[prev in list] [next in list] [prev in thread] [next in thread] 

List:       koffice
Subject:    [Bug 187681] shape shadows are not handled correctly
From:       Thomas Zander <zander () kde ! org>
Date:       2009-03-24 23:43:24
Message-ID: 20090324234324.AAFC715CD6 () immanuel ! kde ! org
[Download RAW message or body]

https://bugs.kde.org/show_bug.cgi?id=187681





--- Comment #7 from Thomas Zander <zander kde org>  2009-03-25 00:43:23 ---
On Wednesday 25. March 2009 00.22.05 Jan Hambrecht wrote:
> > Imagine us inserting 2 objects. One after the other.  We place them next
> > to each other.  The user never choose which one is over the other and
> > frankly doesn't care.  But in our code we force one to be above the other
> > because we auto-assign them indices.
> > If we then end up casting a shadow of one object on the other that is
> > next to it, then that breaks the illusion of them being next to each
> > other. In effect we are showing an implementation detail.
>
> Right it is an illusion, which gets destroyed as soon as you have these two
> shapes overlap. As i already said, i agree that two shape having the same
> z-index should not cast a shadow on each other.

Well, this kind of illustrates my point as we can not have two shapes with the
same z-index since it would be impossible to paint properly.
So its is impossible for the user to get two shapes *not* cast a shadow on each
other.
And the reason is purely an implementation detail.

I don't really care that OOo doesn't handle the properly either, there are
plenty of other places where shadows are done and the research that Trolltech
[1] did shows clearly that a per item shadow as we have now is only optimal
from an implementation perspective.

To be honest, I don't see how you can argue that its correct to have this;
really. Do the fresh-user test and load a random svg from the net and give
everything a shadow. Ask the user if he expects that there are shadows on top
of some elements and below another.
I bet the user expects the shadows to be completely behind *all* elements.

The z-indeces are a red-herring. They just don't exist in the minds of anyone
outside of the developers.

> > Maybe we should have a shadow per layer, but I think that its wrong to
> > have one per shape that cast a shadow on another shape.
>
> Do you mean a shape layer or a z-index-layer" ?

KoShapeLayer

Or, in other words; the layers that a user explicitly groups shapes in to make
them lay above other shapes.

1)
http://labs.trolltech.com/blogs/2008/05/30/decoration-items-light-and-shadow-effects/

-- 
Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
____________________________________
koffice mailing list
koffice@kde.org
To unsubscribe please visit:
https://mail.kde.org/mailman/listinfo/koffice
[prev in list] [next in list] [prev in thread] [next in thread]