►
From YouTube: 20200813 SIG Architecture Community Meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Hello,
everybody:
this
is
the
kubernetes
league
architecture
meeting
for
august
13th
2020..
We
have
a
good
crowd
here,
so
please
just
want
to
remind
everybody
to
follow
our
code
of
conduct,
treat
each
other
with
respect
and
assumption
of
good
intent,
and
I
think,
with
that
we'll
get
started.
We
have
just
a
couple
things
on
the
agenda
today.
First
is
the
discussion
about
the
cube
add-on
manager.
A
I
can
give
some
context
on
this
or
dims.
If
you
wanted
this
concert
or
jake,
I
see
you're
on
online
as
well,
so
I
have
to
let
one
of
you
take
it.
C
C
C
You
know,
additionally,
we
just
you
know
figured
well,
I'm
in
there
might
as
well
start
testing
it.
You
know
novel,
that's
that's!
That's
pretty!
That's
pretty
much!
It.
B
The
part
of
the
problem
here
is
cluster
is
not
maintained
right
in
the
sense
that
there
is
no
sig
ons
that
owns
the
cluster
itself
directory
and
then
all
the
sub
directories
under
it,
and
it's
like
the
sprinkler
sprinkling
of
people
from
here
and
there
branding
together
to
keep
it
alive.
So
that's
part
of
the
part
of
the
year.
B
So
the
other
thing
that
comes
to
mind
is
you
know
if
we
turn
some
of
this
into
the
go
based
binaries,
do
then
the
questions
that
come
around
and
like
do
we
need
to
build
them?
Do
we
need
to
publish
container
images?
Do
we
need
to
you
know,
add
it
to
the
release.
It
there's
like
a
whole
slew
of
maintenance
stuff
that
comes
in.
B
So
if
it's
just
scripts,
it's
easy.
Nobody
cares
right,
but
then,
once
you
turn
it
into
a
binary,
then
we
have
to
talk
about
the
rest
of
the
things
that
come
associated
with
it.
So
well,
some
of
the
questions
were
around
I
mean.
Does
it
need
to
live
here?
Is
there
a
better
home
for
it?
Is
it.
A
A
A
Else
if
we
know
where
to
move
it,
the
justin
had
some
ideas
on
this.
Since
he's
involved
in
cluster
life
cycle.
I
don't
know
justin.
You
wanted
to
comment
on
that.
E
Sure
there
is
a
there
is
a
potentially
overlapping,
at
least
thematically
a
subproject
of
c
cluster
lifecycle
called
the
cluster
add-on
subproject.
It
is,
you
know,
relatively
early
in
its
maturity
life
cycle.
I
don't
think
we
haven't
done
like
even
an
alpha
type
release.
Yet,
however,
we
do
have
the
repo
structure
set
up.
We
have
you
know
a
bunch
of
build
type
mechanics
there
and
I
think-
and
we
don't
have
a.
E
I
would
call
this
a
bootstrapping
tool,
so
there,
the
cluster
islands
is
more
or
less
pursuing
the
idea
of
operators
as
the
way
of
installing
things,
but
we
don't
have
a
way
to
actually
install
the
operators
themselves
and
this
the
cube
item
manager,
former
script
now
go
binary,
could
be
seen
as
that,
and
so
from
that
point
of
view,
I
think
cluster
add-ons
could
be
a
great
place
for
it.
I'm
one
of
the
approvers
I
saw
I
can't
I
would
certainly
be
in
favor
of
it
going
there.
E
I
can't
speak
for
everyone,
but
I
I
can't
imagine
anyone
objecting
if
that,
so
it
seems
like
that
would
meet.
It
seems
like
that's
an
option.
B
C
B
So
the
the
other
question
I
had
for
you
jake
was
other
than
the
cube
of
scripts.
Is
anybody
using
this
in
the
field
is?
Do
you
know
of
any
end
users
that
pick
up
the
crypt
that
we
need
to
notify
or.
C
B
G
C
I
can
I
can
speak
to
that.
