►
From YouTube: IETF105-RATS-20190725-1550
Description
RATS meeting session at IETF105
2019/07/25 1550
https://datatracker.ietf.org/meeting/105/proceedings/
B
A
D
F
G
E
G
E
J
K
K
K
K
K
This
is
a
use
case,
the
most
complicated
ones,
so
they
even
find
the
that's.
The
party
there's
nothing
particular
to
the
actual
service
and
it
hasn't
advice
wants
to
have
across
policies
the
deviations,
what
we
call
so
PVA
ssin
until
the
tails,
okay
and
finally-
and
this
is
the
first
muscle
to
speak
of
the
rest
architecture
and
the
right
back
options
in
there-
simple
ones.
K
N
M
E
N
Q
O
Q
O
O
S
T
A
Was
another
architecture
director
when
what
we
have
to
do
to
you
know
the
glory
with
it
and
the
architecture
intruders.
We
had
waited
for
publishing
architecture
directions
that
later
on
in
the
process,
so
could
be
a
document
that
hangs
on
for
a
while
to
before
it
goes
from
final
publication
because
our
intentions
can
evolve.
What's
where
it
starts
great,
so
that's
an
option
as
well
and
even
wants
to
talk
to
that.
You
know
it
evolve
over
time.
It
becomes
an
evolving.
E
Right
so
we
get
to
the
question
and
the
group
decided
that
we
needed
to
recapture
document
and
so
much
like
the
process
that
we
went
to
eat.
We
need
an
equals
one
to
think
there
were
issues,
but
the
roots
are
very
haunted
and
with
the
condition
that
those
issues
we
can't
resolve.
So
you
know
we
can
clarify
that
here
or
we
can
take
that
to
the
discussion.
But
I
guess
one
of
the
things
that
I
would
like
to
see
is
not
civilized
process.
The
less
won't
wait
until
the
use
case.
E
O
Q
Q
Q
K
K
K
E
B
E
E
R
D
K
K
K
Q
Q
R
R
In
the
end,
we
have
fewer
data
objects
than
we
are
going
to
variations
of
the
objects
based
on
the
new
space.
So
if
we
don't
find
some
things
that
they
all
down
to
common
across
some
we're
able
to
read
a
lady
information
bottle
at
high
level
just
years,
it
provides
a
basis
for
the
DVP
options
are
going
to
be
out
there.
We
have
to
find
something
when
we
should
angle
over
there
Asians.
That's
why
I'm
a
mister.
A
A
K
Q
O
O
O
O
Right
so
section
three
house,
nutrition
of
each
claim,
each
flight
and
then
support
strategy,
a
general
way
to
circulated
staffing
to
either
C
or
R
Jason.
So
really
when
I
was
after
here,
whether
you'd
call
an
information
in
a
lot
or
whatever
you
wanna
call
it
is.
The
description
is
neutral
to
the
serialization.
O
O
O
Okay,
so
here
bone,
this
is
one
of
my
own
compositions
is
I.
Think
it's
really
interesting
to
distinguish
the
description
of
the
claims
where
city
center
consent.
They
were
less
interaction,
model
exchange
because
the
yang
dr.
blank
James
is
actually
going
into
a
whole
substantial
stage
of
work.
That's
called
a
plan
by
DG
immunity,
so.
O
So
go
number
one
part
of
the
ethos.
All
the
industry
work
is
in
claims
definition
sign
up
for
math.
We
picked
the
modern
one
and
then
transport
the
pork
off.
This
is
anything
you
want
no
constraint
that
right
in
the
TPM
TCG
world,
the
claims
are
locked
down
just
because
TPM
devices
only
cite
certain
things
and
actions
of
things,
but
in
the
end,
the
truth.
Instead
of
the
back
side,
inform
activists
lockdown,
it
has
to
be,
but
the
PM's
dude
can't
be
because
they
arose
a,
but
there
isn't
working
right.
O
O
O
So
many
of
these
three
flavors
okay,
so
my
proposal
here
is
that
we
split
the
information
model
state
information
house,
probably
to
one
of
it,
is
or
even
write
of
mine
claims.
So
we
have
to
be
an
information,
well,
the
finest
claims
and
we
used
a
similar
technique.
Yeah
just
do
three
like
them
and
that
suits
most
of
the
work,
but
it's
not
necessary
to
keep
getting
worked
up
so
much
because
you
know
my
lockdown.
O
M
P
M
F
F
F
F
M
F
So
we
basically
have
two
drafts
that
are
defining
information
models.
Currently,
there's
the
the
draft
or
cold
strats
information
model,
0
0.
It
basically
contains
two
kinds
of
information,
meta
characteristics
that
applies
to
assertions
and
also
appliances,
7
surgeons.
We
also
use
the
word
claim
as
synopsis
of
assertion.
So
it's.
F
Q
W
W
Than
and
I'd
like
to
see
something
more
along
the
lines
of
the
student
manifest
work
where
it
describes
a
CDL.
