►
From YouTube: [OCI-WG] Reference Types - 2022-01-18
B
A
In
there
yeah
good
good
cold
over
there,
marina
on
the
east
coast.
I
saw
a
few
snow
videos
over
the
weekend
from
some
friends
out
there
and
I
was
like
woof.
C
A
It's
lovely
ominous
time
where
it's
like.
Is
this
the
last
night
that
we're
going
to
get
or
is
there
going
to
be
a
little
bit
more
in
in
late
january
february?
I
think
I've
actually
seen
it
as
late
as
late
march.
I
think
one
year
like
2017
getting
snow
in
new
york
city
and
it
was
like
whoa.
This
is
wild.
C
A
Hello
to
everybody
for
those
who
just
joining
we'll,
give
it
another
minute
or
so
for
folks
to
come
on
in
and
then
we'll
get
started.
If
there's
anything,
you
want
to
add
to
the
agenda
or
feel
that
you
want
to
discuss
today
feel
free
and
we
can
go
into
it.
I
don't
think
we
have
a
full
agenda
of
needing
the
full
hour,
meaning
that
so,
but
hey
we
have
the
time
we
need
it.
We
can
use
it.
D
A
I
had
put
it
I'd,
put
it
there
if
anybody
wants
to
boil
boil
it
down,
I'd,
take
a
stab
over
it
over
the
weekend,
but
if
you
want
to
take
a
few
moments
vanessa
and
see
if
you
want
to
bubble
it
down,
I
had
thrown
one
out
there
on
the
agenda
there
just
to
see
how
it
felt
and
what
we're
missing,
but
yeah
go
ahead
by
all
means.
A
Okie
doke
I'll
share
the
agenda,
the
link
to
the
agenda
one
more
time
and
chat
because
I
know,
as
you
roll
in,
I
don't
think
you
get
the
back
scroll
for
folks
who
are
joining
and
feel
free
to
add,
add
your
name
there
as
an
attendee,
but
let's
get
into
it.
I
think
we're
about
normalizing
here
on
attendees,
so
hello
and
welcome.
This
is
the
second
working
group
for
reference
types
in
oci
today
is
tuesday
january
18th.
A
A
I
think
there's
a
little
room
on
the
agenda,
but
first
I
wanted
to
go
through
the
action
items
we
agreed
upon
as
a
group
last
week
to
go
through
the
first
one
being
that
people
were
trying
to
actually
work
on
delivering
a
mission
statement
for
the
working
group
and
the
purpose
and
goal
here
of
actually
having
a
mission
statement
is
so
that
everybody's
clear
about
the
goal
of
the
working
group-
and
I
wanted
everybody
to
have
an
opportunity
to
actually
say
what
their
hopes
and
goals
are
for
the
working
group.
A
So
I
know
we're
gonna,
there's
actually
an
action
item
there
and
an
agenda
item.
Additionally.
If
we
want
to
talk
about
that,
so
I've
carved
some
things
out
there
for
us
to
discuss,
but
my
goal
would
be
to
land
on
a
mission
statement
that
we
can
actually
all
agree
upon
and
covers
what
we
think
we're
going
to
achieve
as
a
group
here.
So
we'll
take
some
time
next
up
to
do
that.
Once
we
get
into
the
agenda.
A
The
second
piece
was,
I
had
assigned
myself
to
scaffold
out
just
adding
links
to
the
details
of
this
working
group
to
the
repository
and
provide
links
to
comms
there.
So
you
can
see
pr2
under
working
group
reference
types,
I've
scaffolded
out
a
design
there.
If
you
have
feedback,
feel
free
to
add
your
comments.
A
It's
still
missing
the
mission
statement
for
the
working
group,
so
once
we
do
that,
I
can
submit
that
as
a
subsequent
pr,
but
I
wanted
just
a
single
place
to
have
all
the
information
where
people
are
communicating
how
to
connect,
how
to
get
access
to
the
meetings
and
have
that
in
github
form
on
the
repository.
So
that's
what
that
pr
is
there,
and
if
anybody
wants
to
review
and
help
with
that,
that
would
be
very
much
appreciated.
A
A
Nope,
okay,
we
will
continue
here,
excellent
and
so
finalizing
the
mission
statement,
so
I
basically
just
took
a
read
through,
took
a
look
through
everybody's
feedback
and
tried
to
boil
it
down.
I
do
see
nisha
just
left
a
comment.
Thank
you,
nisha!
What's
wrong
with
the
purpose
on
the
proposal,
the
only
thing
I
felt
that
it
wasn't
clear
enough
in
that
proposal
to
rally
a
mission.
I
didn't
feel
like
that
we
had
an
elevator
pitch
for
what
the
purpose
of
the
wg
was
doing.
A
I
felt
it
was
a
little
too
fluffy
having
wrote
that
so
I
was
hoping
the
rest
of
the
team
could
actually
come
together
and
write
a
mission
statement
that
we
could
all
agree
on
and
we
could
go
and
update
the
original
proposal
as
well.