So
the
answer
is
indirectly,
I
mean
it
is
consumed.
C
The
answer
is,
maybe
it's
definitely
it's
definitely
gke.
It's
definitely.
You
know
the
cube
up
script
for
the
you
know
open
source
tests
in
terms
of
like
actual
customers
using
it.
B
Probably
not
directly,
then
right,
so
the
only
other
thing
that
I
I
could
think
of
was
the
cube
of
scripts
were
being
duplicated
inside
one
other
repository.
So
that
was
the
only
other
thing
that
I
could
think
of,
but
this
you
know
justin's.
The
report
that
jackson
pointed
out
feels
way
better.
You
know
in
terms
of
being
aligned
with
the
work
that
is
getting
started,
so
I
think
that
that's
the
better
sure
this.
F
C
Where
where'd
you
use
it
tim,
I
know
that
there's
like
a
whole
bunch
of
configuration,
mismatches.
H
I
don't
know
which
configuration
broke,
but
I
was
trying
with
a
bunch
of
different
configurations,
just
to
see
how
the
different
parameters
worked
or
didn't
work
with
it,
and
I
think
it
was
like
vpc
native
plus
network
policy,
like
I
turn
on
a
bunch
of
different
parameters
and
my
cluster
came
up
and
none
of
the
nodes
are
schedulable,
because
it's
complaining
that
the
routes
haven't
been
programmed.
But
it's
on
supposed
to
be
on
the
vpc
native
mode,
and
I
didn't
really
dig
into
it.
But
I
just
wanted
to
say:
like
yeah,
I
tried
it.
C
B
C
B
Okay-
and
the
last
question
for
me,
is
the
you
don't
see
foresee
any
specificity
around.
You
know
the
older
releases
that
we
have
currently
16
17
18.
The
same
version
will
work
for
all
three
that
you
know.
C
It
yeah
it
should
like.
I,
I
intentionally
implemented
it
just
like
as
directly
as
possible,
just
a
direct
translation
so
that
if
it
worked
before,
hopefully
it'll
continue
working,
you
know
if
it
doesn't,
then
you
know,
that's
probably
like
a
good
indicator
that
I
screwed
something
up
kind
of
thing
rather
than
you
know
like
us,
introducing.
B
Yeah
yeah,
typically,
we
don't
go
around
deleting
the
scripts
from
the
older
releases,
but
this
was
just
in
a
just
in
case.
A
question.
C
Oh
yeah,
I
mean
the
older
releases
aren't
going
to
be
changed
at
all,
because
the
the
image,
the
add-on
manager
image
that
they're
using
is
just
baked
into
one
of
the
manifests.
B
You
know
that
justin
pointed
out
and
merge
that
in
and
then
do
all
the
things
that
we
need
to
do
in
pro
to
build
the
image
and
push
it.
We.
We
have
a
lot
of
other
images
that
we
can
copy
stuff
from
to
do
that,
and
once
we
use
a
image
promoter,
we
can
use
the
image
promoter
and
promote
it
to
the
broad
repository
and
once
it
is
there,
we
can
go
clean
up
the
existing
kk
master
to
use
the
image
and
remove
the
script.
A
Okay,
great,
then,
why
don't
we
move
on
to
the
sub
project?
Readouts,
I'm
actually
one
thing
I
usually
have
on
the
agenda
that
I
forgot
is
sort
of
follow-up
from.
A
I
Sure
let
me
put
that
now:
yeah
yeah.
So
basically,
to
sum
up,
like
I
wrote
the
proposal
for
creating
reliability
working
group,
I
also
pasted
the
link
in
in
the
email
thread
discussion.
Sometimes
so
so
we
all
there
are
already
couple
discussions
there,
some
of
them
we
already
resolved.
Some
of
them
are
still
open,
but,
like
I
would
appreciate
any
other
feedback.
I
would
like
to
propose
that
more
officially
in
the
next
week
or
two
and
actually
start
doing
some
real
work.
There.
