proposal: feature freeze

[Erik Ostrom <eostrom@research.att.com>]

Here's an attempt to formalize what I started to say in my previous post.

Proposal:  Let's not add any new features to MCP 2.1.  Let its
functionality be set now.  Bug fixes, wording changes, okay, but no
new ideas in MCP 2.1.

This is not a blow against innovation.  We can keep innovating for a
hundred years.  But at some point, we have to take a snapshot and call
it "done, for now", or it won't do any good.  No one can implement a
spec if there's never a spec to implement.

And without implementations, experience, practice, we're flying
blind.  We need to _use_ these systems to figure out what needs to be
fixed, what could be done better, what new concepts to add.

With that in mind, let's keep talking about new ideas, but let's talk
about them as candidates for MCP 3.0, or better yet, as candidates for
experimental implementation.  Can we agree that there is nothing so
urgently needed that it must be in this first version?

--Erik