So
we
have
that
have
that
there
if
there's
anybody
else,
yeah
thanks
for
taking
notes.
A
There
appreciate
the
folks
taking
notes,
but
when
I
read
the
purpose
statement,
it
wasn't
clear
to
me
what
the
goal
of
the
group
was,
and
I
you
know
I
had
a
hand
in
writing
that
so
I
felt
just
having
the
wider
group
and
folks
here
who
are
working
on
it
to
have
a
state
in
there.
So
let
me
go
ahead.
I
will
give
you
the
draft
that
I
put
together,
and
this
is
by
no
means
my
own
words.
I
basically
mixed
up
a
few
different
proposals
I
saw
from
last
week.
A
So
if
you're
looking
at
the
minutes
from
last
week,
you
can
see
everybody's
ideas
there,
so
I
came
up
with
just
a
one
liner
here
that
says,
define
the
semantics
and
the
mechanism
for
describing
and
querying
relationships
between
architects
stored
in
an
oci
registry,
and
I
took
that
as
I
think
it
was
a
lot
of
john's
words.
Specifically,
I
felt
they
were
very
succinct,
but
that's
not
to
exclude
anybody
else.
I
felt
like
that
statement
captured
the
breadth
of
what
everybody
was
saying.
A
So
that's
that
was
what
I
was
was
hoping
to
get
to
something
that
was
very
succinct
and
captured
the
breadth
of
what
we
were
trying
to
do
here.
So
did
anybody
else
have
any
other
ones
they
want
to
propose,
or
I
guess
the
other
way
we
could
do.
It
is
looking
at
the
one
that
I
proposed
there
and
I
can
pop
it
in
chat,
but
it's
in
the
minutes
here
you.
A
What
are
we
missing?
What
doesn't
it
cover?
What
doesn't
that
statement
cover
would
be
my
my
question:
do
we
feel
like
that,
encompasses
the
work
that
we
need
to
do
or
not
adequately,
and
I
will
pass
it
to
the
room
here.
If
you
can
raise
your
hands,
I
see
a
few
folks
doing
that
thanks,
nisha
and
vanessa,
so
I
will
start
with
nisha.
B
That
I'm
attending
welcome
thanks
a
lot
for.
B
This
yeah
yeah.
I
am
wondering
if
just
from
previous
conversations
that
I've
had
about
reference
types
does
the
does
semantics
include
data
structures.
B
Does
it
include
creating
data
structures
and
like
management
of
the
data
structures?
Does
it
include
any
of
that?
B
So
we,
if
there
is
mechanism,
but
I'm
specifically
asking
for
semantics.
F
In
my
mind,
yeah
that
the
mechanism
would
be
like
how
do
we
either
create
new
or
modify
existing
data
structures?
I
mean
I
have
ideas
for
that,
but
what
what
I
think
is
going
to
be
important
is
the
semantics
around
that
mechanism
so
like
what
does
it
actually.
C
F
If
we
have
something
referring
to
another
thing
and
and
how
do
clients
and
registries
interpret
that
so
yeah,
I
think
what
you
said
is
accurate,
that
the
mechanism.
C
B
So
would
that
mean
user
stories
or
what
exactly
is
missing
from
the
current
the
current
architecture
that
needs
to
be
enhanced,
that
sort
of
thing.
F
F
Refers
to
this
other
object.
What
is
that
relationship?
Can
you
define
like
how
I
should
interpret
that
relationship
and
what
is
safe
to
do
or
unsafe
to
do
that?
That's
what
I
think
is
missing.
There's
also
like
different
kinds
of
relationships
that
I
think
are
useful
and
interesting
and
aren't
necessarily
like
the
the
most
obvious
thing.
F
Like
I
think
the
implied
semantic
of
most
of
these
is
oh,
this
is
metadata
about
the
thing
I'm
referencing
right,
but
there's
there's
also
the
potential
for
like
this
is
an
aggregate
of
things
or
like
I'm.
F
Something
that's
maybe
not
inside
the
oci,
dag
or
in
a
different
name
space,
or
something
like
these
relationships
have
a
lot
of
properties
that
maybe
we
decide
aren't
interesting
to
express.
But
I
want
to
know
the
exact
exactly
like
what
it
says
about
this
relationship
when
you
put
one
in
an
object,
because
I,
in
order.
B
So
before
I
hand
off,
is
this
a
topic
that
folks
are
interested
in
following
up
on?
I
have
I
mean
I
have
some
frameworks
around
how
we
can
how
we
can
talk
about
this.
How
we
can
drill
down
to
a
specific
thing
is
that,
okay,
with
everyone.
A
Yeah,
I
think
that's
we're
definitely
going
to
have
to
have
those
conversations
nisha.
So
if
you
have
some
ideas
be
willing
to
to
have
you
facilitate
those
conversations,
I
try
to
capture
your.
What
you've
got
here
in
the
notes
as
well,
so
feel
free
to
add.
Add
it
there
as
well.
