►
From YouTube: [OCI-WG] Reference Types - 2022-02-15
B
D
E
A
No
worries
just
been
a
lot
been
a
lot
of
zoom
bombs
over
the
last
few
years.
Josh.
A
A
A
Do
it
do
it?
Do
it,
okay,
hello
and
welcome
to
the
oci
reference
types
working
group
today
is
tuesday
february
15,
2022.,
hello
and
welcome,
and
I've
just
placed
a
link
to
the
google
doc
where
we're
logging
attendance.
If
you
could
add
your
name
there,
we
very
much
appreciated,
and
then
we
will
ask
for
the
honorary
position
of.
Is
there
anybody
who
would
like
to
be
note
taker
for
this
meeting.
F
A
Take
notes,
okay,
anybody
else
back
up,
just
because
I
see
you
as
a
having
a
few
agenda
items
here
nisha,
so
I
think
probably
just
separation
of
concerns
might
be
good,
so
you
don't
have
to
represent
and
type
the
notes.
At
the
same
time,.
A
Okay,
please
place
your
name
in
the
note
taker
place
and
if
there
are
any
silver,
medalists
and
bronze
medalists,
who
can
help
just
just
to
add
any
backup.
I
know.
Last
week
my
internet
died
mid
mid
meeting
and
I
was
note
taker,
so
I
appreciate
brandon.
I
believe
it
was
who
stepped
up
and
took
some
very
good
notes.
A
I
feel,
like
you,
were
one
of
those
legal
legal
clerks
in
the
courtroom
that
typed
the
whole
transcript
out
because
you're
a
very,
very
good
note
taker
he's
got
the
books
behind
him
to
make
it
look
like
it
yeah
yeah
and
he
actually
he's
he's
on
a
typewriter.
So
he's
he's
a
g-doc
enabled
typewriter.
A
Okay,
so
excellent.
Is
there
anything
before
we
go
into
the
agenda
item?
So
if
you
have
anything
to
add,
please
add
it
to
the
agenda:
there's
a
room
there
and
if
we
can
get
to
it,
we
will
in
the
time
that
we
have
today
see
what
just
a
little
decorum
we're
using
hands
up
so
raising
your
hands,
either
physically
or
in
software.
A
To
indicate
that
you
have
a
question
and
we'll
try
and
speak
to
the
right
order.
That
the
hands
were
raised,
so
in
that
spirit
I
see
that
josh
or
smiley
face
colin
smiley
face
emoji,
not
sure
if
you're
still
going
by
josh.
So.
E
Hey
you
can
call
me,
like
I
my
hand,
did
not
mean
to
be
raised
there.
I
was
offering
to
take
notes.
Misha.
I
didn't
know
what
she
meant
by
that.
If
you
want
to
drive
the
user
story,
talk
we
can
you
can
do
that.
I
mean
I
could
take
notes
while
you're
talking.
F
E
E
So,
a
few
weeks
ago
we
went
through
the
user
stories
and
everyone
was
providing
them
and
I
think,
end
of
last
week
the
the
thing
on
the
agenda
that
we
talked
about
is:
is
there
a
way
for
us
to
capture
these
and
distill
them,
and
jason
was,
I
think,
two
weeks
ago
we
were
going
through
these
and
making
notes
and
de-duplicating
things.
So
I'm
trying
to
recall
where
we
left
off.
I
think.
E
We
made
it
to
here,
okay,
all
right,
I
guess,
and
if
someone
doesn't
have
a
better
idea,
we
can
just
walk
through
each
of
these,
and
I
don't
know,
jason
was
doing
a
really
great
job.
So.
B
D
Yeah
last
week
we
only
really
talked
about
the
governance
process
and
upgrading
and
sharing
discussion
topics.
But
from
two
weeks
ago
we
sort
of
went
through
these
and
tried
to
cluster
them,
and
josh
did
a
great
job
of
clustering
for
this
one.
I
think
it
is
worth
mentioning
that
this
is
a.
D
This
is
more
of
a
cosmetic
issue
that
a
ui
can
paper
over,
not
necessarily
that
we
must
put
that
work
onto
the
ui,
because
there
are
many
uis
and-
and
you
know,
the
ui
resources
are
constrained
basically
everywhere,
but
being
able
to
filter
out
tags
is
possible
to
solve
by
a
ui
or
a
cli
that
filters
or
see
or.
