Successfully reported this slideshow.

More Related Content

A macroscopic web accessibility evaluation at different processing phases

  1. 1. A macroscopic Web accessibility evaluation at different processing phases Nádia Fernandes, Luís Carriço 9th International Cross-Disciplinary Conference on Web Accessibility – 16/17th April 2012 – Lyon, France.
  2. 2. Motivation 9th International Cross-Disciplinary Conference on Web Accessibility – 2 16/17th April 2012 – Lyon, France.
  3. 3. Introduction • The Web is being used by all kinds of people; • Web sites must be accessible; • Modern Web development transcends static HTML; • A more real evaluation is necessary. 9th International Cross-Disciplinary Conference on Web Accessibility – 3 16/17th April 2012 – Lyon, France.
  4. 4. Previous work 9th International Cross-Disciplinary Conference on Web Accessibility – 4 16/17th April 2012 – Lyon, France.
  5. 5. QualWeb evaluator • Accessibility evaluation of Web pages: – using WCAG 2.0 , • 18 HTML techniques; – at different processing Before phases • BEFORE and AFTER Browser processing; After 9th International Cross-Disciplinary Conference on Web Accessibility – 5 16/17th April 2012 – Lyon, France.
  6. 6. QualWeb evaluator: execution 9th International Cross-Disciplinary Conference on Web Accessibility – 6 16/17th April 2012 – Lyon, France.
  7. 7. Some improvements... 9th International Cross-Disciplinary Conference on Web Accessibility – 7 16/17th April 2012 – Lyon, France.
  8. 8. PhantomJs • is a command-line tool that uses WebKit , • works like a WebKit-based Web browser (simulation); • can be controlled using Javascript. 9th International Cross-Disciplinary Conference on Web Accessibility – 8 16/17th April 2012 – Lyon, France.
  9. 9. Problems Solved! 1. It avoids data injection at the browser level; 2. Evaluated Web page before and after browser processing is exactly the same; 3. Integration with a crawler. 9th International Cross-Disciplinary Conference on Web Accessibility – 9 16/17th April 2012 – Lyon, France.
  10. 10. Research Question How do the macroscopic properties emerging from Web accessibility change in respect to the processing phase of delivery? 9th International Cross-Disciplinary Conference on Web Accessibility – 10 16/17th April 2012 – Lyon, France.
  11. 11. Experimental Study • We evaluated a set of Web pages from a list provided by the Portuguese Web Archive (version of 2008); • We used the QualWeb evaluator. 9th International Cross-Disciplinary Conference on Web Accessibility – 11 16/17th April 2012 – Lyon, France.
  12. 12. Metrics • Results of the evaluation are presented in terms of: – PASS, WARN, and FAIL (Applicable = PASS + WARN + FAIL) • Metrics used: – rate conservative = – rate optimistic = – rate strict = • The results are between accessible (100%) and not accessible (0%). 9th International Cross-Disciplinary Conference on Web Accessibility – 12 16/17th April 2012 – Lyon, France.
  13. 13. Results • 24, 462 Web pages evaluated. • Total Number of HTML elements: – before processing - 24,918,720 – after processing - 41,967,072 – (Ratio ≈ 1.7). • Average Number of HTML elements per page: – before processing - 1010 – after processing – 1710 – (Ratio ≈1.7). 9th International Cross-Disciplinary Conference on Web Accessibility – 13 16/17th April 2012 – Lyon, France.
  14. 14. The differences of an HTML document between both Processing Phases were observed. 9th International Cross-Disciplinary Conference on Web Accessibility – 14 16/17th April 2012 – Lyon, France.
  15. 15. Results: Average Outcomes • Successes – before processing - 9 elements – after processing - 87 elements – (Ratio ≈ 9.7). • Failures – before processing - 46 elements – after processing – 176 elements – (Ratio ≈ 3.8). • Warnings – before processing - 262 elements – after processing – 451 elements – (Ratio ≈ 1.7). 9th International Cross-Disciplinary Conference on Web Accessibility – 15 16/17th April 2012 – Lyon, France.
  16. 16. Results: rates Before and After processing • Conservative rate: – The average quality increases after processing. – Accessibility quality between 60% and 90% disappeared after processing. • Optimistic rate: – The average quality decreases after processing. – Some results lower than 20% disappeared after processing. – Decrease of higher accessibility quality, leading to a lower accessibility average. 9th International Cross-Disciplinary Conference on Web Accessibility – 16 16/17th April 2012 – Lyon, France.
  17. 17. Results: rates Before and After processing • Strict rate: – The average increased after processing. – Results higher than 85% disappeared after processing. – Worse pages before processing get higher scores after processing. – Better pages before processing are ranked lower after processing. 9th International Cross-Disciplinary Conference on Web Accessibility – 17 16/17th April 2012 – Lyon, France.
  18. 18. Limitations 1. Techniques coverage 2. Dynamic content 3. Automatic evaluation 9th International Cross-Disciplinary Conference on Web Accessibility – 18 16/17th April 2012 – Lyon, France.
  19. 19. Discussion • Evaluations before processing clearly is not the best option! • We used/percept/interact with the after processed version. • Considering the rates… – Web pages possess higher uniformity – That can be explained with reusable code 9th International Cross-Disciplinary Conference on Web Accessibility – 19 16/17th April 2012 – Lyon, France.
  20. 20. Discussion • Impact on Designing Accessible Web Pages – Importance of sharing reusable code; – High quality reusable code produce better quality pages. • Impact on the Perception of Accessibility – Each metric identify different perspectives; – It is important that evaluations assess what user perceive. 9th International Cross-Disciplinary Conference on Web Accessibility – 20 16/17th April 2012 – Lyon, France.
  21. 21. Conclusion • We presented a large-scale study of accessibility on the Web. • We were able to characterize some accessibility properties of the Web, pointing some differences between processing phases. • The results obtained on the evaluation of pages after browser processing tend to be more homogeneous than before. • Considering that the end-user interacts with the after processed pages then most studies about Web quality should be redone. 9th International Cross-Disciplinary Conference on Web Accessibility – 21 16/17th April 2012 – Lyon, France.
  22. 22. Future Work 1. Enlarge the coverage of WCAG 2.0 implemented tests 2. Evaluate Rich Internet applications 3. Perform a comparative set of studies 9th International Cross-Disciplinary Conference on Web Accessibility – 22 16/17th April 2012 – Lyon, France.
  23. 23. nadiaf@di.fc.ul.pt 9th International Cross-Disciplinary Conference on Web Accessibility – 23 16/17th April 2012 – Lyon, France.

×