background Note that words are
deliberately presented in an overlapping fashion to discourage the user from
reading the words.
2. As each word appears, the user could select it if he/she finds it interesting or
inspiring. This is done by either typing the word at the box at the bottom of the
screen or clicking the save button (see Fig. 2a).
3. When the maximum number of words is displayed, the screen pauses. The user
is given time to work on any ideas brewing in his/her mind.
4. If there are still words to be displayed, the user can repeat step (2) or if the user
has had enough ideas, he/she could go to the next step.
5. The user-selected words are then re-beamed to the user together with some
randomly generated words. The latter is added partly to add some random ideas
and partly to increase the number of ideas in case the user has selected
insufficient words.
6. When all the words are displayed, they will be rotated for a few seconds and redisplayed
in random positions. During this time, the user is supposed to continue
developing his/her ideas. He/she might record his/her ideas by typing a succinct
sentence at the bottom of the screen.
7. When the user has had sufficient time developing his/her ideas, he/she will press
the continue button to move to the next step.
8. Creative-Pad will then display sentences to the user, moving them from the top
right corner to the bottom left corner of the screen. These sentences contain the
user selected words and each sentence is numbered (see Fig. 2b). The user could
(a) Save button (b)
184
select those sentences of interests by entering their number at the box at the
bottom of the screen.
9. Finally, the user views all words and sentences that he/she has selected or
created. He/she continues to develop his/her ideas further.
10. When the session concludes, a report will be generated which contains all the
words and random words generated during this session and all the words,
sentences and idea sentences generated by the user during this session.
Fig. 3 shows our basic model for Creative-Pad. Step 1 is the ideation process which
consists of retrieving information from the web using some keywords from the
message and then processing the information for relevant ideas. In Step 2 ideas are
presented to the users in four phases. Firstly, the retrieved words, W1…Wt, are
presented to the user to help evoke possible ideas for advertisement. User at this stage
only needs to select the words that are of interests to him/her. Secondly, the selected
words, Ws1…Wsm, will be re-displayed and if necessary, together with some randomly
generated words, R1…Rp. User at this stage will have a chance to refine his/her
thoughts further and describe each of them with single sentences. Thirdly, sentences,
S1…So, which contain the selected words, Ws1…Wsm, are displayed as exemplar of
ideas using those words. User at this stage can select sentences which are of interests
to him/her. Finally, both selected words and sentences are displayed and user can
continue to refine his/her ideas.
Fig. 3. Basic model of Creative-Pad
Several experiments were conducted where creatives were asked to use Creative-
Pad to develop ideas for some imaginary advertisements. Details of these experiments
are reported in [12]. One such experiment with two creatives from two different
advertising companies is reported here. The experiment was conducted as follows:
Duration - 30 minutes; Instruction given - You will need to enter the phrase “car +
family + space” into Creative-Pad and during the experiment, develop ideas for an
advertisement with the message “a car with more family space”. At the end of the
Step 2:
Message WWW
Step 1:
S1….Sn W1….Wt
Ws1….Wsm W1….Wt
+ R1….Rp
report S+ Ws S1….So
185
experiment, use an A4 sized paper and a black marker to draw or write whatever you
think you need to describe your concept.
Note that although ideas generated using Creative-Pad is not intended to be used
immediately to generate a graphic display, we nonetheless asked the creatives to do
so. This might help us to understand better the ideas currently in the mind of the
creatives. Ideally though, they should put the ideas in the drawer and re-visit them
later to produce the advertisement required. Figs. 4a and 5a show the words and
sentences selected, and the ideas generated by the user (arrows indicate a possible
connection between the idea generated and the words/sentences selected). Figs. 4b
and 5b show the graphic output of a possible ad.
Fig. 4. Result obtained using keywords: car, family, space
(a) (b)
Selected
words
User’s ideas Selected
Sentences
Buy
Interior
Buying
Camping
Rental
Comfortable
Compact
Feel
Top
Easy
Speed
Owned
Search
Appeal
Set
Cheap
National
Personal
Excellent
Ideal
automotive
Sleeping in car
Cheap
accommodation
Built in stove
Has a shower
Sleeps 2 people
Double bed
Seats fold away
Large space
Camper car
Roof expands
Transforms to a
camper car
Electric hybrid
Pros: Great
Price,
Comfortable
Ride, Lots of
space, Great
family car
You feel very
little movement
in the car and it
wouldn’t be a
terrible
exaggeration to
say that you
could easily fall
asleep in the
backseat.