I
guess
the
question
I
have
before
I
hand
off
to
vanessa
is:
is
semantics
and
mechanisms
clear
enough
in
that
message
statement
or
is
it
do?
Should
we
drill
down
on
that
specifically
because
if
there
are
questions
around
what
it?
What
does
that
mean?
A
I
Yeah
thanks,
I
think
one
one
thing
I
think
is
missing
from
the
mission
statement
is
a
an
assumption
of
a
limit
on
the
amount
of
time
that
we
will
work
on
this
I
mean
not
like
not
like.
We
need
like
a
deadline.
That's
not
what
I'm
saying
but
more
like
we
want
to
define,
propose
to
the
distribution,
spec
or
image
spec
a
set
of
things
instead
of
changes
and
then
stop
and
then
not
have
these
meetings
again.
I
I
love
you
all
I'd
love
to
have
these
meetings
forever,
but
I
think
we're
still
having
them
in
five
years.
We've
made
a
mistake,
so
I
would
like
to
I
mean:
there's
not
like
an
amendment
to
the
mission
statement,
but
I
think
it
is
missing
that
we
are.
We
have
a
specific
goal.
I
Do
this
thing
and
then
stop
have
a
you
know,
high
five
party
and
then
work
on
other
stuff.
A
I
I
agree,
so
how
would
we
serialize
that
into
you
know
some
kind
of?
Is
that
an
additional
statement
on
the
back
of
that
mission
statement,
because
I
think
the
definition
of
done
is
really
critical.
How
do
we
exit
this
thing
and
how
do
we
know
when
to
disband
this
working
group?
Because
the
idea
is
that
we
spin
out
some
work
product
and
spin
down
the
working
group,
and
it's
not
forever,
so
I
think
calling
out
something
really
important
there.
A
Jason
is
what
is
the
definition
of
done,
even
if
that's
what's
the
definition
of
good
enough
and
solving
the
problem
that
we're
setting
out
to
to
solve,
which
is
why
I
think
us
all
coming
and
talking
about
this
mission
statement
is
important
that
we
capture
all
these
things.
So
we
know
within
ourselves.
How
are
we
saying
we're
done
so
yeah?
How
would
we
serialize?
Is
that
an
additional
sentence
here,
or
should
we
try
and
cram
it
in
I,
I
propose
that
sentence
as
a
starting
point.
A
It's
by
all
means
you
can
tear
it
up
and
move
it
around.
I
Yeah,
I
think
I
mean
I
think
it
could
be
an
edit
to
the
mission
statement.
It
could
also
just
be
a
tacit
assumption
among
all
of
us
here
that
when
we
have
when
we
feel
like
we
have
defined
the
semantics
and
the
mechanisms
that
we
you
know
move
to
stop
doing
this
or
or
you
know,
eventually,
I
think
eventually
our
progress
will
slow.
Until
eventually
we
realize
we're
not
talking
about
anything
new
and
then
we
can
just
sort
of
like
look
around
and
high
five
party
and
then
leave.
I
A
A
I
I
think
like
defining
is
maybe
never
done,
but
proposing
is
an
event
that
will
pass
once
we
have.
C
I
A
J
So
I
was
gonna
say
the
relationship.
I
think
we
could
define
that
a
little
clearer,
because
we
already
we
already
do
model
relationships
in
an
oci
registry
right.
We
have
image
to
layer
and
image
index
to
image,
but
the
relationships
that
we
are
trying
to
describe
are
different
than
those,
and
it
seems
like
the
mission
statement,
should
be
specific
enough.
That
somebody
who
hasn't
been
participating
in
this
for
the
last
year
would
understand
why
we
need
this,
and
what
is
new
about
this.
A
J
J
I
opened
other
suggestions,
though,.
B
So
I
can,
I
can
kind
of
bring
in
some
context
here
the
the
specific
kind
of
relationships
we
were
thinking
about
were
signature
off
or
s
bomb
off.
B
Let's
say,
let's
see
in
any
other
like
what
I
would
call
supplemental
objects
that
are
attached
to
a
container
image,
and
this
is
why
actually,
I
asked
a
question
in
the
in
the
document,
which
is
what
are
objects?
Are
they
containers
pieces
of
containers
or
anything
stored
in
an
oci
registry?
B
So
we
probably
want
to
have
a
definition
for
that
we
want
to
be
able
to.
You
know,
actually
describe
the
thing
we
are
working
with
and
I
think
from
my
discussions.
A
Yeah
great
thanks,
nisha,
I'm
I'm
thinking
about
this
a
little
more
and
how
we
might
digest
this
down,
because
we
could
either
add
it
all
to
the
mission
statement
or
add
a
glossary
of
terms
as
a
supplemental
thing
afterwards,
when
we
say
object
when
we
say
relationship.
So
I'm
just
thinking
about
that,
but
I
do
like
that
people
are
saying
well
what
relationships,
what
objects
we're
defining
proposing?