A
D
That
filters
by
default
tags
that
match
this
format,
ideally
we
wouldn't
have
to,
but
I
don't
want
to
constrain
our
solution
space
by
overly
specifying
our
problem
space.
G
I
mean
you,
you
could
absolutely
put
the
filter,
I
mean
you,
can
it's
very
easy
to
use
the
lipstick
conversation
on
that?
I
think
the
this
is
just
one
of
the
places
that
you
want
to
be
able,
the
the
pattern
kind
of
surfaces,
whether
it's
I
want
to
manage
life
cycle,
others
so
yeah.
This
is
just
teasing
out
that
one
particular
item
that
yes,
if
this
was
the
only
last
thing
you
could
fix
this
in
the
ux.
G
I
think
you
have
to
kind
of
also
think
about
the
scalability
of
this.
Also,
this
is
one
of
the
things
that
I've
just
been
seeing
has
struggled
a
little
bit
on
it's
one
thing
when
the
tag
was
used
just
to
represent
a
signature
when
we're
trying
to
represent
attestations
and
security,
scan
results
and
adherence
to
policies
and
a
whole
bunch
of
other
things
that
I
think
it
stretches
the
the
challenge.
It
stretches
the
issue.
A
I
was
just
going
to
say
josh
just
yeah
for
everything,
that's
highlighted,
maybe
just
scroll
to
the
right
and
see
what
we
had
there.
I
think
nisha's
comment
was
very
similar
to
to
jason's.
D
Yeah,
I
think
I
mean
like
to
to
abstract
away
even
that
abstraction.
We're
talking
about
filtering
and
default
filtering
right,
like
is
it
possible
to
hit
then
api
endpoint
that
returns
only
you
know,
quote
unquote
actual
tags,
I'm
making
finger
quotes,
but
you
can't
see
me
and
then
a
you
know,
url
parameter.
That
says
no
show
me
everything
equals
true,
which
shows
you
all
of
the
other
stuff.
D
G
There
are
other
things
where
the
tag
is
meant
to
represent
a
specific
thing,
so
I'm
trying
to
figure
out
how
to
balance
the
simplicity
of
yeah,
let's
put
a
filter
on
the
tag
of
regex
on
the
tag
as
opposed
to
there
is
a
general
pattern
for
how
tags
are
represented
and
used
across
the
ecosystem,
and
I
don't
want
to
hang
up
by
just
this
one.
I'm
just
this
is
one
of
those
that
gets
interesting
on.
A
E
A
I'd
accept
I'd,
I'd
propose,
but
for
now,
let's
just
collate
and
go
through
quickly
and
are
there
any
questions
about
what
this
use
case
is
saying
that
steve
has
sounds
like
everybody
understands
what
it
is
and
then
you
know
if
brandon
has
one,
we
collapse
them
into
one
and
then
later
we
make
a
decision
as
to
how
we're
going
to
treat
it.
That
would
be
my
suggestion.
E
E
I
I
guess
like
okay,
let
me
let
me
try
to
like
better
frame
what
I'm
saying.
I
think
we
have
like
these
several
stories
right
and
how
do
we
know
before
we
have
the
boxes
or
the
buckets,
what
we're
going
to
do
with
them,
or
are
we
going
to
throw
them
in
a
trash
bucket,
like
I
think,
do
we
want
to
spend
time
creating
the
buckets
for
the
requirements
versus
going
through
them,
or
I
just
feel
awkward
that
I'm
I'm
sharing
my
screen
and
I'm
not
sure
what
the
what
the
goal
is.
F
Asia
yeah,
so
I
would
actually
agree
with
lucky
over
here.
It
might
be
a
we'd
go
we'd,
be
able
to
make
a
decision
on
whether
we
want.
Well,
I
mean
it's
a
process
thing,
so
I
think
going
through
them
gives
us
gives
us
all
a
chance
to
think
about
some
larger
patterns.
F
So,
for
example,
this
after
steve
had
explained
about
what
he
meant
by
like
a
selection.
We
all
decided,
okay,
that
means
filtering
and
that
could
easily
fall
under
like
a
category
of
management
or
image
management
or
reference
management
or
tag
management,
but
in
general
that
can
fall
under
management.
