►
From YouTube: NuPIC Sprint Planning - Jan 17, 2014
Description
Planning Sprint 14. https://github.com/numenta/nupic/issues?milestone=8&page=3&state=open
B
A
A
B
A
A
The
this
is
just
an
administrative
task,
the
home
of
AWS
stuff
into
the
right
account
running
a
new
pic
on
a
Mac,
so
one
of
things
I'm
trying
to
do
right
now
is
create
a
sort
of
comprehensive
installation
guides
or
build
guides.
For
you
pick
I've
worked
through
have
been
too
I'm
working
through
fedora
and
I
want
to
work
through
mac,
OS
x,
2.
I've
got
a
new
machine
that
all
should
have
access
to
next
week.
A
A
Ian
has
agreed
to
work
on
cerebro
a
bit
the
sprint,
hopefully,
and
so
that'll
be
good.
There's
a
lot
of
people
that
have
asked
questions
about
it.
It's
it's
really
poorly
documented
and
the
hard
to
get
running
so
I'm
hoping
you
can
approve
that
history
goes
a
really
excellent
tool,
especially
for
understanding
how
things
working
within
the
CLA
in
real
time.
A
A
There's
an
open
these
four
or
five
down
here,
unassigned
I
had
assign
them
to
some
of
them
to
myself,
but
I'm,
not
sure
if
I'm
going
to
work
on
them,
so
just
on
assigning
them.
These
are
probably
the
most
important
issues
that
I
think
anybody
could
grab
and
try
and
work
on
so
I
want
is
to
improve
hot
gym.
There's
some
details.
In
there
the
soups
I
wrote
up
on
how
that
could
be
improved,
and
this
would
be
sort
of
documentation
for
the
hot
zone
example
and
maybe
even
a
improving
an
example
code.
A
Although
there's
not
a
whole
lot
of
code,
there
are
a
tutorial
or
something
like
that,
something
to
walk
people
through
exactly
how
that's
working
on
what's
happening
there,
and
so
there's
a
another
doc
ticket
about
being
more
explicit
about
what
description
pie
is
and
defining
the
prediction
output.
This
sort
of
a
goes
along
with
hot
gym,
because
some
of
the
questions
we
get
is
what
it,
how
does
description
down
by
work
or
what
is
it
defining
there
and
a
lot
of
people
ask
how
do
I
get
to
the
predictions
coming
out
of
hot
gym?
A
So
we
do
have
a
little
bit
of
documentation
about
the
production
output
format,
but
we
might
want
to
give
some
more
guidance
on
how
to
interact
with
it.
What
to
do
with
it?
You
know
once
you
get
a
hold
of
it,
you
have
to
kind
of
write
some
code
to
get
that
out,
perhaps
even
provide
some
some
tools.
You
know
to
attach
to
the
output,
to
write
it
to
files
or
certain
formats,
or
something
like
that.
So
we.
A
A
But
if
we
got
a
good
test
coverage
tool
in
place,
that
would
sort
of
take
care
that
but
I'd
also
like
to
run
some
of
our
sample
applications
or
typical
use
cases
of
new
pic
and
with
some
type
of
instrumentation.
That
will
show
us
what
is
not
being
called
what
code
paths
are
barren,
so
we
can
look
at
deprecating
a
roof,
removing
cruft,
as
we
move
toward
pardoning
an
API
around
that
core,
so
sort
of
a
initial
task.
A
Moving
that
direction
these
these
two
C++
unit
tests
should
generate
X
unit
style
reports
and
there's
one
for
Python
as
well.
I
ran
the
Python
tests
and
I
didn't
I,
didn't
see
any
output,
except
for
some
text
files
so
really
like.
If
we
get
X
unit
style
reports
coming
out
of
these
and
at
some
point
than
we
converge
them,
I
think
when
you
once
you
have
a
key
and
report
this
you
can
merge,
merge
them
into
one
big
report
that
would
show
all
of
our
tests
and
a
publishable
fashion.
A
Once
we
have
that,
and
you
can
store
it,
we
can
track
trends
and
metrics
and
stuff
like
that.
So
that's
that's
just
good
practice.
I
think,
and
then
scott
had
an
idea
based
on
a
message
that
came
through
recently
to
create
some
exam
Apple's
of
how
you
could
build
a
hierarchy
with
the
current
new
pic
with
the
owned
by
a
tempo.
