Hi Attila:

You can add this feature for icons simply by adding the following lines to your ~/.orca/user-settings.py or ~/.orca/orca-customizations.py file:

import orca.formatting
import pyatspi
orca.formatting.formatting['speech'][pyatspi.ROLE_ICON]['unfocused'] = \
    'labelAndName + imageDescription + roleName + positionInList'

What the above is doing is overriding the default formatting string, allowing you to customize the content to your needs.

Will

PS - You need the latest code from git master -- I had to make a change to _generatePositionInList to fallback to the parent of an object if the object wasn't of a role that we were expecting.

Unfortunately, your suggestion is not work. Nothing changed. I try add this code with orca-customizations.py file and src/orca/formatting.py.

Attila

(In reply to comment #2)
> Unfortunately, your suggestion is not work. Nothing changed. I try add this
> code with orca-customizations.py file and src/orca/formatting.py.
> 
> Attila
> 

With the changes in place (and making sure you have pulled/built/installed from the latest git master), try:

1) Ctrl+Alt+d to display the desktop
2) Arrow down to select an icon

You should hear the position information spoken.  If not, you might have a syntax error in the file(s) you've edited.

I forgot update. :-(
Works this method, now I insert default with src/orca/formatting.py file, but result long spokened text.
For example the hungarian translation file have following long string, for example with where am I usable is good, but not this situation:
#. Translators: this is an item in a list.
#.
#: ../src/orca/speech_generator.py:1904 ../src/orca/speech_generator.py:1984
#, python-format
msgid "item %(index)d of %(total)d"
msgstr "%(index)d. elem, összesen %(total)d"

Where am I not this string use when say the user what the current item index and how many items on the list?

So:
What your openion? We do  would like toggle this feature with check box, or insert this method with all controls with have list index with src/orca/formatting.py file? If all specification is clear, I begin work this function.

Attila

(In reply to comment #4)
> Where am I not this string use when say the user what the current item index
> and how many items on the list?

I'm having trouble understanding what you're asking.  :-(

> So:
> What your openion? We do  would like toggle this feature with check box, or
> insert this method with all controls with have list index with
> src/orca/formatting.py file? If all specification is clear, I begin work this
> function.

My opinion is that this feature would add annoying chattiness to focus tracking no matter what word choice is made.  Unless you have a very strong use case for having this information spoken during focus tracking, I think this belongs in only the Where Am I information.  Doing it there would also help prevent adding more clutter to the configuration GUI.

Ok, I try longer write what I would like hear when the icon is unfocused or focused:
When I add src/orca/formatting.py with positionInList for example with pyatspi.ROLE_ICON with focused amd unfocused lines, now result following output when the icon is selected:
"computer icon on item 1 of 6"
This is good in english, but when I use hungarian locale, I hear following long text:
"számítógép ikon 1 elem kijelölve, 6 elemből."
When the icon selected or unfocused, I would like hear following result with normal navigation, not where am I or detailed where am I:
"Computer icon 1/6"
When the user using where am I or detailed where am I operation, this result is good, not need change:
Where am I: "Icon panel. Computer icon. 1 of 6 items selected. On item 1 of 6."
Detailed where am I:
"Iconpanel. Computer Icon. 1 of 6 items selected. On item 1 of 6. Computer"
In hungarian translation:
Where am I: "Ikonpanel. Számítógép. 1 elem kijelölve, 6 elemből. 1 elem, összesen 6."
Detailed where am I: "Ikonpanel. Számítógép. 1 elem kijelölve, 6 elemből. 1 elem, összesen 6. számítógép."

How can replace the "on item 1 of 6" string only with focused or unfocused situations with "1/6, 2/6" form, but not change where am I or detailed where am I output?

Attila

Created attachment 136902 [details] [review]
This is a first rewision this function.

This is an experimental first version of this function, please not commit now.
What can I do?
1. Now I nod do yet a checkbox with configuration GUI.
First I separate with position in group (1904. line with speech_generator.py) output and position in list output (1984 line in speech_generator.py). Need change the 1984 line output string because after change this line, possible translate another text with this output.
Now I translate with this output with 1/6 form with hungarian translation file.
2. Insert positionInList with all important role object with focused and unfocused lines with src/orca/formatting.py file. Please no  surprise, because I not end with the pause bug (bug 585417), now I add the pause patch with src/orca/formatting.py file before I begin work this function. Of course, final patch not contains another bug patches.
One problem with this first patch: if any object where am I or detailed where am I using positionInList with src/orca/formatting.py file, the changed output spokening with short form for example with hungarian language. For example if I press where am I command with a menu item, Orca now says following output (I translating english):
"main menu. accessibility menu. a 1/13"
 I don't known this is problem with another not english language users or not.

Attila

Created attachment 136903 [details] [review]
Second revision.

This is the second release. I mistake the menu and menu items where am I item spokening, now fixed.