A
This
is
kind
of
fidelity
I
wanted
to
get
to
in
a
mission
statement,
so
I'm
going
to
stop
by
see
a
few
more
folks
with
their
hands
up.
So
thanks
for
working
through
this
and
and
using
the
hands
gesture,
I
also
see
your
comment,
marina.
I'm
going
to
respond
to
that
and
chat
kiss
kisik
your
hand
is
next.
A
Nope,
okay,
just
put
your
hand
back
up,
I
saw
it
go
down,
so
I'm
not
sure
if
it
was
just
by
accident
I'll
pass
it
to
brandon
mitchell.
K
Yeah
I
was
just
going
to
follow
along
with
michael's
comment
there
in
terms
of
defining
the
relationship.
If
we
were
defining
another
parent
to
child
objects
inside
of
a
registry,
we
would
just
make
that
you
wouldn't
have
to
go
through
this
whole,
creating
a
new
api
and
other
things.
So
I
think
the
the
complicated
part
here
is
we're
defining
a
new
child
or
sibling
to
an
existing
object
in
a
registry,
not
sure
the
best
way
to
phrase
that,
but
that's
kind
of
my
thought
process.
K
A
Sorry
for
the
silence,
I'm
typing
what
you
said:
brandon
just
make
sure
we
capture
it.
If
anybody
else
would
like
to
jump
in
otherwise
I
can
pass
it
to
josh.
I
I
think
that's
that's
a
really
good
point
there
brandon
is:
there
are
relationships.
What
type
of
relationship
is
this
as
that
name
that
you
came
up
with,
which
is?
Can
you
say
it
again?
It's
new
new
child.
K
B
B
Yeah,
so
one
of
the
things
I
want
to
be
careful
of
is
the
the
fact
that
you
know
we
have
something
called
like.
We
have
a
rudimentary
way
of
extending
the
the
cast,
but
we
also
need
to
allow
ways
to
apply
enhancements
to
the
rudimentary
thing,
maybe
like
the
inheritance
model
or
something
like
that.
K
So
you
can
extend
it
by
creating
a
new
kaz
entry.
That
no
longer
is
the
same
old
object
you
had
before,
or
you
can
extend
it
by
creating
a
new
parent
object
that
points
to
whatever
is
existing
in
there,
but
you
can't
really
extend
it
to
create
a
new
child.
It
goes
against
the
design
of
what
the
directed
a
cyclic
graph
does
for
you.
So
that's
where
the.
K
B
K
I'm
sure
if
you
ask
five
people,
you
get
six
answers,
but
I
think
we're
attempting
to
find
the
best
way
to
do
it
with
the
tools
that
we've
got,
and
hopefully
we
structure
in
a
way
doesn't
break
everything.
H
I'm
going
to
go
ahead
and
say:
I
think
that
the
mission
statement
is
general
enough.
I
think
it's
generally
enough
for
someone
to
understand
what
it
is,
we're
trying
to
do,
and
I
don't
mean
to
down
like
say
that
the
conversations
we're
just
having
are
not
important.
It's
just
like
more
of
a
can.
We
get
the
mission
statement
out
and
then
from
there
dive
down
into
talking
about
dags,
and
if
we
need
clarification
on
mechanism
versus
semantic,
I
feel
like
we're
falling
into
a
trap.
A
I
think
that's
good
good
feedback
josh
and
it's
always
just
how
far
do
we
get
into
technical
details,
but
I
think
somebody
said
earlier-
and
I
can't
remember
my
apologies
if
somebody
was
driving
by
and
they
don't
have
six
months
worth
of
context
in
their
mind.
Could
they
read
this
and
know
exactly
what
you
know
at
a
high
level,
what
we're
trying
to
achieve,
and
that's
kind
of
where
I
wanted
to
get
to,
and
I
think
you
know
that
the
purpose
of
this
conversation
was
just
to
make
sure
we're
all
you
know.
A
We've
talked
through
what
we
think
we're
here
for,
and
at
least
we
capture
most
of
that
and
then
we
can
get
into
the
the
details.
So
is,
what's
there
now
good
enough
to
meet
the
need?
Do
we
all
agree
on
that?
You
know
solving
the
the
drive-by
situation.
I
don't
know
if
there's
a
better
term
for
that
somebody
somebody
casually
looking
at
what
what
are
they
trying
to
do
here
incremental
non-breaking?
I
see
a
lot.
C
A
What
are
the
words
we
need
to
change
like?
How
do
we
drive
this
because
I
also
agree
with
josh.
I
don't
think
we
need
to
talk
about
this
for
three
weeks
and
I'd
love
to
leave
the
room
with
a
this
is
good
enough
and
if
we
need
need
to
we,
we
need
to
iterate
on
it
later.
That's
always
up
for
grabs,
but
as
long
as
we
all
feel
good
about
what
we've
come
up
with,
I
think
the
shared
work
work
here
is
important
to
make
sure
we're
doing
this.
