[Development] Dictation Issues with Qt-built software

[Aleix Pol]

On Thu, Nov 3, 2016 at 2:50 PM, Morten Sorvig <Morten.Sorvig at qt.io> wrote:
>
>> On 1 Nov 2016, at 09:41, Robert Iakobashvili <coroberti at gmail.com> wrote:
>>
>> Dear Qt-Management and Developers,
>>
>> People cannot dictate to Qt-software at Mac as filed:
>>
>> https://bugreports.qt.io/browse/QTBUG-56811
>
> Hi,
>
> I see two possible ways to solve this:
>
> 1) Add cross-platform speech-to-text capabilities to Qt’s text input classes.
> This would be implemented using native API such as NSSpeechRecognizer, or an
> open source speech recognition library bundled with Qt. The behavior we get from
> this option may not be exactly native.
>
> 2) Use NSTextField in Qt applications. This gives us the exact native behavior,
> for speech recognition and everything else, including future NSTextEdit features.
> However, NSTextEdit would integrate on the QWindow level, and not for example
> as a Qt Quick scene graph item.
>
> Neither of these are straightforward, which is one reason why the bug remains
> open.
>
>>
>> The very reason to bother you and write this email
>> is because a similar previous dictation related issue at Windows filed in 2014
>> is still pending without being resolved:
>>
>> https://bugreports.qt.io/browse/QTBUG-43046
>
> This is an iOS (not Windows) bug. Also worth looking at though :)
>
> In general a good way to improve and maintain the accessibility implementation
> in Qt would be to give it a second user. UI testing and automation comes to
> mind as a good candidate.
>
> Morten

There's also this approach in QtSpeech, is it a dead end?
https://github.com/qt/qtspeech/tree/wip/speech-recognition

Aleix