I. INTRODUCTION
The impact of software systems on social and economic
activities increases each year, which makes the analysis of
sustainability requirements of the software systems more and
more important. Becker et al. [
Requirements engineers still have a narrow understanding of sustainability; they focus on one or two dimensions of sustainability such as environmental and economic sustainability, or on the non-functional requirements for technical sustainability such as maintainability and reusability requirements. Requirements engineers should take into account sustainability requirements, which implies additional analysis in the early stages of software development to maximise the positive impact and to minimise the negative impact of all sustainability dimensions. To support these activities a tool is required. Contributions: In this paper, we analyse the core features of
The tool has to be easy-to-use and allow involvement of SuSoftPro (abbreviated from Sustainable Software Profiling) in
stakeholders from diverse groups in the process of software comparison with two approaches which employ Multi-Criteria
development, as empowering more participants with a diversity Decision Analysis (MCDA) in requirements engineering. We
of perspectives leads to more sustainable systems [
There are many requirements engineering (RE) tools to elicit, (SCIS), which is a clinical software to store patient health analyse, model, trace, document, manage, as well as verify records, to optimise SuSoftPro and to increase its usability. and validate software requirements. Some of these tools are Outline: The rest of this paper is organised as follows. web-based, which allows collaborative access to resources. In Section II, we analyse and evaluate the core features of However, none of them has the ability to analyse sustainability SuSoftPro by comparing against two RE approaches which a requirements by involving stakeholders. Thus, the goal of our different methodology using MCDA. Section III illustrates the ongoing work is to evaluate Sustainability Profiling for Software work flow using a case study from the eHealth domain. Further (SuSoftPro) for analysis of sustainability requirements. More discussion of the SuSoftPro features is presented in Section IV. precisely, we are aiming to analyse the SuSoftPro method in- Section V covers related works. Finally, Section VI summarises depth and evaluate the tool-support through comparison and the paper and outlines the directions of our future work.
In our previous work, we introduced a methodology with
supported tool to analyse sustainability requirements for
longliving software systems [
Our new tool presents sustainability as a five-star rating label, a
visualisation for the degree of the five sustainability dimensions,
and a bar graph which illustrates the overall sustainability
level for each requirement. The tool enables requirements
engineers to defining stakeholder groups allotted to one or more
of the five sustainability dimensions, building a fuzzy rating
scale-questionnaire with regard to a sustainability dimension,
specifying the high-level requirements and assign them to
created groups, assigning stakeholders and allow them to rating
requirements, analysing sustainability, and generating software
sustainability profiling [
To analyse and evaluate our methodology and the SuSoftPro tool, we compared the SuSoftPro against two approaches that developed a methodology using MCDA and used for requirements engineering domain. Below, we address the procedure, analysis, and result of the comparative evaluation.
To perform comparative evaluation against SuSoftPro, we defined three criteria for selecting frameworks from literature studies as follow:
Scope: Developed for requirements engineering context, Process: Involved MCDA and stakeholders, and
Objective: Analysed sustainability.
Because there was no study meet the three criteria, we removed the objective criteria (analysing sustainability) because no MCDA technique has been used to analyse sustainability within RE as well as sustainability is a new growing topic in RE. Thus, only two frameworks were found: ReproTizer and sureCM.
Then, we specified 9 sub-criteria to analyse the three frameworks (SuSoftPro, ReproTizer and sureCM) including the purpose of the methodology, collection method, weight scale, analysis method which is one type of the MCDA, participant, rank updates that the methodology can instantly re-compute results, supported tool, computational complexity, error-proneness, prevent imprecision inherent in human responses, number of criteria.
We analysed SuSoftPro’s core process and features with regard to two other frameworks for requirement analysis: ReproTizer and sureCM.