A
A
So
if
you
have
anything
there
that
you
want
to
stay,
please
go
and
put
it
in
the
minutes
as
well,
and
I'd
appreciate
that,
because
there
is
a
lot
of
good
feedback
here,
so
I've
I've
tried
to
put
words
in
brackets
about
words
that
were
being
people
were
picking
on.
Do
we
want
to
not
picking
on
just
saying
that
they're
a
little
vague,
do
we?
A
A
We
also
had
other
wording
like
a
child,
new
child
or
sibling
and
then
objects,
and
then
I
saw
sanjay
and
chat
say:
are
we
only
talking
about
within
a
single
registry?
A
B
I
personally
think
we
have
enough,
but
if
I'm
missing
anything
sorry.
D
A
D
To
describe
and
query
reference
relationships
between
objects
stored
in
a
single
or
oci
registry,
so
just
get
rid
of
semantics
on
a
mechanism.
It's
implied
that
in
the
how
we'll
figure
out
we'll
have
to
discuss
something
like
that,
but
I
don't
think
we
want
to
explicitly
say
semantics
and
a
mechanism,
at
least
at
this
point.
If
it
really
is
this
drive-by
description,
okay,.
A
H
I
am
in
support
of
what
vanessa
said.
I
think
we
start
really
simple.
Like
the
minute
we
start
talking
about
single
registry,
multiple
registry
s-bomb
signature.
I
think
we're
missing
it.
We
need
to
enable
all
those
things,
but
I
think
I
I
think
we
just
need
to
start
very
simple,
so
people
don't
get
tripped
up.
A
Has
how's
that
look?
The
only
questions
I
have
left
it's
in.
It's
in
highlighted
in
the
document
propose
how
to
describe
and
query
that's
got
to
go,
and
yes,
please.
Somebody
else
delete
my
lack
of
clean
up
there.
Thank
you
anonymous
wolf
for
reference.
Do
we
need
reference
relationships.
B
Either
pick
reference
or
relationships,
I
would
choose
relationships,
but.
B
Yeah,
I'm
open
to
whatever
other
people
think
a
relationship
sounds
more
generic
than
references.
A
A
A
B
For
now,
I
think
so,
a
before
before
folks
go
on,
I
would
like
to
say
that
we
do
have
some
requirements
that
I
think
need
to
be
added
to
the
to
the
mission.
A
I
would
just
add
them
outside
yeah.
We
definitely
need
requirements.
A
A
A
Ship,
it
cool
stupid,
all
right,
okay,
so
we
can.
We
can
move
on.
Thank
you
very
much
for
doing
that.
Exercise
really
appreciate
all
the
time
there.
I
will
go
and
add
that
to
that
document.
So
don't
that
pr
that
I
have
out
there
is
the
mission
statement
and
then
we
can
fill
in
other
pieces
as
we
make
decisions
as
a
group
here
so
I'll.
Take
that
as
a
as
an
action
item.
A
So
as
for
next
steps,
I
want
to
get
into
the
the
next
steps
for
here
and
what
I
wanted
to
do
was
just
start
a
discussion
about
how
we
might
break
this
up
and
understand
what
we
think
is
the
best
path
forward,
and
then
we
can
make
make
some
steps
in
that
direction.
So
I
guess
I
will
just
seed
seed
the
conversation
here
and
then
we
can
move
move
on
it
from
there.
A
So
what
I
was
thinking
might
be
a
good
next
step,
and
this
is
just
my
my
opinion
here
to
break
this
down
is
we
could
either
start
with
requirements
or
we
could
take
the
proposals
as
they
are
at
the
moment
and
see
if
we
can
peel
out
pros
and
cons
to
each
implementation,
as
they
are
described
by
individuals
that
have
reviewed
any
of
the
proposals
and
bring
that
conversation
to
this
group,
and
I
think
what
hasn't
happened
in
the
past
has
been
if
people
have
views
as
to
why
they
have
a
pro
or
a
con
for
their
specific
point
or
angle.
A
How
do
we
actually
have
a
conversation
about
that?
Because,
even
as
I
read
it,
it's
hard
to
grok
or
what
does
somebody
mean
by
this
case
isn't
met
or
this
challenge
and
I'm
wondering
if
having
that
conversation
with
asynchronously
to
at
least
describe
it,
have
the
individuals
that
have
either
a
pro
or
a
con
with
the
implementation
describe
it
document
it.
And
then
we
figure
out
the
best
path
forward
and
agree
with
that
together,
because
I
find
it
challenging
just
to
go
through
all
those
things.
A
So
we
could
either
start
from
that
angle,
taking
the
proposals
that
are
already
out
there,
which
you
know
people
have
done
a
lot
of
work
on.
There
are
a
lot
of
comments
and
distill
that
into
a
document
with
you
know,
I
kind
of
feel
like
it
is
the
word
drive
by
or
a
very
high
level
summary
of
what
this
proposal
is
and
then
some
pros
and
cons
for
each
each
proposal,
and
then
we
can
talk
through
those
situations.
So
that's
kind
of
what
I
was
thinking.
