University Collaboration in Toulouse

[Boudewijn Rempt]

On Friday 10 August 2012 Aug, Sven Langkamp wrote:
> On Fri, Aug 10, 2012 at 10:52 AM, Boudewijn Rempt <boud at valdyas.org> wrote:
> 
> > Hi!
> >
> > Kevin Ottens has managed to reboot the very succesful university
> > collaboration scheme he managed in Toulouse for several years. He's looking
> > for KDE projects that can provide mentoring for several students who will
> > collaborate on a project that is more intensive than the usual summer of
> > code. In Calligra, we've had students working on Stage, with great results.
> >
> > I would like to propose a Krita project here. Specifically: selections.
> > Selection creation tools in krita are primitive, our selection handling is
> > primitive and in one respect just plain _wrong_ -- so, in my opinion, this
> > would a be a good, self-contained subject for a group of bright people to
> > dive into and redo, from design to implementation.
> >
> 
> Could you give a bit more details on which things you want them to work?
> 

This is one proposal:

   
Krita is the full-featured painting application for artists who want to create professional digital art from start to finish. Krita was started in 1996 as an alternative to GIMP, but has developed into a specialized painting application. While the painting features are extensive and impressive, currently, a weak point of Krita is selection handling.

Code Mentor: Boudewijn Rempt, CTO of KO GmbH, maintainer of Krita (www.krita.org)
Customer: David Revoy, artist, krita user (www.davidrevoy.com)

A selection in krita is a single-channel mask (one byte per pixel) or a group of vectors, rendered into a single-byte mask.

With selections, there are the following issues:

* selection creation is primitive. There have been attempts at magnetic outline selection (lasso) and siox-based selection, but none of those are available yet. The current selection tools are also quite primitive.
* selections are not easily modifiable, for instance through transformation.
* selections and masks are alpha-channel based, not gray-channel based. This makes painting on masks problematical.

The project would consist of two parts:

* The design of a set of sophisticated, modern tools for creating and modifying selections. Reference both to the existing capabilities of Krita and to other applications such as GIMP and Photoshop should be made here. The goal of Krita is to enable artist productivity, not originality for its own sake! David Revoy is available to discuss work flows and artist needs.
* Implementation of these tools based on a common framework, to be designed by the project group. This entails both the algorithmic work for e.g. magnetic selection tools, as well as design the implementation of the tool GUI. The tool gui consists of two parts: the on-canvas representation of the selection and tool manipulation handles, and the tool settings panel.

Krita has a set of unittests, and students are expected to extend these tests when writing new code.


I also had a bunch of smaller ideas which I put to David and Ervin, but they are on my other computer.

-- 
Boudewijn Rempt
http://www.valdyas.org, http://www.krita.org, http://www.boudewijnrempt.nl