Change core browsers for JAWS?

[mfairchild365]

The new screen reader survey from WebAIM has been published!

Currently, we are treating IE, FF, and Chrome as "core" browsers for JAWS. This means that every test must be run against all 3 of those combinations. Some notes from the new survey:

• JAWS with FF now has a 5.9% usage. That falls below the 20% threshold for this project. In the previous survey, FF was at 15.1%, which also fell under the 20% threshold.
• JAWS with Chrome now has 21.4% usage. In the last survey, JAWS and chrome only had 6.5%.
• IE now has 11.5% usage. In the previous survey, IE had 24.7% of usage and it appeared that Firefox would become the primary browser for JAWS. However, it became clear in late 2018 that most JAWS users might be using Chrome instead. This is how we wound up with 3 "core" browsers for JAWS.

We should strive for the fewest number of combinations as possible, as each combination requires more manual testing and input, and will result in slower progress with this project with potentially very little added value.

Questions

• Should we drop FF as a "core" browser for JAWS? 5.9% usage is very low, however, the browser landscape for JAWS appears to have been very volatile in the last couple of years.
• Should we drop IE as a "core" browser for JAWS? 11.5% is below the 20% threshold for the project, and IE is no longer being updated. However, I know that especially in corporate networks where software updates can move slowly, many JAWS users might be stuck with IE.

Note: if a browser is not considered "core", it is still considered to be "extended". Results for "extended" browsers can still be populated, but are not included in summaries of support and are buried in the interface to reduce clutter.

[scottaohara]

the concept of a core browser at this point is a bit murky with there now only being 3 major browser engines on the market, and then ie11 being a legacy application that microsoft has even said “please don’t consider this a browser anymore”

therefore i think, regardless of usage numbers, it’s important to test with the two browser engines that would be used with JAWS/NVDA, as surfacing support/gaps between the two can be important to provide evidence for a browser/screen reader to pick up the slack.

regarding testing / noting IE11 as a core browser.... eh. Results there are largely going to be indicative of whether the AT is filling in the a11y gaps or not with IE11.