Two suggestions for speechd-el

[Milan Zamazal]

>>>>> "CCU" == Cleverson Casarin Uliana <clever92000 at yahoo.com.br> writes:

    CCU> First, a "read continuously" command would be nice, meaning a
    CCU> command to start saying the contents of a buffer from the
    CCU> current cursor position to the end, and make the cursor follow
    CCU> the spoken text, so that when I stop speech, the cursor gets
    CCU> positioned near to the word or the line where the speech
    CCU> stopped. All of traditional screen readers for Linux and
    CCU> Windows have such feature.

We've discussed it yesterday.  I've been told that Orca does it by
splitting the text into pieces, sending them to Speech Dispatcher
separately and not before the previous piece of text gets spoken.  This
approach has some drawbacks.  I'd like to use index marking in
speechd-el instead.

This requires some changes in speechd-el: Support for handling event
notifications, support for detecting event notifications any time,
support for inserting index marks to suitable places in the synthesized
text (including user configuration) and support for moving the point to
the index mark place.  It shouldn't be very hard but it's not trivial
either, I'll try to look at it at the end of September.

I don't know whether any software uses Speech Dispatcher index marking
and how well it actually works.  We are aware of possible index marking
improvements in Speech Dispatcher.  Let's see how it works in speechd-el
once implemented and whether Speech Dispatcher developers would be
interested in making those improvements.  It might be useful not only
for speechd-el but also for improvements of continuous reading or other
processing in other screen readers.

    CCU> My second suggestion is about porting speechd-el to MS
    CCU> Windows. Because it's completely written in Elisp, I think the
    CCU> only requirement for it to work under Windows is to write a
    CCU> driver for outputting speech to a Windows-based technology,
    CCU> such as Sapi 5. Indeed, I suggest using Sapi 5 as it is present
    CCU> on all current flavours of Windows, and most synthesisers use
    CCU> it, like ESpeak itself.

There is some SSIP implementation for Windows:
http://www.remoteaccessbridge.com/ssip/.  You may want to try it.