Extracting usability information from user interface events Export

@article{citeulike:233339, address = {New York, NY, USA}, author = {Hilbert, David M. and Redmiles, David F.}, citeulike-article-id = {233339}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=371593}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/371578.371593}, comment = {the issue of difficulties in takeup comes into play here, partly attributed to costs (p. 414), but I also wonder if the stage of development and appropriate level of evaluation (formative vs summative) also influence types of tools needed. Tawanna: One question I had while reading this paper was if the framework would scale for ubiquitous technologies. For example, what happens in the case of social interaction between individuals? What about other dimensions of the environment that may need to be recorded such as temperature, sound, etc... How do other interactions such as text messaging or twittering come into play? Reasons for evaluating: to understand user behavior and performance, user thoughts and experience, to compare design alternatives, compute usability metrics, and certify conformance w/standards. Extracting usability- related information is difficult due to high and low frequency band events, issues with grammatical structures across hi/lo freq, context (For instance, when seeing the event MOUSE\_PRESSED PrintToolBarButton, what is the user trying to do?), and user interaction at various levels of abstraction. state of the art: synch and search furthest along; ui events index into video (typically) doesn't mention: webquilt, experiscope; etc (pre-dates them) processing data: 1. Techniques for synchronization and searching: The most powerful and simplest technique for exploiting user interface events. Allows searches in one medium to find supplementary information in others. 2. Techniques for transforming event streams: Selecting, or filtering events; abstracting, or relating events to higher-level concepts of interest in analysis, recoding, or generating new event streams based on selection and abstracting. 3. Analysis: counts \& summary stats; detecting sequences; characterizing \& comparing sequences 4. Visualizing techniques 5. Integrated evaluation support (stuff that does all the above) Evaluation categories: 1. Predictive evaluation - making use of modeling to predict results. 2. Observational evaluation - Allow investigators to identify aspects of user behavior that may be difficult to identify using other techniques. 3. Participative evaluation: user self-reports. Captures aspects of users that are difficult to obtain otherwise. }, doi = {10.1145/371578.371593}, issn = {0360-0300}, journal = {ACM Comput. Surv.}, keywords = {evaluation, survey}, month = {December}, number = {4}, pages = {384--421}, posted-at = {2009-03-23 23:22:06}, priority = {2}, publisher = {ACM Press}, title = {Extracting usability information from user interface events}, url = {http://dx.doi.org/10.1145/371578.371593}, volume = {32}, year = {2000} }

TY - JOUR ID - citeulike:233339 L3 - citeulike-article-id:233339 IS - 4 JF - ACM Comput. Surv. SN - 0360-0300 AD - New York, NY, USA EP - 421 TI - Extracting usability information from user interface events PB - ACM Press VL - 32 SP - 384 KW - evaluation KW - survey N1 - the issue of difficulties in takeup comes into play here, partly attributed to costs (p. 414), but I also wonder if the stage of development and appropriate level of evaluation (formative vs summative) also influence types of tools needed. Tawanna: One question I had while reading this paper was if the framework would scale for ubiquitous technologies. For example, what happens in the case of social interaction between individuals? What about other dimensions of the environment that may need to be recorded such as temperature, sound, etc... How do other interactions such as text messaging or twittering come into play? Reasons for evaluating: to understand user behavior and performance, user thoughts and experience, to compare design alternatives, compute usability metrics, and certify conformance w/standards. Extracting usability- related information is difficult due to high and low frequency band events, issues with grammatical structures across hi/lo freq, context (For instance, when seeing the event MOUSE_PRESSED PrintToolBarButton, what is the user trying to do?), and user interaction at various levels of abstraction. state of the art: synch and search furthest along; ui events index into video (typically) doesn't mention: webquilt, experiscope; etc (pre-dates them) processing data: 1. Techniques for synchronization and searching: The most powerful and simplest technique for exploiting user interface events. Allows searches in one medium to find supplementary information in others. 2. Techniques for transforming event streams: Selecting, or filtering events; abstracting, or relating events to higher-level concepts of interest in analysis, recoding, or generating new event streams based on selection and abstracting. 3. Analysis: counts & summary stats; detecting sequences; characterizing & comparing sequences 4. Visualizing techniques 5. Integrated evaluation support (stuff that does all the above) Evaluation categories: 1. Predictive evaluation - making use of modeling to predict results. 2. Observational evaluation - Allow investigators to identify aspects of user behavior that may be difficult to identify using other techniques. 3. Participative evaluation: user self-reports. Captures aspects of users that are difficult to obtain otherwise. AU - Hilbert, David AU - Redmiles, David PY - 2000/12// UR - http://dx.doi.org/10.1145/371578.371593 ER -