►
From YouTube: Kubernetes Community 1 3 Retrospective 20160701
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
This is the Special Edition 1.3 Retrospective --
https://docs.google.com/document/d/1COLrHw619SXrUU1hlqaVzZel2o2nUvAlEM0xLeOHkc4/edit?ts=577640df&actionButton=1
A
Hello
and
welcome
to
the
Cuban,
a
nice
community,
1.3
retrospective
and
I'd
like
to
introduce
Jay,
King
Jamar's,
who
is
going
to
lead
and
facilitate
our
discussion
here.
There
is
a
link,
or
he
is.
If
he's
not
present,
there
is
a
link
in
our
chat
window
that
has
a
it
is
a
document
that
Jason
is
going
to
be
leading
from
and
we
will
get
started
as
soon
as
Jason
comes
back.
Oh
no.
B
I
guess
my
feeling:
demeanor
flows,
people
all
right,
so
I'm,
first
of
all,
I'm
thrilled
to
be
in
this
position
be
able
to
facilitate
this
meeting
I.
What
an
amazing
community
I'm
relatively
new
to
in
the
last
few
months.
I
I
am
astonished
by
how
well
everybody
works
together
and
is
amazing.
B
So
just
a
quick
overview
of
what
a
retro
is
I'm
sure
most
people
have
been
in
them
or
participated.
A
retrospective
is
a
little
bit
different
than
a
post-mortem.
B
So
in
that
regard,
one
of
the
things
that
I
feel
strongly
is
should
be
part
of
our
respective.
Is
not
just
words
and
saying
this
or
that,
but
to
really
take
ownership
wherever
possible
things
that
we
can
change
and
to
be
accountable
for
those
in
the
next
release
cycle,
so
just
a
quick
overview.
The
document
at
the
bottom.
There
is
a
section
that
says
action
items
I'll,
try
and
get
some
of
these
as
we
go
through
the
retrospective
and
document
them
it.
But
if
there's
something
missing
there,
just
click
on
the.
B
What
and
add
a
comment
to
the
document
if
there's
something
that
you
want
to
take
ownership
of-
or
you
feel
should
be
in
that-
and
that
goes
for
anything
in
this
document.
If
you
see
something,
that's
maybe
word
in
a
way
that
you
disagree
with,
or
you
want
to
add
comments
to.
Please
add
comments.
Liberally
to
document
and
I
will
be
taking
notes
as
we
progress,
so
that's
pretty
much
it
for
the
document,
just
as
a
general
rule,
I
think
a
retrospective
works.
B
Well,
when
we
have
a
set
of
common
agreements
about
how
we
want
to
run
the
meeting
I'm,
simply
a
facilitator,
I'm,
not
dictating
anything
about
the
meeting
other
than
just
to
make
sure
that
everybody's
heard
and
respected
so
in
terms
of
working
agreements,
I'd
like
to
open
the
floor
for
the
next
few
minutes
just
to
get.
If
anybody
has
anything
that
they'd
like
to
see
come
from
this
meeting
or
rules
of
engagement
as
we
progress
through
this
topic,
I
would
be
really
helpful.
B
A
Okay
and
to
your
point
about
forward-looking.
A
B
For
sure
okay,
so
it
looks
like
we're
doing
pretty
good.
It
looks
like
some
people
are
adding
two
with
the
sections
in
the
document,
which
is
great.
Thank
you
very
much.
It
means
less
that
I
have
to
that
type
home.
So.
B
Being
an
investor
okay,
so
let's
go
ahead
and
progress
into
what
went
well.
I
feel
like
this
should
be
a
pretty
long
discussion,
because
I
feel
like
a
lot,
went
really
well,
especially
when
you
look
at
the
deltas
between
12
and
where
we
are
now.
So
that's
really
exciting,
go
ahead
and
feel
free
to
to
try
manage
unmute
yourself.
If
you
got
stuff,
you
want
to
add
and
I'll
type.
A
C
B
So
we're
getting
some
actually,
people
are
typing
grab
it,
which
is
great
so
far,
a
speed
of
response
to
pull
requests.
I.
B
Think
that's
great!
That's
one
of
the
harder
things
to
accomplish.
So
that's
nice
to
see
looks
like
Robbie
wrote
using
zoom
for
meetings
works
well,.
B
D
B
All
right
so
verbally
is
anybody
want
to
either
chime
in
on
some
of
the
things
that
we're
looking
at
here
in
the
web?
Well,
look
circa
getting
a
lot
of
what
had
gone
better,
which
is
good
we'll
get
to
that
in
a
second
but
I'm
really
interested
in
things.
