Ok, I think I've digested the material enough to ask for suggestions: [1] Sam listed a number of tradeoffs. Before we can make intelligent decisions about tradeoffs, we're going to need some measurements of the costs and the benefits.
Measurements themselves will require some significant resources [or maybe we can find that someone else has already done the relevant work]. [2] Sam also talked about "a global direction". Right now, debian is an ad-hoc mix of binary and source packages (with some source packages being in deb form because of the limitations of our "dependency mechanisms"). We're also seeing limitations [mostly potential but some quite real] of our archive architecture and dependency architecture. Maybe it's time to sit down and lay out a strategy for where we're going. I'm interested in how the rest of you see the problems, and the implied requirements. [Ok, I think I've a pretty good idea of how Herbert Xu and Sam Hartman think -- I'm looking for other viewpoints.] Thanks, -- Raul
