Comment 1 by Ahiiron on 2011-06-25 18:39
I think for this to work and still be snappy and such, it should only buffer the last 5 or so non-user invoked speech strings. That is, if you do a Say Line, NVDA should just speak the text without buffering it in the recently spoken strings list.. If, however, a event triggers this speech, it should be buffered.
I just wonder how much this would degrade performance with the constant updating and deleting of spoken text. Good idea though.

Comment 2 by briang1 on 2011-07-01 17:39
Perhaps it should just buffer the text in a kink of scratchpad and these should be cursorable through, with a description of what generated them. Could be useful for troubleshooting.
One of the main problems with some interactions are.

1. the message gets replaced by another one too fast. Very likely to happen with multiple tasks runnig.
   2 bubble balloon thingies. Most annoying.

3 Installers that flsh tings up so fast you just get a stuttered few sylables!

Comment 3 by m11chen on 2011-07-02 03:27
Just a thought, but doesn't the log already store the previously spoken text when level is set to input/output, so it wouldn't be too hard to access these previous speech outputs?

Comment 6 by Agent Golder on 2014-11-14 13:08
Can you please implement it?

@elliott94, this is now possible with the Speech History addon. I do not know if this will become a core feature of NVDA though. Is this what you were looking for?

This is exactly what I was looking for. :) Thank you, happy for this to be closed.

Closing because of #1609 (comment).