Anybody
want
to
give
props
to
people
who
they
were
interacting
with
on
this
project,
maybe
that
they
hadn't
done
before
or
a
process
that
we
added
this
time
in
terms
of
the
management
of
the
project
itself
or
any
of
that
good
stuff
would
be
great.
A
F
F
3
release
at
the
end
got
involved
in
some
sort
of
sort
of
urgent
areas
and
paul
was
already
neck
deep
in
things
and
really
I
think
you
know
just
put
in
a
lot
of
effort
with
clear
communication
and
I
think
what
was
exceptional
was
that
he
not
only
did
a
lot
of
work
and
checked
in
a
lot
of
code
and
then
reviewed
a
lot
of
code,
but
he
also,
you
know,
was
part
of
a
discussion
where
we're
trying
to
figure
out.
You
know
when
to
stop.
F
You
know
when
is
too
much
too
much,
and
you
know
we.
We
all
were
part
of
that
discussion
and
he
was
very
open
and
you
know
I
think
in
the
heat
of
the
moment
when
we
wanted
to
do
more
on
both
sides.
Both
you
know
some
of
the
storage
areas
on
google
and
on
RedHat.
You
know
we
were
able
to
say
you
know
what
we
we
need
to
put
an
end
to
this
and
make
sure
that
we
do
exactly.
We
need
to
do
for
one
dot,
three
and
no
more
and
I.
F
G
So
sorry
I
hit
power
ism,
but
so
this
mike
from
Korres,
so
I
would
like
to
give
big
kudos
to
Mike
Denise
and
to
Eric
tuned.
Both
of
them
were
very
helpful
or
in
various
areas
where
the
workplace
jumped
his
code.
Reviews
talvez
code
reviews
helped
his
guidance
with
the
process
reached
out
to
other
people
inside
Google
and
Reggie,
coordinated
to
kind
of
tinkerers
and
red
hat
when
needed
for
us
all
to
work
together.
B
G
A
B
All
right
else,
I
have
a
big
one
ad.
As
far
as
the
right
I.
H
Want
to
I'll
throw
out
a
a
big
thanks
to
Dan
Williams
and
Casey
Davenport
from
red
hat
and
tie
Guerra
early
calico
respectively.
They
did
great
work
in
the
network
sig
getting
the
sig
to
get
together
all
the
time
and
actually
driving
forward
on
a
bunch
of
the
issues,
including
the
network
policy
design.
So
I
just
know
if
we're
putting
out
thanks
I
think
thank
those
guys.
Yay.
A
B
F
Yeah
up
until
the
last
minute,
yawn
and
Justin
were
also
yes
all
over
the
place
on
support,
validation
and
fixing
bugs
and
just
rewriting
large
pieces
of
code,
and
always
doing
it
with
high
quality.
For
someone
like
me,
who
hasn't
had
a
chance
to
really
meet
any
of
them
and
person
until
near
the
end,
bit
more
like
ghost
in
the
machine
making
everything
better
I'm,
so
yeah,
big
plus
12,
that.
B
Don't
we
all
okay,
let's,
let's
go
ahead,
move
on
to
what
could
have
gone
better,
let's
go
ahead
and
I'm
going
to
go
ahead
and
just
for
the
sake
of
the
meeting
will
read
through
these
I
and
I'm
going
to
go
ahead
and
approve
the
things
that
have
already
been
out
there.
And
then
we
can
add
more
comments.
B
Assuming
Ford
so
need
a
clear
set
of
priorities
in
progress
and
if
I
read
these
and
you
are
the
person
who
wrote
it
and
have
more
to
add
that
would
be
super
helpful,
because
some
of
these
maybe
I
have
nuances
that
I'd
like
to
understand
more
in
documents.
So
who,
who
is
the
person
who
wrote
this?
And
would
you
like
to
speak
to
it.
A
B
Yes,
Joe
will
be
typing
since
he's
at
the
starbucks
in
love.
B
A
B
E
B
A
Did
the
this
sort
of
retrospective
a
little
bit
smaller
and
then
a
lot
of
it
was
changed
in
one
point
and
you
know
how
we
released
1.2
and
then
again
after
1.2
and
1.3.
So
for
those
people
who
see
a
little
bit
of
chaos
and
drama
now
it
was.
It
was
differently,
chaotic
and
more
difficult
to
follow
with
process
and
priorities,
then,
but
with
fewer
people.
So
where
we
have
improved
some
process
and
accountability,
we
have
gotten
more
people
trying
to
help
which
has
made
this
also
a
bit
a
bit
more
chaotic
in
1.3.
B
Excellent
anybody
else
want
to
make
a
comment
on
clear
set
of
parties
and
progress
before
we
move
on
to
the
next
one.
