Joker: A Unified Interaction Model For Web Customization
Name
Katongo-kkatongo-meng-eecs-2021-thesis.pdf
Description
Thesis PDF
Size
1.68 MB
Format
Adobe PDF
Checksum (MD5)
7b68b4d4bc0f1a2db45f4712508f033c
Author(s)
Katongo, Kapaya
Advisor(s)
Jackson, Daniel
Date Issued
June 2021
Publisher
Massachusetts Institute of Technology
Abstract
Tools that enable end-users to customize websites typically use a two-stage workflow: first, users extract data into a structured form; second, they use that extracted data to augment the original website in some way. This two-stage workflow poses a usability barrier because it requires users to make upfront decisions about what data to extract, rather than allowing them to incrementally extract data as they augment it.
In this thesis, we present a new, unified interaction model for web customization that encompasses both extraction and augmentation. The key idea is to provide users with a spreadsheet-like formula language that can be used for both data extraction and augmentation. We also provide a programming-by-demonstration (PBD) interface that allows users to create data extraction formulas by clicking on elements in the website. This interaction model allows users to naturally and iteratively move between extraction and augmentation during the customization process.
To illustrate our unified interaction model, we have implemented a tool called Joker which is an extension of Wildcard, a prior web customization system. Through case studies, we show that Joker can be used to customize many real-world websites. We also present a formative user study with five participants, which showed that people with a wide range of technical backgrounds can use Joker to customize websites, and also revealed some interesting limitations of our approach. Finally, we present a heuristic evaluation of our design using the Cognitive Dimensions framework.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright MIT
Persistent DSpace Link