1) SuSoftPro: SuSoftPro1 is a methodology and tool-support
to analyse sustainability requirements within sustainability
dimensions having individual, social, economic and
environmental dimensions. The general idea of the SuSoftPro process
is presented in Figure 1 and discussed in [
Define stakeholder groups: through creating stakeholder group based on stakeholders' role or expert, and then
assign this group to one or more of the five sustainability dimensions, so the group will be allotted to stakeholders and requirements; Define questions: that will be generated automatically as five instructions with regard to a sustainability dimension for FRS questionnaire; Define requirements: via the specifications of the highlevel requirements and allotting them to related groups affected stakeholders and requirement ownership; Assign stakeholders: to related groups based on stakeholders' role in the system and their areas of expertise after defining them; Rate requirements: through enable stakeholders to use ratio quantity approach as FRS responses; Analyse sustainability: with MCDA using TOPSIS approach to determine the level of sustainability dimensions and sustainability requirements measurements; and
five-star sustainability rating label, visualisation of sustainability dimension levels, and bar-chart graph for each sustainability requirements level.
Sustainability profiling provides insight and identifies the predictability of sustainability to enable requirements engineers and stakeholders to analyse and break true requirementsdependencies, and the interaction and overlapping of sustainability dimensions by predicting the outcome value before developing software systems.
2) ReproTizer: ReproTizer was elaborated by Achimugu
et al. [
tisation automatically are determined using WADM, after scoring requirements by stakeholders; and Artifact exports from requirements projects
requirement prioritisation is presented in ordered list.
3) sureCM: sureCM was introduced by Mairiza et al. [
As shown in Table I, both SuSoftPro and ReproTizer work
with more than two criteria for analysis, and are supported
by a tool, providing a fully systematic computation to prevent
errors. The sureCM framework is based on a semi-automatic
computation and data collection (requirements rating), which
are more error-proneness than a fully automated solution.
Another advantage of SuSoftPro is utilising the FRS, which
allows better precision of requirements' rating. Although the
FRS application provides more accurate scale than Likert scale
to capture real-valued responses, the FRS is not fully
friendlyto-use scale [
Neither ReproTizer nor sureCM support the sustainability context whereas SuSoftPro supports and utilises a fully systematic and comprehensive discovery methodology to analyse sustainability requirements.
The core results of the comparison are summarised in Table I.
To illustrate how SuSoftPro can be applied, we used a case
study from the eHealth domain, based on a real-life project,
a Skin Cancer Information System (SCIS), cf. [
2) Nurses, 3) Receptionists, 4) Administrators and Managers, and 5) Developers and IT Support.
Requirements engineers have selected 14 stakeholders (two physicians, two nurses, four receptionists, three administrators and managers, and three developers and IT supports). Five groups (corresponding to the stakeholder roles) are created and assigned to sustainability dimensions, cf. Table II. Groups are used not only to group stakeholders with related sustainability dimensions but also to associate requirements with related stakeholder groups. For instance, nurse group is assigned to individual and social sustainability dimensions as well as each requirements affecting or related to nurses are assigned to nurse group.
SuSoftPro generates questions/instructions according to the following format: “Rate the influence of the requirement on the X sustainability”, where X is replaced in a concrete case by the corresponding sustainability dimension: individual, social, technical, economic, and environmental. There is an option to adjust each question, but we decided to continue with the generated questions for our case.
A 23 high-level requirements specification of the system
in [
The user profiles for the stakeholders are created and then assigned to the groups, as shown in Figure 2. Therefore, each group is assigned to related sustainability dimensions, requirements and stakeholders. In other word, stakeholders are grouped and designated to related sustainability dimensions and requirements. Adjusting stakeholder details are automatically prevented when stakeholders start responding to the questionnaire. For example, a nurse begins answering the questionnaire, the change of the group and other related details are frozen.
After building the questionnaire, generating and sending autosign-in link to the stakeholders to access the questionnaire, the status of all the stakeholders in the project becomes waiting, until they begin to respond to the questionnaire. For each high-level requirement to be rated, the stakeholder can rate its influence on the sustainability dimensions using interface presented on Figure 3. In SCIS case to illustrate the flexibility of the tool, nurses have 30 questions to answer, where 15 questions are on the individual sustainability perspective for the 15 allotted requirements to physician and nurse group in the SCIS, and 15 questions for the social perspective of the same requirements.
Physicians have 45 questions: 30 questions are the same as for the nurse group, additional 10 questions on the economic perspective of the same requirements.
SCIS REQUIREMENTS WITH SUSTAINABILITY RATING WHERE 1 IS THE HIGHEST AND 0 THE LOWEST POSSIBLE RATING 16 There are 24 questions covering the individual, social, and economic perspective for requirements related to receptionists.
Administrators and managers are assigned 52 questions to answer for administration and managements requirements covering the following perspectives (13 questions each): economic, technical, social, and environmental perspectives. Developers and IT people have 92 questions for all the requirements covering 23 questions on each individual, technical, economic, and environmental sustainability perspective, see Table III.
Guidance on how to use the FRS is provided for stakeholders, so stakeholders such as nurses or physicians, who have not seen or used the FRS before, will be easily guided. They also had
Assigned Group the ability to save their responses and return back to continue.
An option for skipping any question for certain requirements within particular sustainability dimension is implemented. For example, a physician was asked to rate the influence of Req. 6 "Insert procedure" on economic sustainability; the physician was able to skip this question. However, the question has a probability to be answered by other stakeholders such as other physicians and developers who are assigned to rate Req. 6, for the economic dimension.
Only the individual dimension is in a satisfactory range which is more than 0.60 (the corresponding bar in the chart is light green).
The technical, social and economic dimensions are between 0.54 and 0.58 (the corresponding bars in the chart are yellow).
The environmental dimensions is in an unsatisfactory range which is around 0.35 (the corresponding bars in the chart are orange).
The sustainability value of each requirement is indicated in
Table III. The value in the result is between 0-1 where in the
TOPSIS method 0 represents the worst ideal solution and 1 is
the best ideal solution [
As the next step, SuSoftPro applies the TOPSIS method and creates the sustainability profiling of the system. A systematic computation of TOPSIS is performed and recalculated when IV. DISCUSSION each stakeholder submits their response. Also, rated require- SuSoftPro is an automated solution in the sustainability ments with its questions are automatically locked when any context to analyse sustainability requirements based on quesstakeholder begins to rate it, so engineers can not amend them. tionnaire, in which quantity data gathers via FRS questionnaire and analyses using TOPSIS. The result presents as sustainG. Generating Software Sustainability Profiling ability profiling for software having a five-star rating label,
The created profiling presents in the dashboard in Figure 4. visualisation of the degree of sustainability dimensions, and
Based on the simulated responses we used to illustrate the bar graph of overall sustainability level for each requirement.
example (where only 13 out of 14 stakeholders submitted their From the comparative evaluation result, both SuSoftPro and
ReproTizer approaches are based on individuals perspective.
responses), the overall sustainability of the SCIS has FFF
three-star rating (3 out of 5). The five sustainability dimensions The perspective is important to change sustainability of software
are presented by a bar chart: when users’ opinions are addressed and taken into account.
Scholars of social practice theory believed that practices and
perspectives of individuals in the performance of daily tasks
stimulate social, economic and environmental changes [
The SuSoftPro tool aggregated all stakeholders' requirements.
This enables the recognition of diverse visions and voices into decisions that are needed to develop sustainable software.
Thus, the point of sustainability perspective while analysing requirements could be the main force in providing sustainable software in the early stages.
Besides, providing FRS in SuSoftPro to capture
individuals views was necessary to prevent imprecision. However,
there is the need for reconciling plurality through supporting
stakeholders with the diversity of points of view that ensure
sustainability [
As the case study demonstrated, practitioners were supplied
with information related to sustainability aspects. The
sustainability profiling presented sustainability scores for each
requirements and sustainability dimensions. These scores will
improve the sustainability attention and allow practitioners
to provide sustainable software. For example, the lowest
sustainability scores in SCIS was Req. 9 "Allocate pathology
report to procedure", so practitioners could give more attention
to improve this requirements and acceptance as well as increase
users satisfaction which lead to sustainability [
Additionally, the tool allows requirements engineers to create groups with regard to stakeholders diversity or role.
