Maciej Pilichowski ([EMAIL PROTECTED]) wrote:
> From: Simon Budig <[EMAIL PROTECTED]>
> > But even taking him into account the current model
> > has served us well for the last years and there has been very few
> > discussion about this specific problem, if it cropped up it was in
> > the context of indexed PNGs with "alphacolors" or generating
> > textures for renderers. For the general public the current model
> > seems to generally work and I don't see a compelling reason to
> > change this.
> Well, one thing is history -- everything (in general) worked in the 
> past -- why introduce cars when horses... etc etc etc :-)
Well, historically new stuff always had to proove that it is better than
the old stuff and I personally don't think that is a bad thing.

I guess I have not yet understood what you actually want to achieve.

> The second thing is potential benefit -- and I believe less time spent 
> on creating an image _is_ benefit. Note, that it will not introduce 
> new features per se, it only could only change the way you work (if 
> one wishes so). You can work with colors with color tools and you can 
> only work with alpha channel with alpha channel tools. It is 
> narrowing possibilities -- why not pick up brush, shape, pressure and 
> polish a picture edges with transparency (as color)?

So how do you want to work with transparency? What should happen if you
have a semi-transparent red and you paint on top of blue? Currently it
gets blended on top of the blue resulting in some kind of violet and
that is a widely used feature to do natural looking paintings. I
understand your proposal, that you actually want to have a
semi-transparent red in the image after painting?

How is your new feature supposed to interact with tablets with varying
pressure devices like tablets? Right now you can map the pressure
information pretty naturally to the opacity. Your "replacement" approach
for alphacolors would directly influence the images alpha channel,
making it pretty tricky to lift off the pen without leaving a
transparent spot (tablets tend to add some events at the end et the
stroke with very low pressure).

How is your "replacement" approach supposed to work with multiple

I see lots of open questions here and I have doubts that your desired
workflow is an improvement over the current one. Also I think that most
of the tasks you mention in other mails - like cleaning up edges - can
nicely be done with the current gimp, provided that you don't need 100%
exact control over the alpha values (like the texture guys seem to need
it, but they are not the main target audience). So I still don't see the
problem here that needs to be solved.


              [EMAIL PROTECTED]    
Gimp-developer mailing list

Reply via email to