Who should buy
this car
Best car I’ve
ever owned
186
Fig. 5. Another result obtained using keywords: car, family, space
For Fig. 4, two themes emerge from the words selected by the elective: room and
them. Interestingly, the word “room” appears in one of the selected sentences. The
number of words displayed was: 234 adjectives, 43 verbs, and 30 random words. The
creative selected 9 words and 2 sentences, and generated 4 ideas. For Fig. 5, the
emerging theme is a camper car. The number of words displayed was: 283 adjectives,
78 verbs, and 20 random words. User selected 22 words and 4 sentences, and
generated 2 ideas.
4 Discussion and Conclusion
Given that creative thinking is still very much a mystery process, we argued that
the development of creativity support tools should focus on:
1. Ideation – the tool must be able to generate a new set of ideas when used
repeatedly and the ideas generated must somehow be able to inspire the user to
then work towards a creative solution. To achieve the latter, we find words
which have some relevance to the problem on hand.
2. Empowerment – the tool must be designed to support the user to be creative in
deriving a solution to his/her problem. It aids the “thinking” part as opposed to
the “action” part in solving a problem and it should afford minimum interference
to the user’s thought process.
With the above properties, a creative support tool could generate ideas that are more
relevant than one which supports creative thinking in general and yet it supports
finding a solution but not solving the problem itself.
The web undoubtedly provides a rich source of information for the ideation process
in Creative-Pad. However, unlike many of the earlier approaches which use this
resource, we do not involve the user to search and filter the information retrieved. The
ideation process automatically extracts what it believes is useful and the user’s role is
to focus on developing a creative solution based upon the information presented to
Selected word User’s ideas Selected Sentences
More
Extra
Larger
Designed
Organize
Stowed
Uses
Groundbreaking
full
More room for
your precious
cargo
Designed with
them in mind
(family)
It’s all about
them
It’s like a spare
room
The firm claims it
combines the
practicality of a
larger car with the
appeal of a
supermini
An extra flex room
gives you plenty of
creative ways to use
this space
(a) (b)
187
him/her. In this implementation, a simple algorithm is used for extracting words from
it. Nonetheless, the resource is so rich that our simple mechanism proves to be
adequate, in the sense that the creatives found the information useful and interesting.
In addition to the words, we also present sentences that contain the words that the
creatives have chosen. One surprising finding is that many of them found that the
sentences generated interesting. One possible explanation is that these sentences are
like tidbits of news or opinions or commentaries related to the product on hand and
thus they kept the creatives informed. They enable the creatives to “stay tune with the
world” as if he is gathering the information from a stroll down the street. The
creatives could develop an advertisement to position the product in light of these
tidbits of information.
Our initial experiments with Creative-Pad are not intended to test the effectiveness
of Creative-Pad in helping the creatives to generate creative ideas. As noted earlier in
the introduction, such a test is ill-defined, at least for the moment. Rather, they are
designed to test whether the ideas generated are sufficiently interesting for the
creatives. This, we argue, proves to be the case. Creative-Pad could now serve as a
platform for developing and testing these algorithms further. However, a closer
inspection of the words generated shows many of the words are judged not
interesting. Furthermore, if we treat each word generated as an idea for the creatives,
Creative-Pad has generated on average 200 words per experiment and the creatives
chose, on average, about 10-15 words. Does the ideation process need to be more
“thoughtful” in generating ideas? The initial phase of our ideation process closely
resembles a brainstorming session; an extremely popular idea in creative thinking and
which has influenced our implementation of the ideation process. Is brainstorming,
where quantity of ideas is of essence, the only and best way to interact with the
creatives? Or, would it be better to present fewer but better developed ideas?
Creative-Pad would provide a suitable platform to experiment with these different
alternatives in the future.
