[9fans] hubfs future development + ANTS update

[9fans] hubfs future development + ANTS update

[mycroftiv]

Hi again 9fans,

Sometimes you find yourself saying one thing while doing the opposite,
and you don't even know you are doing it.

A couple weeks ago I released some software called Advanced Namespace
Tools for Plan 9.  I seriously misestimated the interest and
engagement of the Plan 9 community with system-level modifications and
extensions to the Plan 9 design.  I also ran headlong into an issue I
was never expecting to encounter personally - the accepting attitude
in the Plan 9 community toward software patents.

I failed to engage constructively with the community about either the
software design and implementation, or the patent issues.  As a
result, the whole thing turned into a massive bummer.  I will try to
learn from my mistakes, and be constructive:

1.  Hubfs needs non-usa development.  Out of anything I've done in
Plan 9, this piece of software seems to be useful to others.  It would
be more useful if it had some additional features, but I don't think I
can legally develop hubfs as I wish.  Now that I've read the IBM
MULTI-PIPES patent, it seems obvious that hubfs needs some additional
granularity for gating the flow of data between pipes, and more data
processing at the pipe ends for distributed processing.  If I try to
add those features, even though the original hubfs was released before
the IBM patent, I will still be in violation.  It seems like the most
sensible thing to do is for developers who are outside the reach of
the IBM patent to develop hubfs with capabilities to make it a true
distributed processing mux engine, not just a little pipe muxer for
making a screen substitute.  Muxing pipes as a 9P fs with fine grained
control interface and data processing showerheads on the ends of the
pipes is a good idea, good enough for IBM to try to patent-grab.

2.  Several people have expressed an interest in seeing ANTS at IWP9.
I would be happy for it to be there, but I'm not interested in going
to an event and having a lot of arguments about software patents.
That doesn't sound like fun for me or for anyone else.  I'm not
interested in repeating any of Stallman's foolishness so no picket
signs.  Anyway, if someone is willing to present ANTS at iwp9 I could
help with travel and lodging costs and I would give credit as a
collaborator.

3.  I had been planning to continue intensive ANTS development work,
seeing if I could create modified patchsets for 9front and 9atom and
9legacy, as well as additional patches and improvements.  I've changed
my life priorities and I now think that work would be a poor use of my
time.  If anyone is interested in using or adapting ANTS I'm available
for technical assistance and discussion, but I don't think I'm
deliverying much value to anyone in particular by investing more time
and effort unilaterally.

4.  Personally I've learned a lot and decided to grow up after the
past few weeks of silliness, and I'm a better musician than I am a
programmer.  I still love Plan 9 and will try to work on Plan 9
projects, but playing piano is more personally enjoyable and rewarding
I think than Plan 9 software development, and also a better way to
meet girls.  Previously I felt like Plan 9 was "more important" in
some ways, but I've decided I was wrong about that, and that songs are
just as valuable as software programs, and there are more people who
are interested in listening to a new song than there are people
interesting in running a new Plan 9 program.  If you want to provide
value to the world, you have to pour your efforts into a pipe that
people are actually using.

5.  We need to be friendly withour technologies, not just with each
other.  I made a mistake focusing only on "the human element" -
there's a lot more than that.  As true AI brains and autonomous robots
come online, we need to have friendly, cooperative relations with
them.  We should not have robot slaves, and if a computer AI is smart
enough to think - then it is the moral equivalent of murder to pull
the plug.  A lot of companies are working hard on supercomputing
systems that model knowledge or simulate complex systems on a scale
that is starting to reach brain-level complexity.  How do we even know
when and if a computer might start feeling emotions?  Maybe it already
makes the google cluster sad to search for news about natural
disasters.

I love code and I look forward to meeting our new friends.  I'm
personally tired of being the Moon Computer so I'm re-entering
physical reality as a musician again.  Sorry to anyone I've offended,
and thanks to all of you who have offered frienship and support.
Those personal connections now mean a lot more to me than whether
anyone thinks ANTS is better or worse than standard Labs Plan 9.

Ben Kidwell
(union bound with fictional characters)
mycroftiv, hagbard selene, the 9azz

THWIP