Comment 1 by expressABLE on 2010-10-07 23:06
In Zoomtext the feature is called Focus Enhancement, which can in turn be customized with various "Schemes", under the Magnifier toolbar heading. A 60-day demo of Zoomtext is available from AI Squared here: http://aisquared.com/zoomtext/more/download_zoomtext_trial/

Comment 2 by jteh on 2013-08-27 02:40
@nishimotz: I recently learnt about your focus highlight add-on, which covers much of this functionality. Would you be interested in getting this integrated into the core? Failing that, are you happy for us to do it based on your code?

Comment 3 by nishimotz on 2013-08-27 05:03
Please use my focus highlight add-on code for enhancement of NVDA core.

My implementation sometimes fail to follow current focus or navigator object, and I think it would be nice to be integrated with core system for further improvements.

Comment 4 by briang1 on 2013-08-27 07:42
The version of the add on as downloaded from the web site produced continuous nvda errors, but the tweaked version as on the dropbox link seemed to stop that. I myself cannot see how well it actually works though, so some sightded input might be worthwhile here before its merged into the main code I think.

Comment 5 by nishimotz on 2013-08-27 07:51
Work around the issue is as follows:

https://bitbucket.org/nishimotz/focushighlight/commits/d30e9de2d58e7e9975122a88403fe7fb1d1287a8

Comment 6 by k_kolev1985 on 2013-08-27 08:05
Hello Mr. Nishimoto,

After my e-mail to you from yesterday, I've found another bug. It seams that reading of opened combo-boxes is broken if using the "Focus Highlight" add-on. Examples of such combo-boxes are the ones in NVDA's interface (settings) and the ones for choosing a output file type in a "Save as" dialog (I think - no matter in witch application). Actually, it may as well affect all combo-boxes - I haven't managed to test.

See also: #6172

@feerrenrut: Is there still interest in having this integrated into NVDA core?

@nishimotz: I recall having some issues with screen review while using focus highlight, but it could have to do with the fact that I was using an older version. Could you elaborate on the what, the why and the how things have been changed in your add-on since your last comment of around 5 years ago?

The stateChange and valueChange events, regarding ComboBox windows, are specially handled in current focusHighlight add-on to avoid the issues.
However, I am not sure it is always safe.

Regarding display resolution and multiple monitors, the version 5.0 uses physical coordinates consistently in NVDA, as discussed in #7065.

So far, focusHighlight uses its own thread and message loop.
The event mechanism of NVDA can be improved for better implementation.
I cannot remember, but there are discussions regarding this.

Lastly, I have selected the colors of red and green.
It was not the best choice in terms of color universal design, so I am using dotted lines in addition to colors.

@nishimotz: Thanks for your information. I also read that you are now using GDI Plus since version 5, but what were you using before?

Have you considered using wx to draw the rectangles? That way, you can avoid interfacing with GDI at the library level, and if that works, I'd really prefer that for a core implementation.

@LeonarddeR before using GDI plus, I have been using legacy Win32 API, such as CreateWindow and FillRect.
In order to use physical coordinates consistently, I thought GDI Plus is preferable.
I have not tried using wx so far.

A screen-reader isn't only for the blind and I think that beeing able to see where on the screen focus is should be a part of the core product. It isn't a special feature for a certain application and should therefore qualify as a core feature.
I'm happy that I found the add-in. VoiceOver has its VoiceOver cursor that is a great help when partially sighted and the same goes for the focus highlight in ZoomText that I've been using set to inverted color.

I think there is some consensus within the developer space that a highlight feature should be part of NVDA core. Therefore I'm marking this as feature. Personally, I'd vote for exploring how we could do this with WXPython, since it is much more pythonic than interfacing with GDI or GDI Plus directly.

CC @feerrenrut

@nishimotz: From what I've seen, I think using wx might be the most elegant way within NVDA's core.

Instead of using one window per highlight, have you also considered using only one window that contains the several highlights, thus one transparent window that doesn't have to be repositioned, only cleared and painted again?

@LeonarddeR I am investigating how to use GDIPlus from wxPython.

Ok, yet wx it will be. I have a prototype implementation that will be released later this year as part of a bigger vision framework that also adds a screen curtain.

I am aware that my FocusHighlight add-on doesn't work with NVDA 2018.4 anymore, regarding navigator object within the Web browsers, such as Chrome and Firefox.
FocusHighlight just shows focuses on the web browsers even in the browse mode.
Until NVDA 2018.3.2, FocusHighlight could show both the navigator object and focus on the web contents.

As far as I have investigated, #8678 is related.
I don't think it is the regression of NVDA side because NVDA doesn't have visual highlighter itself.
I would rather think a more robust method is needed.

New vision framework i.e. PR #9064 snapshot is working regarding this as expected.

Now that #9064 has been merged, thanks to Leonard, Reef and Takuya for their good work on this, I think the basis for this feature is implemented. I am closing this issue now but please feel free to create new issues or new discussion based on the current implementation on last NVDA Alpha.

@Adriani90 I think I would prefer to keep this open until the Gui is implemented. While it is possible to enable this at the moment if you make the right changes in the config file, it is not really usable.
This should be addressed with #10082