Advice for path forward

To date, our first public working draft (FPWD)
<> has
received comment from two important groups: assistive technology (AT)
vendors and browser vendors. And there are two major areas for concern:
authoring and parsing of SSML in HTML.

The comments from the browser perspective strongly advise in favor of the
"multi-attribute" approach described in our document. This is important
because of how browsers parse HTML, currently. While the AT perspective has
shown a preference for the "single-attribute" approach. The AT-related
comments express a preference for ease of parsing through scripting, not

This is a fundamental misunderstanding of our goals. We want AT to be
onboard with the authoring practices and, once the browsers have agreed to
our strategy, help move the pronunciation data into the axTree. At this
point, our work would need to change into a coordination of updates to the
accessibility APIs. Then AT should not have a concern over parsing.

The group that should express concerns over parsing would be the extension
vendors. Browser extensions must usually make use of scripted parsing
techniques to synthesize speech. Yet because they do not have an interest
in the accessibility APIs, they most likely will not show continued support
in our second phase. Additionally, extensions can already make use of
pronunciation information; it's just not specified or normalized.

It's my contention that we should first satisfy the browser vendors with an
open ear to the extension and AT vendors as it relates to authoring
practices. This path forward will allow us the most success in our ultimate
goals of bringing pronunciation information to AT in addition to extensions.

Received on Saturday, 7 August 2021 18:50:35 UTC