D
You're
really
quiet
sure
so
jorge,
who
I
think
is
actually
here,
has
been
developing
the
ci
signal,
subproject
plan
and
there's
a
perception
of
overlap
between
these
two
efforts.
So
I
just
want
to
point
that
out,
but
there's
rfc
that
jorge
has
developed
around
the
subproject
idea,
which
would
be
well.
I
mean
he
can
weigh
in
on
if
you'd
like,
but
essentially
federated
ci
signal
and
different
sigs
to
help
manage
that
aspect
of
work
and
it
gets
it
a
lot
of
the
reliability
goals
that
are
in
vortex,
email.
I
D
We
were
hoping
to
make
it
go
live
this
week,
but
with
the
the
issues
with
the
releases,
it
was
not.
I
Possible
yeah,
I
think
my
goal
isn't
really
to
for
for
reliability
to
own,
like
how
we
track
how
we
track
like
six
test,
individual
six
tests,
healthiness
and
stuff
like
that,
but
rather
give
us
whoever
the
act
will
be
here:
the
powers
to
actually
somewhat
enforce
that
as
part
of
the
sick.
So
I'm
I
still
can
easily
imagine
that
it
will
be
fully
driven
by
by
the
ci
signal
sub
project
and
we
yeah.
I
A
D
F
A
B
Yeah,
absolutely
I
I
need
I,
I
need
you
and
direct
to
look
john,
but
I
definitely
need
other
people,
especially
who,
who
have
been
through
this
process
or
who
will
go
through
this
process
to
take
a
look
to
see
what
we
are
in
for
what
we
are
looking
at
in
terms
of
like
the
life
cycle
of
the
whole
thing
and
where,
where
we
are
looking
for
information
where
we
are
looking
for
help
how
to
get
integrated
into
the
community
so
how
to
get
integrated
into
the
testing
and
release
sigs,
you
know
so
there's
a
bunch
of
information
there.
B
So
so
I
definitely
need
people
from
mips
and
somebody
somebody
else.
I
think
ilumas
was
the
one,
but
that's
not
really
an
architecture
that
was
an
os
so
who
whoever
needs
to
integrate
and
get
things
in.
Should
please
take
a
look
at
it
tina.
Please
ask
your
folks
to
take
a
look
at
it.
B
A
J
Readout
is
unfortunately,
but
I
could
speak
to
that
if
you
want
to
join
absolutely,
please
do.
Thank
you.
Thank
you,
thoughts
and
conformance.
We
had
a
very
robust
meeting
in
this
week
in
the
office
hours
meeting
and
it's
really
taking
shape
around
the
profiles,
and
one
of
the
thoughts
is
to
make
it
go,
live
without
being
a
block,
but
more
of
a
data
collection
of
information
to
see
where
we're
going
in
the
next
releases
to
get
it
going.
J
I
think
I
summarize
this
probably
somebody
on
the
pool
that
can
summarize
it
better
than
I
do
so
that
is
for
that
and
then
on
api
snoop
on
conformance
coverage.
If
you
go
to
the
ad
guys
new
page
you're,
not
sharing
your.
J
J
We
did
very
well
for
this
quarter.
There
was
end
points
added
by
the
community
with
conformance
test
41
endpoints,
all
the
endpoints
that
got
the
conformance.
There
was
58
added
which
took
us
over
50.
J
We
have
177
endpoints
that
do
not
have
conformance
tests,
but
we
have
229
at
the
end
of
190,
so
it
was
really
a
very
good
effort
from
both
killing
all
endpoints
and
bringing
all
all
end
points
that
came
in
on
119
actually
came
in
with
tests.
So
it's
really
a
very
good
performance
by
everybody
and
then
on
the
conformance
gate
for
checking
automatic
checking
and
labeling
of
performance
requests.
J
We've
got
a
thanks
for
that,
john.
If
you
slide
it
up
to
the
top
you'll
see,
basically
the
dark,
green
you
see
in
119.
That
is
the
point
that
was
carried
over
from
118.
That
was
conformant
just
above
that
the
darker
green,
that
is
38,
end
points.
