Adaptation of User Interface Based on Contextual Feedback Robertas Damaševičius, Paulius Paškevičius Department of Software Engineering Kaunas University of Technology Kaunas, Lithuania robertas.damasevicius@ktu.lt Abstract—The spread of social networks and other online The users are no longer the consumers of media content, but collaboration-related practices changes the target of software also want to act as producers of content or even co-designers of products from a single user to virtual communities. Such content delivery platforms [4] to have impact on the face of the communities view user interfaces of social websites as virtual community. communication partners (facilitators) rather than mere communication medium. Effective communication requires The strength of relationships that bind a member to a proper feedback from community-driven systems that create the community can be influenced by the impact a member can illusion of active participation and control. Furthermore, make as well as a feedback that a member can receive from a feedback combined with community effort and evaluation community. The success of a virtual community relies on the mechanisms such as crowdvoting can be used as a vehicle for voluntary contribution of valuable intellectual property of adaptation of interfaces to the requests of community. The individuals to a community without explicit compensation [5]. implementation of community-driven interfaces requires the Even if an individual does not receive any explicit reward for extension of existing user interface development architectures his/her contribution, he/she often wants his/her contribution to and design patterns. In this paper we analyse known user make impact or at least be seen. Capturing and understanding interaction and user interface models and present the contextual feedback received from users also is critical in business feedback based adaptation (CFBA) meta-model, the four-tiered information systems and customer relationship management [6] user interface (4TUI) architecture, and the Model-Control-View- as well as in intelligent systems and intelligent user interfaces Adapter (MCVA) pattern. A case study in the community-driven [7]. This practice has been recognized for long now and interface adaptation is presented. content sharing sites such as Flickr or Youtube have been Keywords—user interface design, feedback modelling, successful very much due to this practice. However, the need to adaptable interface, interface evolution, crowdvoting. say or show something to a community is paired with a need to obtain answers or feedback from it. INTRODUCTION1 In this paper we analyse feedback models and methods that The spectacular rise of social networks and other are aimed to increase the role of feedback in community collaboration-based practices such as crowdsourcing [1] building and support efforts. To introduce the contextual underlined the importance of effective communication in conditions into user interface evolution process, the contextual virtual communities. The term “virtual community” refers to a model is required that maps contextual requirements from a large group of individuals that regularly share and exchange community of users received through feedback mechanisms to information through computer-mediated mechanisms such as e- the adaptations of user interfaces. Building user interfaces that mail, weblogs, or forums [2]. Many research studies dynamically adapt to the context is not new [8-12]. Similar investigated what motivates people to participate in virtual approaches include mediation strategies for integrating the communities (e.g. see [3]). Out of many contributing factors, input of multiple crowd workers in real-time [13], the the most important ones are common interests (e.g., extension of the model-view user interface architecture with an professional networks such as LinkedIn), status seeking and intelligent layer that handles interface events as commands and reputation (e.g., question-and-answer websites such as allows a user to evolve an interface in a way that is entirely StackOverflow), and affiliation (e.g., friend networks such as independent of applications [14], and the website plug-in that Facebook). All these factors depend upon supporting effective makes use of crowdsourcing to collect context-aware activity communication between a member of the virtual community data based on which suitable user interface adaptations for and the community represented by its other members and the different target devices are inferred [15]. user interface of the social platform. In virtual communities Our novelty is the interpretation of the feedback from the this communication is mostly computer-mediated, i.e., the user community of users of a system as the context of the system’s interface of the platform that supports virtual community acts user interface. This interpretation does not contradict the as the „face” of