[Athen] Perusall update

[Deborah Armstrong via athen-list]

No good news. I have now tried both JAWS and NVDA with Firefox, Chrome and Edge. Same problem. It has a ton of keystrokes for moving to different areas of the interface. But you must press those keys when your virtual cursor or browse mode is off. And once you turn the mode back on to actually read, you are then positioned at the top of the page.
I think the problem is that the virtual cursor or browse mode is looking at the HTML document model and does not pay attention to the position of the focus when caret browsing mode is on. Web pages in the days of the static web had no cursor.
But now, a web app is just like a desktop app; it has focus and apparently the screen readers are unable to track that.
But what's weird is that in apps like Office online and google docs they do. Also in web apps like Visual Studio Code and Discord, the focus is also tracked. So are these screen readers writing special scripts and add-ons for specific web-based apps instead of simply supporting the concept of a web-based app?
If that's so, that's band-aid access.
I wrote to Freedom scientific support, and they said they don't support Perusall, which of course isn't shocking. But as more and more apps transition to being browser based, the screen readers need to do a better job.
I really think this app is WCAG compliant with all the keyboard shortcuts and read out loud features it offers. And it has adjustments for fonts and colors and lets you resize all its panels and supports the dyslexic font. So I think the problem is with the screen readers that fail to understand the web is now an operating system.


From: athen-list <athen-list-bounces at mailman12.u.washington.edu> On Behalf Of Deborah Armstrong via athen-list
Sent: Thursday, April 10, 2025 8:54 AM
To: 'Access Technology Higher Education Network' <athen-list at u.washington.edu>
Subject: [Athen] Perusall update

So the folks at Perusall are very supportive. Even their CTO contacted me about my accessibility issue.  I am going to screen share with one of their developers in just a few minutes.

Perusall does automatically OCR any document the teacher embeds in its interface. But if it is a low-res image, such as a camera-phone photo of a page, the OCR is lousy. Hence professors really need to put accessible content in to the app.
For selecting text, the keyboard user turns on caret browsing mode and selects text with the shift and arrow keys. To annotate (comment) on it) enter is pressed and the user is now in an edit field, able to easily type and correct text, just as if they were editing a discussion in Canvas.
Screen reader users need to be sure to toggle the JAWS virtual cursor or the NVDA browse mode off. Mac users need to use Chrome or Firefox because they say, Safari doesn't have caret browsing. And I'd guess VoiceOver users would also need to disable the quick key navigation.
Anyway, the problem I believe is really one with the screen readers, as they don't support caret browsing. Once you turn off browse or virtual cursor mode, you get no feedback.
I've written to Freedom Scientific, saying this is a broader issue than just one web-based app. Caret browsing should be fully supported with feedback when that cursor is moved and focus changes.
The screen readers are stuck in this 20th-century paradigm of the static web page, where you occasionally use a Forms or Focus mode to fill in a form. But the reality now of web-based apps means they need to fully support the browser as if it is an operating system.
--Debee


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20250415/87785896/attachment.html>