So
we
can
only
get
that
if
we
go
through
all
of
the
user
stories
quickly,.
E
A
All
right,
I
was
just
going
to
say:
josh,
it
might
be
worth
just
prefixing.
Everyone
we've
looked
at
with
the
general
theme,
so
we
can
make
collation.
So
if
this
is
like
query
filtering
or
querying,
you
know
at
least
then,
when
we
come
back
to
it,
we
can
start
duping
and
then
bucketing,
because
I,
like
the
buckets
you've
done
josh,
but
I
think
let's
synthesize
a
list
that
we
all
understand
and
then
we
can
start
to
pick
and
choose
which
ones
are
in
it,
which
group
that
would
be
my
suggestion,
josh.
E
That
sounds
good
I'll
I'll
continue
going
that
route.
If
anyone
else
wants
to
like
add
notes
as
we're
going
through
each
item
go
ahead,
but
I'll
try
to
capture
this.
Like
theme
idea:
okay,
steve
lasker.
As
a
registry
operator,
it
wants
to
help
users
understand
how
they
can
manage
lifecycle
of
their
container
images
and
their
references.
B
So
this
one,
I
guess,
steve
you're,
going
to
not
only
that
a
a
reference
could
be
deleted,
but
that
it
has
been
deleted
in
a
certain
situation.
G
G
If
I
want
to
archive
that,
I
can,
but
I
want
to
make
sure
that
the
unit
that
I
deal
with
it's
kind
of
like
if
I
deleted
files
from
my
operating
system,
I
don't
want
the
operating
system
to
keep
track
in
some
other
metabase
of
which
are
read-only
attributes,
which
were
you
know,
various
policies
assigned
to
those
files
that
that
should
go
as
a
unit.
B
G
G
And
I
I
realized
this
from
the
previous
conversation
of
like.
Let's
I
think
I
was
jason.
I
started
debating
different
ways
to
solve
it
and
rather
than
try
to
solve
it,
do
we
understand
the
problem?
So
here
it's
just
a
matter
of.
Does
there
has
been
discussions
around
which
should
be
client
problems
versus?
Can
a
registry
know
about
it?
G
So
the
pro
the
concept
here
is
registries,
don't
need
to
know
about
the
specific
types,
there's
a
generalized
pattern
that
it
can
see
relationships
and
if
users
delete
the
root,
then
the
dangling
relationships
have
the
option
to
go
with
it.
So
it's
really
around
life
cycle
management,
on
how
the
the
registry
can
understand
a
graph
without
knowing
the
specific
type
so
yeah
we've
done
things
around
the
artifact
type
and
the
subject
and
so
forth
that
enable
it,
but
I'm
not
trying
to
jump
to
solutions.
B
Yeah
and
I'm
not
trying
to
solve
it
here
either
and
sorry
chase.
I
see
you
put
your
hand
up
and
I
should
have
done
the
same
myself
always
go
ahead.
Go
ahead,
yeah,
so
I'm
not
looking
to
solve
it.
I
just
want
to
make
sure
we
clarify
that
when
we
come
out
with
our
solutions,
I
don't
think
any
solution
is
going
to
solve
this
because
nothing
on
oci
does
life
cycle.
But
as
long
as
we
give
you
the
tools
to
do
it
yourself,
I
think
we
might
have
solved
this
for
your
use
case.
A
A
D
I
agree
with
the
theme:
are
we
just
going
through
and
assigning
themes
to
everything
as
in
a
lightning
round
or
there's
no
discussion?
I'm
fine
with
that
too.
I
have
discussion.
If
we
have
discussion
on
the
table.
A
E
G
G
G
C
Yeah,
I'm
not
sure
how
this
the
theme
aligns
with
the
reference
type
working
group,
because
we
seem
to
be
discussing
about
tags.
I
I
would
say
we
should
probably
like
call
out
how
this
feature
would
be
used
by
references
rather
than
limiting
or
like
defining.
C
F
It
seems
to
me
that
the
specific
use
case
you're
talking
about
is
immutable
tags,
making
attack,
making
a
tag
immutable
or
not,
and
that
sounds
like
content
management
to
me
or
life
cycle
management
to
me,
but
that's
whether
whether
we
should
you
know
think
about
it
when
we're
making
proposals
it's
another
matter.