the community or as a partner of conversation. definition of context provided in [16]: “context is any information that can be used to characterize the situation of an 1 Copyright © 2016 held by the authors entity“. Feedback conveys context information (e.g., the 1 interests, preferences, opinion of user) that has influence over people can participate and contribute their solutions [32]. This the presentation and functionality of user interface. shift represents a transition from a world in which a small number of experts define rules, create static products, and make We propose the crowdvoting-based contextual feedback decisions for many consumers toward a world in which meta-model, the four-tier community driven architecture and everyone has interests and opportunities to actively participate the extension of the MVC pattern for implementing the in the development of dynamically evolving products [33]. community-driven adaptation of user interfaces at the use stage rather than at the design stage (as, e.g., in [17]). We describe its The essential role of feedback in natural communication application in the community-driven website user interface makes it a crucial issue in the development of human-computer project aimed to engage community members in the controlled user interfaces [34] where users communicate proactively evolution of website interface. rather than passively or reactively. An example of the proactive role of the user is so called Split Interfaces, where frequently I. ROLE OF FEEDBACK IN USER COMPUTER INTERACTION used functionality is automatically copied to a specially designated adaptive part of the interface [35, 36]. Altered The term ‘feedback’ originates from the area of cybernetics Prominence is another approach to interface adaptation that and refers to the information that a system receives from its highlights recently used elements of an interface [37]. Without environment about the effects or consequences of its actions feedback, a human-computer dialog quickly breaks down while [18]. In communication, feedback is used for a broad range of proper feedback can create the illusion of a dialog partner responses at various levels of communication. Commonly, listening [38]. feedback is understood as any information about reaction to a product or a person that can be used as a basis for improvement According to [19, 28, 29], in order to be effective, feedback [19]. Allwood et al. [20] claim that feedback is a central must be 1) persuasive (i.e. influencing future state of functional subsystem of human communication. It consists of community and behaviour of community members), 2) methods that allow providing, without interrupting the dialog, contextual (i.e. include context information by default), and 3) information about quality of communication such as ability and informative (i.e. convey useful information), 4) contributive willingness to have contact, ability to understand (i.e. contribute towards benefit of a community as a whole), 5) communicated information as well as the emotions and continual (i.e. to support conversation as narrative of attitudes triggered by the information in the recipient. community), 6) expressive (i.e. demonstrate polarity using According to Kotzé [21], feedback has three main elements: 1) affective means such as emotions), and 7) effortless (easy to response, which serves to confirm that the recipient has use). In any case, feedback comes as a response to a previous received information, 2) modification of behaviour, which communicative act [30], i.e., in reaction to the status or reassures the user that his input is relevant and has power to opinions of a community members or an entire community in change, and 3) intelligence (see, e.g., social creativity [22], order to achieve consensus or alignment [31]. collective intelligence [23] and “wisdom of crowds” [24]) that Techniques for collecting user feedback in software the opinion or understanding of the community can lead to systems cover a wide spectrum, ranging from error reporting improved quality of communication and usability of a product. facilities to the content-related feedback mechanisms of social The main aim of feedback is to induce the change of a networks [28]. Examples of such feedback mechanisms are the software product while the direction of the change itself Facebook “Like” button or the YouTube’s thumbs-up/thumbs- depends upon the polarity of feedback: positive feedback down, which allow evaluating content, linking members while (agreement) reinforces the change in the same direction; require only a minimum amount of effort on the users’ side. negative feedback (disagreement) causes a change in the However, the amount of “likes” and “don’t likes” do not have a opposite direction, while homeostatic feedback maintains direct influence how information is presented, i.e., the platform equilibrium [25]. In the long term, such change leads to the of a virtual