That
is
all
end
points
that
is
technical
depth,
that
received
their
test
in
testing
discord,
release
and
then
the
41
at
the
top
is
all
community
endpoints
from
ligit
and
from
project
team
that
brought
in
a
lot
of
endpoints
with
new
endpoints
from
beta
to
ga.
J
Then,
if
we
go
to
the
graph
just
below,
we
have
now
a
new
graph
that
shows
technical
debt,
which
is
basically
on
the
left
hand
side.
The
orange
that
you
see
is
all
endpoints
that
came
in
without
this
and,
as
you
see
from
115,
we've
really
been
good
as
a
community
to
not
bring
in
anything
without
this,
but
only
three
endpoints
that
sneaked
in
over
the
last
couple
of
releases
with
our
tests.
So
slowly
was
in
knocking
down
technical
debt
working
backwards
as
good
as
we
can.
J
Then,
let's
find
the
agenda
again.
Okay,
the
next
thing:
okay,
the
conformance
gate,
we're
doing
very
well
in
the
conformance
gate.
We've
got
an
instance
running
on
a
form
of
this
cncf
release.
You'll
see,
there's
a
cncf
infra
and
we
can
actually
run
prs
against
it
and
actually
there's
just
a
layout
of
the
process
flow,
and
it
actually
successfully
gives
you
feedback
on
conformance
of
your
pr.
J
It
will
give
you
some
feedback.
It's
an
australian
description
of
the
actions,
basically
taking
you
through
with
all
the
different
options,
if
you're,
if
you're
going
from
alpha
to
beta
or
into
alpha,
it
will
warn
you
that
you
do
need
performance
test
to
go
to
ga,
but
by
the
time
you
get
to
ga,
which
will
actually
block
you,
you'll
not
be
able
to
merge.
Unless
you
have
performance
tests,
the
idea
is-
and
we
would
like
to
get
some
community
feedback
on
that.
J
J
The
thought
we
have
is
not
be
the
actual
api's
new
page,
but
it
would
be
a
page
with
several
documents
explaining
based
on
performance
testing
requirements,
that
you
must
have
tests
for,
going
to
ga
the
documentation
for
conformance
data,
writing
all
those
type
of
information,
and
then
it
will
actually
have
some
hype
links
to
api
snip
itself
and
then
we'll
show
some
explanation
about
the
access
to
different
areas.
What
the
sunburst
does
list
of
endpoints
that's
per
release
and
basically
giving
the
community
a
good
overview.
J
A
All
right
well,
thank
you,
ryan,
any
any
questions,
or
should
we
move
on
to
to
the
enhancements
sub-project.
A
I
will
take
that
because
it's
time
to
move
on
all
right,
lori
you're
up.
D
All
right,
so
I
added
some
notes
to
the
agenda,
so
the
note
taker
can
take
a
little
break,
but
basically
just
the
major
points
I
wanted
to
cover
with
you.
So
there
are,
there
are
basically
two
main
topics
are
one
and
one.
A.
The
first
is
just
an
update
on
some
prioritization
efforts
that
we've
had
underway
in
sig
release
and
the
the
119
release
team
around
prioritizing
existing
backlog
items.
D
So
what
I
did
to
facilitate
that
was
collect
a
lot
of
existing
items
from
github
and
then
also
some
enhancements,
related
documentation,
links
and
then
shove
them
into
a
spreadsheet
and
then
had
about
two-thirds
of
the
release.
Team
members
actually
joined
small
breakout
sessions
where
they
prioritized
the
items
and
we
focused
on,
must
items
and
should
items
so
those
sessions
happened
in
june
and
were
across
the
whole
release
team.
D
So
all
the
sub
teams,
representatives
from
all
of
those
joined-
and
so
we
had
a
lot
of
this
healthy
discussion
in
these
one-hour
sessions
and
ran
through
everything
and
got
some
sense
of
what
the
group
thought
should
be
worked
on
most
immediately
and
what
was
holding
them
up
as
release
team
members
from
actually
managing
the
cycle
as
easily
as
they
could
and
then
in
july.
