Askus: Amplifying Mobile Actions

Shin'ichi Konomi, Niwat Thepvilojanapong, Ryohei Suzuki, Susanna Pirttikangas, Kaoru Sezaki, Yoshito TobePervasive 2009 · DOI

TL;DR

This paper introduces "Askus," a mobile platform designed to facilitate lightweight, networked actions by connecting users who need assistance with those who can provide it. By extending the concept of Web 2.0 to mobile devices, Askus aims to overcome physical limitations and enable real-time collaboration for everyday tasks.

What Problem Does This Paper Address?

In our increasingly connected world, information access is ubiquitous, but acting on that information in real-time, especially across physical distances, remains a challenge. Imagine needing to know if a train car is crowded, or if an elevator is waiting, or even asking someone to grab a flyer for you – these are small, everyday tasks that are difficult to accomplish remotely. The authors identify a gap in current systems that focus on information sharing but not on facilitating immediate, distributed actions. They propose Askus to bridge this gap by enabling users to request and receive help for "lightweight actions" from nearby or remote individuals.

Methods

The core of Askus is its Task Matching Protocol. When a user (requester) submits a task, the Askus server uses a scoring system to identify potential helpers (agents). This score considers:

  • Distance (d): Euclidean distance between the task location and the agent.
  • Past Success Rate (p): How often the agent has successfully completed tasks.
  • Average Task Completion Time (t): How long tasks typically take the agent.
  • Reputation Score (r): A dynamic score that increases or decreases based on feedback.
  • Status (s): Whether the agent is "busy" or "available."

The system offers two modes for agent selection:

  • Manual Mode: Lists candidates ranked by their score, allowing the requester to choose based on familiarity (friends vs. strangers) or past satisfaction.
  • Automatic Mode: Selects the agent with the highest score.

Task forwarding can be done serially (one agent after another) or in parallel (to multiple agents simultaneously). The system also incorporates elements of Distributed Cognition and Social Translucence to understand and support human activity in context and make social interactions visible.

Datasets

The paper describes two primary data collection efforts:

  • Diary Study: 11 male undergraduate/graduate students (ages 19-30) maintained hourly diaries for a day, recording events, requests, and feelings. This study informed the initial design of Askus.
  • Field Trial: 20 participants (mostly male, ages 19-25) used a mobile phone-based Askus prototype in a 1km x 1km area of central Tokyo. This trial involved a "structured session" where participants responded to pre-defined requests and a "free-form session" where they used the system as they wished.

No pre-existing datasets were used for training; the data was collected through these user studies for evaluation.

Key Insights and Ideas

  • The "Lightweight Action" Gap: The study highlights that many everyday needs involve simple physical actions (e.g., pushing an elevator button, finding a trash can) that are not well-served by current information-centric platforms.
  • Context is King for Matching: The success of Askus relies heavily on matching requests to agents based on a rich set of contextual factors beyond just location, including availability, past performance, and reputation.
  • Social Dynamics Matter: The study revealed that users are comfortable interacting with both friends and strangers, with nicknames and the server-based nature of the system contributing to this comfort. Privacy concerns and the potential for "request overload" are significant considerations.
  • Usability Challenges on Mobile: Text entry on mobile phones was identified as a burden, and frequent switching between applications for notifications was a usability issue. Battery life is also a concern for mobile deployments.
  • Incentives and Intrinsic Motivation: While material rewards might play a role, the study suggests that intrinsic motivations like mutual support and shared purpose are crucial for sustained participation.

Key Results

  • Diary Study: Revealed that 50% of messages were requests for actions, indicating a strong need for assistance beyond information retrieval.
  • Field Trial - Structured Session:
    • 21 users responded to 10 requests 180 times.
  • Requests R1-R4 (checking crowds, recommending restaurants, finding trash/restrooms) had high acceptance rates (>70%).
  • Requests R6 and R8 (getting free items, checking iPhone prices) had low acceptance rates (<10%), suggesting difficulty in accessing people or things.
  • Average response times were often within several minutes.
  • Field Trial - Free-form Session:
    • 23 participants generated 54 requests, receiving 165 requests and responding to 113.
  • 10 requests required physical actions, generating 21 responses.
  • 70% of participants felt comfortable sending requests to strangers.
  • 56% of participants felt comfortable receiving requests from strangers.
  • User Satisfaction: 55% of participants were satisfied with the system, though usability concerns (difficulty of text entry, notification management) were noted. 55% also found the system useful, particularly for navigation and information exchange in unfamiliar cities.

Limitations and Future Work

Limitations: * Short Field Trial: The trial was brief, with a specific participant pool, potentially leading to novelty effects and limited authenticity. * Location Resolution: The prototype's location resolution was considered coarse, and finer identification of places is desirable. * Privacy Implications: The understanding of privacy implications is still limited. * Battery Life: Mobile interaction can quickly drain batteries.

Future Work: * Long-term Investigations: Extend studies to various population segments and conduct long-term investigations. * Improved Usability: Develop more effective user interface tools to reduce cognitive burden, explore alternative input mechanisms (cameras, touch screens), and optimize battery usage. * Urgency and Prioritization: Implement mechanisms to handle request urgency and prioritize responses. * Incentive Systems: Explore points-based systems or other incentives to encourage participation. * Privacy Boundary Regulation: Develop better support for privacy boundary regulation and plausible deniability. * City Image Integration: Consider the "image of a city" to better articulate and understand place-relevant requests. * Technological Substrates: Explore implementing Askus on technological substrates that support spontaneous interactions and context-rich sensing, such as networked wearable devices.

Summary updated

BibTeX
@inproceedings{konomi2009askus,
  author = {Shin'ichi Konomi and Niwat Thepvilojanapong and Ryohei Suzuki and Susanna Pirttikangas and Kaoru Sezaki and Yoshito Tobe},
  title = {{Askus: Amplifying Mobile Actions}},
  booktitle = {Pervasive},
  year = {2009},
  doi = {10.1007/978-3-642-01516-8\_15},
}

Loading related content...

Figure