Layer usuability idea

[Casper Boemann]

On Monday 07 November 2005 06:09, you wrote:
> Hello,
>
> I'm mostly just lurking on the krita mailing list, following the
> development of krita passively. However, when I saw the thread on the layer
> usability idea, and the skepticism expressed by Boudewijn and Casper, I
> thought I should make a comment. I think the idea is great, and would work
> very well.
Thank you for not just lurking anymore ;-) We need feedback on everything 
(well not clutter but constructive feedback)

> A similar implementation exists in xfig (see attached 
> screenshots). xfig desaturates the inactivated layers and shows them in
> light gray behind all currently active layers (because of the way layers in
> xfig work, more than one layer can be active). This feature is great and
> makes xfig a joy to work with. I have always wondered why other
> drawing/painting programs don't offer a similar feature.
I don't think neiher Boudewijn nor I was totally dismissive of the idea, but 
as I said there are several problems, and using xfig as an example doesn't 
convince me, as pixel apps add to the complexity.

> Also, the desaturated layers could be painted not just in grayscale, but
> also in intensity levels of any user-chosen color ("redscale",
> "greenscale", etc). I would default to gray, but there is no reason why a
> user shouldn't be able to choose another color for an inactivated layer,
> and no need to restrict this choice to any particular color.
Well I believe you are wrong. In doing our current selection visualisation I 
had one combat the requirement that:
   Any pixel should look different when inactivated

Simply going to grayscale won't work. For starters a grayscale image would 
simply look the same. And letting the user choose a color might get to close 
to gray. But the 6 combinations of max r,g or b that is not grayscale does 
work faily well. Oh and I also had to implement som contrast limiting. All in 
all it's not possible to use any color and still fulfill the requirement.

This is where the xfig example is too simple. There you have clearly defined 
objects, wheras on a pixel image it all blurs together. So while it does to a 
degree work in xfig I'm far from convinced that it'l work with pixels.

But as both Boudewijn and I said, avoiding to mistake the active layer is an 
important matter, and we should come with a solution. So thanks for your 
feedback, and don't hesitate to think/write more about this issue.

btw did you really suggest bringing the active layer to front? With composite 
ops from layer to layer this could really distort the image. Again a thing 
that works in xfig but hardly with pixels.

Again; I'm not totally dismissive, but we need to work through these issues to 
make it work.

-- 
best regards / venlig hilsen
Casper Boemann