[Digikam-devel] [Bug 112799] Hierarchical tag selection on right pane does not work correctly

[Thomas McGuire]

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
         
http://bugs.kde.org/show_bug.cgi?id=112799         




------- Additional Comments From Thomas.McGuire gmx net  2007-02-15 01:49 -------
> if you look at bug #140675, it says:
> > Now i have picture taged by animals and not birds or
> > insects and i'm not able to show only this animals-taged.
>
> Exactly this situation is resolved now, as I showed in the screenshot:
> > Now i have a pictured tagged by Obst and not Rot or
> > Grün, and i AM able to show only this Obst-tagged.

I don't think this is the same. According to comment #6, you did not tag a 
picture with Obst and Grün at the same time.
Then, it would still not be possible to see only Obst which is not Grün.

Same with the animals. The reporter said
>Now i have picture tagged by animals and not birds or insects and I'm not 

able to show only this animals-tagged 
The problem is that he additionally tagged the bird and insect pictures with 
the "animal" tag. Now, when only selecting the animal tag in the tag filter, 
the bird picture is of course shown, because it also has a (superfluous) 
animal tag.
With this, it is not possible to show all animals which are NOT birds or 
insects. Same as you described below really.


> But I do understand, what you are talking about. You say, each tag checkbox
> should not only have *two* states, i.e. "checked => has to be tagged with
> this" and "unchecked => doesn't matter", but also a *third* state "third
> checktype => must not be tagged with this". I am not sure, if Qt provides
> for such a checkbox type. If yes, it wouldn't complicate the user
> interface:
>
> [x] Family
>   [-] Sisters
>   [-] Brothers
>
> as additional possibility to
> [x] Family
>   [ ] Sisters
>   [ ] Brothers
>
> Instead of using a [-] which causes confusion with expansion/de-expansion
> symbols, one could probably use a red check for the negative selection, and
> a green check for the positive selection. If this is also what you have in
> mind, and Chris is satisfied with the current solution, we should carry
> this to a separate wish report.

Yes, you understand me, thanks :) 
I think this was possible in previous versions (at least the checkbox was 
tri-state), but had many bugs and was weird. The current tag filter is much 
cleaner and nicer, I am really not sure how it all could fit in. I'll think 
about it.
Green and red marks sound like a good idea though.
I'll open a new bug report in the next days and try to formulate it as clear 
as possible.
And I still think bug #140675 wants the same, but the reporter probably did 
not formulate it that clearly.

Thanks for looking into this. I know it is complicated :)