►
From YouTube: NuPIC Development Progress Review - Dec 1, 2014
Description
Rescheduled from last Friday because of the Thanksgiving holiday.
A
We
are
at
75
percent
done
for
0.1
milestone,
and
that
means
that
by
the
time
we're
done,
we
will
have
a
complete
release
or
pre-release
of
new
pic.
Most
of
the
remaining
issues
for
the
release
process
are
in
my
court.
We've
got
a
lot
of
good
stuff
done
over
the
past
couple
of
weeks,
even
with
the
holiday
in
place.
A
So
these
are
all
of
the
issues
that
have
been
closed
over
the
past
couple
weeks,
basically
on
this
screen,
so
that
includes
including
lubricore
binaries
as
a
part
of
the
new
pick
build
improving
a
lot
of
the
build
framework
stuff,
organizing
the
cmake
file
and
new
core
a
lot
of
nice
cleanup
stuff.
A
We
go
back
to
the
roadmap.
We
can
see
exactly
how
many
things
have
been
depleted
at
this
point,
so
all
of
our
build
improvements
are
done.
A
So
the
initial
temporal
memory
work
is
done
and
yeah
this
is
included
as
a
build.
Our
build
process
itself
has
been
updated,
there's
still
quite
a
few
stand-alone
tasks
down
here.
I
would
probably
venture
to
say
most
of
these.
A
If
we
get
to
the
point
where
all
this
the
big
initiatives
are
complete
and
we
we
were
coming
up
to
the
line
on
getting
our
lease
done
by
the
end
of
the
year.
I'd
probably
bail
on
some
of
these
and
just
move
them
to
the
next
milestone.
A
Maybe
a
couple
of
exceptions
in
here
that
are
that
are
important
enough
to
try
and
keep
around,
but
they're,
definitely
not
as
important
as
these
super
issues.
So
scott
is
still
working
on
the
model
serialization,
so
you've
got
code.
That's
already
been
pushed
to
that.
Does
the
captain
proto
serialization,
but
we're
not
using
it
yet
right,
okay
right
so,
let's
say
the
major
work
has
been
done
so
at
this
point.
B
I
think
a
lot
of
work
to
move
powders
over
to
it
is
that
what
that
milestone
is,
is
it
all
of
it
yeah?
I
was
interested
there's
two
super
issues.
This
one
and
one
of
the
next
box
sound
and
then
one
of
the
next
milestone
is
the
one
for
getting
everything
over
to
it.
So
this
one
is
the
initial
work,
so
we
can
use
it.
The
next
one
is
the
one
to
make
sure
we
switch.
B
A
All
right
sounds
good
to
me.
Sorry
about
scrolling.
The
rest
of
most
of
the
rest
of
the
major
work
that
needs
to
be
done
is
just
preparation
for
a
proper
release
process,
which
I
am
working
through
right
now.
A
So
that's
where
we
are
on
the
road
map.
We
haven't
talked
much
about
what's
in
0.2,
which
is
fine,
let's
not
talk
about
it.
Yet,
let's
focus
on
0.1.
A
Okay,
so
I
do
want
to
go
over
our
open
pull
requests,
starting
from
the
oldest
at
the
top.
The
windows
build
implementation
is
still
open,
I'm
still
not
sure.
If
this
is
actually
going
to
merge
or
not,
it
looks
like
it's
been
cleaned
up
a
bit,
but
as
long
as
it
has
third-party
windows
libraries
as
a
part
of
this
pull
request,
I
don't
want
to
merge
it,
so
it
may
end
up
as
another
pr
like
I
said,
but
richard
crowder
is
still
working
through
some
issues.
A
On
that
it's
like
a
lot
of
it's
done.
Yeah
he's
he's
done
a
lot
of
work
and
it's
pr.
For
the
most
part,
the
builds
have
been
passing
and
travis
even
on
his
pr
all
along.
So
he
can
run
windows,
he
can
build
new
core
in
windows
and
he's
got
an
app
conveyor.
Ci
system
set
up
that
it's
building
and
running.
A
A
Me
see
here,
I
think
this
is
a
core
issue
yeah,
so
I
think
he's
started
working
on
core
yeah.
I
don't
think
anything's
been
done
in
new
pick
directly
yet
so
this
is
all
just
core
stuff,
but
he's
he's
been
keeping
up
to
date
with
master
and
it's
a
lot
of
work
in
this
pr.
A
So
it's
impressive
so
keep
it
up
richard
thanks
again.
A
Okay,
so
there's
also
this
pr
open
that
does
some
more
sort
of
inheritance
for
the
from
the
encoder
classes
and
some
cleaner
inheritance
from
the
base
encoder
class.
This
is
in
mark's
court.
All
three
of
these
four
of
these
are
in
mark's
court,
so
I'm
not
going
to
go
over
them,
but
I've.
I
ping
them
every
two
weeks
on
these.