A
A
I
A
question,
if
you
don't
mind
regarding,
for
example,
pet
set
when
and
where
was
it
designed?
It's
one
thing
to
say:
here's
what
we
decided.
It's
maybe
even
more
important
to
get
the
right
people
involved
in
the
design.
So
my
god.
H
This
Tim
I'll
speak
to
that
and
that
one
in
particular
is
a
I,
don't
know
if
it's
a
success
story
or
a
dismal
failure.
It
started
in
some
like
a
1000
level
issue
literally
two
years
ago.
It
accumulated
probably
on
the
order
of
a
thousand
comments.
Then
it
got
split
to
a
proposal
which
accumulated
probably
a
thousand
more
comments
at
which
point
I
forget
exactly
if
it
we
just
started
designing
the
workout
code
or
what,
but
it
was
a
man
or
miss
backlog.
H
The
problem
is
with
I
think
we
will
always
have
these
sorts
of
enormous
design,
problems
that
have
a
very
wide
space
and
are
very
hard
to
come
to
consensus
on
and
I'm,
not
sure
how
to
do
that
better,
I'm
very
open
to
ideas.
I,
don't
think
that
was
a
smashing
success,
but
at
the
same
time
we
got
a
lot
of
input
from
a
lot
of
people
and
we
didn't
rush
into
any
decisions,
and
so
you
know
in
some
senses,
I
think
it
really
went
well
on.
I
I
J
I
could
just
give
him
just
a
teeny
bit
more
clarity
to
this
I
to
be
clear,
a
you
know,
there's
without
question.
There
was
a
some
improved
communication
that
we
could
have
done
this
time
around.
There's
no
question
about
that.
That
said,
I
was
watching
this
transpire
and,
and
you
know
around
the
12
timeframe,
you
said
hey.
This
is
something
really
big
and
we
found
some
red
hat
folks
who
are
interested
in
working
on
it
as
well
as
Googlers
as
they
double
click
on
it
to
try
and
understand
what
was
realistic
for
a
milestone.
J
We-
and
this
is
where
the
communication
could
have
happened-
said
jeez.
This
is
actually
really
really
complicated.
You
know
Tim's
point
about.
You
know
a
thousand
comments
on
an
issue
and
things
like
that
is
probably
even
understating
how
bad
it
was,
and
we
said
well,
what
we're
going
to
do
is
we're
going
to
go
at
we're
going
to
go
forward
with
with
kind
of
just
a
test.
You
know
around
a
given
proposal.
J
Like
you
know,
here
are
the
things
that
we
think
we're
going
to
like
try
and
complete
in
this
milestone,
because
it's
clearly
beyond
a
milestone.
It's
clearly
only
going
to
make
it
to
alpha
because
we
need
miles
on
this
feature
before
we
lock
anything
down
or
any
api's
down
that
to
be
clear.
That
was
the
point
where
we
should
have
shared
with
the
community.
That
was
a
failing.
We
are
well
aware
of
that.
That
said,
what
you
see
in
pet
set
to
really
be
considered
hey.
This
was
our
first
go-round.
J
We've
got
it
working
with
this
set
of
apps
we'd
love
miles
on
it.
We'd
love
your
feedback
from
this
point,
think
of
it
as
an
MVP
that
you
know
will
will
use
as
the
basis
to
move
forward
from,
but
but
yeah.
You
know,
there's
no
question.
When
we
made
that
decision,
we
should
have
been
more
clear
about
that
and
and.
H
I
want
to
I,
don't
want
to
try
to
sound
like
I'm,
justifying
working
in
private
part
of
the
reason
to
just
say
we're
going
to
pick
a
couple
and
try
to
experiment
was
just
to
reduce
scope
and
remember
of
voices,
because
we
could
honestly
continue
to
argue
about
this
thing.
Ever
it's
just
an
applique
t'adore,
a
problem
so
yeah
at
some
point.
A
Joke
I,
just
own
without
explicit,
set
set
of
decision
makers,
which
is
my
project
for
this
next
week,
is
to
define
to
define
what
the
scope
and
charter
is
for
elders
and
then
bring
that
to
the
community
meeting
next
week
and
talk
about
it
in
the
upcoming
weeks
week
or
weeks.
Given
that
tomorrow
or
next
week
might
be
light.
Yeah.
G
I
If
I
may
compare
and
contrast
with
the
the
network
policy
design
discussion,
which
also
went
on
for
a
long
time,
not
as
long
course,
but
there
was
an
explicit
agreement
at
some
point
to
say,
look,
we
know
this
is
not
the
end
of
the
road
we're
going
to
pick
some.
You
know
a
first
step
on
the
road
map,
understanding
that
we
could
sit
here
and
argue
for
a
lot
longer,
but
we
instead
want
to
get
something
out
and
get
experience
with
it.