A
The
other
route
we
could
take
is
putting
requirements
together
to
things
that
we
have
and
then
trying
to
draft
a
proposal
together.
So
they
were
the
two
options
I
had
and
if
there
are
others,
please
let
us
know
or
what
we
think
is
going
to
be
most
beneficial
for
this
group,
because
again
keeping
that
mission
in
mind,
we
do
want
to
come
up
with
a
proposal
and
do
it
in
a
timely
manner
as
well.
A
So
I
want
to
make
sure
that
all
the
prior
art
out
there
that's
been
invested
into
this
space
is
that
we
discuss
and
make
a
decision.
I
see
so
hands
up
if
you'd
like
to
feed,
provide
feedback
to
that
or
or
let
me
know
before
we
go
off
and
and
start
this
endeavor.
So
I
see
nisha
has
their
hand
up
first
and
then
brendan
after
nisha.
Anybody
else
throw
you
throw
your
hands
up
as
well.
Please.
B
So,
just
from
previous
conversations
around
this
subject,
I
think
that
we
ought
to
start
with
requirements.
First.
K
Second,
your
proposal
of
saying,
let's
just
throw
a
bunch
of
spaghetti
on
the
wall,
see
what
sticks
and
then
farm
out
of
that
come
up
with
what
the
requirements
should
be,
what
the
goal
should
be
stuff
like
that,
sometimes
it's
easier
to
evaluate
those
and
to
come
to
add
those
the
list
when
you
have
something
you're
looking
at
and
so
to
look
at
some
of
the
existing
proposals.
B
A
So
does
it
make
sense
to
to
take
the
the
proposals
that
we
have
and
cleave
out?
I
tend
to
operate
more
right
now,
given
the
situation
with
proposals
already
been
out
there
more
with
brendan's.
Let's
take
the
what
what
the
work's
already
been
done
and
cleave
out
the
proposals
and
make
sure
that
we're
all
in
agreement
with
the
requirements
and
then
rationalize
the
the
proposals
that
are
out
there
with
those
requirements
and
see
if
they
meet,
because
that's
going
to
give
us
a
good
foundation
for
making
a
decision.
A
I
feel
from
doing
that
so.
I
Yeah
there's:
actually
there
was
a
doc
that
dan
wrote
it's
dated
june
11th
2021
I'll
share
it
in
the
chat
that
goes.
B
I
That
goes
over
two
of
the
proposals.
I
don't
know,
if
that's
all
the
proposals
we
have
now
or
if
there's
more,
but
I
don't.
C
I
It's
at
least
a
good
refresher
review
of
some
of
the
proposals
that
I
think
we
could
benefit
from
going
over.
I
also
wanted
to
mention,
I
think,
in
addition
to
spelling
out
the
requirements.
I
think
we
should
also
keep
track
of
a
list
of
things
that
we
know
are
unsolved
and
un
perhaps
unsolvable
problems.
I
think
garbage
collection
belongs
in
there.
I
Garbage
collection
is
not
specified
by
oci
in
any
way
at
all,
and
so
and
in
the
past
we
have
like
fallen
into
a
bat
of
oil
matar,
something
I
don't
know
my
analogies
are
terrible.
We've
fallen
into
a
tarpit,
that's
what
I
was
looking
for,
trying
to
specify
the
semantics
around
garbage
collection
in
the
face
of
reference
types
when
there
is
no
specification
of
garbage
collection
in
general
at
all.
I
So
unless
we
want
to
like
pin
this
and
go
solve
garbage
collection
in
the
spec
and
then
come
back
and
solve
garbage
collection
in
this
addition
to
this
pack,
which
I
don't
think
we
should
do,
I
think
we
should
spell
out
that
garbage
collection
is
out
of
scope
for
this
proposal
unless
other
people
disagree.
But
I
think
that
is
a
useful
narrowing
of
our
scope
so
that
we
can
make
progress
on
things
that
we
can
actually.
I
It
exists
in
the
real
world
and
we
shouldn't
ignore
it,
and
we
should
even
I
think
that
sentiment
or
that
statement
is
perfect
to
put
in
there
that
like.
If
you
know
real
registries
will
need
to
do
garbage
collection,
but
we
should
not
include
garbage
collection
in
the
spec
of
this
thing,
because
unless
we
want
to
go
solve
garbage
collection
in
all
of
oc,
I
would
but
don't
please.
I
I
think
I
think,
because
garbage
collection
is
not
specified
by
oci
at
all.
We
should
not
make
a
statement
about
whether
we
will
or
won't
break
or
what
breaking
means,
because
that
term
doesn't
exist
in
the
context
of
oci.
If
somebody
wants
to
go,
do
a
survey
and
figure
out
like
what
will?
I
What
will
ecr
do
with
this
data
right
now
today
I
mean
you
can
go,
you
can
add
references.
Reference
objects.
Do
things
right
now
because
of
the
unknown
fields.