I
did
a
few
two
sessions
with
roughly
the
same
items
in
a
spreadsheet
with
jorge
stephen
augustus,
tim
pepper
and
sasha,
and
we
they.
D
They
also
gave
their
ideas
around
what
is
high
priority,
medium
and
low,
and
so
this
week
I
actually
took
the
notes
from
the
sessions
with
the
four
I
was.
The
release
leads
and
the
previous
set
of
notes
from
the
release
team
members
and
combined
them
into
the
single
spreadsheet,
which
is
in
the
agenda.
D
It's
the
prioritization
efforts
link.
So
if
you
click
on
that,
you
will
see
a
tab,
and
this
is
maybe
a
little
bit
frightening,
because
it's
rather
large,
but
the
idea
is
to
actually
get
rid
of
the
spreadsheet.
It
was
just
basically
a
dumping
ground
in
a
sandbox.
For
us
to
have
discussions
around
what
is
highest
value
high
prior
highest
priority
work
and
so
from
these
sessions.
You'll
note
that
there's
in
column
n,
what
I
did
to
create
these
totals
was
that
every
person
here,
steven,
tim,
jorge
and
sasha
had
a
vote.
D
So
one
would
be
least
important
and
four
would
be
most
important,
and
then
I
took
the
release
team
members
and
weighted
their
musts
as
fours
and
their
shoulds
as
threes.
So
anybody
who
had
spoken
up
during
those
sessions
and
said
we
must
do
this,
I
would
add
four
and
then,
if
they
should,
we
should
do
this.
It
should
be
three
and
then,
if
they
just
noted
anything
at
all,
but
weren't
really
specific.
That
would
be
a
one,
and
so
these
are
not.
You
know
scientific
and
these
kind.
D
This
kind
of
prioritization
strategy
has
a
lot
of
flaws,
but
I
I
just
want
to
bring
it
to
attention
like
this
isn't
a
perfect
system
yet,
but
what
it
does
do
is
it
actually
gives
us
some
ideas
about
what
people
think
is
most
important.
D
So
I
noted
four
items
in
the
agenda
on
this
and
these
will
be
familiar
to
you,
I'm
sure,
but
basically
enhancing
the
cups,
implementation
and
the
related
tooling.
So
kep
cuddle
or,
however,
we're
pronouncing
that
it
was
of
high
interest
and
you
will
find
in
the
spreadsheet
lots
of
contexts
and
comments
about
how
people
in
the
release
team
felt
about
that
tool
and
then
the
second
one
is
make
caps
transparent.
So
again,
this
has
a
lot
of
comments
that
were
generated
during
those
breakouts
they're
all
dumped
in
that
spreadsheet.
D
D
And
I
don't
know
if
you
talked
about
that
at
the
last
enhancements
meeting,
because
I
wasn't
there
for
that.
But
it
was
in
july
7th
and
the
link
here
is
the
notes
from
their
findings.
So
I
have
a
question
here.
If
we
could
perhaps
review
their
work
at
the
next
meeting
next
week,.
A
Well,
yeah:
let's
definitely
review
it
there,
okay
and
then
bubble
it
up
here.
Oh
cool.
D
Cool
and
then
two
other
items
I
just
want
to
call
out
because
they
were
heavily
weighted
in
these
prioritization
sessions,
is
tracking
enhancement,
slip
across
multiple
releases
and
then,
finally,
just
drilling
down
further
into
the
criterion
process
for
alpha
beta
ga
and
a
lot
of
these
items
seem
to
have
been
parked
for
a
while.
I
mean
there's
a
lot
going
on,
so
it's
completely
understandable,
but
this
means
that
we
might
have
to
review
and
update
some
of
the
items
because
there
has
been
work
done
on
some
of
them.
D
So
and
also
maybe
scope
has
changed,
and
also
we
have
found
out
new
information
during
119.