I
believe
designers
are
capable
of
reaching
such
an
agreement.
I
My
problem-
and
it
may
just
be
because
of
the
peculiar
way
I
slouched
into
this
I-
started
getting
involved.
I
think
during
the
13
time
frame
when
pets
that
had
already
been
underway
for
a
while,
nobody
told
me
it
was
underway,
I
just
kind
of
heard
about
it.
Occasionally
people
would
refer
me
to
these
PRS
with
a
thousand
comments
and
cos
not
humanly
possible
to
track
it
with
that
I
myself
have
issues
with
the
design
and
so
I
guess.
Maybe
the
constructive
thing
is
to
say
going
forward.
I
I
think
what
would
make
sense
to
say
is
you.
We
should
have
some
forum
where
the
design
is
discussed
and
you
know
hopefully
most
people
are
mature
enough
to
understand
that
we
don't
want
to
argue
forever.
We
want
to
get
things
out
incrementally
to
get
experience
with
them
and
it
sort
of
should
be
known
where
this
forum
is
and
people
who
care
to
can
participate.
Yeah.
H
Like
I
think,
that's
I
think
that's
a
great
idea,
and
it's
I'm
personally,
sorry
that
you
didn't
know
it
existed
because
I'm
sure
it
could
have
benefited
from
your
input.
I
think
now
that
we
have
a
toe
in
the
water
design
that
the
feedback
is.
Actually.
This
is
the
right
time
for
it
now
like
what
works,
what
doesn't
work,
what
can't
possibly
work
in
the
future?
It
is
an
alpha,
great
API,
let's
eat
the
heck
out
of
it
and
throw
it
away.
F
Or
pre
happy
I
think
the
other
thing
to
notice,
too,
is
that
this
is
still
alpha
and,
as
I
understand
it,
that
means
that
it's
still
open
for
input
and
changes
that
correct
five
hundred
percent,
so
I
think
that
to
some
degree
it's
it
isn't
great
that
the
early
design,
the
way
I
look
at
things
is
and
maybe
again
I
said
I'm
new.
The
early
design
process
was
probably
a
little
under
under
the
water
more
than
it
should
be
it
under
the
visibility
but
I.
H
A
David
dropped
into
the
chat,
the
link
to
the
design
doc
for
it,
so
there
is
a
design
doc,
although
most
people
have
referenced
the
epic
issues
instead
generally,
so
this
will
be
in
another
spot
to
be
that's
a
fair
point
that
is
not
in
the
Cooper
tanisha
repo
would
have
been
hard
to
find.
I
agree.
G
Kind
of
sorry
to
return
it
is
it
possible
to
set
up
like
hey
a
proposal
will
take
around
two
or
four
weeks
of
discussion
have
clear
kind
of
guidelines,
timelines
communicated
to
the
group,
so
that
I
mean
it's
a
large
distributed
project.
Right
people
come
and
go.
They
have
their
own
things
outside
of
coordinators
right,
so
it
doesn't
become
awesome.
Buddy,
return
on
the
day
before
the
deadline
and
started
subject
and
everything
sure.
A
C
G
C
B
Could
we
the
half
past
the
hour?
So
what
I'd
like
to
do
is
move
on
to
the
next
ones,
because
there's
really
big
things
coming
up
here,
like
the
next
one,
for
example,
better
documentation
on
how
the
automated
tests
are
done
and
helping
them
sent
down
flaky
tests.
What
we
did
that,
because
that's
a
really
huge
deal.
B
I
I
had
an
experience,
I
think
a
week
or
two
ago
and
again
I'm
kind
of
new
here,
so
I
there's
just
a
lot
of
stuff.
I
don't
know
right
so
I
submitted
a
PR
and
one
of
the
automated
tests
failed.
You
know
and
it
produces
a
bunch
of
stuff.
In
this
case
it
looks
like
even
what
failed
was
something
really
basic
in
the
setup.
I
didn't
even
see
like
functional
testing
happening.
It
was
something
just
pulling
something
something
really
basic
and
I.
I
Just
don't
know
enough
about
the
framework
to
have
any
idea
what
failed
and
second,
if
a
functional
test
did
fail.
How
do
I
find
the
test?
I
mean
there's
just
basic
stuff
that
you
know
there's
evidence
there,
but
I
don't
know
enough
to
go
work
with
it
right,
I
mad.
If
I
can
read
the
code
if
I
know
how
to
find
the
code.
C
Maybe
I
can
answer
that
a
little
bit,
we've
had
a
large
problem
with
flaky
tests
and
we've
been
working
on
fixing
it
and
making
the
framework
more
tolerant
to
flakes.
