After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 111853 - "switch to window" -> "switch to: <window label>"
"switch to window" -> "switch to: <window label>"
Status: RESOLVED DUPLICATE of bug 107638
Product: gnopernicus
Classification: Deprecated
Component: general
unspecified
Other Linux
: Normal normal
: ---
Assigned To: remus draica
remus draica
Depends on:
Blocks:
 
 
Reported: 2003-04-29 17:42 UTC by Jeff Waugh
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Jeff Waugh 2003-04-29 17:42:48 UTC
the "switch to window" message would be heaps cooler if it said "switch to:
<window label>", for example:

  "switch to terminal"
  "switch to galeon"
  "switch to x-chat"

Thanks!
Comment 1 Adi Dascal 2003-04-30 14:33:35 UTC
This kind of presentation should be present, indeed. I think that due
to some priorities that were introduced some time ago, this feature
was lost. We will investigate more on this bug and hopefully solve it ;-).
Comment 2 remus draica 2003-05-05 13:51:02 UTC
Jeff: Behaviour described by you is already implemented.

It is posible not to work for you if you are using festival. Can you
tell me what speech driver is installed on your computer?

Comment 3 Jeff Waugh 2003-05-05 15:42:15 UTC
Using festival, latest tarball releases of the a11y stack.

Thanks!
Comment 4 remus draica 2003-05-06 10:39:32 UTC
In case of using festival, the problem is same as one described in bug
107638 (speech markers are missing for some speech drivers).

For better results you must use ViaVoice.

So, this bug is a duplicat of bug 107638.

*** This bug has been marked as a duplicate of 107638 ***
Comment 5 bill.haneman 2003-05-06 11:05:09 UTC
I do not think the resolution of this bug (i.e. "must use viavoice or
other driver") is satisfactory.

We *must* be able to support festival since it is the most commonly
available TTS engine.

Comment 6 korn 2003-05-06 22:22:14 UTC
It is not clear from the thread on this bug whether the problem is a 
mix of TTS engines with one of them being festival (the known bug 
#107638), or whether this problem is occuring when the only TTS engine 
on the system is Festival.  Jeff - could you clarify?

Also, please note that the general statements:

 1. "For better results you must use ViaVoice."
and
 2. "We *must* be able to support festival since it is the most 
commonly available TTS engine."

are not necessarily at odds.  We know that there only certain things 
we can do if we have speech markers (like making the magnification 
ROI track the words being spoken by the TTS engine).  Also, it is 
generally accepted that Festival does  not speak as clearly or as 
quickly as other engines (e.g. ViaVoice).  "For better results" use 
something that isn't Festival should be OK.

The question is whether we can have a core feature - like this one - 
not work with Festival and be OK with that.  There, I think the answer 
is as Bill says: this feature should work with Festival.

But, must all features?  As with the magnification case above, I don't 
see how we can make that happen short of major modifications to 
Festival which is outside the scope of the Gnopernicus project.
Comment 7 Jeff Waugh 2003-05-07 04:06:42 UTC
I only have festival. :-)
Comment 8 bill.haneman 2003-05-07 10:41:35 UTC
I agree with Peter; while we can't expect gnopernicus to provide the
same quality user experience/same featureset with Festival as with
other TTS engines, the user experience should be reasonably "intact"
in both cases.

If we are relying so heavily on things like speech callbacks that
major features of gnopernicus break when Festival is used, the end
result is that gnopernicus+festival is not a useful combination.  I
think we should continue to target Festival as a usable TTS engine for
gnopernicus, thus we need to address problem like this bug.

In general we need to make sure that gnopernicus is not _assuming_
that callbacks work, in terms of its output, but rather that
gnopernicus is _aware_ of whether callbacks are working, and adjusts
its behavior accordingly.

The main problem seems to be with sequential/preemptive utterances.  I
think the desired behavior should be that sequential utterances should
not preempt one another unless the client explicitly requests that
they do so.
Comment 9 remus draica 2003-05-07 11:33:30 UTC
Soory for my bad phrase "For better results you must use ViaVoice."
My intention was to say that "with ViaVoice you will achieve the
expected results".

My english is not so good, so some phrases doesn't show my real
intention (what I really want to say) :-).

Also I want to say that in my opinion all discutions about support of
speach markers for festival and how gnopernicus works with festival
(and all speech engines which has no support for speech markers) is
the subject of bug 107638. This was the reason why I closed this bug
as duplicate of the above one. This bug is only a special case of bug
107638. If that bug will be fixed, then this will be fixed too.
Comment 10 bill.haneman 2003-05-07 11:59:39 UTC
Hi Remus:  don't worry about your English, I think I understand what
you meant to say.

I think you are right, perhaps we should re-close this as a duplicate.
 However I will mark bug 107638 to mention this discussion.



*** This bug has been marked as a duplicate of 107638 ***
Comment 11 Jeff Waugh 2003-05-07 14:05:46 UTC
Thanks - just reading your comments here is giving me a much better
understanding about the challenges and issues involved with a11y
support. Great to read! :-)