Re: [Gimp-developer] GSoC project query

[Jehan Pagès <jehan marmottard gmail com>]

Hi!

On Mon, Mar 5, 2018 at 11:09 PM, Elle Stone
<ellestone ninedegreesbelow com> wrote:
> On 03/05/2018 03:54 PM, Jehan Pagès wrote:
> > Now we have to decide of a good project for GIMP/Darshan since GSoC
> > needs us to decide this first. Note that, as already said previously,
> > I am more in favor of feasible and more concrete projects for GIMP,
> > even though maybe less fancy looking that what we used to have for
> > GSoC. Something which will definitely end up in GIMP in the end.:-)
>
>
> Some suggestions focused on painting and color management:
>
> * Better LCH support including high bit depth LCH palettes (to replace
> current 8-bit sRGB palettes) and gradients, plus for the color pickers
> outlines showing LCH vs the user-selected RGB color space at any given
> Lightness value, and showing complements/color harmonies.
>
> * Improved soft proofing to compensate for deficiencies in LCMS soft
> proofing.
>
> * Improved painting assets in general, including brushes and brush dynamics,
> building on Americo's work on improving painting assets, and addressing a
> bunch of painting-related bug reports.

This is typically the kind of projects which are both "easy"
technically, but actually difficult to do right. Because it is not
only about stacking up various brushes which look cool and have as
many as possible in GIMP (especially in current situation where brush
management is so shitty that it is not a good idea to have too many).

We are discussing the topic with Aryeom and Americo and this is quite
harder than I thought it would be (I am the one who originally
proposed Americo to work on it).

So I definitely don't want to give this to a student. We will stick to
working with the artists who use them.

> * High bit depth brushes - Cinepaint supported 16-bit brushes, but it would
> probably be more efficient for GIMP to go straight to 32f brushes.

Interesting. Since everything is 32-bit internally already, will it
contribute to make painting faster as well?

> * An HDR viewer and improved HDR editing capabilities.
>
> * OCIO color management
>
> * Improved support for AnyRGB and AnyTRC.
>
> * Adding JCH and CIECAM02 support, following the lead set by RawTherapee's
> extensive and incredibly useful (for artistic purposes as well as
> compensating for viewing conditions) CIECAM02 support.

If you want to contribute ideas, you can add them to the wiki:
https://wiki.gimp.org/wiki/Hacking:GSoC/2018/Ideas
Please if you do so, if you could add some explanatory links too, that
would be awesome, because I would not be able to implement several of
your propositions without some serious reading myself.

Remember that this is for a student. The main reason why GIMP stopped
doing GSoC is because we were giving the students very awesome
projects, and even though the final results were very cool
proof-of-concepts (so the students did do very good work), they were
not finished (either extremely buggy, with bad standard code, or so
slow that it was unusable on real data, etc.) and they had been a dead
weight to the GIMP project for years (even nowadays, some older
student projects are still making issues).

Well that is from a time where I barely started contributing myself,
so I am just telling what I gathered, and basically I want to try and
take into account the "elder's" experience (old farts! ;p) to really
get something good out of such a project. That may mean less ambitious
ideas (which does not mean they are not just as important for GIMP!).
:-)

Jehan

> Best,
> Elle
>
> _______________________________________________
> gimp-developer-list mailing list
> List address:    gimp-developer-list gnome org
> List membership: https://mail.gnome.org/mailman/listinfo/gimp-developer-list
> List archives:   https://mail.gnome.org/archives/gimp-developer-list



-- 
ZeMarmot open animation film
http://film.zemarmot.net
Liberapay: https://liberapay.com/ZeMarmot/
Patreon: https://patreon.com/zemarmot
Tipeee: https://www.tipeee.com/zemarmot