A
Fooling
there's
a
lot
of
people
interested
in
this
and
a
lot
of
problems
that
come
across
the
list
that
we
just
probably
would
not
do
very
well
without
the
hierarchy
in
place,
so
at
least
giving
people
some
guidance
to
attempt
some
things
and
as
we
move
forward
thanks
and
create
the
new
temple
following
algorithm
of.
However,
we
change
that
if
we
could
try
and
harden
what
the
interfaces
for
that
or
these
have
an
example
of
how
it
works,
so
people
can
try
it
out
once
we
have
pieces
in
place
to
construct
a
hierarchy.
A
We
look
at
sprint,
13
our
current
sprint,
I'm
just
going
to
go
through
each
one
of
these
tasks,
but
so
I'm
going
to
look
at
the
ready
tasks.
First.
These
are
all
PRS,
except
for
this
one,
which
is
someone
reported
in
region,
test,
failure
in
linux
mint
and
I'm
waiting
for
them
to
confirm
that
the
new
installation
instructions
work.
So
all
of
these
are
basically
ready
for
review
a
code
review
on
a
PR.
Some
of
them
there's
still
something
of
a
discussion
going
on,
but
I'm
going
to
move
all
of
these
into
sprint.
A
A
So
we've
got
quite
a
few
components
now
that
that
have
docs
for
them,
including
the
network
API.
Does
these
are
by
no
means
complete
when
we're
strings
and
documentation
strings
existed?
I
translated
them
into
a
doctor
to
format,
but
there
are
still
a
lot
of
places
where
there's
a
missing
documentation
and
those
are
noted
in
the
code.
A
A
B
A
B
B
A
C
I
guess
I
mean
that
it's
essentially
resolved
and
that
we
gettin
a
successful
docker
implementation
until
someone
says
that
they
can't
do
it
and
it's
not
working
for
them.
My
understanding
is
it's
working.
I
think
the
follow
up
step
would
be
to
create
some
automation
for
linking
it's
the
most
recent
successful
build
with
new
pic
so
that
it
continually
stays
updated
automatically.
It
doesn't
fall
too
far
behind
ok,.
A
C
A
Should
be
installing
if
using
docker
there
we
go?
Ok,
oh
great
great,
so
I
must
have
missed
this.
So
what
I'm
going
to
do
is
I'm
going
to
take
this
and
assign
it
to
myself,
because
I
want
to
work
through
the
wiki
from
scratch
and
get
it
good
at
all
working
and
then
once
I
confirm
that
it
works
for
me,
then
I
will
close
it
out
and
we
can
talk
about
maybe
creating
another
ticket
for
some
automation.
C
B
C
Ok,
so
it
should
I
think,
but
the
docker
is
kind
of
new
and
they
are
essentially
warranting
it
or
not
warranting
for
anything
really,
but
they
are.
Their
intent
is
for
their
main
OS
that
it's
used
in
as
you
want
to
and
I
think
arch,
gen,
2
and
I
think
sent
OS
has
linux
containers
but
I'd
have
to
check
into
it.
Okay,
in
theory,
you
could
get
that
all
to
work.
A
A
B
It
looked
like
he's
being
very
careful
of
it
to
be
honest,
I,
don't
really
understand
the
details
well
enough
to
see
a
yay
or
nay
on
it,
maybe
mark
or
Scott
or
someone
else
it
was
very
familiar
with
was
my
beam.
Maybe
another
pair
of
eyes
might
be
good,
but
it
you
know
just
understanding.
What's
going
on,
there
is
a
little
bit
hard.
Okay.
A
B
B
Potentially,
algorithms,
you
know
maybe
memory
leaks
or
you
know,
memory
use.
You
know
some
hidden
crash
that
might
be
there
or
whatever.
This
is
this
string
stream,
yeah
yeah,
it's
it's
I
mean
it
seems,
like
he's,
really
thought
through
it
and
all
the
tests
pass.
I
think
it
should
be
fine,
yeah.
Okay,
if
you
want
to
maybe
just
take
a
quick
just
a
second
of
the
ad,
but
I
can
really
add
anything
other
than
it
seems
like
he
thought
through
it
carefully
with
the
pea.
Remember
that
oh.
A
A
I'm
going
to
the
duplicate,
build
artifacts
is
not
extremely
important,
so
I'm
probably
going
to
leave
that
out,
so
the
utility
encoder.
This
is
bumped
along
for
the
sprint
sprint
for
a
while
and
as
I
thought
there
was
a
commit
associated
with
yeah.
This
is
the
pr.