Created attachment 136974 [details] [review]
This is the new, I hope final version.

This is the 3. release of this function, i hope this is the final release. Please look if you have a little time.
Doed modifications and differences with prewious version:
1. Have checkbox and possible toggle this function with Orca GUI preferences/speech page.
2. This feature is turned off by default with src/orca/settings.py file, if the User would like the automatic spokening position indexes function, check the checkbox and use this function. Because this feature is turned off by default, not disturb another users if they would like old Orca speech method, similar with tutorial messages check box or mnemonic spokening feature.

3. Where am I work all time, because I insert forceList parameter with src/orca/where_am_I.py file.
4. I change speech order with positionInList with src/orca/formatting.py file by importance with 'focused' and 'unfocused' lines. For example if the user navigate a menu, higher priority with the mnemonic and accelerator information, the positionInList spokened after this informations spokened.

Attila

Created attachment 139280 [details] [review]
Actualized patch for latest Orca source code

Will, I actualized my feature patch with new Orca source code.
This is the final release. If you does'nt found critical problem my patch, I am ready with this feature.
The patch use the useExperimentalSpeechProsody setting determining pauses needs or not.
Now default enabled my feature (enablePositionindexSpeaking setting is true), but fel free modify this setting to false if you would like.

Attila

Created attachment 139300 [details] [review]
Slightly revised patch

This is a slightly revised patch. I'm leaning towards not enabling the feature by default since it adds a bunch of verbosity we've never had.

In addition, should we also consider adding this to braille?

If I understand right, you would like add this feature both speech and braille formatting output?
Very good ydea, but now I have'nt got a braille display device. But good news, wednesday I get an 20 cell braille display with me blind collage, and this display have me with 2009 september.
What important to do if we would like this feature both speech and braille output? Speech output is fullready.
We done this entire feature (speech and braille function) before UI freeze? What the UI freeze date?

Attila

Created attachment 139309 [details] [review]
Patch that was committed.

(In reply to comment #12)
> If I understand right, you would like add this feature both speech and braille
> formatting output?
> Very good ydea, but now I have'nt got a braille display device. But good news,
> wednesday I get an 20 cell braille display with me blind collage, and this
> display have me with 2009 september.
> What important to do if we would like this feature both speech and braille
> output? Speech output is fullready.
> We done this entire feature (speech and braille function) before UI freeze?
> What the UI freeze date?

I just committed the speech portion.  I think we can open a new bug for the braille portion and close this one.

I agree with Will that this can be closed with the default being off.  

If anyone wants to file a braille bug I feel that this information should be at the end of the braille context.  

Nice feature.
As a small suggestion, could be possible omit the word item when verbose is set to brief?
For example:
Instead of "file menu, item 1 of 7" orca could announce "file menu, 1 of 7".

Another place wher this feature would be useful, is in thunderbird. Would be nice to hear the  message position and the number of messages.

Jose, if I understand right, you would like hear this format when the verbosity set in brief:
"File menu, 1 of 7" so, disturb you with "item" word?
I think understand your problem.

Now, I edit my hungarian translation file, because too long the original hungarian output of this string and chhange following form only the translated string:
#. Translators: this is an item in a list.
#.
#: ../src/orca/speech_generator.py:1231 ../src/orca/speech_generator.py:1313
#, python-format
msgid "item %(index)d of %(total)d"
msgstr "%(index)d/%(total)d"

The original long hungarian translated output of this string was following and disturb me always when this feature toggled:
msgid "item %(index)d of %(total)d"
msgstr "%(index)d elem, összesen %(total)d"
So, the problematic part the "item" word with original text.
I have simple  ydea, I think this is good all people and not need verbosity toggle:
Need cut this string with two strings, a shorter and the longer form. The shorter form not contains the "item" word and equals with Jose wish when this feature is active, the longer form using where am I output command when user do where am I command (item %(index)d of %(total)d) string.
Will, what line number toggle with where am I output in
src/orca/speech_generator.py file? the 1231 line or 1313 line producing where am I message?

Jose, you wroted your prewious comment:
"Another place wher this feature would be useful, is in thunderbird. Would be
nice to hear the  message position and the number of messages."
Very good ydea, Ok, I do. One question: you better would like the position index information with all tables (Evolution, Pidgin for example), or enough this function with Thunderbird only? Will, not do this modification problem with Openoffice.org Writer table navigation, Firefox, etc?

Attila

If the word "item" is to verbose regardless of the locale, we should just remove it and then update all the regression tests accordingly.  We need to do so soon, however, since string freeze is approaching.  Mike - what are your thoughts?

Created attachment 139379 [details] [review]
Cut the item word with src/orca/speech_generator.py file with 1313 line number.

Jose, please test this patch. Now nod do your Thunderbird request, but I not forgotten.

