iStar models allow users to capture high-level social requirements, including alternatives and qualities. Typical approaches advocate discovering model content through traditional Requirements Engineering (RE) elicitation methods, such as interviews, observations, and domain documents. For the typical, familiar space of requirements, such techniques may be sufficient, but stakeholders may not always know the best way to meet their needs, and may have difficultly articulating their goals. Creativity theories and activities have been successfully applied to RE in order to discover both creative and more typical requirements. In this work we describe the web-based Creative Leaf tool that facilitates a combination of established creativity techniques with iStar modeling. Divergent techniques help the discovery of ideas, while selection, evaluation and modeling techniques help to converge over ideas, integrating them into the model and system design.
Goal models, such as iStar, provide a concrete graphical way to help ensure that
potential system requirements meet user goals and qualities. But where do potential
requirements come from? How do we find the goals and desired qualities of systems
stakeholders? Traditional Requirements Engineering (RE) approaches advocate for
various forms of elicitation (interviews, observations, domain documents). For the typical,
familiar space of requirements, such techniques may be sufficient. But users and
stakeholders may not always know the best way to solve their problems, to meet their needs,
or even to identify what those needs may be. Stakeholders are biased by their own
experiences, and are often too-grounded in the use of existing tools [
The role of creativity theories and techniques has been investigated and applied in
an RE-context for more than ten years, e.g., [
Copyright © 2016 for this paper by its authors. Copying permitted for private and academic purposes.
seen success when applied to real contexts such as Air Traffic Control and Food Safety,
workshops are expensive to run, require highly-skilled facilitators, and produce output
which is often fragmented and only lightly structured (ideas, scenarios, use cases) [
In order to address many of these drawbacks in both goal modeling and creativity approaches as used in RE, our work has combined the two approaches, producing a concrete tool and guiding methodology: Creative Leaf 1. Our overall intention is to facilitate the structured capture of creative requirements – those requirements which are not obvious, typical, or which do not immediately come to the mind of the stakeholders.
Creative Leaf is an output of a design research process involving a series of
formative and summative studies conducted over the last 1.5 years. We leave the reporting
of these studies to further work, and in this short paper focus on describing the
resulting tool. We have previously published initial visions of how the tool and method may
work, using historical examples [
Section 2 describes the design and functionality of the tool. Section 3 concludes the paper and makes a call for participation in the use and evaluation of Creative Leaf. 2
A screenshot of Creative Leaf is shown in 1. The left hand panel contains a palette with
iStar elements. The included elements and links were based on an early version of the
iStar 2.0 core [
The iStar modeling capability of Creative Leaf is based off the Leaf Beta tool developed at the University of Toronto by Alicia Grubb et al.2. The Leaf code has been modified and adapted, e.g., to add ideas and assumptions. (Creative) Leaf uses both the JointJS and Rappid3 HTML 5 diagramming frameworks. JointJS is available via an open
2 http://www.cs.toronto.edu/˜amgrubb/leaf
license, while Rappid is a commercial product, but provides free academic licenses. Code is written in JavasScript, using CSS and HTML files for formatting and display. Creative Leaf code is stored on a private GitHub repository, forked from the Leaf code base. Access can be made available to other institutions which acquire a license for Rappid (all other sections of code apart from the Rappid libraries are Open Source and are available on request).
In addition to modeling and creative functions, the tool has an embedded tracking function which tracks user actions with time stamps. Users have the option to turn off or on this tracking at first use, and later via the Options menu. Creative Leaf uses a series of unobtrusive prompts to guide users to the next likely activity, particularly when the user is inactive for some period of time. These prompts can also be disabled via Options. 2.2
The Creativity panel on the right of the tool allows users to work through a number of
established creativity activities. Following the model of the Creative Problem Solving
(CPS) method, creative activities are aimed to support either divergent creativity,
generating ideas, or convergent creativity, selecting, combining and developing ideas [
Exploratory creativity activities in Creative Leaf include Brainstorming, CRUISE
creative search, Pairwise Comparison, Creativity Triggers, and Bright Sparks (Hall of
Fame). More information about the general form of these activities can be found within
the BeCreative Creativity Support tool [
Brainstorming is the very general activity encouraging people to capture all the “obvious” ideas, which they already have in mind. Studies showed that without this initial activity, users ignored the structure and prompts of further activities and only listed the ideas they already had. The CRUISE activity calls an external creativity service, a result of the Collage Project 4. The input search string is a selected iStar element, actor or idea. The CRUISE service returns a cloud of images, links and text which are tangentially related to the search string. Ideas prompted by the output are captured at the bottom of the window. An example screenshot of the CRUISE activity can be seen in Fig. 2.
The Pairwise Comparison activity, inspired by Poincare’s emphasis on creativity
through connections [
makes a call to an external service 5. This activity takes an actor as input and asks the user to think of the actor being played by a famous persona (e.g., Batman, Joan of Arc, Henry Ford), allowing the user to iterate through personas. 2.3
Once ideas have been discovered, the user needs help selecting amongst ideas and
developing the selected ideas by integrating them into the iStar model. We offer a selection
activity which guides users through the process, see the text in Fig. 3a. Part of the
selection activity introduces the Hover Evaluation feature included in Creative Leaf. This
feature is based on qualitative evaluation from the first author’s previous work [
The final activity encourages users to model their best ideas in iStar, manually converting their idea to a number of iStar elements and links, connecting them to the existing model. This step is currently fully manual, but we are looking into providing users with suggested mappings of their ideas into iStar elements. Once an idea has been modeled, it is marked as modeled and removed (automatically) from the canvas. All ideas, including rejected and modeled ideas can be found via the idea list, accessible via a button on the top right of the tool. More information, including tutorial videos and a user guide can be found in the following support site: https: //sites.google.com/site/creativitygm4re.
(b) Hover Evaluate Feature: Evaluating Idea “Sell adds for shipping” (a) Selection Activity Instructions
We have introduced the reader to the Creative Leaf tool, combining goal modeling with established creativity techniques. We encourage users to try out the tool, and provide feedback to the authors at creativeleaf@city.ac.uk. After extensive in-lab evaluation of the tool, we are looking for organizations with real problems willing to try out the tool and provide feedback. We hope to use this realistic evaluation to improve the functionality and general applicability of the tool.
Acknowledgments. This research is supported by an ERC Marie Skodowska-Curie Intra European Fellow-ship (PIEF-GA-2013-627489) and a Natural Sciences and Engineering Research Council of Canada Postdoctoral Fellowship (Sept. 2014 - Aug. 2016).