I
have
changes
in
the
works
to
sort
of
address
that
there's
also
a
test
grid,
basically
to
make
a
long
story
short.
This
area
isn't
a
lot
of
flux
and
it's
time
consuming
to
keep
the
documentation
up
to
date
with
the
this
stuff
that
you
doing
rapidly,
and
so
that
hasn't
happened.
C
A
This
is,
this
is
definitely
something
that
is
a
known
issue
and
that
we're
working
hard
against
Daniels
point
keeping
the
documentation
up
as
it
is
changing
getting
from
very
broken
through
many
iterations
too
many
fast
iterations
to
improving
improved
new
structure.
We
have
been
lagging
on
documentation
for
sure,
but
this
is
we
really
need
and
know.
We
really
need
to
get
a
lot
of
this
information
out
of
the
early
deaths
heads
in
order
to
bring
on
board
new
devs
and
make
them
more
successful
and
useful.
I
Yes,
my
point,
which
is
I,
don't
know
really
what
they
was
coming
from.
I
I,
don't
know
how
to
look
at
the
code
that
was
doing
the
the
polls
or
whatever
allocating
the
VMS.
Whatever
you
know
the
basic
stuff,
I
don't
know
how
to
find
what
tests
are
run
you
know
is.
The
problem
is
not
finding
the
word
error
in
the
output.
The
problem
is
okay.
Why
is
that
occur
there?
What
was
being
tested?
What
what
is
the
test
to
produce
the
error?
Yeah.
K
So
I
think
that
you're
running
into
a
situation
where
just
the
amount
of
time
that
we
have
to
dedicate
to
clear
and
explicit
documentation
all
of
this
as
it's
changing
so
rapidly,
is
just
not
really
a
priority
for
us.
This
is
aaron
from
sick
testing
by
the
way
I
don't
if
that
was
a
clue
that
Sarah,
oh,
is.
K
K
I
can
say
that
Rob
I
apologize
RMH
is
the
avatar
I,
don't
know
the
last
name
put
together
this
goober
nadir
app
that
like
actually
comments
on
the
PRS,
the
specific
test
that
failed,
the
specific
assertion
that
failed
and
there
is
I-
think
a
link
to
the
actual
logs
of
that
particular
thing.
Thank
You
Ryan
see
I
totally
messed
that
up
even
but
so
there's
active
work
on
going
here
and
if
you
have
specific
needs
that
aren't
being
addressed
and
there's
not
an
issue
that
we
have
filed
about
this.
K
F
There
a
way
for
people
who
need
to
understand
the
workflow
like
this
is
something
that,
when
I
joined,
was
a
little
dizzying
but
now
makes
sense
of
you've
checked
in
your
code.
These
are
the
Edes
that
run
on
it
and
when
they
have
errors,
and
especially
if
you
you
get
an
error-
and
you
don't
know
if
it's
it
doesn't
seem
to
make
sense
that
you're,
you
know,
check
your
PR
at
anything
to
do
with
it.
F
You
can
search
issues
find
that
editor
and
then
you
know,
hit
the
cakes
robot
ignore
retry
because
of
this
flaky
bug,
that's
already
known,
especially
now
that,
due
to
all
of
the
great
work
that
I'm
seeing
you
know,
this
is
so
much
easier.
Even
in
the
one
month
since
I've
joined.
Is
that
sort
of
workflow
documented
anywhere
that
we
could
add
a
link
to
its.
I
K
F
Think
I
think
this
is
that
there's
only
a
small
piece
that
I
think
I
needed
as
a
developer,
which
was
when
I
get
my
PR
and
the
test
says
it
failed
in
the
cave
spot
to
have
a
link
that
just
says:
hey
here's,
the
step
of
search
issues
for
your
error
text,
because
it
sounds
like
from
what
we
just
heard.
It's
very
easy
to
now
know
the
error
text
and
if
you
see
a
flake,
you
know
if
you
see
a
test
of
that
error,
flake
Kate
spots
and
so.
C
C
K
F
C
K
I
This
I'm
a
dip
I,
don't
need
a
run
book.
I
just
need
to
understand
how
the
system
works.
So
my
car,
you
just
ask
Malcolm.
My
problem-
may
have
been
peculiar
to
what
was
going
wrong
on
the
day
that
I
had
my
flake,
but
it
looked.
He
saw.
I
saw
a
couple
of
warnings
that
were
really
generic
and
then
some
errors
that
might
have
been
just
a
consequence
of
the
warning,
wasn't
sure,
but
the
things
that
said
warning
an
error
were
so
generic
that
I
did
try
searching
for
them.
I
H
Know
Mike,
if
we,
if
we
had
a
high-level
doc.
