Maybe I'm just missing something but ... I can't help but feel somewhat cynical about this. I released the Linux Trace Toolkit on July 22nd 1999, over 10 years ago - and yet still, this was quite some time before DTrace came to be FWIW. The value of static tracepoints seemed obvious at the time (for me at least). I can't believe this debate is still going on. In fact, I can't help but think that Linux in this case could have been way much further ahead of DTrace. Of course the worst part is that early on all this static tracing was turned down because it would result in unmaintainable bloat. The irony is that the vast majority of initial trace-points suggested are still valid today.
As to the argument that nobody wants to "use" this stuff, I've never bought this. You can't expect users to come asking for tools they've never seen before -- that's rare. That doesn't mean they won't find those tools very useful if they were made available to them. It just so happens that those feeling the need for these tools have no way to show mass user adoption of these tools because they can never get those tools to the users in the 1st place (if it's not mainlined, it's likely not if your latest distro ...) So one can only point at other OSes delivering same functionality ... The fact of the matter is that in this specific case users don't get enough credit for acting smart when given the right information. Even Windows allows me to get more information about what's going on than Linux does. There has to be a point where users are given the tools to find out what *they* want to know about what's going on, not what some maintainer somewhere decides they should see through /proc/foo.
I seriously hope this issue can be settled to Linux's benefit at some point in time in the future. Though I've stopped maintaining LTT quite a few years ago, I still hope one day being able to have a tool in Ubuntu to have the tools that give me the power to get full control over the information *I* want to see.