RE: how do screenreader users actually select, copy text?

Hi,

On Windows, both Jaws and NVDA read from a virtual buffer. So when the person is selecting the text, it is from that buffer. It is difficult for the application to know what is being selected, because that virtual buffer is not closely attached to the traditional DOM. At least that is my understanding. For this reason annotations are a serious problem. I believe Hypothesis has engaged with NVDA developers on this issue, but I have not heard of a work around as of yet.

Normally what I do in EPUB reading systems like Thorium is copy the text to the clipboard, create a bookmark using Thorium's bookmark function, and then copy the text to the note field associated with the Bookmark. When I want to go back to that location, I activate the bookmark to get me to the right page, and then copy the text of the note and search for it using the screen reader  search function. It is a lot of work, and I don't expect most people would want to go through this trouble, but it is all that I have found that works until the annotation problem is solved on Windows.

Not sure of this is an issue using VoiceOver, because I believe VoiceOver is pushing the mouse around instead of using a virtual buffer.

Best
George


-----Original Message-----
From: bryan rasmussen <rasmussen.bryan@gmail.com> 
Sent: Tuesday, October 5, 2021 3:17 AM
To: WAI IG <w3c-wai-ig@w3.org>
Subject: how do screenreader users actually select, copy text?

Hi,

I'm trying to make an annotation solution for visually disabled screenreader users, as such the closest workflow I can imagine is similar to that of selecting and copying text. reading through the description of how various screenreaders implement select text I have to image it works like this:

Screenreader user listening to text being read - 'saves a lot of space. Chromosomes are always found in the nucleus of eukaryotic cells; all eukaryotic'
decides I want to copy that text I just listened to, must go back to beginning of line, and then start listening word by word until they get to 'leaves' , then go back one word again to before leaves and then select until the end of the text.

Is this the correct workflow / how the experience actually is for people selecting text? Or is there another way of selecting text that a screenreader user would probably use to make it significantly easier to select text?

Thanks,
Bryan Rasmussen

Received on Wednesday, 6 October 2021 03:26:22 UTC