So,
let's
just
bucket
it
and
move
forward.
E
G
E
That's
a
great
that's
a
great
that
those
are
that's
a
great
combination
of
words.
E
E
Disagree:
okay,
oh.
F
I'm
going
I'm
going
back
and
forth
as
well.
Sorry,
so
I
I
failed
to
understand
why
we're
referencing,
why
we're
talking
about
layers
here?
G
The
concept
is,
as
should
reference
types,
be
limited
to
not
support,
multiple
layers,
which
is
what
the
image
spec
does
today.
That's
one
of
the
great
flexibilities
of
a
registry,
so
this
is
more
of
just
making
sure
that
we
are
capturing
that
we
don't
constrain
reference
types
beyond
what
the
schemas
already
support.
F
Okay,
so
does
that
mean
like
backwards
compatibility
or
sounds
like
sounds
like
it?
Maybe.
F
Or
is
it
like?
I
don't
know,
that's
how
that's
how
I'm
seeing
it
is
like
it's
a
backwards,
maintaining
backwards
compatibility.
F
So
there
are
many
situations
where
the
the
blobs
and
the
layers
are
not
necessarily
being
used,
but
separate
manifests
like
point
to
other
manifest.
I
don't
know
like
some
people
may
be
using
blobs
and
some
people
may
not
be
using
blobs.
Is
it
something?
Is.
G
And
maybe
ask
the
question
after
the
second
the
next
one
that's
highlighted,
and
this
is
kind
of
the
opposite
that
says:
hey
I
just
want
to
push
an
annotation.
We
can
call
it
an
attestation,
there's
a
bunch
of
debate
around
whether
attestations
claims
and
so
forth,
but
the
idea
is,
I
want
to
be
able
to
push
a
name
value
pair
to
the
registry
as
a
reference
to
the
other
thing,
with
no
additional
blob
data,
because
the
annotation
can
carry
enough
information
and
that
is
also
a
reference
type.
G
This
is
something
that
the
image
spec
doesn't
support.
Today,
we've
done
an
era's
artifact
spec.
G
So
one
is
the
niches
point
this,
the
the
one
above
it
was
hey
today
manifest
support,
multiple
blobs
and
I
can
use
it
for
deduping.
I
can
use
it
for
parallel
downloads.
I
can
use
it,
for
I
only
need
one
of
the
blobs
and
not
the
other,
depending
on
the
scenario.
The
the
next
one
says
I
don't
need
blobs
at
all.
The
annotations
is
enough.
A
E
D
Brandon
yeah
from
from
before
I
mean
from
two
weeks
ago
or
whatever,
when
we
started
writing
this
doc,
I
just
took
a
pass
through
and
tried
to
dedupe
or
cluster
them
as
dupes
as
well
as
I
could.
This
is
not
an
authoritative,
dupe
signal
if
other
people
disagree.
B
G
G
That's
really
kind
of
what
this
one's
about
sure
people
can
spoof
things
anything
they
want.
I
can
put
random
stuff
in
a
jpeg,
you
know,
but
is
there
a
pregate
that
and
that
then,
that
same
thing
is
used
for
other
areas
also,
so
it's
just
an
example
of
one
of
them.
E
So
I
think
I
I
bolded
are
not
affected
from
brandon's
thing
which
we'll
bring
up
later,
but
I
think
maybe
the
way
to
look
at
this
is
these
changes
could
have
negative
effects
and
one
of
them
is
security.
E
G
G
Maybe
it's
is
there
infrastructure
that
allows
me
to
do
the
right
thing
and
not
overly
put
burden
on
run
times
or
any
artifact
type
like
the
helm,
the
helm
seal?
I
shouldn't
be
confused
when
it
tries
to
pull.
A
I'm
starting
to
play
the
the
walk
off
the
stage
music,
steve
sanjay
has
his
hand
up.
C
B
E
Then,
okay.
E
G
E
Okay,
as
a
security
scanner
product,
I
want
to
know
the
type
of
each
artifact,
so
I
know
how
to
scan
it
for
vulnerabilities
and
then
another
dupe
as
a
user.
I
would
like
to
know
the
type
of
of
each
artifact
without
pulling
each
artifact
separately
is.