community has full control over the presentation evolution of a product or its interface and user feedback acts a while the function of the user is reduced to evaluating other main driver of such evolution. Software evolution has been users’ content rather than making influence over its recognized as a key issue in software development and use a presentation. However, if properly implemented feedback long time ago: as software application is released for use, the could increase affiliation, loyalty and immersion of the world in which it is situated changes, and therefore new community members beyond simple collection of “likes”. demands constantly arise [26]. Traditionally, software Examples of such “socially advanced” user interfaces are a evolution has been dealt with offline using the version-based crowdsourcing interface that collects user-generated mappings approach as follows: a version is released, user response is between pairs of web pages [39], an adaptive user interface that collected, a new (updated, corrected) version is released, and is constructed using consensus methods [40], and socially- the cycle is repeated again. However, in a modern world of adaptable interfaces, interfaces that crowdsource the creation of software development, software evolution has an task-specific interface customizations and instantly share them unprecedented speed [27], and feedback can be seen as a with all users of the application [41]. The development of such means to accommodate and drive the change at the use time. socially advanced interfaces requires adequate models of interaction, which we discuss in Section 3. The understanding of increased importance of feedback mechanisms signifies a shift from consumer cultures (specialized in producing finished artefacts to be consumed passively) to the participation-based cultures in which all 2 II. MODELING INTERACTION AND INTERFACE ADAPTATION Computers and internet are the media of social interaction in virtual communities. Therefore, the social interaction in virtual communities is mainly guided by the principles of human-computer interaction. When humans interact with computer, they first formulate their goals and then develop a Fig. 2. Seeheim model emphasizing the three different levels of visual series of steps required to achieve that goal. Such mental model feedback of action is known as Norman’s Interaction Cycle (see Fig. 1) [42], which has been used to evaluate the efficiency of a user The Seeheim model (see Fig. 2) [47] reveals the linguistic interface. The model includes both cognitive and physical nature of the visual feedback identifying three main software activities, and includes feedback, which is called “Evaluation” modules (or layers): 1) Dialogue Control module handles the in the model. The Norman’s model does not distinguish syntactic aspects of the interaction and is responsible for the between the content of the message delivered, its presentation dynamic behaviour of the system; 2) Application Interface form and its affect (i.e., emotions associated with the message). module provides a semantic interpretation of the information Therefore, it should be considered only as the simplest received for the dialogue component; 3) Presentation Module approximation of human-computer interaction. Furthermore, handles the lexical aspects of the interaction such in input as the interface in the Norman’s model can be interpreted as a well in output and is only aware of the presentation technology. metaphor of dialogue between interface designer and its users. Visual feedback can be formulated at three different levels: lexical (Presentation), syntactical (Dialogue Control) and semantic (Application Interface Model). The Bezold’s model [48] (see Fig. 3) deals with interface adaptation, i.e., the ability of the interface to improve itself for an individual user based on an observation of the user's behaviour. Adaptation to user behaviour comprises two steps: 1) reasoning on the user-system interaction, and 2) adapting the user interface accordingly. The user-system interaction is considered as a linear sequence of basic events, which are emitted by the interactive system. User modelling algorithms Fig. 1. Interpretation of the Norman’s Interaction Cycle [42] as a dialogue between designer and user extract new knowledge from the user-system interaction and trigger interface adaptations. A semantic layer is introduced as The extension of the Norman’s mental model is the Isatine an abstraction of the interactive system that allows model [43] that is also based the Dieterich’s taxonomy of user implementing reasoning on the user-system interaction. The interface adaptation [44]. The model states that three entities system-independent logic is defined on the top of the semantic are involved in the adaptation of user interface: the user, the layer. The adaptation layer decides which adaptations can be interactive system, or any third