RE: Costs of testing with Silver

Hi Everyone,

I think we would struggle to put together a document that provides well defined levels for types of organisation, or even size of project. Many project are updates to an existing web-estate, so lots of small project could then avoid requirements.

I’ll try and be solution focused, and suggest that:


  *   We lead with the user-requirements as ‘guidelines’ (as I suggested previously), with general and per-technology specific criteria underneath that guideline.

  *   Each guideline could have levels, like A/AA/AAA, except that it cuts the criteria into levels instead of the guideline. E.g. WCAG 1.3.1 for HTML could be split into:
     *   Guideline: The design is represented with appropriate structure and metadata.
     *   HTML Gold: Every element uses the right tag/attributes, and are appropriately nested (manual test).
     *   HTML Silver: Headings and lists are used and correctly nested, labels and for/ID relationships are valid.
     *   HTML Tool Bronze: The CMS provides a headings feature for content authors, and warns about full lines of bold text.
     *   HTML bronze: Headings and lists are used (with some pre-defined auto-wcag style tests)
(A quick, off-the-top-of-my-head example.)

  *   The requirement is not split between levels, but the amount of effort needed to achieve it might be.

  *   Some requirements are weighted more towards user-agents and authoring tools. If Wix/Squarespace/Wordpress et al provided options for (more) accessible output, smaller organisations would have less testing to do.

  *   One for John: At bronze the requirement for focus styles could be placed on the user-agent, but for silver/gold the requirement could be for the site.

  *   There could be a ‘slice’ of the criteria that are aimed at sites using a good tool provider, reducing the testing ‘surface area’.
NB: The tool provider would need to say that they fulfil the other requirements, so it becomes a marketing & procurement issue rather than a site development issue.

  *   I take John’s point that we have little or no leverage with the user-agents, but if we lead with the user-requirement, and provide ‘techniques’/  methods across websites/UA/authoring tools, it will make it much clearer where the effort needs to be applied!

  *   If we go down the route of levels for organisation capability, then it should be tied to other activities they are doing. For example, usability testing could be a valid method if the organisation already runs such testing in general, and that puts them above the small scale. This supports my recurring point that some things should be process-based rather than content based.

Cheers,

-Alastair