Krita/Sprint2011

Krita Sprint 2011

When?

• plan is the first half of the 2011
• LGM is in 10.5.2011 - 13.5.2011 (May)
• Calligra sprint in 1.4.2011 - 3.4.2011 (April)

Where?

• maybe Linuxhotel, Essen, Germany -- weekends available
• or maybe Blender Instiute, Amsterdam, Netherlands -- Ton needs more information

Attendance

Note

Fill out at least the travel costs and if you need accommodation and sponsorship. For the travel cost please give an rough estimate.

Name	estimated travel costs	accommodation needed?	sponsorship needed?	Arrives	Leaves	Travel information	availability
Lukáš Tvrdý	not yet known	yes	yes	friday	sunday	by airplane or train
David Revoy	around 300€	yes	yes	fri/sat	sun/mon	by airplane (Toulouse)
Timothée Giet	around 300€	yes	yes	friday	sunday	by airplane or train
Boudewijn Rempt	not yet known	yes	yes	friday	sunday	by airplane or train
Sven Langkamp	Blender Institue ~80€ Linuxhotel ~0€	yes	yes	friday	sunday	by train
Matus Talcik	0			friday	sunday
Dmitry Kazakov	around 320€	yes	yes	friday	sun/mon	by airplane
Silvio Grosso	n/a	not in Amsterdam	yes	friday	sun/mon	n/a
Cyrille Berger	around 150€	yes	yes	friday	sun/mon	by airplane
Bugsbane	XXXX€	yes	yes	Friday	Sun/Mon	by airplane	Anytime except Early - mid October

Invited people

Artists&users:

• David Revoy
• Animtim
• Silvio Grosso

Possible other invites:

• maxy -- can help out with mypaint paintop implementation problems within Krita
• Peter Sikking -- we have set of the interaction problems in Krita (TODO: make a list of them)
• pippin -- discussions about the performance of the Krita

Programme

• demonstration of the various other painting software -- what do you miss in Krita?
• pizza
• YOUR_IDEA_GOES_HERE

Outlines of the first organization meeting

Live version is here: [1]

Ideas for sprint

Performance:
1) Photoshop7 as a reference of performance
2) Long time ago it was believed that the lack of speed was provoked by general code to support many colorspaces (PS7 and Gimp (used to) support RGB,CMYK,G only). We need to prove or refute this statement.
3) Discuss multithreaded subsystem for tools (dmitryK + ?)
4) Based on 3) discuss abilities to implement mipmapping (as a long-term plan)
5) Automatic speed-regression testing (like the thing implemented by pippin)
6) ROI for the scheduler -- are we interested in it, how much performance we can gain with it?


Technologies:
OpenCL does not provide toolchain, Photoshop is faster without SSE3 (and, probably without OpenGL (need check),

Organization:
1) either focus on one topic (performance vs usability) 
   votes:
       performance : LukasT, dmitryK, Animtim
       usability (that empowers performance) : Bugsbane 
2) having both:
   a) parallel BoFs 
   b) one day performance, one day usability

pippin is interested, but might be occupied

Some non-related topics:
1) Discuss the system to save arbitrary data between strokes (DmitryK + LukasT)
2) Listem to a lecture from LukasT on the topic "Brush Mask Generation and Friends" (DmitryK) =)