RAPD: Rapid and Participatory Application Development of Usable Systems During COVID19 Crisis

Software development methods have been evolved to enable producing usable systems rapidly while considering all requirements. Several studies have focused on the need to balance between rapid development and capturing requirements related to user experience and business workflow. This balance has become more urging during COVID19 because many businesses want to quickly transfer to usable electronic systems that are accurate, efficient, easy to learn, satisfy users and support remote work. Therefore, this paper proposes a framework by integrating Rapid Application Development (RAD) method with Participatory Design (PD) method for enabling rapid production of usable systems. Both RAD and PD consist of design stages that can overlap and generate new phases where users participate in the design process and accelerate the production. Five usability tests are also added to the framework to validate the usability of the design at all stages. The Action Research method is used to assess the framework empirically in a context of an urgent need to an electronic system, and qualitative data analyses were conducted. The results show that the framework can be adopted by software companies because it satisfies the requirements of adopting software development methods. Also, the system developed using the framework is usable. The paper concludes that COVID19 affects software development by emphasizing rapid development while maintaining workflow. Also, using video conference for remote design assists in meeting users more frequently and in creating concise requirement documentation.

concise user stories, eliminates conflicts and reduces post-100 development modifications. We found that RAPD satisfies 101 the factors stated in [13] and [14] that enable the adoption in 102 the company. The users stated that they accepted the system 103 developed using RPAD because it supports remote work, and 104 it is a usable system satisfying all factors affecting usability 105 stated in [15] and [16]. Both, the company and the users stated 106 that participating in the design even through video conference 107 was time-saving and effective design process. They were able 108 to make design meeting at homes more frequent, review and 109 refine the design, and focus on the important elements with-110 out including unnecessary conversations. This research also 111 found that COVID19 has affected the software development 112 process and created a need for methods that focus on the new 113 work style while urgently responding to the crisis. 114 The contributions of the paper can be summarized as: -Post-crisis quantitative analyses identifying the usability 126 of RAPD as a development method and the usability of 127 the system developed by RAPD.

128
The rest of the paper is organized as follows: Section II 129 presents the theoretical background, Section III explains the 130 methodology, Section IV presents and discusses the results, 131 Section V presents the limitations of the research, Section VI 132 concludes the paper.

134
COVID19 has imposed new challenges to the software devel-135 opment by creating a critical situation for businesses due to 136 new work environment. This situation has forced businesses 137 to search for electronic systems that can be employed quickly 138 and support remote work while maintaining workflow. This 139 situation has also forced software development companies to 140 search for methods that balance between rapid development 141 and addressing all requirements. Therefore, the design frame-142 work proposed in this paper integrates three fundamental 143 methods that are Rapid Application Development (RAD), 144 Participatory Design (PD) and usability testing. These meth-145 ods were successfully used in [17], but the integration details 146 and influence of these methods on the development team and 147 end users were not discussed. process during all software development four stages: require-154 ments planning, user design, construction, and cutover [18]. 155 Iterative development means creating increasingly functional 156 versions of a system in short development cycles. Each verend-users and developers [25], [26]. So, PD is a user centered 207 design method in which users express their requirements 208 for building software that is usable [4], user friendly and 209 fits well with user culture, age, educational background and 210 communities [27], [28], [29], [31]. Designers interacts with 211 users through design meetings, interviews and participa-212 tory observation for collecting requirements and evaluating 213 designs. The design process in PD passes through three main 214 stages [26]: 215 1-Pre-design stage: designers and users decide the project 216 plan, objectives and schedule, and they select represen-217 tatives to perform the coming design tasks.  However, PD is criticized because it neither provides a 244 fully specified design process [32] nor structured and sys-245 tematic assessment of the design concepts [33]. Therefore, 246 PD requires integration with other practical development 247 methods, such as RAD and agile, to achieve a sufficient influ-248 ence on information system applications [5], [8]. PD also can 249 be integrated with usability testing methods to enable assess-250 ing products at each stage effectively [33]. Also, PD puts 251 emphasis on the early systems development phases which 252 delays production of ready-to-use system and causes design 253 sessions to consume much time [34]. The usability testing is borrowed from usability engineering 256 and allows users to perform real tasks so that they can evaluate 257 each developed block, and this helps developers produce 258 a ready to use system earlier [ show that an innovative solution should be compatible with 318 existing skills and practices, triable where results are exper-319 imented without extra effort and expenses, simple and used 320 with low complexity, observable where results can be seen, 321 and technically superior to its predecessors [14]. We also 322 analyzed the factors affecting the acceptance of users to the 323 systems developed by the integrated framework. We focus 324 on the usability of websites as we aim to support remote 325 work during crisis through web interface. Website usability 326 include efficiency, learnability, memorability, accuracy and 327 satisfactions [15]. We also focus on user satisfaction since 328 we aim to make the system acceptable. User satisfaction can 329 be measured by the website ease of use, content, delay, and 330 customization [16].

