Thanks Steve. We've done several other releases too, including Fedora 15. I would instead say that this is the "most complete" release so far, where we are within a hair of having every package available on ARM (obviously excluding architecture specific stuff). By Fedora 18, we hope to be even more feature complete, and using standard image creation tools (for which support already is in progress). Many people did great work to make this release happen, and the net result is something we are proud of.
On the subject of compatibility, Steve is referring to a cross-distribution effort to standardize a new dynamic linker path between distributions. This began when Debian/Ubuntu decided to go to multi-arch, necessitating changes to various things (such as the linker path). In the interest of compatibility, we had many rounds of discussions about how to reconcile around a common path, and eventually we came to a conclusion everyone can be comfortable with. Alas, to avoid a mass rebuild, it is necessary to take a kludge that our tools guys were not comfortable with fully testing prior to final F17 release. Rather than risk any stability problems, we decided to make the change in F18 (which is done) and then figure out the plan for F17. We'll either release an update, or wait until F18. In either case, users won't notice the impact because the transition uses a symlink on the existing path to keep older binaries running. The net result is that, today, if you want to build a binary that will run on Debian, Ubuntu, and Fedora, you'll need to do some heavy lifting to the target system to make it work, but in not very long this problem will go away. We're on a 6 month cycle, and so it was important that the guy shipping a 5 year LTS had the fix. It would have been nice if we did too, but we'll be switched over at most in 6 months, which isn't all that long, and then you'll have cross-distro compatible binaries between ARM distros (at least as far as the basics go, we have other work to do beyond that).