A
A
A
E
I'm
wondering
if
someone
can
volunteer
to
start
like
simultaneously
writing
this
portion
and
and
pulling
things
that
we
decide
on
into
these,
so
that
you
know
in
the
next
call-ins
and
25
minutes.
Maybe
we
can
josh.
A
F
Okay,
fine
yeah,
keep
team
as
filtering.
It's
fine.
E
Okay,
next
nisha,
as
a
user,
I
would
like
to
be
able
to
map
monotonically
increasing
product
versions
to
container
images.
So
I
have
an
idea
of
deployment
progression.
F
Okay,
then
I
would
say
this
is
content
management.
E
As
a
developer,
slash
devopser,
I
like
that
word-
I
would
like
to
buy
sec,
builds
based
on
container
images.
I
have
deployed.
B
F
B
E
Okay,
this
one
was
me
as
a
user.
I
want
to
store
images
in
one
registries
and
signatures
s-bombs
attestations
in
a
separate
registry.
G
E
Yes,
but
I
guess
like
this,
isn't
anything
to
do
with
management
right.
This
is
just
I
want
to
upload
one
thing
here
and
one
thing
here
and
make
sure
I
can
link
them
so.
C
If
we've
spoken
about
this,
like
detached
references,
if
you
want
to
think
about
it,
like
you
have
in
the
manifest,
you
can
actually
have
blobs
coming
from
different
places.
So
I
think
the
schema
can
take
a
requirement
saying
that
you
want
to
be
able
to
support
a
schema
that
can
store
images
here.
Team
seems
okay,
if
that's
that's
good
enough
to
kind
of
like
start
with
and
roll
it
into
the
scheme
as
a
requirement
for
the
scheme
to
support
these
things.
E
Okay,
that
sounds
good
to
me,
brandon
and
by
the
way,
if
we're
running
into
brandon's
dupes,
then
please
someone
stop
me
as
an
artifact
producer.
I
want
to
be
able
to
create
more
multiple
artifacts,
referring
to
the
same
manifest
and
upload
them
separately.
E
E
E
From
the
themes
that
we've
already
said,
which
I
think
is
for
right
now
or
three
four
backwards
combat,
I
would
say
management
unless
there's
a
new
theme
for
that.
G
E
A
I
was
just
gonna
say
the
only
thing
to
call
out
there
that
is
there
any
impetus
on
the
container
image
itself
existing
first
or
do
you
want
to
push
all
everything
at
the
same
time
and
establish
references
afterwards?
Let's
say
you
land
a
signature
and
then
you
land
a
container
image,
and
then
you
link
them.
So
can
a
can
a
signature,
not
reference
a
container
image
and
exist
in
the
registry,
because
what
I
was
just
wondering
about
that
in
the
multiple
artifacts
concurrently.
B
My
thought
behind
that
is
probably
not
probably,
the
image
will
be
up
there
first,
but
someone
could
always
stretch
this
different
way.
A
D
I
have
a,
I
have
a
use
case
for
wanting
to
push
a
signature
before
the
image
that
it
is
signing
and
that's
that
I
want
to
enforce
a
policy,
perhaps
that
my
registry
disallows,
unsigned
images,
and
so
the
only
way
to
achieve
that
is
to
push
the
signature
and
then
push
the
image.
There
is
no
time
at
which
the
image
exists
before
the
signature.
I
think
the
way
the
way
that
we
solve
this
is
the
signature
points
to
the
image
that
is
signed
by
digest.
D
There's
no
validation
in
the
registry
that
that
digest
exists.
Just
that,
like
I
am
a
signature
for
digest,
abc
and
then,
when
image
abc
exists,
the
world
is
right
and
everyone's
happy,
but
that
doesn't
require
me
to
push
a
signature
free
floating.
You
know,
sort
of
like
unreferenced,
unreferencing
signature
and
then
the
thing
is
it
references
and
then
update
it
to
point
to
the
thing
we
just
say:
I
sign
digest:
abc
abc
doesn't
have
to
exist
for
me
to
exist
and
then,
when
abc
exists,
everything's
happy.
D
G
I
think
I
think
brandon
there's
two
parts,
one
just
to
go
in
reverse
order.