331
The theory of AR states that the solution of a specific 332 problem for the sake of any situation improvement requires 333 understanding on both the real-world context and the meth-334 ods used for achieving the improvements. So, this section 335 presents the context in which AR is used, and the activities 336 performed to understand the context and build a practical 337 solution to the problem in the context. This section also 338 explains the actions that should be made to integrate RAD, 339 PD, and usability testing for producing a product that is usable 340 and accepted by end users.

342
The research context includes the company which is respon-343 sible for developing the software and the market which needs 344 the software. 345

346
We refer to the name of the software company which partici-347 pated in the research and in the development as 'the company' 348 to ensure anonymity. The company is a software development 349 company located in Palestine and started this business in 350 2003. The company produced several software products and 351 participated in transforming the local commercial and finan-352 cial business from paper-based systems to electronics sys-353 tems. The company business size is small due to the limited 354 software market in Palestine. In 2017, the company shifted its 355 work from developing desktop applications to cloud services, 356 and started focusing on software services.

357
The company previously depended on the waterfall soft-358 ware engineering method for developing most of its prod-359 ucts. The company has started to adopt agile frameworks, 360 mainly RAD, due to its simplicity compared to other agile 361 frameworks. Also, RAD can be an answer to call for software 362 solutions that support remote work due to the COVID19 363 pandemic. In the last three years, the company started to add 364 PD to the development process, and PD was brought to the 365 company by the manager who lived and worked in software 366 development in other countries. The managers decided to use 367 PD as the company found that RAD alone did not capture 368 all requirements and the detailed workflow and therefore, the 369 company had to perform extra RAD cycles to add new tasks 370 to products and fix bugs. the company allocated a team for managing the development 373 process, and the team includes: a manager for signing con-374 tracts and participate in setting up plans, schedules and objec-375 tives; designers whose jobs focus on collecting requirements, 376 designing prototypes, testing and validations; and developers 377 who converts designs to working prototypes and products. All    [42], [43], [44], [45]. The clients in the 424 vegetable market justified that by the inability to deal with 425 new workflow imposed by new software. They stated that 426 farmers, traders, and other suppliers will be confused when 427 shifting from the existing work style, which they are used to, 428 to another style. These results agree with those in [46] about 429 software that fits with the industrial countries may not suit the 430 developing ones. While investigating this case, we found that 431 clients, farmers, traders, and suppliers want the same format 432 of bills, financial balance, and other documents. The clients 433 stated that is more useful if the company let them participate 434 in the design.

435
On the other hand, clients insist to see a prototype as fast 436 as possible for two reasons: they want to ensure that their 437 requirements are satisfied before paying any money for the 438 software producers, they find learning a complete system very 439 difficult so they prefer a step by step training learning fashion. 440 The need for rapid development is also emphasized during 441 crisis such as COVID19 which has imposed restrictions on 442 the market as shop workers, farmers and traders can not 443 meet face to face inside the shops to discuss transactions, 444 billing, and payments. These processes usually depend on a 445 long discussion between all parties and include paper printing 446 of all documents. For example, farmers and traders always 447 request paper bills containing all transactions of that day. So, 448 all these processes have to be transformed to digital format 449 during COVID19.

450
These findings were shared with the clients, and again 451 the company and the clients agreed on these finding and 452 decided to start the development of the software. The com-453 pany explained that the framework, which will be used during 454 the development, requires the client participation and involve-455 ment, and users agreed on that. The company team formulate the software development 458 framework that will be used in the context based on the AR 459 theory and on the initial process findings. The team names the 460 framework 'Rapid Application Participatory Development' 461 (RAPD) because it integrates both RAD and PD by overlap-462 ping the design stages so that advantages are strengthened 463 and disadvantages are eliminated. Also, RAPD allows adding 464 testing at any design stage. RAD requirements analysis stage is integrated with PD pre-467 design stage and PD stage 2-A to form RAPD stage 1. In this 468 stage, designers and users determines objectives, plans, and 469 schedules, and they form a team of designers and user rep-470 resentatives to collect the requirements, identify all entities 471 and draw all diagrams and interactions between functions and 472 data. To avoid delay in this task, a team leader is selected 473 to coordinate tasks among other team members who are 474 divided into members for collecting requirements, members 475 for putting the initial designs and diagrams, members for 476 documenting all details, and others for reviewing related 477 projects and using similar features for the new design. representative users are involved in the design and testing, 507 no training is needed from development team for the client 508 side.