This
is
a
view,
so
this
latest
status
of
this
is
mark
is
working
on
it
based
on
some
of
scott's
feedback,
so
I've.
This
is
currently
assigned
to
mark
and
he's
working
on
it.
So
I'm
going
to
put
this
end
sprint
routine.
A
A
B
The
meat
one
of
the
big
issues
here
was
I
think
he
marked
it
a
really
good
job
of
removing
a
bunch
of
obsolete
stuff
and
but
when
I
ran
it
against
the
rock
pipeline
and
ran
into
some
issues
and
a
couple
of
weeks
ago,
as
part
of
the
random
distributed,
encoder
I
I
fixed
some
of
those
issues
on
the
groc
side
and
so
that
one
of
those
changes
has
been
already
merged.
In
so
probably,
we
need
to
really
verify
the
rest
of
the
changes
against
the
groc
pipeline.
Make
sure
I'm
cancel.
A
So
one
of
Brock
employees
are
going
to
have
to
do
this,
so
I'd
like
to
make
an
assignment,
and
if
this
is
not
something
we
think
we're
going
to
do
in
this
sprint.
I
would
rather
keep
it
out
of
the
sprint
so
Scott
for
soon,
as
I
thinks
they
have
time.
For
this
look,
let
me
know
and
I'll
sign
it
to
you.
A
Okay,
that's
good
enough
for
me.
There's
still
discussion
on
this
math
unit
test,
one
between
scotton
and
Mark
I
think
he's
going
to
review
it
absolved.
That's
ongoing
work
that
I'm
the
this
yes
I'd
still
need
Joe.
To
give
me
admin
rights
before
I
canna
end
to
any
administrative
stuff
on
our
AWS
server
or
AWS
account:
okay,
refactor
spatial
pooler
unit
test
by
unifying
duplicate
code.
This
looks
like
zubat.
I
created
this
and
I
don't
know
what
the
status
is.
I've.
B
A
A
I
would
rather,
if
you
have
time
for
this
print,
I
would
I
would
rather
you
work
on
reviews
and
any
other
more
important
task
to
him.
Okay,
look
david
is
here,
and
I
don't
remember
what
the
status
of
this
was
yes,
so
I
gave
this
to
mark
to
investigate
because
he
had
the
most
comments
about
about
this.
This
is
the
single
seeming
file.
I
then.
B
A
A
Ok,
api
docs
for
encoders
I'm
going
to
keep
that
in
because
I
want
to
keep
asking
people
to
do
that.
Also
I
was
assigned
to
that,
but
I
took
myself
off,
and
so
we
have
this
ticket
that
Chris
Simon's
worked
on
to
rename
this
environment
variable,
but
there's
some
repercussions
in
the
graph
pipeline.
So
it's
reassigned
to
Joe
Black
is
their
system
reliability
engineer
who
will
confirm
the
pipeline
as
we
not
dated
before
this
gets
merged,
so
I'm
going
to
keep
that
in
too
all
right,
let's
go
to
spend
13
and
close
it.
A
So
there
are
a
few
tickets
in
here.
Anybody
can
come
in
and
grab
and
the
ones
that
are
currently
unassigned.
I
would
I
don't
know
if
I
call
them
newbie
tickets,
but
you
don't
have
to
know
a
whole
lot
about
new
pic
to
do
this.
You
just
need
to
know
about
either
C++
or
Python
or
code
general,
so
I'm
going
to
mark
these
as
movie
tickets,
because
you
don't
need
to
know
about
new
big
to
do
them.
A
Okay,
we
still
have
a
ton
of
tickets
issues
with
no
milestone,
56
tickets,
with
no
milestone
and,
as
always,
if
anybody's
looking
for
work
and
there's
nothing
in
our
current
sprint
that
catches
your
fancy-
you
can
run
through
here
and
pick
something
that
you
would
like
to
work
on.
If
you
cannot
assign
yourself
signer
contributors
licensed
and
all
update
your
permissions.
A
B
That
what
one
quick
with
the
encoders
I
noticed,
while
I
was
going
through
the
random
the
stupid
encoder
is
that
on
I
wasn't
very
familiar
with
the
encoder
code
before.
But
there
are
a
number
of
methods
in
the
base
class
that
are
somewhat
tricky
to
implement,
but
are
actually
not
used
anywhere.
So
I
can
maybe
I
can
create
an
issue
that
at
least
I
can
list
those
out,
and
then
someone
could
take
them
out
because
it
involves
taking
them
out
and
all
the
encoders
okay.