As far as I know, screen-readers that use the MSAA
buffer, such as NVDA, don't display the page or let you move around in it before
the page has fully loaded.
I''m not sure what is causing the behavior you are
describing. If a page fully refreshes, the expected behavior is that you
will be returned to the top and reading will begin from the top. If some
certain part of the page is being refreshed, my speculation is that that may
account for what you are experiencing if you are reading that part of the
I'm not sure if you can control such
behavior. But you can try changing a setting in options.
Do you know how to find the checkbox that says
something like warn me when sites attempt to redirect or reload the
If not, I'll tell you. But I have
no idea if that setting would have any effect on this problem because if
anything, only a part of a page is being refreshed, such as a region
containing updating information such as a stock ticker or certain text on a page
that may need to be kept current. I doubt this setting has any effect on
such behavior, though I don't know.
----- Original Message -----
Sent: Wednesday, July 27, 2016 12:45 AM
Subject: Re: [nvda] Some observations on usability aspects with
Firefox and NVDA
I remember there is an item in about:config to set when
will the webpage be shown, fully loaded and shown or immediately shown while
loading the remaining.
Felix G. 於 27/7/2016 13:41 寫道:
I'd like to ask if anyone else is observing the following
with recent versions of Firefox. When loading a page, whose title we shall
define as T, NVDA announces chatter of the following structure:
Mozilla Firefox. Unknown. Unknown. T Mozilla Firefox. T document
if you substitute anything substantial for T you can see how this would amount
to a lot of sylables with no useful information content.
Alt+Tabbing to Firefox produces:
T Mozilla Firefox, T
Finally, are there plans to implement a way of finding out
if a page has been fully loaded? I quite often find myself in the middle of
reading a paragraph when suddenly some asynchronous update happens and the
review cursor is thrown back a few lines or to the previous heading, to name
but two instances.
Any suggestions on any of those?