That
would
cause
us
to
adjust
the
original
goals
and
that
gets
into
the
topic
of
the
project
board,
which
I'm
wondering,
if
there's
a
way
to
set
up
any
sort
of
regular
process
like
walking
the
board
so
just
track.
D
How
we
make
progress
on
the
priorities,
and
I
want
to
call
out
one
thing
that
would
actually
help
us
would
be
to
if
we
have
some
of
these
big
items,
breaking
them
down
into
smaller
bits
of
value,
which
would
we
be
great
if
these
product
managers
that
want
to
help
out
would
actually
help
us
do
that
using
their
expertise
and
also
thinking
about
what
contributors
might
be
able
to
help
with,
and
then
also
applying
triage
principles
to
this
incoming
work,
because
we'll
always
have
new
things
to
talk
about.
A
You
no,
I
think
there
has
been
like
let's
okay,
this
is
great,
so
we're
getting
a
lot
more
organized
in
the
enhancement
subproject.
Thank
you
laurie
for
all
this
super
helpful.
Okay,
and
you
know
you
can
see,
we
have
a
lot
a
lot
to
do
so.
Anybody
who's
on
this
call
and
wants
to
join
us
next
week
for
our
enhancement,
subproject,
call
and
help.
You
know
some
of
the
concrete
things
are,
as
you
have
in
here,
like
this
kept
control
kept
cuddle.
Whatever
you
want
to
call
it
is.
A
It
is
a
cli
tool
to
you
know,
discover
caps
and
also
to
create
new
caps
that
sort
of
have
the
right
metadata
set
up
and
and
promote
them
from
you
know,
when
you're
doing
a
promotion
from
alpha
beta
it'll
be
a
tool
that
does
that
and
it
can
sort
of
validate
that
you've
met
certain
gates
and
things
like
that.
So,
if
you're
interested
in
helping
out
with
that,
we
definitely
can
use
the
help.
It's
the
enhancement.
A
Some
project
is
new,
and
so
we
just
have
a
few
people
involved
and
could
really
could
use
more
eyes
on
it
and
more
fingers
on
keyboard
as
well
so
awesome.
Thank
you.
Any
questions
about
this
stuff.
D
I
put
as
a
note
like
it
would
be
great
to
have
subject
matter
experts
to
help
review
some
of
this
text.
Like
again,
I'm
going
to
say,
break
things
off
like
the
big
spreadsheet
into
smaller,
manageable
bits.
So,
if
you
look
in
that
spreadsheet
you'll
see
things
like
column,
a
which
has
lots
of
colors
or
like
the
main
themes,
so
whether
it's
enhancements
tracking
or
the
alpha
beta
ga
process
or
the
templates
anything
like
pulling
one
of
them.
D
A
People
doing
kubernetes
development
so
we're
all
in
some
sense
subject
matter
experts,
but
it's
not
like
you
know
a
subject
matter:
expert
in
storage
or
something
I
think
you're.
Probably
the
such
encounter
expert
for
us
here
in
program
management
so
definitely
helpful.
Yes,
please
anybody
who
has
opinions
and
ideas
join
in.
B
It's
a
quick
question,
john
for
you
and
tim
thanks
laurie.
So
we
had.
I
guess
you
you
you,
you
were
doing
a
book
club
of
caps.
Is
it
still
on
or
did
it
get.
A
What
we
did
in
tim,
you
could
jump
in
if
you
disagree
with
what
happens,
we
met
a
few
times
and
we
found
that
it
was
extremely
difficult
to
discover
caps
that
needed
attention.
A
So
you
know
some
of
the
goals
of
that
were
like,
let's
find
caps
that
maybe
are
really
interesting,
good
features,
but
are
stuck
because
of
whatever
reasons,
or
sometimes
there
are
caps
that
it's
like.
You
know
this
sounds
interesting,
but
but
there's
actually
this
subtle
thing.
That's
going
to
cause.
