Citation Context

...ing descriptive elements within the documentation (either text or images) with the actual application UI. Furthermore, users may find it difficult to keep track of where they are in the document (see =-=[1]-=- for a discussion). This is particularly problematic when there is control logic in the procedure (e.g., a branch based on state of the UI or on the user’s goal), or when unanticipated events occur. F...

Citation Context

... documentation require minimal resources to generate, they present a number of challenges for users, who struggle to keep track of instructions when switching between the tutorial and the application =-=[1, 15, 17]-=-. As a result, many researchers have worked on improving external software documentation by incorporating screenshots, graphics, animations, and annotations into textual content [9, 12, 13, 21]. Despi...

Citation Context

...ace algebra, CHINLE handles inconsistent training data differently from most machine-learning systems, which strive to be tolerant of noisy data. Systems based on decision trees [29], HMM derivatives =-=[16]-=-, support-vector machines, or similar methods continue to predict the highest probability action, even when the user’s actions contradict all representable hypotheses. While noise-tolerant methods are...

Citation Context

..., though our feature was never released. During the development process, we solicited user feedback on the resulting system and learned that poor usability was the key barrier to acceptance. Sheepdog =-=[1]-=- is a PBD system for learning to automate Windows-based system administration tasks based on traces of experts performing those tasks. For example, based on several demonstrations of experts fixing th...

Citation Context

... spend the time and effort reading through text [17] or may be unable to understand the terminology used in the text help. They have difficulty translating text from the help to widgets on the screen =-=[12]-=-. They can lose their place with the text instructions and miss crucial steps [11]. Creating text help documentation is also costly, magnified by translation costs if the application is targeted at in...

Citation Context

...rk on building mashups by example in Karma. In programming by demonstration, methods and procedures are induced from users’ examples and interaction. This approach can be effective in various domains =-=[13, 23, 33]-=- where users understand and know how to do such tasks. In CopyCat users may not know how to formulate queries and only interact with the system through the data. The interaction is in a table similar ...

Citation Context

...ical user interfaces. General terms: Human Factors; Design Keywords: Help; Contextual Help; Pixel Analysis INTRODUCTION Contextual help has been shown effective for learning graphical user interfaces =-=[3,4,13,16,17,18,25]-=-. Unlike traditional help based on screenshots or screencasts, contextual help allows users to receive help in the actual interface they are interacting with, rather than in another help interface suc...

Citation Context

... closely related to our work is the use of PBD for gathering collective procedural knowledge from multiple demonstrations and distributing it as executable procedure models, as in the Sheepdog system =-=[14]-=-. Sheepdog, however, is not suitable for documentation generation because it relies on a complex statistical model of the task that cannot be converted into a human-readable format. There are a number...

Citation Context

...y contain world-changing steps or steps with no effects known to the system. Most work in procedure learning relies on demonstration from the user (Lieberman 2001; Oblinger, Castelli, & Bergman 2006; =-=Lau et al. 2004-=-). The user steps through solving a task and the system captures the steps of the procedure and generalizes them. This is intuitive for users, but in some cases several examples may be required to fin...