That
said
this
is
what
happens
when
a
PR
gets
submitted.
We
first
we
spin
up
a
small
cluster
and
then
we
push
the
current
head
with
your
patch
and
then
we
run
this
battery
of
tests
against
it,
and
these
are
all
the
places
that
it
might
fail
and
we're
sorry
if
it
fails
at
this
point,
because
it's
not
your
fault,
but
sometimes
that
happens,
would
that
be
helpful.
That'd.
I
H
I
think
we're
getting
wrapped
up
in
the
details
of
like
which
bug
failed,
which
test
and
I
think
actually
what
might
be
missing
here
is
a
very
high
level
picture
of
the
test
flow
that
people
who
aren't
familiar
with
the
fact
that
every
PR
runs
against
a
real
GCE
cluster
would
be
able
to
look
at
and
be
like.
Oh
well,
GCE
shat,
the
bed,
and-
and
so
it's
not
my
fault,
yeah.
H
C
B
We
do
need
red
line,
actually
just
real
quick,
I'm,
really
happy
that
we're
getting
so
much
feedback
as
compared
to
last
one.
This
is
a
ton
of
useful
feedback,
so
so
the
next
one
is
really
an
easy
way
to
on
docs
labels
for
issues
for
new
developers
to
contribute
kate's.
This
actually
kind
of
ties
into
what
was
just
mentioned
was
we
want
to
have
a
very
low
barrier
to
entry
for
people
to
be
able
to
start
contributing.
B
So
if
it's
that,
if
the
testing
process
isn't
clear,
we
need
to
make
sure
that
that's
document
and
this
really
sort
of
falls
in
line
with
that
is
to
make
sure
that
developers
can
easily
do
that,
and
this
is
something
we
need
for
sick
drops
as
well,
because
we
want
to
make
sure
that
when
people
want
to
contribute
to
the
docs
repo
that
we
can
do
that.
So,
first
of
all
who
added
this
and
in
what
it
would?
How
do
we
make
this
better.
L
Hi,
I'm
serge
so
I
entered
this
being
on
YouTube
Covenant
is
starting
to
contribute
to
it.
I'm
also,
so
so.
I
understand
that
it's
a
great
it's
a
it's
a
lot
of
work
for
making
us
to
label
each
thing,
which
is
like
easy
for
new
piece,
but
then
it
will
help
for
you
to
get
more
people
into
it
into
the
project
and
it
is.
There
were
also
what
Sarah
is
going
right
now.
L
H
Have
that
label
we
just
don't
apply
it
well
and
that
label
is
called
help
wanted.
Maybe
that's
the
wrong
name,
but
it
is
not
applied
consistently
or
with
any
predictable
meaning,
but
I
try
to
use
it
when
I
can,
whenever
I
file
an
issue
that
I
think
somebody
community
could
totally
tackle
in
less
than
a
week
of
work,
I
file
it
that's
my
sort
of
rule
of
thumb,
I,
don't
know
how
other
people
use
it.
We
should
probably
clarify
that,
but.
A
Suresh
to
your
point,
I
believe
that
we
also
need
to
make
that
more
obvious.
I
mention
it
frequently
when,
with
the
work
that
I
do
but
making
a
hey,
are
you
new
de
Kate's,
where
we
want
you
to
help
us
develop
things
page
and
the
getting
started
with
developing
page
within
case
is
probably
something
that
needs
to
be
more
obvious.
I
know
we
have
some
of
that
documentation,
but
make
it
more
obvious
in
more
clear
and
make
easier
onboarding
polluters
there's.
E
The
other
this
is
done
from
the
null
team
and
there's
the
other
way
we
do,
because
for
each
of
your
knees,
doing
the
doing
the
planning,
so
we
mark
the
p0,
p1
and
p2
you
so
our
last
two
capellinis
I
literally
say
the
P
to
you
is
kind
of
less
too
high
and
we
don't
have
the
engineering
resources
working
on.
No
matter
is
the
partner
or
the
google
nur,
and
we
want
to
come
community
can
help.
E
If
someone
can
drive
started
from
the
design
discussing
and
the
test
development
and
testing
and
release,
then
we
totally
okay
to
you,
but
we
need
an
active
working
planning.
Next
that
you
want
to
work
on
those
project.
You
need
the
kind
you
have
to
make
the
commitment
and
say:
oh
I,
want
to
drive
this
holder
feature
from
the
beginning
to
end,
and
if
there
anything,
you
cannot
make
such
commitment
to
you.
The
upfront
tell
us
so
something
like
that.
E
B
Okay:
let's
go
to
move
on
the
next
item,
which
is
need
some
tea
bleeds
for
areas
who
did
this
because
I'm
not
sure
I
completely
understand
what
that
means.
J
J
You
know
what
stealing
from
the
Linux
contributors
or
the
Linux
kernel
system.
Some
team
excuse
me
ownership.
They
have
owners
for
USB
and
net,
and
so
on
and
so
forth
and
I'd
love
to
you
know,
offer
up
to
the
community
to
figure
out
how
we
work
towards
a
path
to
getting
there
where
someone
outside
of
Google
specifically
potentially
owns
an
entire
sub
team
area
and
thereby,
you
know,
accelerates
that
area's
PR
review,
you
know
bandwidth
and
so
on.
J
J
Subject
matter,
expert
and
more
about
developing
a
kind
of
tree
of
trust
between
various
people
where,
as
things
flow
up
the
chain,
they
have
been,
you
know
fully
reviewed
by
someone,
rather
than
kind
of
like
in
the
design
and
contributions
as
phases,
it's
much
more
about
as
a
person
reviewed
this
entire
section
and
and
for
the
you
know,
whatever
in
this
case
that
you
know
stealing
from
the
Linux
kernel
mod
mod
methodology.
You
know
the
USB
sub-team,
you
know
owner
when
they
you
know
approve
that
entire
thing,
the
entire
USB
tree
comes.
H
F
H
A
H
A
H
I
mean
what
makes
the
Linux
model
possible.
Is
that
literally
you
know,
Dave
Miller's
net
tree
corresponds
to
the
net
directory
and
everything
underneath
it
and
lynus
respects
Dave's
opinion
on
it
and
Dave
respects
that
his
boundaries
are
net
and
basically
any
change
in
the
neck
directory
is
if
it's
approved
by
dave
is
approved
by
linis,
and
you
know,
dave
has
his
set
of
people
that
he
trusts
and
they
have
their
set
of
people
that
they
trust
and
everything
flows
upstream,
and
it
merges
well
because
of
the
structure
there.
F
A
Other
thing
that
helps
with
that
is
that
those
sub
areas
then
know
if
this,
the
net
thing
that
also
has
to
talk
to
USB.
They
know
where
to
make
those
interconnections
as
well,
and
that's
that
is
one
of
the
pain
points
of
the
growing
community
right
now
is
a
making
sure
we
know
where
those
interconnects
are,
who
owns
them
and
also
keeping
that
updated
and
keeping
people
apprised
of
it,
as
as
the
community
girls
letting
having
a
place
that
people
find
it.
A
J
Of
grunt
work
to
be
clear:
I,
don't
agree
that
the
cigs
should
be
the
alignment
for
this
I
think
my
goal
or
my
opinion
is
that
I
think
that
cigs
are
user,
facing
issues
that
may
be
cross-cutting,
so
OpenStack
or
AWS
or
ops,
or
things
of
that
nature.
I
think
that
some
team
leads
similar
to
the
Linux
model.
Our
technical
tree
areas
that
that
have
cleaner
code
boundaries
but
I
mean,
if
you
take
note,
for
example,
you
know
you're
going
to
have
an
upgrade
situation
which
crosses
kernel
or
Samia.
A
B
Let's
move
on
real,
quick
we're
going
to
be
in
a
race
to
actually
get
this
done
so
better
documentation
of
new
1.3
features
some
people
that
both
point
in
there.
This
was
improved
when
we
put
the
river
all,
but
there
is
a
fair
amount
of
outdated
docks,
diagrams
and
best
practices
that
we
could
have
improved
upon.
N
This
is
a
general
call
out
for
bringing
dogs
overall
I
nodded,
there's
way
more
eyes
on
the
project.
That's
this
will
benefit
a
lot
of
meters,
not
only
in
just
how
to
install
and
get
the
stuff
running,
but
in
terms
of
understanding
the
logic
behind
key
to
moving
parts.
B
Okay,
cool
yeah
is
there
any
I
mean
I
feel
like
this
is
a
little
self-evident,
and
everything
we've
talked
about
thus
far
is
really
tied
into
this.
So
without
objections
I'd
like
to
move
on
to
the
next
one,
which
is
no
explicit
set
of
decision
makers
for
big
changes,
IE
what
belongs
in
core
and
what
doesn't
I
think
this
is
really
interesting
and
again
ties
back
to
what
we
were
just
discussing.
So,
if
there's
anything
that
is
not
related
to
the
prior
discussion,
let's
add
that
now
anybody
want
to
take
a
crack
at
that.
K
So
I
guess
just
to
drag
us
too
far
into
it,
but
the
owners
file
that
I
keep
seeing
everywhere.
Is
that
an
attempt
at
implementing
some
of
this
right
now,
like
making
sure
somebody
owns
the
code
base?
