apply a feed pattern to our current semantics.

[tonyfast]

we've learned that cell numbers are important to connect the visual and audible experience with a notebook. a really frustrating part about this effort has been finding canonical attributes for the cell numbers.

this modification addresses some of this frustration:

• introduces the feed pattern inside the main documents
• uses form landmarks instead of region landmarks
• puts the form inside an article. now a cell is main>[role=feed]>article>form~*
• introduces aria-setsize aria-posinset as positional descriptors.

The feature that most distinguishes feed from other ARIA patterns that support loading data as users scroll, e.g., a grid, is that a feed is a structure, not a widget. Consequently, assistive technologies with a reading mode, such as screen readers, default to reading mode when interacting with feed content.

Feed Pattern

the notebook cells are simultaneously interactive widgets and readable objects. the feed pattern complements the interactive form pattern we've been using by explicitly defining a reading mode for the notebook document.

assistive tech users can navigate the forms in the landmark. the article and forms share a label, but the article will not announced in reading mode. the lack of duplication of the article label is the feature that allows us to connect the forms with feed pattern.

[tonyfast]

sample notebooks with this updated pattern https://iota-school.github.io/notebooks-for-all/branch/feed-pattern-smol/exports/reports/configs/#smol