For example, groups in SCIS profiling were divided into user role. Grouping stakeholders and requirements are not only to reduce the number of questions but also to express their opinion
Fig. 4. The SuSoftPro Dashboard for the SCIS profile about what is related to them. Also, there are two ways to invite stakeholders either with a public link to accommodate more stakeholders with self-registration or being registered by the engineers.
Two different colours are provided for practitioners with
colour-deficient vision in the tool. The red colour is replaced
with blue when the colour-deficient vision is opted. This option
ensures better accessibility and equally user experience to read
sustainability profile because one in every 12 people has colour
vision deficiency [
Intuitive design is taken into account during designing the tool. For practitioners, the tool divided into logic sections including a dashboard, questionnaire, requirements, stakeholders, and profiling. A systematic computation of stakeholders responses after submitting is implemented to prevent error.
Icons and colours also are provided for an effortless
understanding of the tool. However, stakeholders may face difficulty
to understand the FRS when they start to respond [
The tool also allow integration with commonly used requirements engineering tools such as ReqMan and Rational DOORS: Its export and import features allows the exchange of requirement specifications using the CSV format.
SuSoftPro has emerged to:
Capture more individuals perspective with the diversity and accurate impression, Analyse software requirements in sustainability context, and
Present the result as a sustainability profiling.
However, a few limitations need to be taken into account.
There is need to provide a standards for sustainability fivestar rating label to specify the minimum level of sustainability performance that software should meet before they can be developed. Also, when the number of requirements is increased and a group has assigned more than two sustainability dimensions, the number of question will be large either double or treble requirements. This large number could lead to take a long time for responding to a questionnaire, so stakeholders might feel more annoying. An initial optimised solution, requirements engineers can divide a group that allotted to more than two sustainability dimensions into two groups and then assign them to one or two different sustainability dimensions. Another solution is to leverage machine learning to assign stakeholders and divide questions between one group. We did optimise the number of questions in the tool through establishing a group and assign stakeholders and requirements to it. This solution assists to reduce the number of questions about 20-50% in some cases.
V. RELATED WORK
Some works on embedding sustainability in the software
development process, e.g. [
Porras et al. [
Mahaux [
Hence, involving supported participants will ensure sustainable software. This argument emerges the need of a tool involving supported participants easily, and the SuSoftPro is developed to involve supported participants vis providing their perspective as support.
Al Hinai [
Chitchyan et al. [
The lack of methodological support was one of the identified barriers to the engagement with sustainability design in RE practice. The SuSoftPro is a solution to overcome this barrier through engaging practitioners and stakeholders to analyse sustainability.
Becker et al. [
A number of requirements engineering tools with general
or specific features for eliciting, analysing, modelling,
tracing, documenting, managing, and verifying and validating
requirements [
VI. CONCLUSIONS
We evaluated SuSoftPro with two approaches that developed a methodology using Multi-Criteria Decision Analysis (MCDA) and used for requirements engineering domain. The evaluation demonstrated a number of advantages of SuSoftPro for the sustainability analysis: such as tool support, FRS to allow better impression of requirements' rating, and last but not least a systematic methodology to analyse the sustainability of the system under development.
Also, we explored the viability and demonstrated the usability and feasibility of SuSoftPro by conducting a case study from the eHealth domain, based on a real-life project: a Skin Cancer Information System to store patient health records.
Future work: We are conducting an empirical assessment of
SuSoftPro. The evaluation methodology is designed to capture
the views of professional practice experts in sustainability
requirements through qualitative approach, using qualitative
methods [
Then, these data will be analysed using predefined themes such as the usefulness of the framework and developed tool, and the potential of adapting the framework and tool. Furthermore, we are currently conducting two other case SuSoftPro studies from education domain, having two different eLearning systems and large number of participants in higher education institution from two countries.
ACKNOWLEDGEMENTS
The first author is supported by a scholarship from Umm Al-Qura University, Saudi Arabia.