party. The adaptation is applied to an interactive system. The advantage of the model is performed as follows: 1. Goals for user interface adaptation are a multi-layered architecture that allows separation of semantic, formulated; 2. The user or third party initiates request for interaction and adaptation aspects of user interface. adaptation; 3. The adaptation is specified as a sequence of commands issued to the interactive system for execution; 4. The adaptation is applied using the adaptation support mechanisms (e.g., through user interface options, personalization); 5. Transition between the interface before and after adaptation is performed; 6. Information about adaptation is issued to the interested parties; and 7. Adaptation is evaluated. The advantage of the Isatine model is a detailed guideline for performing adaptation of user interfaces. The Taylor’s Layered Protocols (LP) model and its elaboration in [45, 46] are based upon the cognitive principle that humans use superimposed layers of abstraction in perception. From this principle the LP model arrives at the Fig. 3. Bezold’s model of interative system adaptation [48] architecture for structuring user-system interaction. The model distinguishes between the system’s interpretation of their The Baxley's model of user interface [49] applies the messages (I-feedback), and information the system expects to separation of concerns and decomposes user interface into receive from the user (E-feedback). The advantage of the three tiers as follows: Structure (conceptual model, task flow, model is the classification of feedback types (user’s feedback and organizational model), Behaviour (viewing and navigation, and system’s feedback). editing and manipulation, user assistance) and Presentation (layout, graphic design style, text). Here, the conceptual model supplies the ‘metaphor’ that helps users to interact with an application, and the organizational model provides 3 classification schemes to group and associate application All analysed models of interface interaction and adaptation information and interface objects. The model’s advantage is a emphasize the role of feedback in the interface adaptation clear separation of the different aspects of user interface. process. For interface customization, a separate dedicated interface (or interface layer) is required. We call this interface, The RUX (Rich User eXperience) model [50] is used for the meta-interface, since it is overlaid on top of the software the systematic adaptation of user interfaces over the existing product’s interface and allows making configuration choices on web applications. The user interface specification is divided the product’s interface. One example of such meta-interface is into four levels: 1) Concepts and Tasks, 2) Abstract Interface, the Facebook’s Like button (see Fig. 5, a), which allows the 3) Concrete Interface and 4) Final Interface. Concepts and Facebook users to express their opinion on the content of a Tasks are taken from the underlying web model. Abstract website. The button provides a one-click shortcut to express Interface provides a common representation to all devices and and externalize the affective reactions of a user. Another interface development platforms without any kind of spatial example of meta-interface is provided by Usabilla arrangement or behaviour. Concrete Interface optimizes the (www.usabilla.com), which is a service for real-time visual presentation of user interface for a specific device or group of user feedback tracking. The users of the website click the devices, and has three Presentation levels: Spatial Presentation feedback button and can select any part of the page to evaluate allows the spatial arrangement and interface style of to be it (see Fig. 5, b). Google provides a similar mechanism, where specified; Temporal Presentation allows the specification of users can highlight any areas of web interface, black out behaviours which require a temporal synchronization; and personal information, comment on relevant issues and send it Interaction Presentation allows modelling the user’s behaviour. to Google (see Fig. 5, c). Final Interface provides code generation of the modelled application. The advantage of the RUX model is a hierarchy of interface entities from most abstract to specific ones, which could be easily mapped to the hierarchy of models according to the model-driven architecture (computation-independent, platform-independent, and platform specific models). Currently commonly used software interface patterns (such Fig. 5. Feedback in (left to right): Facebook, Usabilla and Google. as MVC or PAC) are derived from the Seeheim model. The Presentation–Abstraction–Control (PAC) pattern [51] separates Summarizing, two possible implementations of feedback an interactive system into three types of components are usually considered [53]: 1) Emoticons-based feedback: responsible for specific aspects of the