Closed nvaccessAuto closed 5 years ago
Attachment jmc.py added by aleksey_s on 2010-11-11 09:59 Description: the appModule in question Update: File added from Trac jmc.py.txt
Comment 1 by jteh (in reply to comment description) on 2010-11-16 06:13 Haven't looked at this yet, but some quick replies: Replying to aleksey_s:
- It automatically sets navigator to the output window whenever input field gets focused, because it is very inconvenient for user to activate flat review each time he/she wants to review text from server
Could a script perhaps be added to make the navigator jump to the output window? I'm not sure this automatic behaviour is ideal, as it breaks consistency. However, I agree that having to activate flat review (or better still, find the object with object navigation) is inconvenient.
- It automatically manages !ReviewCursorFollowsCaret option that it becomes disabled when user activates program window, so review always is on the output control even if user types something in the input field.
Again, I'm concerned that this breaks consistency. The user can set this with one keystroke if desired.
I find current implementation of last two options neither elegant, but since NVDA doesn't have some concept like favorite objects I do need to emulate it for user convenience.
I agree we need some sort of generic implementation for this kind of thing, as it's going to become more common now that we have !LiveText.
Comment 2 by aleksey_s (in reply to comment 1) on 2010-11-16 08:29 Replying to jteh:
Could a script perhaps be added to make the navigator jump to the output window? I'm not sure this automatic behaviour is ideal, as it breaks consistency. However, I agree that having to activate flat review (or better still, find the object with object navigation) is inconvenient.
Activate the script is not shorter than activate flat review. About consistency: i'd argue that this is an expected behavior for MUD client.
- It automatically manages !ReviewCursorFollowsCaret option that it becomes disabled when user activates program window, so review always is on the output control even if user types something in the input field.
Again, I'm concerned that this breaks consistency. The user can set this with one keystroke if desired.
One needs to turn it on/off each time he/she switches apps, because having this option turned off on regular basis isn't convenient either. While I understand a bit of inconsistency, I've implemented it that way to be competitive with commercial readers. With jaws, for example, you can enable reading all new text (insert+s), turn on jaws cursor, then type and review the output without switching modes. Remember, that each key press counts in real time game environment.
Comment 3 by jteh (in reply to comment 2) on 2010-11-16 09:42 Replying to aleksey_s:
Activate the script is not shorter than activate flat review. About consistency: i'd argue that this is an expected behavior for MUD client.
Perhaps, but it is certainly incosistent with normal NVDA behaviour, which makes it confusing for new NVDA users at least.
While I understand a bit of inconsistency, I've implemented it that way to be competitive with commercial readers. With jaws, for example, you can enable reading all new text (insert+s), turn on jaws cursor, then type and review the output without switching modes.
Using JAWS as the example, if you have your JAWS cursor tethered to your PC cursor, it behaves like NVDA's default. Granted, NVDA's default for locking review to system cursor is the reverse of JAWS's default, but nevertheless, it would require a toggle if a JAWS user had tethering enabled.
Comment 4 by jteh (in reply to comment 1) on 2010-11-20 07:11 Replying to jteh:
I agree we need some sort of generic implementation for this kind of thing, as it's going to become more common now that we have !LiveText.
I think this can be solved very simply by adding a reviewPositionOnFocus attribute. Normally, this will be the caret of the focused object, but it will be possible to override it to specify any !TextInfo. In this case, this will be the bottom of the output window. Does this sound okay?
The other problem is that caret moves review cursor currently moves the review cursor even if the user has navigated to a different object. Perhaps we could change this so that this doesn't happen unless the review cursor is actually in the focus object. I can't even remember whether the current behaviour was intentional. :) A focus change will still cause the review cursor to move due to focus moves navigator object.
Comment 5 by aleksey_s (in reply to comment 4) on 2010-11-20 09:52 Replying to jteh:
I think this can be solved very simply by adding a reviewPositionOnFocus attribute. Normally, this will be the caret of the focused object, but it will be possible to override it to specify any !TextInfo. In this case, this will be the bottom of the output window. Does this sound okay?
Definitely yes.
The other problem is that caret moves review cursor currently moves the review cursor even if the user has navigated to a different object. Perhaps we could change this so that this doesn't happen unless the review cursor is actually in the focus object.
I am for this change.
Comment 6 by mdcurran on 2010-11-22 23:17 As of abde7519d02c2e76753d79cc1fed3259037f8e87 review follows caret option now only allows review position to move with the caret if the caret moved within the same object as the review position. If the caret moves outside of the object nothing happens. For this appModule and other situations like it, Jamie and I have agreed that the best thing to do is to have the specific appModule/NVDAObject set the review position where it wants with in an event_gainFocus. We were considering things like a reviewPositionOnFocus property etc, but in reality, custom code moving review position on gainFocus is good enough anyway. Especially now that the review follows caret bug noted above is fixed.
This ticket looks quite abandoned. Is there still desire to include this appModule into core?
We cannot find the author of this issue and we do not have an update on this since many years. Meawhile, the most current version of Jaba mud client is 3.5. We also don't have any further requests related to this mud client at all from other users. I am closing this as abandoned.
Reported by aleksey_s on 2010-11-11 09:58 This is an inclusion request. Some notes about the module:
References: There are a few forks of this client, One of those you can download here: http://jmc.alfuken.net/en/