OpenSolaris ponderings

As we near the release of OpenSolaris this quarter, I’ve been increasingly curious, anxious even, about a few things. The first is to just see where the thing is gonna go right off the bat. Are we going to see a GNUification of the userland? "Forget /usr/sfw, let’s just replace the old crufty stuff with shiney new GNU bits!" A non-broken tar would be nice, for example. A GNUified OpenSolaris would certainly be easier to get into Linux shops than Solaris currently is. I’m expecting we’ll see a bunch of driver work. Solaris still lacks support for lots of x86 hardware. Obviously code couldn’t be pulled from GPL’ed Linux, but the BSDs certainly have a lot of good code that could be at least referred to. How about moving to a better packaging format and wrapping the entire system under portage, apt (or even blastwave)? Will GCC be able to build the thing? Ever? Or will Sun release their C/C++ compiler under the CDDL in order to bootstrap the OpenSolaris build process? What impact would a second open source C compiler have on the world? Not taking the steps to align with GCC from the start might be a mistake. Or an advantage. Will the advisory board and community process be able to keep control and provide direction, or will we see “forks” early on taking OpenSolaris in different directions?

Whether or not we see forking, how will OpenSolaris relate to Solaris going forward? How does code from Solaris get into OpenSolaris and vice versa? Or is there going to be One True Code Base to rule them all? If OpenSolaris becomes the defacto Solaris, how long will Sun have to continue with the traditional Solaris releases and support? A few years? A decade? Decades perhaps? How different might Solaris and OpenSolaris get?

Man. This is gonna be interesting.