Checking for similarity or difference: Is the list of applicable success criteria complete? Otherwise a phrase like "for example" should be added.
Closed comments and questions
This page is a general introduction to different sampling methods. It does not address the specific requirements related to multi-page testing that we have identified in auto-wcag.
The challenge is: Some SCs can only be judged by inspecting multiple pages. There seem to be two types of scenarios:
- The SC requires all web pages to be similar (e.g. consistent navigation).
- The SC requires all web pages to be different (e.g. page titled).
The questions are:
- How many pages have to be compared?
- How are the results reported? (Is there only one outcome for the whole set of web pages? Or is there an outcome for each page?)
- If there is an outcome for each page: Do the pages get different outcomes? How do we define thresholds for the two scenarios mentioned above?
Comments on the sampling ideas:
- Why is the approach from "A sampling method based on URL clustering for fast web accessibility evaluation" recommended? (Is this a W3C endorsement? In the study of related work carried out for EIII, I compared several approaches and this one didn't look particularly promising. Mainly because it relies on strong assumptions that will only hold for a subset of web sites.)
- In statistics "weak" is a technical term related to the statistical power of a statement. Is this what you had in mind when introducing the term "weak sample"? If not, please use another word instead.