A
At
some
point
he
kind
of
does
work
in
in
bursts,
so
he
may
come
in
and
just
clean
all
this
up
next
week
or
next
month.
I
don't
know
we'll
see.
None
of
these
are,
although
I
think
this
is
a
0.1
task,
it's
something
that
I
wouldn't
mind
removing
to
0.2.
So
none
of
this
is
crucial
to
getting
for
the
release.
A
I
have
a
couple
of
anomaly
benchmark
issues
that
I'm
not
going
to
go
over.
This
is
good
to
merge.
Austin
joe
had
one
comment
that
you've
missed
a
url
simplified
test.
Folders,
that's
progress
is
being
made
on
that.
I
think
right.
All
I
had
to
do
is
merge
and
master.
So
all
he
has
to
do
is
press
the
merge
button
and
it
should
be
good
to
go
same
with
these
cleanups
david
ragazzi
has
been
doing
some
namespace
cleanup
to
get
rid
of
nta
replacement.
A
A
I
don't
know
the
status
of
this
pr,
chad
and
I
think
it's
been
sitting
around
for
a
little
while
yeah
in
review
yeah.
I
think
I
just
pinged
you
guys
so
that
needs
a
review.
It's
ready
to
go,
and
this
is
htm.java
right.
So
those
are
open,
pr's,
nothing
egregious
at
the
moment.
Most
of
it
is
our
works
in
progress
not
on
our
plate
to
review.
A
So
that's
good,
so
I
should
one
other
thing
I
wanted
to
mention
is,
and
this
isn't
directly
related
to
new
pic,
but
the
the
htm
java
code
base
is
getting
a
lot
more
attention
now.
I
think
that's
up
to
six
contributors
includes
myself
and
david
grace.
There's
four
people
in
the
community
that
are
pushing
prs
and
it's
been
moving
quite
quickly.
A
Lately,
david
ray
just
added
a
cla
classifier,
so
he's
he's
attempted
to
call
this
a
complete
stack
of
a
cla
stack
and
he's
going
to
start
working
on
demo
applications
and
that
sort
of
thing.
So
it
would
be
interesting
to
start
testing
this
out.
At
some
point
trying
to
figure
out
if
there
would
be
a
way
to
wrap
both
newbik
and
something
like
this
and
some
type
of
functional
testing
suite.
A
So
that's
something
to
think
about
that.
Something.
I'd
like
to
start
thinking
about
that's
cool
yeah,.
A
C
A
There's
there's
no
opf
at
this
point,
but
something
like
that
would
be
nice
just
reusability
right,
but
I
think
that
he
talked
about
that
and
there's
he
has
some
plans
to
do
something
along
those
lines:
yeah
nothing
for
an
opf
yet,
but
he's
he's
given
himself
kind
of
a
short
lease
as
far
as
issues
upcoming
issues
are
coming
so
once
again,
good
work,
david
ray
he's
really
keeping
on
top
of
this
and
keeping
things
going.
Let's
engage
the
community
to
write
several
encoders.
So
that's
nice.
A
Let
us
see
yeah!
So
that's
where
we're
at.
I
don't
think
anybody's
really
seriously
blocked.
At
this
point,
my
big
initiative
is
to
continue
getting
the
release
process
down
so
that
we
can
have
real
binary
packages
for
new
big
available,
so
we're
kind
of
halfway.
There
we've
got
the
new
core
stuff
going
thanks
to
austin
and
myself
and
the
new
big
one
is
going
to
be
a
little
bit
harder.
A
I
think
just
to
get
a
proper
linux
binary
at
darwin
binary
up
in
the
right
places
that
works
on
and
the
testing
is
going
to
be
hard.
I
don't
think
I'm
going
to
start
off
testing
a
whole
bunch
of
linux
flavors,
but
we
should
at
least
start
with
the
ones
that
we
have
access
to.
So
we
know
the
travis
won't
work,
that's
ubuntu!
A
We
know
that
centos
will
probably
work
because
we
have
pipelines
and
centos,
but
probably
we're
probably
going
to
get
reports
immediately
as
soon
as
we
publicize
an
install
path
that
uses
those
binaries,
that
they
don't
work
on
all
flavors
of
linux
and
mac,
so
I'll
be
looking
out
for
that.
But
that's
that's
the
that's
the
initiative
right
now.
Let's
get
those
binaries
published,
get
a
release
process
that
as
soon
as
we
tag
the
new
repo
with
a
version
number
it
automatically
kicks
off,
pushes
the
binaries
and
makes
them
downloadable.
B
A
It's
pretty
nice
yeah.
I
like
that
better.
So
it
was
your
idea.
Anybody
have
any
issues
they'd
like
to
talk
about,
while
we're
here
concerns.
A
Very
good
always
nice
to
hear
you
we're
making
pretty
good
headway.
It
looks
like
it.