Behavior,
you
can
go
see
what
that
will
do
and
it
will
probably
not
do
what
you
want,
but
I
don't
think
it
will
take
down
a
registry.
If
it
does
that's
a
bug
but
yeah,
I
don't.
I
don't
think
we
should
specify
it
because
it's
just
not
in
oci
at
all.
A
Yeah,
I
think
jason
you're
raising
a
good
point
and
without
you
know,
going
down
the
rabbit
hole.
We
should
have
a
way
just
to
say
this-
we're
absolutely
not
covering
this
and
it's
out
of
scope
and
just
do
that
as
a
team.
So
it
sounds
like
you
know,
there's
a
conversation
to
be
had
about
garbage
collection
and
I
know
it's
been
had
before.
But
let's
just
get
consensus
within
this
group
or
some
agreement
and
there'll
be
you
know
several
things.
I'd
imagine
that'll.
A
You
know
warrant
the
same
discussion,
but
still
you
know
what
is
the
best
way
to
move
forward
with
the
requirements
and
then
you
know
taking
the
proposals
that
are
out
there
and
I
think,
looking
for
a
way
to
rationalize
the
proposals
as
they
are
and
have
the
conversations
about
you
know
what
meeting
requirements
are.
Not
is
the
most
important
way
to
facilitate
how
we
would
make
a
decision
here.
G
I
think
gc
is
a
great
example,
but
I
wouldn't
say:
gc
belongs
or
not
belongs
in
spec.
I
would
kind
of
like
first
call
out
from
a
requirement
perspective.
The
client
has
a
certain
expectation
and
registries
are
free
to
implement
what
they
want
to
so
scope.
It
rather
than
saying
something
is
not
there
or
not.
That's
what
I
would
say
and
that
would
obviously
lead
to
gcs
out
of
scope,
because
clients
don't
really
care
whether
gc
is
done
or
not
so
scoping.
G
How
the
client,
behavior
might
be
from
a
specification
standpoint,
would
be
the
driving
force
or
the
driving
factor
of
why
certain
areas
would
just
be
out
of
scope,
and
that
way
we
don't
get
into
discussions
like
we
had
before
totally
support
that
and
then
yeah.
I'm
not
a
fan
of
putting
gc
into
respect.
Also.
I
Thanks,
I
think
we
we
should
every
time
we
all
agree
on
something.
We
should
ring
a
bell
and
write
it
in
the
doc
and
one
of
the
things
I'm
hearing.
I
I
Agree,
we
don't
want
to
specify
garbage
collection
behaviors,
either
in
oci
or
related
to
references
and
that
well
I'll
I'll
stop
there
ring
the
bell.
We
all
agree.
We
don't
want
to
specify
garbage
collection,
behavior.
B
Yeah,
sorry,
I
don't
want
to
ring
the
bell
just
yet.
I
am
concerned
that
if
we
do
not
consider
garbage
collections,
then
whatever
proposals
we
have
for
our
mechanisms-
and
that
is
in
our
charter-
would
end
up
conflicting
with
expectations
that
implementers
would
have.
A
A
I
see
some
feedback
on
the
doc
sarjay
said:
can
we
update
with
links
to
prior
art
here?
So
I
know
in
the
proposal
there
was
a
bunch
of
links
with
all
the
proprietor
art
that
I
went
and
just
did
a
quick
take
across
the
ecosystem,
but
if
there
are
other
proposals
out
there
that
people
want
to
be
part
of
the
consideration
or,
let's,
let's
collect
them
as
well.
A
A
First,
first
yeah:
let's
see
what
we
agree
on
on
which
bits
we
think
is
valuable
and
then
yeah.
Obviously,
thank
thanks
for
your
enthusiasm
and
and
love
to
have
you
be
a
part
of
it?
But
what
do
we
think
is
the
best
next
step
here,
because
I
don't
want
to
come
back
to
the
group
with
you
know.
If
nisha
goes
off
and
does
a
whole
bunch
of
work,
we
don't
find
it
beneficial
as
a
group.
So
that's
what
I'm
trying
to
you
know.
Let's
not
put
our
energy
into
things
that
aren't
going
to
be
valuable.
A
For
this
group
to
move
forward
sounds
like
you
know,
do
we
take
the
the
proposals
that
are
out
there
and
craft
a
set
of
requirements
and
bring
a
set
of
requirements
back
to
the
group
and
have
them
prior
to
next
week's
meeting,
so
we
can
start
to
riff
on
them?
Is
that
the
best
next
step,
or
is
it
just
collating
all
the
requirements
and
everybody
reads
through
them,
or
is
it
producing
a
doc
with
summaries?
I
need
to
read
dan's
doc
thanks
for
providing
it
there
jason.
A
B
I'm
sorry
about
that
lucky.
I
thought
that
the
next
steps
was
collecting
all
the
documents
and
summarizing
them
and
that's
what
I
volunteered
to
do,
but
that's
okay,.