I
think
there's
another
way
to
approach
that
is
there's
this
quarantine
pattern
thing,
so
I
think
capturing
the
root
requirement
of
we're
trying
to
make
sure
there's
no
unsigned
artifacts
in
a
registry
is
a
good
category,
because
there
are
there's
other
ways
to
think
about
just
like
with
most
of
these,
but
the
other
one.
I
think,
brandon.
G
What
you're
trying
to
capture
was
again
kind
of
just
making
sure
that
solutions
around
doing
an
update
to
a
an
index
as
multiple
things
are
being
pushed
that
there's
this
one
index
that
gets
updated
to
capture
all
of
those,
and
that's
where
we
had
the
concurrency
locking
issue.
So
I
think
that's
what
you
were
really
trying
to
poke
at
is
that
is
that
accurate.
G
E
B
E
All
right,
brandon,
you
wanna,
go
ahead
again.
B
Yeah,
as
an
artifact
producer,
I
wanna
be
able
to
push
an
artifact
to
a
registry
with
a
tag,
so
they
can
be
pulled
directly
in
addition
to
having
references
to
other
manifests
in
the
registry.
So
this
is
the
inverse
of
above.
So
I
don't
know
if
that's
just
backward
compatibility
so
that
we
can
tag
it
if
we
want
to,
or
does
that
just
fall
under.
A
D
B
B
B
A
B
A
F
A
B
And
it
joshua's
right
here,
I
think
we
might
be
able
to
get
them
even
faster,
because
I'm
agreeing
with
a
bunch
of
them
as
a
tool
writer,
I'm
looking
to
be
able
to
query
these
things
efficiently
and
picking
specific
artifacts
by
type.
I
think
that
one's
probably
a
duplicate
from
before,
let's
see
being
able
to
identify
when
a
registry
doesn't
support
the
apis,
that
I'm
going
to
call
backward
compatibility.
B
Yep
no
problem
as
a
tool
writer,
I
want
to
be
able
to
include
reference
types
within
the
image
layout
file
system.
So
this
one,
this
one
probably
breaks
everybody's
idea
of
everything
you're
doing.
If
I
push
something
that
has
reference
types
to
the
file
system
format
that
oci
has
for
the
image
layout,
how
do
I
push
the
reference
types
with
it,
content,
management
or
backward
compatibility?
B
There
is
a
file
system
format
for
pushing
images
to
the
file
system
and
storing
it
same
format.
That's
used
by
various
other
oci
compatible
tools
and
all
the
registry
api
stuff
we're
thinking
of
is
api
specs
for
querying
registries.
How
do
you
take
that
same
logic
and
put
it
into
that
file
system
format.
B
B
Basically
make
sure
the
new
apis
don't
break
the
existing
features
we've
already
got
and
as
a
user,
I
would
like
to
be
able
to
walk
the
cash
tree
and
reverse
this
one's
getting
a
little
crazy.
So,
basically,
I
would
like
to
be
able
to
query
all
manifest
point
to
a
blob
digest
or
all
manifest
lists
that
are
pointing
to
a
manifest
digest
effectively.
B
This
takes
whatever
we're
thinking
of
for
the
reference
types
and
pumps
it
up
a
notch,
so
I
don't
know
if
we
want
to
even
be
able
to
do
that,
but
I'm
throwing
it
out
there
in
case
people
agree,
and
then
that
would
be
awesome
if
we
do
if
this
gets
lost
on
the
cutting
board
later
on.
That's
fine,
too.
B
A
B
All
right,
I'm
going
to
keep
reading
and
I'll
go
back,
and
I
might
retheme
a
couple
of
these
when
we
jump
to
niches
here
in
a
minute.
So,
as
a
registry
operator,
I
want
attempts
to
push
an
artifact
to
an
older
registry
to
have
minimal
impact.
This
was
a
dupe
of
one
of
steve's
earlier,
so
I
just
don't
want
to
break
garbage
collection.
B
F
Yeah,
so
these
two,
these
three
were
after
I
spoke
to
the
tough
community
about
what
they
would
like
to
see
so
assurances
that
the
any
oci
object
really
did
come
from
a
supplier,
the
claim
supplier.
So
this
may
be
artifact
schema,
but
it
could
also
be
under
content
management.
I
don't
know,
then
the
second
one
is
as
a.
