From Search Engines to Augmented Search Services
0 likes100 views
This document discusses an approach to augmenting web searches by enabling users to customize searches and perform ancillary searches without leaving their current web context. The approach defines Augmented Services that match the user interface of existing search engines using DOM annotation and automation. A Firefox extension allows users to perform on-demand, in-context searches that reuse existing search mechanisms and present results without opening new tabs. This reduces the time and interactions needed for searching. The approach was validated on the top 20 sites from Alexa and demonstrated success in integrating search results into the original web context. Further work is needed to experiment with non-technical users and define metadata for search services.
From Search Engines to Augmented Search Services
- 1. LIFIA, Facultad de Informática, UNLP - Argentina ICS, IRIT, Université Toulouse III - France G. Bosetti, S. Firmenich, A. Fernández, M. Winckler & G. Rossi From Search Engines to Augmented Search Services: An End-User Development Approach ICWE 2017
- 2. Motivation
- 3. 3Primary search 1. Motivation 2. Web Objects Ambient 3. Future work Google ICWE to Google WWW to
- 4. 4Ancillary search 1. Motivation 2. Web Objects Ambient 3. Future work 1 2 3
- 5. 5How users are supported? 1. Motivation 2. Web Objects Ambient 3. Future work 1. Manual interaction with Web apps and Browser 2. Web Browser support for new search engines
- 6. 6Just an example 1. Motivation 2. Web Objects Ambient 3. Future work
- 7. 7Just an example 1. Motivation 2. Web Objects Ambient 3. Future work not supported✗ supported✓
- 8. 8The Underlying problem 1. Motivation 2. Web Objects Ambient 3. Future work Not all search engines in the Web are directly available from the browser’s ancillary search mechanisms a. Disadvantage for «not popular» Web sites b. End users have to wait for an extension to be implemented 1
- 9. 9And anyway... 1. Motivation 2. Web Objects Ambient 3. Future work Results are presented in another context
- 10. 10The Underlying problems 1. Motivation 2. Web Objects Ambient 3. Future work Not all search engines in the Web are directly available from the browser’s ancillary search mechanisms a. Disadvantage for «not popular» Web sites b. You have to wait for an extension to be implemented 1 Results are presented in a second context (a new tab) a. Extra interactions for comparing b. Time consuming 2
- 11. Our Approach
- 12. 12What? 1. Motivation 2. Augmented Search Services 3. Future work An inside-in approach empowering users to customize their searches, by ➔ defining Augmented Services matching the existing search engines’ UI components ⬝ by means of DOM annotation and automation of interaction with the DOM ➔ performing ancillary searches without leaving their current Web context ⬝ by means of Web Augmentation
- 13. 13How, when, where, who? 1. Motivation 2. Augmented Search Services 3. Future work ➔ A Firefox extension ➔ On-demand ➔ In-situ, almost any Web site ➔ End-users Benefits ➔ reuses existing search mechanisms & content ➔ transforms search results into domain objects ➔ reduces the user’s efforts for searching ⬝ time ⬝ interactions ➔ Integrates the resulting domain instances in the current Web context
- 14. 14Defining Augmented Services 1. Motivation 2. Augmented Search Services 3. Future work search input search input filters results spec.
- 15. 15Defining Augmented Services 1. Motivation 2. Augmented Search Services 3. Future work
- 16. 16Defining Augmented Services 1. Motivation 2. Augmented Search Services 3. Future work
- 17. 17Executing ancillary searches 1. Motivation 2. Augmented Search Services 3. Future work
- 18. 18Executing ancillary searches 1. Motivation 2. Augmented Search Services 3. Future work
- 19. 19Preliminary results 1. Motivation 2. Web Objects Ambient 3. Future work ➔ A quantitative Assessment based on GOMS-KLM
- 20. 20Preliminary results 1. Motivation 2. Web Objects Ambient 3. Future work Validation by construction ➔ Success definition in 18/20 first sites of the top 500 by Alexa ⬝ the ones with a search engine ⬝ in a language we could understand (not chinese or russian, sorry!) ⬝ with different top-level domain are considered just once ➔ composed by different UI controls. E.g. ⬝ 17 of the full list of sites have a trigger element ╶ 10 buttons, 6 inputs and 1 anchor ➔ With different “loading-data strategies” ⬝ full results in the html response ⬝ ajax call ⬝ https://sites.google.com/site/webancillarysearches/
- 21. Further work
- 22. 22Further work ➔ Experiment focused on non-technical end-users ⬝ potential of adoption ⬝ the usefulness of in-situ visualizations ➔ Consider defining metadata for Search Services ➔ Deep into domain-specific visualizers ➔ Collaboration in ... ⬝ Repo and communication platform for sharing templates 1. Motivation 2. Web Objects Ambient 3. Future work
- 24. 24Disadvantages ➔ We did not create yet a solution involving multi-interaction for selecting a single control (e.g. hotmail) ➔ Resilience & security are the common challenges in Web Augmentation 1. Motivation 2. Web Objects Ambient 3. Future work
- 25. 25Executing ancillary searches 1. Motivation 2. Augmented Search Services 3. Future work
- 26. 26Executing ancillary searches 1. Motivation 2. Augmented Search Services 3. Future work