[rfk-dev] i'm in my underwear at school

[Peter A. H. Peterson]

So, you guys know that classic anxiety dream, where you were somehow
enrolled in a class that you never attended, and now it's the day of
the final so it's too late to drop OR to prepare and you have to take
the test anyway?

Well, the discussion of moving the NKI out of the binary triggered
something in the back of my mind... haven't we had this discussion
before? Didn't we already do this...?

Turns out, yeah, we kind of did.

You may remember that Alexey Toptygin rewrote the POSIX code with a
number of enhancements that were discussed between about August 2005
and October 2005. At the same time, we were looking for a new Debian
Maintainer, but due to the long lag, it wasn't looking like Alexey was
going to get there soon enough.

http://robotfindskitten.org/pipermail/rfk-dev/2005-August/000516.html

We had a comment period on the new codebase, and we voted (4 to 1!) to
go with the new codebase. I collected the extant NKI in a file. I
added Alexey to the devteam page with an annotation. Then, there were
some open questions about including some base NKIs or not, while
supporting a system-wide NKI directory and user ~/.nki directories...
etc. -- it's worth looking at the list archives to see what the issues
were.

But then (while it's not described here) in October of 2005 I got an
insane job in LA, then went to grad school and forgot about this issue
completely. Ryan Finnie took over the maint. And in my MEMORY, we had
transitioned to the new code base years ago. But it turns out we never
actually did that, and I haven't thought or heard anything about it
since.

So, first of all, a colossal apology to Alexey (and to the rest of you)
for dropping the ball on this. However, here is his code:

http://alexeyt.freeshell.org/code/rfk/rfk-20051013.tgz

If we are going to make the first significant changes to the
codebase in years, I suppose we ought to straighten this stuff out.

Should we finally transition to Alexey's code? Or do we go with the
newly ESR-improved legacy code and consider Alexey's version to be the
cult classic album that was intentionally buried by the studio? Or is
there a third way? 

In any event, I think that Alexey's version deserves its own place on
the site. Alexey, I am truly and deeply sorry.

Another thing that comes up in those threads from way back is that the
only thing we use Sourceforge for is RCS, and that's a bit silly... it
just makes things confusing. So I would be happy to host the canonical
git repo at robotfindskitten.org if people think that would make more
sense.

Anyway, thoughts, ideas, pitchforks?

pedro

-- 
Peter A. H. Peterson
Graduate Student Researcher
Laboratory for Advanced Systems Research
University of California, Los Angeles