Thanks to Louis Desjardins and other people I was able to attend Libre Graphics Meeting in Montréal this year and meet face to face many magic people I’ve been working with for last several years.
I’m in the middle of getting full article on LGM in Russian out of the door, into the wild 
now (Edit: read it here), so here is a snippet of it — a short interview with Ã?yvind KolÃ¥s — GEGL‘s mastermind.
Q: Ã?yvind, your presentation wasn’t much different from the one at FOSDEM last January. Is that because not much has changed since then?
A: Yes, the FOSDEM and LGM presentations were approximatly the same, with only the name of the conference, and some items stricken off the todo list at the end and some of the code examples reduced a bit. I think I might even do a third revision of that presentation at some point, expanding some sections, as well as adding some more pieces as soon as I consider them ready for public consumption. The presentation didn’t need to change much is that it is a presentation primarily about GEGL, and it’s public API, the public API and the nature of GEGL not having changed between FOSDEM and LGM is probably a good thing 
Q: How are your gggl based projects doing? Are we going to see some of your own tools based on GEGL too?
A: Bauxite has never moved from gggl to GEGL, but portions of oxide, as well as gggl live in GEGL after the merge. The presentation tool used for FOSDEM/LGM is probably the most complex GEGL using app. The other semi-large thing using GEGL is the test app in the SVN repository, but that is more of a test scaffolding that I didn’t really want to put in svn, since people might mistake it for a real application. That thing is written in C, which IMHO is the wrong language for such an app. If it was rewritten in Ruby or Python it would be 1/20th the amount of code, and most probably be bug-free.
Q: What contribution to GEGL would you consider most anticipated?
A: Some application built on top of GEGL for doing a fixed list of global changes to an image or set of images (white balance, levels, unsharp-masking etc) adjustments. Ideally this application would be written in python to properly exercise that binding. It would also be extremely neat if someone made GPU processing in GEGL happen. Either by providing an alternate core with completely separate plug-ins, or by providing the needed infrastructure for a hybrid approach without cluttering up the existing reference implementations of the operations.
Q: How much useful was LGM for GIMP and GEGL this year to your opinion?
A: I cannot speak for GIMP, though it was interesting for the GEGL project to meet the GIMP developers One person I had scheduled to meet that I never had met face to face before was Enselic (Martin Nordholts). I basically told Joao (who did LGM related organization work for GIMP this year) that Martin should get sponsored to get there — he had already been doing some nice work in GEGL and I suspected that the best way to trick him into becoming a regular contributor was to set up a trap: give him the opportunity to meet all the cool people and projects in person. And it seemed to work even before LGM, if you look closely at svn log
Q: Have you had some interesting meetings with other teams during LGM?
A: One of important things for me was my meeting with Krita developers. I was disappointed that they were only two (doubling of last year though), I missed quite a few of the Krita regulars. See, I’m nowhere close from being a code contributor, but I do get asked by opinion on technical matters of the Krita architecture. One of the reasons is that I’ve infiltrated their irc channel Sure, they were a bit skeptical of me at first. I think I’ve been in the Krita IRC channel for 14 months or so now, probably spent more time there than any of the actual Krita developers since I never sign off. The other reason is my work on OpenRaster with Cyrille Berger.
Q: And what’s you overall impression of the conference?
A: One of the interesting and vitalizing aspects of LGM, both for individual developers as well as projects as larger entities is that it brings together a lot of like minded people, working within overlapping domains. Comparing notes and socializing with other people that have aligned views and interests makes the meeting a rewarding experience. I think I mostly spoke with people I’ve met before at other conferences/meetings, I probably should have tried making more new contacts as well, but I had a great time, met old friends, and saw old friends meet other old friends; together becoming new friends, overall a very rewarding experience.