509
At the end of this stage, a validation testing is performed 510 to evaluate the final system and ensure that all objectives and 511 plans have been sufficiently addressed. Further, each stage 512 also contains comparison testing for allowing participants to 513 look for alternative designs and solutions.

514
At the end of stage 3, all major requirements are addressed 515 in the design, and functions that are necessary to complete 516 tasks are ready. So, end users can perform their work and send 517 feedback to the design team. The team also revisits stage 1 to 518 consider the requirements that were postponed and consider 519 any emerging requirements. This section presents how the RAPD framework is applied 522 to the Jericho Central Vegetable Market. The process started 523 in April/2020 when the market made a call for a system that 524 transforms the traditional work routine to electronic one and 525 supports remote work. Four auctioning stores were selected, 526 and two of them have electronic desktop system and the 527 others are paper based stores. Three levels of users; managers, 528 accountants, and data entry employees, cooperate to build the 529 system. We refer to these users as direct users because they are 530 responsible for the management of the system. Meanwhile, 531 other users such as traders, farmers, workers, and supply 532 stores are indirect users because they can view transactions 533 and interact with the system after the direct users initiate 534 transactions. In total, the direct users include nine auctioning 535 shop owners and accountants, and indirect users include 120 536 farmers, 70 traders and nine agriculture supply stores.

537
COVID19 has also imposed restrictions on designers and 538 developers as they can not meet users face to face. So, 539 designers and developers shift to digital forms of communi-540 cation such as video conferencing for most of meeting and 541 workshops. This was not easy at the beginning due to lack 542 of technical experience of end users. But, once a prototype 543 is shown to users, they started interacting with designers 544 remotely for testing and providing feedback.

545
The RAPD framework passed through the three stages:      This stage lasted for 14 days including three virtual meet-622 ings, and the development team was able to build a prototype 623 for 90% of the entire system. The services related to indirect 624 users were also added to the prototype, and the services 625 developed in this stage are shown in Table 2.

626
Other low priority services, such as data analytics and 627 forecasting, were postponed because these services require 628 data. So, the development team waited until users inputted 629 some data. This postponing also ensured that the end users are 630 VOLUME 10, 2022 representatives, designers and developers     Five questions tested the reputation of RAPD among 687 developers' peers, their skills in PD and RAD, and their 688 motivations to learn RAPD even if they do not know PD 689 and RAD. 690

4-Willingness to participate in remote design based on 691
RAPD: four questions tested if the team is confident 692 that they master RAPD and ready to practice it for 693 developing software for other markets. This is impor-694 tant because most companies in Palestine develop soft-695 ware for other international companies. This also tested 696 the willingness of developers to cooperate with users 697 face to face or by video conference after the end of 698 COVID-19.

699
On the other hand, the usability analyses targeted direct 700 and indirect users, particularly users who did not partici-701 pate in the development process. This tested if the system 702 developed using RAPD during COVID-19 is still usable after 703 COVID-19. The questionnaire focused on: 704 1-testing the willingness of user to work remotely. including farmers, traders, and supply store workers partici-717 pated in the usability analysis questionnaire. We ensured that 718 all end user respondents knew about RAPD and experienced 719 the system developed by RAPD. For example, each user has 720 at least three months of experience in using the system.

722
This section presents and discusses the effect of RAPD on 723 the company and end users. The section discusses the factors 724 affecting the adoption of RAPD in the company. Then, it dis-725 cusses the factors affecting the usability of system developed 726 using RAPD. Finally, the section discusses the quantitative 727 analysis results for the use of RAPD in normal conditions.  the team members stated that they experienced less pressure 795 during the RAPD processes, and they were fully aware of 796 the purpose of tasks under-development and the final goals. 797 This clarity of objectives, tasks, and final system made the 798 development cost less than before in terms of team size and 799 project time. This made the company adopt RAPD easily, 800 which agrees with [13]. The authors in [13] show that the communication with users 803 is one of the factors affecting adoption of software develop-804 ment methods. So, the interaction between the development 805 team and the users was also studies through observation 806 during the development and interviewing participants. The 807 analysis shows that the interaction was limited at the first 808 meeting but after that the interaction became smoother. The 809 project manager revealed important issues about the inter-810 actions with users. In older development project, developers 811 used to resist new changes demanded by users. In some cases, 812 conflicts occurred between developers and users and between 813 developers and designers because developers do not want 814 to keep changing the product. The conflicts resulted from 815 erroneous assumptions made by designers who depended on 816 verbose user stories that were interpreted differently from 817 time to time. Also, users expressed 'what' functions they 818 need without caring about 'how' these functions should be 819 made, and designers had to extract 'why' these functions were 820 needed and then build a design to show how functions could 821 be in the digital form.