application's aiming at expressing the emotionally affected satisfaction functionality. The abstraction component retrieves and degrees among the end-users via picking an emoticon (virtual processes the data, the presentation component formats the facial expression) for judging his user experience [6]; 2) visual and audio presentation of data, and the control Recommendation frames: a simple interaction illustrated component handles things such as the flow of control and differently (e.g. pop-up window, sliding area), which is mainly communication between the other two components. In the used in e-commerce to provide client recommendations. The Model–View–Controller (MVC) pattern (Fig. 4), the Model implementation of such meta-interface together with the need consists of application data and business rules, the Controller for handling community requests and implementation of acts as mediator like the Dialogue component in the Seeheim conflict resolution and opinion aggregation mechanisms for architecture, and a View can be any output representation of crowdvoting, requires the extension of existing user interface data [52]. Model–View–Presenter (MVP) is a derivative of development architectures and design patterns. MVC, where the presenter assumes the role of the Controller, retrieves and formats data for the View, the View is III. FRAMEWORK OF COMMUNITY-DRIVEN USER INTERFACE responsible for handling the user interface events, which is the DEVELOPMENT controller's role in MVC, and the Model is strictly a domain model. In Model-View-ViewModel (MVVM), the ViewModel The proposed framework of community-driven user is responsible for providing access to data objects and backend interface development consists of 1) the contextual feedback logic from the Model. View is all elements displayed by the based adaptation (CFBA) metamodel, 2) the four-tiered user user interface. Model is either a domain model which interface (4TUI) architecture, and 3) the Model-Control-View- represents the real state content, or the data access layer that Adapter (MCVA) pattern. represents that content. The CFBA metamodel (see Fig. 6) describes the relationship between different entities and models in the modelling and implementation of adaptable and evolvable user interfaces. The metamodel is based on the Norman’s Model [42], Taylors Layered Protocols [45], Seeheim model [47] and its implementations as user interface design patterns (PAC, MVC and their variants), Bezold’s model [48], Baxley’s model [49] and RUX model [50]. The CFBA metamodel actually features two interaction cycles: 1) a traditional cycle, where a Fig. 4. The MVC pattern user and a system exchange with messages and feedbacks during the system’s use, and 2) a community-driven cycle, 4 where a crowdvoting entity collects feedback from a judgments of the community of user interface, the crowdvoting community of users and changes the presentation of the [23] mechanism is used. The implementation of the user interface according to the needs of the majority of users. interface is described by the adaptation of the Seeheim model. The community (crowd) is treated as a part of context that depends on the Context Model. To collect the opinion and Fig. 6. Contextual feedback based user interface adaptation metamodel However, where is a difference, the Presentation entity is The MVCA pattern (see Fig. 8) is an extension of the MVC variable and different variants of interface presentation can be family of patterns with an additional class for managing the selected according to the requests of the community aiming to community-driven requests for user interface modification. guarantee usability while achieving adaptation to the changing Since the community may include users with conflicting needs of the community of users. The relationship of interface interests, a mechanism for solving these conflicts is required. to other models is given by the adaptation of the Baxley’s Community Layer model: Structure depends upon Task Model, Behaviour depends upon Dialog and Navigation Model, and Presentation depends upon interface Metaphor. Modelling of user interface at different levels of abstraction is represented by the adaptation of the RUX model, where a hierarchy of interfaces is used to represent interface independence and specificity with respect to different platforms and / or user groups, while allowing to implement automatic generation of interfaces [54]. We adopt the elements of crowdsourcing, i.e. crowdvoting, as a model [55] for solving a problem of interface adaptability and evolvability at use time. The 4TUI architecture (see Fig. 7) describes the structural organization of the community-driven user interface system. The proposed architecture is an extension of the classic three- Fig. 7. Proposed four layered architecture of community-driven business tiered architectures and models (such as MVC pattern), which applications include Persistence layer for