A
You
know
a
lot
of
performance
degradation
like
the
the
so
we're
sort
of
we're
trying
to
help
triage,
but
not
just
triage
sort
of
typical
incoming
kept,
but
the
things
that
were
getting
caught
in
the
cracks,
but
it
was
so
hard
to
discover
things
that
I
think
we
just
kind
of
stalled,
and
so
one
of
the
things
we're
trying
to
do
is
with
the
tooling
here
is
to
improve
the
discoverability
that
that,
for
me,
is
a
big
thing.
I
don't
know
if
tim
has
anything
to
add
to
that.
H
Yeah
we
also,
I
mean
we
did
some
work
to
make
the
caps
better
right.
We
broke
them
into
directories
and
we
broke
the
metadata
out
into
a
more
machine
parcel
format,
ostensibly
so
that
we
could
have
better
tooling.
H
We
got
the
sort
of
unresolved
annotation
clarified,
which
I
think
are
all
net
good
things.
I
don't
know
how
much
people
are
using
them.
I
use
them,
but
those
were
helpful
in
figuring
out
what
has
and
hasn't
been
resolved
the
meeting
sort
of
stalled
out
right
around
the
time.
Everybody
went
into
quarantine
and
we
haven't
really
reinstated
it
since
then.
A
Speaking
of
the
unresolved,
the
the
so
there
is
a
there's,
a
pr
out
there
for
the
query
function
in
the
kep
control
and
I
didn't
add,
looking
for
things
with
unresolved,
but
that's
actually
would
be
really
useful.
I
think
to
examine
the
cats
and
say:
oh,
this
one's
got
unresolved
segments
because
one
of
the
hard
things
about
caps
is:
we
want
to
merge
them
quickly,
but
then
you
lose
the
ability
to
comment
on
them
right.
A
B
So
yeah,
I'm
asking
because
me
and
andrew
are
probably
interested
in
you
know
getting
into
into
this
cycle
of
reviewing
as
well.
That
that's
why
I
was
asking.
A
H
Yeah
I'd
love
to
put
back
together.
I
thought
it
was
super
useful
for
being
exposed
to
caps
that
I
might
not
otherwise
have
clicked
on
and
for
getting
different
people's
perspectives
on
it
and,
in
fact,
through
those
discussions
I
had
my
mind
changed
on
quite
a
number
of
things,
so
it
was
a
useful
process.
H
K
Okay,
so
I
know
this
meeting
sometimes
ends
early.
Could
we
like
just
use
like
extra
time
in
this
meeting
to
just
go
over
whatever
caps
folks
want
to
get
reviewed
here.
H
So
the
problem
with
that
I
mean
it
sounds
good.
The
problem
with
it
is
you
have
to
do
homework.
You
can't
read
the
cap
in
real
time.
You
have
to
have
like
called
it
out
a
few
days
before
made
time
for
yourself
to
go
through
and
read
it.
Some
of
these
caps
were,
you
know,
eight
ten
pages,
long
for
the
complicated
ones
you
have
to
think
about
it
and
and
then
you
can
come
and
have
a
reasonable
discussion
about
it.
H
A
Not
that
we
want
another
meeting
andrew,
but
it's
an
interesting
meeting.
H
I
I'm
willing
to
give
up
an
hour
every
week
or
every
other
week
for
this,
because
I
think
it
is
important
and
useful
yeah
absolutely.
A
A
A
Criteria
to
use
for
discovery
type,
you
know
you
know
at
some
point:
it's
you
got.
A
Just
decide
whether
they're
interesting,
but
but
if
there's
any
filtering
we
can
do.
I
think
it's
helpful.
D
D
It
in
there
was
a
list
of
notes
that
jeremy
and
bob
had
created
around.
It
was
called
the
cuddle
notes,
and
then
I
drafted
a
bunch
of
user
stories
for
it.
So
right
now
the
doc
is
really
long,
because
it's
it
needs
review
and
it
needs
to
be
cut
down.
So,
like
you
know,
how
would
we
want
pep
cuddle
to
work
and
for
which
users
and
what?
D
What
is
most
important
to
do?
First,
I
just.
A
A
Okay,
in
that
case,
thank
you,
everybody
and
we'll
see
you
in
two
weeks
you.