Re: Giving the right amount of feedback


Tyler Spivey
 

I agree with you, and it's not just state transitions.

NVDA really doesn't focus on user experience, and it shows. I can count some things off the top of my head.

1. The focus issues. There are threads about this on this list. I seem to remember there being code in NVDA which will drop focus events if too many of them come in.
2. If you have NVDA going through a specific audio device and you switch your default, NVDA will simply stop speaking, or switch to another audio device.
3. NVDA uses a keyboard hook to intercept all keys. Sometimes it gets into a state where it doesn't pass the keys anywhere.
the result of this is your keyboard no longer works, except for win+l and ctrl+alt+delete. If you press ctrl+alt+delete, you might think you can launch task manager.
As soon as you do, you're back on your normal desktop with NVDA still capturing the keyboard.
The only thing you can do is sign out, losing some or all of your work, or figure out how to kill NVDA either with a second account or with a non-keyboard-based solution.
4. NVDA has a review cursor. In consoles and edit fields, if I press NVDA+backspace, NVDA moves my navigator to the focus, where I can use that review cursor to review my text again if I moved it.
This is useful for, say, finding the top of a function in my editor without moving the system cursor.
However, what happens if you do this same thing in browse mode? Your navigator gets moved to the focus, but you then can't move the review cursor around like you could when you first entered the browser. IIRC this is by design.
5. If a program stops responding, NVDA goes with it. Take Event Viewer.
View one event's details with enter. From there, press escape and be prepared to figure out how to get NVDA back, because it's most likely going to stop speaking.
6. We know that in notepad, WordPad, Microsoft Word, and other applications, we can press ctrl+f to find things.
However, in browsers, we need to use ctrl+NVDA+f. Why? Because NVDA doesn't want to override the inaccessible browser find dialog.

I'm sure I'm missing a few things. But IMO, none of this is good user experience.


On Mon, Jul 22, 2019 at 12:49 AM, Perry Simm wrote:
Hi!
Lately I've been wondering about the research that has been done to figure out the right amount of feedback that NVDA should give in different circumstances. These thoughts were triggered by the following incident:
I opened up Chrome, and it came up "not responding" which, on a slow machine, can sometimes happen and is no immediate cause of concern for me. My Braille display kept showing "New Tab - Google Chrome (Not Responding)". I decided to wait it out before throwing key presses at it. When neither speech nor Braille updated for about two minutes, I decided to press tab and, much to my confusion, found that Chrome had recovered and was fully prepared to take my instructions.
When something goes from not responding to ready, wouldn't it be logical to expect NVDA to tell me about this transition without me having to poll?
I have a nagging suspicion, and I apologize for sounding overly critical here, that these user experience patterns are not researched at all, not even consciously specified. Rather, they are just left to random luck based on how the operating system decides to handle or propagate events. In a world where nearly every aspect of user experiences is designed and micromanaged, I feel we are lagging behind and left with unprofessional hit or miss situations. What we need are reliable, reproducible patterns or sequences which can be taught to students, fully duplicated by them, and then applied consistently, especially in workplace environments. State transitions like the one I described above must be just as obvious to the NVDA user as they are to the sighted user if he or she has been trained to recognize them.
This is not rocket science. If fifty years ago it was possible to program a computer to behave consistently enough to get people to the moon, then today it must not only be possible but trivial to program it to deal consistently with common everyday exceptions while working with email and the web. For people on the autism spectrum, for example, digital systems can be a bedrock of certainty and consistency in an otherwise random sea of white noise.
I am seeing strong indicators that user experience research in the area of screen readers is overdue. Most importantly, however, the user experience should always be a conscious decision, never left to fate or accident.
Cheers Perry

Join nvda@nvda.groups.io to automatically receive all group messages.