data storage and handling, Business layer for business logic; and Presentation layer for content delivery. The additional fourth layer (tier) is proposed TABLE I. FUNCTIONS OF LAYERS IN FOUR LAYERED SYSTEM for managing community requests for interface representation Layer Function and community-driven reasoning based on crowdvoting. This Persistence Stores data and handles requests for data layer performs the functions of the semantic layer in the Bezold’s model [48]. The functions of layers in the four Business Specifies business objects and business logic layered system are summarized in Table 1. rules, and handles interfacing between presented 5 information and stored data The proposed solution is finding sub-communities or Presentation Delivers content to browser groups of users with similar interests based on their profiles and interface preferences, and providing customized variants of Community Manages community requests for interface interfaces to these particular groups. The mechanism for change and provides representation conflict resolution conflict solving is based on the consensus-based user profile determination method [40]. Adapt Interface Adapter Updateinterface Community Fig. 8. Handling interface adaptation requests using the MVCA pattern The proposed approach for the community-driven user quality is ensured by the intelligent voting mechanisms that interface adaptation combines elements of the following harnesses “wisdom of crowds” and retargets interface for a methodologies: Design-for-change [56] for developing specific groups of users. interface that is evolvable and adaptable to anticipated and unanticipated changes; Meta-design [57] and related REFERENCES methodologies (End-User Development [58], Participatory [1] J. Howe, “The rise of crowdsourcing”, Wired 14(6), 2006. Design [59], Collaborative Design [60], etc.) as a theoretical [2] P. Waterson, “Motivation in Online Communities”, in S. Dasgupta (ed.) foundation for involvement of end-users as co-designers of a Encyclopedia of Virtual Communties, 2006. product; and Crowdsourcing [1] for enabling the participation [3] P. Kollock, “The economies of online cooperation: gifts and public of a crowd (i.e., a community of dedicated users) in the user goods in cyberspace”, in M.A. Smith, P. Kollock, (Eds.), Communities interface improvement and evolution process. in Cyberspace, pp. 220-238. London: Routledge.1999. [4] D. Schuler and A. Namioka, Participatory design: Principles and IV. CONCLUSIONS practices. Hillsdale, NJ: Erlbaum, 1993. [5] J. Roberts, I.-H. Hann and S. Slaughter, “Understanding the The social collaboration based approach can be applied to Motivations, Participation and Performance of Open Source Software the process of user interface development. The community of Developers: A Longitudinal Study of the Apache Projects”, users can drive the evolution of user interfaces to increase their Management Science 52(7), July 2006, pp. 984 - 999. flexibility (the interface adapts flexibly to change requirements [6] S. Abeyratna, G.V. Paramei, H. Tawfik and R. Huang, “An Affective of the users), plasticity (the interface’s capacity of adaptation to Interface for Conveying User Feedback”, Proc. of 12th International Conference on Computer Modelling and Simulation (UKSim), pp. 369- cope with changing context), quality (the interface of a product 374, 2010. represents the combined efforts aka collective intelligence of [7] S. Stumpf, E. Sullivan, E. Fitzhenry, I. Oberst, W.-K. Wong and M.M. the community of users), usability (the community itself selects Burnett, “Integrating rich user feedback into intelligent user interfaces”, and adopts via natural selection the best practices of interface Proc. of International Conference on Intelligent User Interfaces (IUI design), cultural acceptance (the interface reflects the culture of 2008), pp. 50-59, 2008. its users), user satisfaction (opportunity to have a say and an [8] J. Eisenstein, J. Vanderdonckt and A. Puerta, “Adapting to mobile impact increases user satisfaction) and product popularity (the contexts with user-interface modeling”, Proc. of the Third IEEE Workshop on Mobile Computing Systems and Applications users have control over what they see and what they get). (WMCSA'00), pp. 83-92, 2000. The advantages of the proposed framework for community- [9] G. Menkhaus and W. Pree, “User interface tailoring for multi-platform driven user interface adaptation are as follows: iterative service access”, Proc. of the 7th international conference on intelligent user interfaces (IUI '02), pp. 208-209, 2002. refinement of user interface enables interface evolution; the [10] P. Repo, “Facilitating user interface adaptation to mobile devices”, Proc. community performs interface evaluation; interface of the third Nordic conference on Human-computer interaction dynamically changes in response to the changing requirements (NordiCHI '04), pp. 433-436, 2004. of the community; the