F
F
E
Yeah
so
lockey,
you
said
you
want
to
run
through
this
list.
Do
we
want
to
give
people
time
to
look
at
this?
Brandon
said
he
wants
to
maybe
change
some
but
yeah,
and
also
we
should
consider
so
once
we
do
have
these
in
the
buckets
it
doesn't
mean
that
we
all
agree
that
that's
the
requirements
right.
So
how
should
we
go
about
for
each
of
these
like
gaining
consensus.
B
Can
I
throw
an
id
out?
I
didn't
raise
my
hand,
but
my
leaning
is
to
not
try
to
gain
consensus
at
this
point,
but
just
to
try
to
capture
we've
got
a
whole
bunch
of
different
competing
interests.
People
are
going
to
want
to
go
different
directions,
and
so
we've
captured
a
lot
of
our
ideas.
B
Let's
go
ahead
and
look
at
potential
solutions
after
we
do
any
filtering
reformatting
collating
of
what
we've
already
put
down
and
then
start
looking
at
how
what
we're
proposing
solves
these
different
questions,
you
know
build
a
grid
or
whatever
we
need
to
do
to
start
putting
that
together,
because
different
people
are
going
to
put
different
emphasis
on
different
id
on
different
items
here
so
rather
than
us
trying
to
solve.
What's
the
most
important
item
across
all
these
things
now
and
getting
consensus,
there
get
consensus
later
once
we
once
we.
E
I
think
we
can
be
very
strategic
with
this
language
in
the
repo
that's
like
these
were
provided
by
the
members
of
the
working
group.
It
does
not
mean
that
all
of
these
will
be
like
addressed
by
the
final
output,
but
that
this
is
a
reflection
of
what
we
all
wanted.
So
I
think,
if
we
bucket
these,
I
don't
know.
If
we
should,
we
keep
the
name
attached
to
them,
and
then
I
have
a
pr
that
I
can
close
and
we
can
replace
with
the
bucketing.
B
Was
gonna
say
maybe
a
real
fast
way
to
go
through
this
is
to
say:
are
there
any?
Are
there
any
things
on
this
list
that
we've
gone
through,
that
people
are
really
opposed
to
being
included
as
a
potential
thing,
we're
comparing
to
our
solutions
against
not
saying
a
solution
has
to
solve
it.
Just
saying
that
we
want
to
compare
it
against
them.
A
A
A
Is
it
to
identify
any
it
doesn't
sound
like
there's
any
contentious
ones
at
the
moment,
but
I
don't
know
if
that's
going
to
change
or
do
we
go
through
each
of
the
ones
that
are
in
that
finalized
list
and
then
start
to
you
know,
add
some
more
detail
to
exactly
what
the
use
case
is
or
how
do
we
want
to
break
that
up
or
look
at
use
these
as
a
way
to
rationalize
the
the
existing
proposals
that
are
out
there
and
actually
say
you
know,
as
you
said,
brandon
here's
a
matrix.
A
A
F
Okay,
so
what
about
the
ice
cream,
people.
F
Okay,
I'll
wait:
we'll
wait
until
there
is
a
conflict
and
then
we'll
convert
it
into
ice
cream
people.
A
Okay,
so
before
the
next
meeting
and
I'll
I'll
do
this
right
now,
josh's
bucket
them
all.
Let
you
know
in
chat
and
then,
if
everybody's
in
agreement
I'll
give
everybody
till
the
end
of
the
week
and
then
I'll
pr
them
up
through
your
pr
josh,
and
then
we
can
start
to
look
at
a
grid
and
rationalize
the
proposals
that
currently
are
there
as
to
whether
they
do
or
do
not
meet
these
requirements.
That
we've
created
sound
good
and
then
we,
nickel
and
dime
over
the
things
that
we
need
to
and
then
leave
everything
else.
A
Coolio
I'll
see
you
all
in
the
chat.
Thank
you
very
much
and
josh
really
appreciate
you
leading
this.
There
was
one
other
item
in
the
agenda:
nisha's
put
up
a
governance
doc
if
people
have
time
this
week
to
actually
go
and
comment
on
that
pr.
That
would
be
very
great.
Thank
you
all
for
your
time
and
sorry
for
going
over
two
minutes
have
a
great
day.
Thank
you
thanks.
Charles
thanks,
nisha.