title-trigger
Title trigger
An alternative to the abbr-design-pattern, to address that pattern's accessibility issues.
Contributors
- Author: Andy Mabbett
- Original idea: Patrick Lauke
Proposal
Use a class name; say "ufusetitle" (for "microformat, use title") or something equally unlikely to otherwise occur in the wild, on any element, to trigger the use of that element's title attribute. e.g.
<span
class="dtstart ufusetitle"
title="2007-08-16"
>
16th August this year
</span>
Conversion
All existing and new microformats should use this pattern; abbr-design-pattern should be deprecated, but parsers should still be required to recognise it on legacy pages (perhaps for a period of, say, three years)
Class names
Other potential class names include:
- hsource
- hvalue
- uftitletrigger
- ...
Advantages
- Simple
- Works on any element
- Usable on CMSs (e.g. MediaWiki) which do not allow use of
abbr - Easy to learn
- Ease of authoring
- Easy for parsers to adapt
Issues
Semantics
HTML is a markup language: its elements and attributes are intended to be descriptive. They are descriptors rather than commands. The proposed usage of the class attribute would be an instruction rather than a description (somewhat like the target attribute, now deprecated). This would be an abuse of the semantics of the class attribute.
Parsers
Existing microformats parsers (such as the Operator extension to Firefox) would need to be updated to handle this proposed alternative to the abbr design pattern. However, this is true of any proposed alternatives.
Complexity
No known or perceived issues.
Assistive technology
The proposed solution assumes that screen readers don't treat title as a special case on elements other than abbr or acronym (or perhaps a and img, which may need to be excluded from the solution) even so, the use of the special class name would be optional in such cases; and suitable guidance should be given to editors.
Is this actually the case? Please create a page for assistive technology title trigger results and add the results of testing there.
Usage in the wild
This design pattern is built on the assumption that the eventually- chosen is not widely used, where it used. Please document examples and references here: