Orv Beach wrote:
In an effort to be more open, SCALE has decided to move some of the discussion (planning - doh) for SCALE 7x to the SCALE-planning list.
Bravo! This is fantastic.
3 main tracks (as usual) 1 beginner track 1 developer track 1 lab tutorial track (??)
So in the past the "main" tracks tended to fall into categories, even if unofficial. In recent years there was usually one that tended towards desktop, one that tended towards kernel/sysamdin/low-level and one that was all the other stuff. Patterns vary for different years.
So how do you see the "3 main tracks?" I can tell you that with the exception of Stormy Peter's talk, my favorite talks are always pretty close to the kernel. At the original SCALE some of you probably remember that despite being very busy as the technical chair and "power balancer," I made very sure I got to see Robert Love's talk on the 2.5 kernel. Andrew Morton's two talks were amazing. The Zumastor talk 2 years ago was a particular highlight. Checkpoints, containers and live migration was a great one this year. Of course some of the devel ones are also on my favorites list: the parrot talk. And by definition anything SA I'm interested in such as Puppet.
But you don't care about my personal interests, so I'll get to the real point. I think it's very important to not just distinguish between "areas" like desktop, devel, user, SA. I think it's also important to distinguish between low level and high level. Lots of high level talks are available at lots of LUGs and conferences. One of my favorite things about SCALE has always been the fact that there's always some extremely low-level talks as well. Something that digs deep into something and doesn't assume you know nothing about it. These talks aren't for everyone, but neither are the overview-type talks.
So, for example, a "developer track" seems like a "low level" track, but if it's "comparing the three big P languages" that could be at a very high level suitable for newbie programmers ("perl is good at string manipulation and python has a more traditional OO approach with a more mature OO feature set", etc.), or it could be very low-level ("perl's internal regex compiler works like this...").
So I think it's very important not to just split up by category but also to ensure that a certain percentage of talks across various tracks are at a very low level for that track. For example, a desktop track may dig deep into the various RPC and message bus systems that various desktop environments use, how they work, and the pros and cons. That's much more low level than your "what's new in Ubuntu" talk. Both talks are important though.
So that's in general. Specifically, I recommend getting in touch with Robert Love again and Greg K-H. Both like giving talks and are fantastic at it. Andrew Morton is obviously always a favorite, although since he's been twice, perhaps hitting up Robert and Greg may be a better place to start.
Hopefully that is some useful feedback.