822
In contrast, the RAPD made the interaction more coop-823 erative, direct and clearer than before because developers 824 were aware of the project objectives, and they also aware that 825 user role in the design would lead to an optimized product. 826 During the workshops and meeting, the researcher observed 827 that the company team members were consistently working 828 on performing tasks and harmonically interacting with each 829 other and with the end-users. According to the development 830 team, the roles of each team member was clearer in RAPD 831 than before. Even the end users were participating in the 832 design according to PD principles and were aware of the final 833 goals. They also benefited from RAD as they could try the 834 prototype, give feedback and suggest alternative solutions. 835 Therefore, the end users started to care about 'how' the 836 function should be made and to suggest ways for doing that. 837 Also, the end users started to appreciate the effort spent on 838 developing every single function, and they could estimate the 839 required time for transforming one task from paper to digital. 840 So, the end users became more considerate about requesting 841 changes or modifications.

842
The video conference also made the interaction more coop-843 erative and useful because all participants knew that the time 844 for the virtual meeting is limited and they avoided unneces-845 sary talks. There were nine virtual meetings with two hours 846 per each. In the vegetable market culture, auction shop owners 847 and customers usually use much conversation about different 848 topics while making a specific deal which costs much time. 849 VOLUME 10, 2022 more effective because users become a major player in the 904 development team. The users were observed very active in 905 expressing their requirements and in adding new changes 906 to the system after testing developed services. The RAPD 907 documentations consisting of concise user stories allowed the 908 developers to refer to these changes and address them one 909 by one. These features of RAPD made it acceptable by the 910 company. The developers at the company initially thought that changing 913 the method, which they are used to, will complicate the devel-914 opment process. Their lack of experience in PD strength-915 ened their argument. However, their behavior towards RAPD 916 started to change when the PD principles were explained to 917 them and when they saw that PD and RAD stages can overlap. 918 Their knowledge of RAD and the usability testing made them 919 easily understand RAPD. Also, the COVID19 conditions 920 motivated the company and the developers to think more 921 seriously about finding a way to incorporate all requirements 922 effectively. The video conference also made the developers 923 less reluctant to use RAPD because they did not need to meet 924 users face to face, which reduced travel burdens and infection 925 possibilities. By the time, RAPD complexity was reduced, 926 and the company accepted it which agrees with [14]. But, this 927 required the existence of a member who has strong experience 928 in PD and RAD.

930
The developed system during the action research through the 931 RAPD framework is a web-based system running on cloud 932 and support several services. These services include data 933 entry, data review, billing, balancing, queries, and creating 934 accounts for customers and users. The interaction of users 935 with the new system was observed during the design and 936 performing tasks from the workplace and remotely. Also  This factor is a new one and emerged due to COVID19 944 to maintain safety. The system utilizes cloud computing 945 technologies for hosting the data and process the functions 946 efficiently. The users interact with the system through a 947 responsive web interface which supports desktop computers 948 and mobile devices. Therefore, the users are able to work 949 ubiquitously; from the office, home or other places at any 950 time. To enable secure access to the system, the users need 951 only to login to the system and start any process based on the 952 privileges given to them. 953 The users stated that the remote work was very useful to 954 them during the COVID19. When the mobility constraints 955 were very tight, the direct users, such as the shop owners 956 and accountants, were able to perform the basic data entry at 957 remote work functionalities, they started to depend on that 967 because the remote work enabled reviewing the billing before 968 adding bills to the total financial balance. In the traditional 969 system, bills had to be written manually and added to the 970 balances and customers had to wait before taking their bills.

971
The RAPD framework allowed the users to participate in the 972 design remotely and build experience of the system step by 973 step. So, the adaptation to remote work was easier than they 974 thought at the beginning. work and more efficient. 987 We also found that the remote work saves users time.

988
In the vegetable market, the remote work saved time because

1013
Accuracy is very critical for the auctioning shop because 1014 any mistake in the calculation causes conflict between the 1015 direct users and the indirect users. Also, mistake causes time 1016 delay to determine the reasons of the mistake and solve it. 1017 The users stated that the system is very accurate because all 1018 calculation errors were solved during the development stages. 1019 The users actually participated in making the system accurate 1020 because they explained the workflow, dependencies and show 1021 all parameters clearly. User participation in the system design 1022 made the system easy to use without causing any mistake.