E
A
I
I
think
that
sounds
good
and
I'd
also
be
willing
to
help
with
that
as
much
in
any
way
that
I
can
nisha
if
you
need
help
or
if
there's
a
doc,
you
are
summarizing
these
things
in
I'm
happy
to
take
a
look
and
help
with
that
before
next
week.
A
Okay,
so
we
can
create
a
doc
with
the
links
to
the
proposals
and
start
summaries
and
see
if
we
summaries
of
each
proposal
and
then
start
to
see
what
the
requirements
are
and
then
next
week
we
we
can
cleave
out
the
requirements
as
you're
reading
through
those
proposals,
and
then
we
can
present
the
requirements
of
the
group
and
start
making
decisions
about
whether
these
requirements
cover
everything
and
if
there's
anything,
we're
missing.
Does
that
summarize?
What
we're
doing
here?
A
I
That
is
also
my
understanding.
We
are
at
least
equally
wrong
if
we
are
both
wrong,
we
are
both.
You
are
not
alone
in
being
wrong.
If
you
are
wrong.
I
A
A
H
Yeah,
I
have
a
comment.
I
I
kind
of
put
this
in
the
doc
too.
I'm
wondering
as
we're
designing
the
purpose.
If
we
can
avoid,
I
guess
I'm
asking,
should
we
avoid
talking
about
some
of
the
very
like
pressing
industry,
specifics
for
certain
people
coming
to
this
working
group
and
that
specifically
being
s-bombs
and
crypto
signatures?
H
L
A
B
H
Yeah,
that
would
be
phenomenal,
because
not
only
tooling,
but
there's
also
the
standards,
there's
spdx
and
cyclone
and
we
have
six
store,
and
you
know
it's
just,
I
think
we're
starting
off
to
a
really
good
foot,
and
I
think
we
can
stay
that
way
and
nisha
I'm
happy
to
work
with
you.
If
you
want
to
come
up
with
something
silly
to.
H
A
C
This
sounds
great
to
me.
Yeah.
I
think
that,
like
once
implemented,
I
think
there'll
be
a
lot
of
potential
uses
for
this,
so
it'd
be
nice
to
focus
more
on
what
we
want
the
thing
to
do
than
how
it
can
solve
one
of
these
many
problems.
So
I
like
that
idea.
G
So
I
see
both
there's
a
positive
and
a
negative
to
it,
then,
the
negative
being
that
anybody
who's
coming
to
this
body
would
have
to
understand
these
fantasy
concepts
before
they
can
actually
get
to
something
real,
but
keeping
them
closer
to
something
generic
like.
Let's
call
it
s
bombs
s
bombs
is
generic
itself
like
this
pdx.
The
cyclone
signatures
have
no
trees
six
store,
so
I
would
say,
bring
them
close
to
the
real
use
case
as
much
as
possible,
but
generic
enough,
where
there's
no
specific
yeah.
G
What
do
you
say
attachment
to
any
conflicting
views
and
things
like
that
might
be
a
model?
Just
don't
call
it
fubar.
That's
gonna,
be
like
totally
unusual.
So
as
long
as
we.
H
K
I
was
going
to
call
out
john
johnson
to
see
if
he
wanted
to
chime
in
or
not
because
I
think
I
heard
him
mention
earlier
in
this.
He
had
some
very
non-traditional
ideas
of
what
this
looked
like,
and
so
I
was
kind
of
going
along
the
lines
of
if
we
did
say
s
bomb.
We
could
also
say-
and
this
other
very
other
different
thing
that
you
might
not
be
thinking
of
and
to
solve
both
of
those
you're
going
to
come
to
this
group
and
find
the
common
solution
that
solves
everybody's
problem.
A
A
It's
okay,
okay,
so
it
looks
like
unless
there's
any
you
know
anybody
that
wants
a
really
good
real
world
use
case
right
now.
I'm
okay,
with
with
coming
up
with
you,
know
the
abstract
concept
here
so
josh
can
you
go
and
put
together?
I
think
ram.
Kumar
said
you
know
a
straw
person
reference
reference
type,
an
example
that
we
could
use
as
the
single
thread
to
go
through
this
whole
working
group
as
a
way
to
rationalize
the
experience.
A
A
It's
a
special
committee
josh
a
committee:
yes,
a
committee
okay,
so
you
do
that
and
and
nisha
and
and
jason
and
everybody
else.
We
can
start
to
to
get
the
boil
down
the
summary
and
make
sure
that
we
start
to
get
requirements
together.
Sound
good.
A
Okay,
I'm
going
to
capture
those
as
the
action
items
I'll
give
everybody
four
minutes
back
before
the
top
of
the
next
hour.
Thanks
for
the
conversation-
and
I
I'm
really
excited
about
the
work
that
this
group
is
doing,
and
thanks
for
indulging
the
mission
statement
I'll
get
that
updated
as
well.
If
there's
anything
in
between
feel
free
to
reach
out
into
the
discussions
or
on
the
slack
channel
in
the
oci
slack
thanks
folks,
we'll
see
you
this
time
next
week,
bye.