It's
kind
of
yeah
I
mean
yes,
the
software
is
organizing
we're
definitely
going
to
be
reorganizing.
The
people,
so
Conway's
law
is
going
to
be,
in
effect,
the
whole
time
I'm
just
trying
to
understand
like
what
does
the
owners
file
not
provide
us
it?
Given
the
context
of
the
previous
discussion.
H
Which
we've
had
lots
of
discussion
on,
and
it's
just
going
to
take
the
dedicated
application
of
hands
on
keyboards
for
extended
periods
of
time
to
move
things
around
and
fix
up
all
the
backlinks
and
the
documentation?
And
then
we
can,
you
know,
do
better
at
this
and
we're
starting
it
and
we
will
keep
going,
but
you
as
ideas
this.
B
A
B
Yes,
cool
is
great
I
love,
putting
improvements
or,
on
the
way,
I'd
like
to
see
more
focus
on
getting
the
test
working
PR
up
speed,
it's
a
bit
slow,
getting
point
where
you're
on
backwards.
Oh,
it's
like
we
can
occur
that
that
is
there
anything
else
you
want
to
add
there.
C
On
the
p
RQ,
what
what
what
are
we
doing
to
to
reduce
it?
Yeah
specifically
and
I've,
sent
out
a
document
to
Cooper
Nettie's
dev
about
this
a
couple
times,
specifically,
we
are
making
sure
all
pre
submit
tests
running
within
half
an
hour,
we're
going
to
start
monitoring
this
and
pinging
humans.
If
that
stops
being
the
case,
where
we're
making
sure
that
the
retest
rate
has
a
success
rate
of
ninety-five
percent
or
more,
that's
the
last
thing
that
I'm
working
on
also.
H
C
And
we've
we've
already
optimized
the
post
emits
to
be
tolerant
of
occasional
flakes
and
we're
firing
filing
issues
for
flakes
automatically
to
make
sure
that
humans
have
visibility
into
the
problem
and
we'll
fix
them
or
automatically
prioritizing
them
we're
still
working
on
getting
like
after
the
fact
reporting.
So
we
get
up
a
backup
system
that
will
ping
humans
or
or
their
managers
or
something
if
flakes
are
sitting
there.
If
important
flakes
are
sitting
there
not
getting
fixed
so.
O
C
That's
a
significant
problem
is,
we
are
there's
like
a
there's
like
a
funnel
pipeline
to
people
like
people
start
out
sending
us
PRS
and
we
review
them
and
then
then
they
like
grasp
and
understand
larger
parts
of
the
code
base
and
we
start
trusting
them
to
review
things
and
and
like
we
have.
We
have
significant
growing
pains,
in
my
opinion,
right
now
on
I'm,
finding
bandwidth
to
review
so.
H
B
A
Onboarding
new
reviewers
is
another
component
that
I've
been
working
on
as
well
figuring
out
how
to
better
on
board
get
more
information
all
right.
So
we
are
six
minutes
from
out
of
time
and
we
haven't
even
gotten
to
the.
What
will
we
be
doing
differently
so
I
want
to
do
a
quick
poll
of
Jason
and
everyone
do
we
need
to
make?
B
A
We
have
a
lot
of
the
team
that
is
out
next
week,
given
it's
a
holiday
week
in
in
seattle
or
sorry
in
seattle,
in
the
u.s.
I
had
to
localize
too
much
so
does
this
look
like
something
the
first
part
of
the
following
week
that
we
could
be
11th
tonight.
J
Since
there
were
kind
of
two
halves
of
this
and
I'm
kind
of
curious,
if
we
could
potentially
do
this
asynchronously,
the
first
half
was
collect
a
whole
bunch
of
the
general
feed
issues
and
I.
Think
we've
actually
done
that
pretty
well
I
wonder
if
we
could
shard
the
issues
remaining
and
basically
allow
people
to
asynchronously,
join
or
discuss
things
off
line
via
email
via
whatever
is
appropriate,
just
basically
+1
yourself
into
a
particular
issue
or
add
a
comment
and
and
rather
than
everyone
get
together
for
a
meeting
like
potentially
you
know
call
it.
J
B
Think
that'd
be
good.
What
I
would
like
to
see
is:
maybe
we
have
a
time
period
where
people
can
add
in
a
synchronously
in
the
comments
and
I
will
actively
manage
the
document.
During
that
time,
I
mean
we
just
have
a
follow-up
for
people
who
are
really
interested
after
that,
after
we've
had
some
time
in
the
holidays,
because
I
feel
like
there's
a
certain
amount
of
reporter
that
happens,
face-to-face
that
can't
really
be
accomplished
in
the
document.
Oh
yeah,
so.