Dear Danielle,

Thanks for the info.

From a programming point of view it is a pity only a series of backspaces is sent....

Given the fact we are speaking about steno, and steno is all about speed, I have following question to you:

would it be acceptable to you that individual letters are completely spoken out when they are deleted? Not very ideal from a speed perscpective (user experience), but from a programming point of view perhaps a reliable solution could be worked out.

Otherwise, I guess we would need a mechanism to detect when the sequence of generated backspaces is over (and not waiting for the next key stroke): my first simplistic idea would be using a hold-off timer going off, but that is probably not a very reliable solution. (if I remember well from your video with steno you do 300 wpm or 5 words per sec or 200 millisec per word, counting 3 or 4 syllables per word means max 50 millisec per syllable or keystroke... that is pretty tight).

An alternative would be to create a plover plugin to be plugged in into nvda, that creates a backdoor open to plover in order to let plover instruct nvda to speak the complete syllable being deleted; obviously the plover code should be modified to give nvda this instrucction via this backdoor. At the positive side, I believe it would work more reliably and simpler to implement; at the negative side, I dislike this option as it requires code on both the nvda and plover side and I am not in favour of creating backdoors.

Please let me know what your opinion as user on the above question is / what behavior you exactly would expect.

Kind regards,

Didier

On 16/01/2018 18:30, Daniele Casarola wrote:

Sorry Joseph Lee, I don't have your contact, I can't ask to add you CC. Can you leave it here? Thanks.

Thank you Didier Colle for your question.

This is the answer of one of the Plover developer:

"When undoing a translation, a series of backspace events are sent. Really, just like a user would do when deleting some previously typed text; doesn't this use case work with NVDA?"

So, it seems to be as you supposed.

About your second question.
As I wrote above, to have audio feedback when I press the "delete" button, the option  “Handle keys from other applications” should be on.
NVDA reads only one letter of the group of letters I delete, it seems to be only the first or the second letter.

For example I type "telefono" using 3 strokes: te, le, fono.
To delete the word I should press the "delete" button 3 times of cours. The first time I hear the letter "f", the second time "l" and the third the letter "t".

In this case it reads the first letter of each group of stroke. But sometimes it reads the second. Honestrly I don't understand the logic, it seems randomly.
Maybe it depends what I've typed before, if strokes with even or odd number of letters, maybe...
Anyway always reads the first or the second letter of the group.