In our framework, one way to empower the user to think freely is to identify
clearly different areas within the problem where creative thinking is needed and then
develop separate creativity-support tools for them. In advertising, there are two
different sub-problems, namely getting an idea for the advertisement and developing
the final advertisement itself. Creative-Pad is successfully developed for the former.
In implementing the interface, we provide ample of time for the users to develop
his/her ideas. However, this is done in an ad hoc fashion; more study is needed to
develop an interface that better suits the way creatives work.
In summary, creative thinking is such a remarkable feat of the human mind that
researchers attempting to develop tools to support such an endeavour must develop a
multitude of approaches for experimentation. One such approach is being presented in
this paper whereby the focus is on developing an ideation process and an interface
which requires the user to almost do nothing except to focus on generating creative
ideas for the problem on hand. Using this approach, a tool for advertising creatives
has been developed and tested successfully. The tool now provides a platform for
future experimentation and discovery about creativity support tools in general and
creativity support tools for advertising in particular. Much more research needs to be
carried out to establish whether these two desiderata are essential for all such tools or
whether we need different desiderata for different tools. If the latter, what are these?
188
<references_biblio/>
References
1. Boden, M.A.: The Creative Mind: Myths and Mechanisms. Weidenfeld and Nicolson,
London (1997)
2. Bonnardel, N.: Creativity in Design Activities. In: Creativity and Cognition, pp. 158—165.
ACM Press (1999)
3. Chen, Z.: Towards a Better Understanding of Idea Processors, Info. and Soft. Techn. 40,
541—553 (1998)
4. Csikszentmihalyi, M.: Creativity – Flow and the Psychology of Discovery and Invention.
Harper Collins Publishers, New York, (1996)
5. De Bono, E.: Lateral Thinking. Harper Colophon Books, New York (1973)
6. Fisher, G.: Domain-oriented Design Environments. Automated Soft. Eng. 1, 177--203
(1994)
7. Johnson, H., and Carruthers, L.: Supporting Creative and Reflective Processes. Int. J. of
Human-Computer Studies 64, 998--1030 (2006)
8. Koh, E., Kerne, A., and Hill, R.: Creativity Support: Information Discovery and
Exploratory Search. In: Proc SIGIR 07, pp. 895--896, ACM Press (2007)
9. Kules, B. Supporting Creativity with Search Tools. In: NSF Workshop Report on Creativity
Support Tools, pp. 53--63 (2005)
10. Lubart, T. How Can Computers be Partners in the Creative Process. Int. J. of Human-
Computer Studies, 63, 365--369 (2005)
11. MacCrimmon, K.R. and Wagner, C. Stimulating Ideas Through Creativity Software.
Management Science, 40, 1514--1532 (1994)
12. Opas, T. An Investigation into the Development of a Creativity Support Tool for
Advertising. PhD Thesis, Auckland University of Technology, New Zealand (2008)
13. Osborn, A.F. Applied Imagination: Principles and Procedures of Creative Thinking.
Scribner’s, New York (1953)
14. Otsubo, G. (2007) Goromi-Web: Browsing for Unexpected Information on the Web. In:
Creativity & Cognition 2007, pp. 267—268. ACM Press (2007)
15. Poltrack, T. Stalking the Big Idea: Some Meditations on the Creative Process. Agency
Magazine Spring (1991)
16. Resnick, M., Myers, B., Nakakoji, K., Shneiderman, B., Pausch, R., Selker, T., and
Eisenberg, M. Design Principles for Tools to Support Creative Thinking. In: NSF
Workshop Report on Creativity Support Tools, pp. 25--36 (2005)
17. Shneiderman, B. Creativity Support Tools. CACM, 45, 116--120 (2002)
18. Shneiderman, B., Fischer, G., Czerwinski, M., Resnick, M., Myers, B., Candy, L.,
Edmonds, E., Eisenberg, M., Giaccardi, E., Hwett, T., Jennings, P. and Kules, B. Creativity
Support Tools: Report From a U.S. National Science Foundation Sponsored Workshop. Int.
J. Human-Computer Interaction, 20, 61--77 (2006)
19. Sternberg, R.J. The Nature of Creativity: Contemporary Psychological Perspectives.
Cambridge University Press (1988)
20. Wallas, G.: The Art of Thought. New York: Harcourt Brace (1926)