user interface evolution process is [11] M. Bisignano, G. Di Modica and O. Tomarchio, “Dynamic User outsourced to the community of users and is fully automated; Interface Adaptation for Mobile Computing Devices”, Proc. of the 2005 community feedback ensures that interface of the product is Symposium on Applications and the Internet Workshops (SAINT-W prevented from ageing and decay so long as the community is '05), pp. 158-161, 2005. interested in the services provided by the product itself; [12] G. Calvary, J. Coutaz and D. Thevenin, “Supporting Context Changes collective intelligence allows to evolve a user interface that is for Plastic User Interfaces: a Process and a Mechanism”, Proc. of the Joint AFIHM-BCS Conf. on Human-Computer Interaction IHM-HCI inherently usable, culturally-acceptable and visually-pleasant; 2001, vol. I, pp. 349-363. Springer, London, 2001. direct real-time user participation in the user interface [13] W.S. Lasecki, K.I. Murray, S. White, R.C. Miller and J.P. Bigham, evolution increases user engagement and satisfaction; interface “Real-time Crowd Control of Existing Interfaces”, Proc. of the ACM 6 Symposium on User Interface Software and Technology (UIST 2011), [36] K.Z. Gajos, D.S. Weld, and J.O. Wobbrock, “Decision-theoretic user pp. 23-32, 2001. interface generation”, Proc. of the 23rd national conference on Artificial [14] J. Dicker and B. Cowan, “Platforms for Interface Evolution”, Proc. of intelligence (AAAI'08), Vol. 3, pp. 1532-1536, 2008. ACM CHI 2008 Conference on Human Factors in Computing Systems [37] K. Gajos, D. Christianson, R. Hoffmann, T. Shaked, K. Henning, J.J. (CHI’08). Long and D.S. Weld, “Fast and robust interface generation for [15] M. Speicher, Crowdsourced Evaluation and Adaptation of Web ubiquitous applications”, Proc. of the 7th international Conference on Interfaces for Touch. Master thesis. Global Information Systems Group, Ubiquitous Computing (UbiComp'05), pp. 37-55, 2005. ETH Zurich, 2012. [38] M.A. Pérez, “Conversational dialogue in graphical user interfaces: [16] A.K. Dey, “Understanding and Using Context”, Personal and ubiquitous interaction technique feedback and dialogue structure”, in CHI 95 computing, Vol. 5, February 2001, pp. 4-7. Conference Companion, 1995, pp. 71-72. [17] S. Hennig, J. Van den Bergh, K. Luyten and A. Braune, “User driven [39] J. Kim, R. Kumar and S.R. Klemmer, “Crowdsourcing Interface for evolution of user interface models - The FLEPR approach”, Proc. of the Collecting Correspondences of Web Pages”, UIST '09 Poster, Victoria, 13th IFIP TC 13 international conference on Human-computer BC, Canada 2009. interaction - (INTERACT'11), Part III. Springer-Verlag, pp. 610-627, [40] J. Sobecki and N.T. Nguyen, “Consensus-based adaptive interface 2011. construction for multiplatform Web applications”, Proc. of 4th [18] N. Wiener, Cybernetics: or Control and Communication in the Animal International Conference on Intelligent Data Engineering and Automated and the Machine. Cambridge: MIT Press, 1948. Learning, IDEAL 2003. Springer LNCS Vol. 2690, 2003, pp. 457-461. [19] M. Norgaard and K. Hornbaek, Exploring the value of usability [41] B. Lafreniere and M. Terry, “Socially-Adaptable Interfaces: feedback formats. International Journal of Human-Computer Interaction Crowdsourcing Customization”, Proc. of ACM CHI 2011 Conference on 25(1), pp. 49-74, 2009. Human Factors in Computing Systems (CHI 2011), 2011. [20] J. Allwood, J. Nivre and E. Ahlsén, “On the semantics and pragmatics of [42] D.A. Norman, “Cognitive Engineering”, in D.A. Norman and S.W. linguistic feedback”, Journal of Semantics 9(1), 1992, pp. 1-26. Draper (eds.), User Centered System Design. Lawrence Erlbaum Associates, Hillsdale, 1986, pp. 31–61. [21] P. Kotzé, “Feedback And Task Analysis For E-Commerce Sites”, Proc. of the ISSA 2002 Information for Security for South-Africa 2nd Annual [43] V. Lopez-Jaquero, J. Vanderdonckt, F. Montero, and P. Gonzalez, Conference, 10-12 July 2002, Muldersdrift, South Africa, pp. 1-17. “Towards an Extended Model of User Interface Adaptation: The Isatine Framework”, in J. Gulliksen, M.B. Harning, P. Palanque, G.C. Veer and [22] A. Warr and E. O'Neill, “Getting Creative with Participatory Design”, J. Wesson (Eds.), Engineering Interactive Systems, LNCS, Vol. 4940. Proc. of Participatory Design Conference, 2004, pp. 57-61. Springer-Verlag, Berlin, Heidelberg, pp. 374-392, 2008. [23] J.M. Leimeister, “Collective Intelligence”, Business & Information [44] H. Dieterich, U. Malinowski, T. Kühme and M. Schneider-Hufschmidt, Systems Engineering, 2(4), 2010, pp. 245-248. “State of the Art in Adaptive User Interfaces”, in: M. Schneider- [24] J. Surowiecki, The Wisdom of Crowds, Anchor, 2005. Hufschmidt, T. Khüme and U. Malinowski (eds.), Adaptive User [25] A. Spink and T. Saracevic, “Human-computer interaction in information Interfaces: Principle and Practice. North Holland, Amsterdam, 1993. retrieval: Nature and