1023
In contrast, the users stated that they have bad experience 1024 with other systems developed by other methods. They were 1025 not used to workflows added by the other systems and they 1026 took much time to learn how to use these systems. Also, they 1027 were not satisfied with the format of the output reports and 1028 bills. So, they did many mistakes during the learning phase 1029 and received many complaints from indirect users. They gave 1030 up using these systems because of the low accuracy and new 1031 tasks needed to be added to make the systems more accurate. 1032

1033
We also analyzed the responses of users towards the factors 1034 affecting user satisfaction, and these factors include: The customization of the system to users and business 1037 requirements makes users more satisfied with the sys-1038 tem [16]. In RAPD, users participated in the design and 1039 expressed their requirements effectively. They refined the 1040 design at each workshop according to the artifacts. The 1041 usability testing also assisted the users in verifying and val-1042 idating the produced services. To users, the most important 1043 points were the sequence of operations and the output format. 1044 Previously when using RAD alone, the developers decided 1045 where to start the programming process. But with RAPD, 1046 the developers had to build the functions that can achieve the 1047 sequence of operations based on dependency priority so that 1048 the users can test them and give feedback. The well consideration of the workflow and output formats 1051 made the users satisfied with the content of the system. 1052 An important issue that was raised during the development 1053 is the number of queries and their output reports. When more 1054 functions were generated and the users saw the strength of 1055 the system, they kept on adding new queries for generating 1056 useful reports. For example, more data management, items 1057 tracking and forecasting functions were demanded by the 1058 direct users. The developers followed the style of expressing 1059 the power of the developed queries, and if the users were 1060 not satisfied, they considered the requested queries in the 1061 new stage. The developers also explained the dependency 1062 priority to users and both side committed to the dependencies. 1063 When the third design stage was completed, the users found 1064 the query services were enough. RAPD helped the company 1065 agree that RAPD does not need modifications to fit with 1120 design in normal life conditions. This is because RAPD rapid 1121 and participatory nature assist developers in building usable 1122 software. A low percentage of respondent think that video 1123 conference tool can be replaced by face to face workshops. 1124 The respondents answered the open answer question that 1125 the video conference supporting remote cooperative design is 1126 the most important feature, particularly nowadays as remote 1127 work has become a daily routine. They added that the least 1128 favorable feature of RAPD is incorporating many users in the 1129 design which may complicate the process, particularly when 1130 users have low computer proficiency degree. Others, who 1131 were using software methods other than RAD and PD, added 1132 that they do not mind learning RAPD but this requires time. 1133 The results of the usability questionnaire show that remote 1134 work has become one of the usability factors as most users, 1135 Table 4, require that the system should support remote work 1136 although they sometimes prefer to work onsite. Most users 1137 also are stratified with the system developed using RAPD 1138 and agree that the system is usable, as shown in Table 4. The 1139 reason why some do not agree is that some users need more 1140  The most important summary of this quantitative results 1144 is that a high percentage of users agree that the system is 1145 usable, and easy to be understood and used with low com-1146 plexity. On the other hand, a high percentage of developers 1147 agree that RAPD is more useful for the company than the 1148 current method, which is RAD, and RAPD can be used for 1149 developing different applications at different times.

1151
The limitations of our research concern the used research accepted systems in a short time. To save time, businesses 1169 adopted on-shelf systems that do not fit with their require-1170 ments. This caused new challenges, such low accuracy long 1171 training time on using these systems. On the other hand, using 1172 agile methods that do not focus on capturing user experience 1173 also made the produced systems inappropriate for business 1174 workflow. Therefore, this paper has proposed a framework 1175 called RAPD that integrates two well know software devel-1176 opment methods. RAD allows for rapid development and PD 1177 allows for user participation in the design process. To make 1178 the produced system more acceptable, usability testing meth-1179 ods are also added to all design phases. For the company, 1180 RAPD was useful as it offers sufficient management plan, 1181 reduces development time and cost, and eliminates conflicts 1182 with end users. The company adopted the new framework 1183 easily as it is a small new company with young development 1184 teams, which agrees with [47]. For user, RAPD made the 1185 developed system usable and acceptable as it is customized 1186 to the workflow, accurate and time saving. We also found 1187 that software development is affected during COVID19, and 1188 video conference enabled PD to create more concise design 1189 documentations. The post COVID-19 analyses show that 1190 RAPD, its tools and process can still be used to develop dif-1191 ferent applications in normal conditions without crisis. Future 1192 work will include using RAPD for software development in 1193 another context so that it can be generalized.