Natural contextual reasoning for end users

G. Metaxas, P. Markopoulos

    Research output: Contribution to journalArticleAcademicpeer-review

    11 Citations (Scopus)
    1 Downloads (Pure)

    Abstract

    The realization and deployment of the Internet of Things require providing to non-programmers some level of programmatic control for tailoring system behaviour to their context and needs. We introduce a simple context-range semantics (CRS) and a context-range editor (CoRE) that support end users formulate and understand logical expressions regarding context. The editor builds on two key ideas (a) contextual information is used to evaluate and minimize logical expressions; (b) logical expressions are presented in a disjunctive normal form (DNF) thus applying a principle established in mental model theory. User tests reveal situations in which the theory regarding the intuitiveness of the DNF needs to be extended with a new element: Logical terms are easier to comprehend and formulate when grouped according to their semantic affinity. We report two experiments that demonstrate the intuitiveness of this approach and how it improves performance of non-programmers in specifying context sensitive system behaviour.
    Original languageEnglish
    Article number13
    Number of pages36
    JournalACM Transactions on Computer-Human Interaction
    Volume24
    Issue number2
    DOIs
    Publication statusPublished - Apr 2017

    Keywords

    • End-user development
    • ambient intelligence
    • context awareness
    • mental-models
    • natural programming

    Fingerprint

    Dive into the research topics of 'Natural contextual reasoning for end users'. Together they form a unique fingerprint.

    Cite this