manifestations of feedback”, Interacting with [45] J.H. Eggen, R. Haakma and J.H.D.M. Westerink, “Layered Protocols: Computers, 10(3), pp. 249-267, 1998. hands-on experience”, International Journal on Human–Computer [26] M.M. Lehman and L.A. Belady (eds.), Program evolution: processes of Studies, 1996, 44, pp. 45–72. software change. Academic Press Professional, 1985. [46] R. Haakma, Layered Feedback in User-System Interaction, Master [27] C. Ghezzi, P. Inverardi and C. Montangero, “Dynamically Evolvable thesis, Eindhoven University of Technology. Dependable Software: From Oxymoron to Reality”, in P. Degano, R. [47] A.J. Dix, B. Russell and A. Wood, “Architectures to make Simple Nicola and J. Meseguer (Eds.), Concurrency, Graphs and Models, LNCS Visualizations using Simple Systems”, Proc. of Advanced Visual vol. 5065. Springer-Verlag, pp. 330-353, 2008. Interfaces, AVI2000, 2000, pp. 51-60. [28] F. Heller, L. Lichtschlag, M. Wittenhagen, T. Karrer and J. Borchers, [48] M. Bezold, “A Semantic Framework for Adapting Interactive Systems in “Me Hates This: Exploring Different Levels of Use”, Proc. of ACM CHI Intelligent Environments”, Proc. of Intelligent Environments 2009, pp. 2011 Conference on Human Factors in Computing Systems (CHI 2011), 204-211. 2011, pp. 1357-1362. [49] B. Baxley, “Universal model of a user interface”, Proc. of Conference on [29] R. Mendoza Gonzalez, J. Munoz Arteaga, F.J Alvarez and M. Vargas Designing for user experiences (DUX '03), pp. 1-14, 2003. Martin, “Integration of auditive and visual feedback in the design of [50] J.C. Preciado, M.L.Trigueros and F. Sánchez-Figueroa, “An Approach interfaces for security applications”, in Workshop on Perspectives, to Support the Web User Interfaces Evolution”, Proc. of the 2nd Int. Challenges and Opportunities for Human-Computer Interaction in Latin Workshop on Adaptation and Evolution in Web Systems Engineering America (CLIHC), Rio de Janeiro, Brazil, 2007. AEWSE'07, 2007. CEUR Workshop Proc. 267. [30] L. Cerrato, “A comparison between feedback strategies in human-to- [51] J. Coutaz, “PAC: an Implementation Model for Dialog Design”, Proc. of human and human-machine communication”, Proc. of 7th International the Interact'87 conference, 1987, pp. 431–436. Conference on Spoken Language Processing, ICSLP2002 - INTERSPEECH 2002, Denver, Colorado, USA, September 16-20, 2002. [52] O. Moravcik, T. Skripcak, D. Petrik and P. Schreiber, “Approaches of the Modern Software Development”, International Journal of Machine [31] R.-J. Beun, R.M. van Eijk and H. Prust, “Ontological Feedback in Learning and Computing, Vol. 1, No. 5, December 2011, pp. 479-487. Multiagent Systems”, Proc. of the Third International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS '04), Vol. 1, [53] N. Mezhoudi, “User interface adaptation based on user feedback and pp. 110-117, 2004. machine learning”, IUI Companion, 2013, pp. 25-28 [32] G. Fischer, “Understanding, fostering, and supporting cultures of [54] V. Štuikys, R. Damaševičius, J. Valančius, G. Ziberkas, V. participation”, Interactions 18(3), pp. 42-53, 2011. Limanauskienė and E. Toldinas, “Generation of Database Interfaces for Nomadic Users”, Information Technology & Control, No. 2(27), pp. 41- [33] G. Fischer, “End User Development and Meta-Design: Foundations for 50, 2003. Cultures of Participation”, Journal of Organizational and End User Computing, 22(1), pp. 52-82, 2010. [55] D. Brabham, “Crowdsourcing as a model for problem solving: An introduction and cases”, Convergence: The International Journal of [34] S. Kopp, J. Allwood, K. Grammer, E. Ahlsén and T. Stocksmeier, Research into New Media Technologies, 14(1), pp. 75-90, 2008. “Modeling Embodied Feedback with Virtual Humans”, ZiF Workshop, 2006, pp. 18-37. [56] V. Štuikys, R. Damaševičius, M. Montvilas, V. Limanauskiene and G. Ziberkas, “Educational Portal Development Model for Implementing [35] A. Sears and B. Shneiderman, “Split menus: effectively using selection Design for Change”, Information Technology and Control, 35(3), pp. frequency to organize menus”, ACM Transactions on Computer-Human 222-228, 2006. Interaction (TOCHI), v.1 n.1, p.27-51, March 1994. 7 [57] G. Fischer, “Meta-Design: A Conceptual Framework for End-User [59] Y. Dittrich, S. Eriksén and C. Hansson, “PD in the Wild; Evolving Software Engineering”, in M.M. Burnett, G. Engels, B.A. Myers and G. Practices of Design in Use”, Proc. of the Participatory Design Rothermel (eds.), End-User Software Engineering, Dagstuhl Seminar Conference (PDC 02), 2002, pp. 124-134. Proc. 07081, Schloss Dagstuhl, Germany, 2007. [60] L. Zhu, “Cultivating collaborative design: design for evolution”, Proc. of [58] H. Lieberman, F. Paternó, M. Klann and V. Wulf, “End-user the Second Conference on Creativity and Innovation in Design (DESIRE development: An emerging paradigm”, in H. Lieberman, F. Paterno and '11), pp. 255-266, 2011. V.Wulf (eds.), End-user development. Springer, pp. 9-15, 2005. 8