Attila

I tested my prewious patch with Us english locale. Correct says for example with Jose would like:
Accessories menu, 1 of 13"
So, I think Jose first problem is solved.
Not disturb this change with another non english language translators, because this string is translatable.
Very near with string freeze? Anybody send me with current important GNOME dates mí e-mail address (UI freeze date, string freeze date etc)?

What happen with Jose Thunderbird request? Will we done this before any important freeze?

Attila

(In reply to comment #19)
> Created an attachment (id=139379) [edit]
> Cut the item word with src/orca/speech_generator.py file with 1313 line number.

There are a couple lines with "item" in them:

wwalker@opensolaris:~/work/orca/src/orca$ find . -name \*.py | xargs grep '"item'
./speech_generator.py:            result.append(_("item %(index)d of %(total)d") \
./speech_generator.py:            result.append(_("item %(index)d of %(total)d") \
wwalker@opensolaris:~/work/orca/src/orca$ cat -n speech_generator.py | grep '"item'
  1231              result.append(_("item %(index)d of %(total)d") \
  1313              result.append(_("item %(index)d of %(total)d") \

The reason I put "[mike]" on this bug is to get an opinion from our UI lead before making this change.  While changing these two lines is trivial in one respect, it causes a significant amount of work in another: all the regression tests need to be rerun, changed, and rerun again.  So, before making this change, I want to make sure it's what we really want to do.

(In reply to comment #20)
> Very near with string freeze? Anybody send me with current important GNOME
> dates mí e-mail address (UI freeze date, string freeze date etc)?

http://live.gnome.org/TwoPointTwentyseven -- please note that I want to be a bit more conservative this release.  That is, I don't want to push things as close as we can to the cut off dates.  The reason for this is that we have had a significant number of high impact changes already and I want us to spend a bit more time testing/hardening those things.

> What happen with Jose Thunderbird request? Will we done this before any
> important freeze?

The Thunderbird request is an enhancement request. I'd like to give priority to bug fixes at this point in the release.  If someone provides a patch, however, I will review it for this release.

(In reply to comment #17)
> Jose, you wroted your prewious comment:
> "Another place wher this feature would be useful, is in thunderbird. Would be
> nice to hear the  message position and the number of messages."
> Very good ydea, Ok, I do. One question: you better would like the position
> index information with all tables (Evolution, Pidgin for example), or enough
> this function with Thunderbird only? Will, not do this modification problem
> with Openoffice.org Writer table navigation, Firefox, etc?
> 
> Attila
> 
With all tables would be very good.

A small problem is that when I press the f10 key to activate the menu bar, I hear the "file menu" but orca doesn't announce the index position. If I press right and left keys, the item position is announced.

(In reply to comment #23)
> A small problem is that when I press the f10 key to activate the menu bar, I
> hear the "file menu" but orca doesn't announce the index position. If I press
> right and left keys, the item position is announced.

Works for me with gnome-terminal.  Which application are you using?

(In reply to comment #24)
> (In reply to comment #23)
> > A small problem is that when I press the f10 key to activate the menu bar, I
> > hear the "file menu" but orca doesn't announce the index position. If I press
> > right and left keys, the item position is announced.
> 
> Works for me with gnome-terminal.  Which application are you using?
> 
Firefox, thunderbird, pidgin, gedit ...

I look your wroted applications and your step to reproduce the problem.
Firefox: I press f10 key, I hear:
"file menu, 1 of 7"
Thunderbird: I press f10 key, I hear:
"file menu, 1 of 7"
Gnome-terminal: I press f10 key, I hear:
"file menu f 1 of 5"
Pidgin: I press f10 key, I hear:
"Buddyes menu b 1 of 4"

So, I can not reproduce your problem. I try this function with Ubuntu Jaunty, GNOME2-26 and 2.27.90pre Orca version.

Attila

(In reply to comment #26)
> I look your wroted applications and your step to reproduce the problem.
> Firefox: I press f10 key, I hear:
> "file menu, 1 of 7"
> Thunderbird: I press f10 key, I hear:
> "file menu, 1 of 7"
> Gnome-terminal: I press f10 key, I hear:
> "file menu f 1 of 5"
> Pidgin: I press f10 key, I hear:
> "Buddyes menu b 1 of 4"
> 
> So, I can not reproduce your problem. I try this function with Ubuntu Jaunty,
> GNOME2-26 and 2.27.90pre Orca version.
> 
> Attila
> 

I believe to be found the guilty! :-)
I am running with a configuration similar to yours. However I am running Ibm ViaVoice as my synthesiser. When I changed to ESpeak the problem has gone. Changing back to Ibm ViaVoice the problem appeared again.

I'm perfectly OK with just removing the word "item" if that is what everyone wants.  

(In reply to comment #28)
> I'm perfectly OK with just removing the word "item" if that is what everyone
> wants.  
> 

I opened the "item" word removal as bug #590070.  Please open the table row index (and possibly table column index?) suggestion as a separate enhancement request.  Thanks!