That
says
this
is
a
format
of
an
attestation,
and
it
says
these
other
data
elements.
These
are
how
they're
nested.
This
is,
how
you
extended
it
and
in
the
client,
specific
elements
that
use
gravity
in
the
attestation
code,
and
so
some
of
that,
my
computation
area.
This
is
how
we
are
back.
This
is
how
we
put.
W
W
W
K
K
Is
operability?
Here's
what
taxonomy
there
so
have
this
meta
items
that
that
group
them
and
whatever
we
do.
You
should
never
forget
that
at
least
very
well.
That
goes
if
anyone's
supposed
to
be
go
this
system
also
somewhere.
This
cannot
just
be
dropped.
It
wasn't.
My
only
may
only
be
concerned
that
that
might
be
a
few
somehow
these
categories
of
information
elements
that
say
this
is
about
extrinsic,
intrinsic
factor
software,
some
movies
or
protocol
parameter.
That
is
highly
important
for
later
people
to
understand
better
look
when
I
must
remind
everybody.
F
E
I
have
someone
who
declare
define
in
summer
X,
which
is
seven
eight
correctly
you're,
saying
regardless
how
we
define
the
individual
elements
of
an
information
model.
We
also
need
to
talk
instruction
either
be
grouping.
This
versus
categorizations
versus
fights
you're
advocating
to
there
needs
to
be
someone
with
structure
and
guidance
in
that
structure
as
well.
This
I
just
want
to
chew.
Meccans
captured
Thank
You.
N
N
J
N
Q
E
You
start
you
stole
our
line,
so
the
intent
here
is
and
again
I
come
back
to.
We
are
going
to
be
guided
by
the
use
cases,
which
is
why
the
Charter
said
we
will
list
the
use
cases
to
tell
us
the
requirements.
Then
it
will
be
up
to
the
group
based
on
those
requirements.
Do
we
choose
to
say
this
is
a
minimal
set,
or
do
we
make
them
optional,
so
met
try
to
capture
that
in
the
this
is
how
we
leave
the
discussions,
but
thank
you.
This
is
all
intertwined.
Q
So
my
point
there
is
about
what
you
need:
an
information
model
for
means
when
you
eat
one
right,
because
the
next
time
somebody
comes
along
and
wants
to
add
an
assertion
by
adding
asking
Ayane
or
whatever
are
you
gonna
say?
Sorry,
you
can't
until
you
give
us
an
information
model,
and
hopefully
the
answer
is
no,
that's
not
the
intent
right.
So
that's
what
I
was
getting
at
by
saying.
I
think
it
is
useful
to
have
a
document
that
says
here
are
patterns
and
structures,
an
example
of
patterns
right,
if
you're,
defining
a
new
information
model.
Q
If
you're
it's
going
to
be
lets,
say
a
numeric
one:
here's
how
to
define
it
in
such
a
way
as
it
can
be
represented
in
JWT
and
CWT
and
so
on,
because
there's
a
mapping
here,
so
you
don't
have
to
specify
that
so
it
becomes
a
lot
easier.
So
you
don't
need
a
new
information
model
because
all
the
mapping
work
is
done
for
you,
so
something
that
specifies
mappings,
but
that
you
can
do
that
without
specifying
claims
right
and
so
so
far.
Q
I
strongly
support
everything
that
is
clean
agnostic,
whether
we
need
the
actual
claims
in
the
document.
I
am
less
sure
about.
Okay,
maybe
maybe
not,
but
what
we
do
need
is
a
way
to
map,
because
one
of
the
main
points
of
an
information
model
is
to
allow
you
to
easily
have
multiple
formats
and
potentially
even
met
between
them
in
an
intermediary
or
support
both
of
them
say
in
any
protocol
entities
right.
So.
E
Q
E
So
I
will
speak
as
team
co-chair.
Sorry,
we
had
discussions
there
as
we're
building
dependencies
in
teep
for
leveraging
the
LLC
sure
for
now
and
one
of
the
discussions
which
I
don't
want
to
open
up
here,
but
we
eventually
have
to
have
that
discussion
and
that
goes
to
the
IANA
structure
and
to
Dave's
comment
and
Oh
Dave
wheeler
and
Dave
Taylor
and
they're.
E
O
So
I'm
primarily
focused
on
definition
of
claims
of
definition
of
assertions.
I
think
every
I
mean
I
think
that's
the
interesting
work
to
figure
out
what
what
assertions
and
claims
we're
going
to
have
and
write
them
out
and
make
them
really
clear
and
pretty
and
so
understand
how
they
should
be
currently
with
the
the
eat
draft.
It
puts
all
the
the
information
about
a
particular
claim
in
one
place:
I
I'm,
not
sure
I.
O
X
O
O
So
if
it
that's
probably
the
most
critical
most
important
thing
for
me
and
the
way
the
current
information
model
draft
looks
is
it
looks
like
it
wants
to
split
that
so
half
of
the
half
of
the
description
of
the
location
claim
would
be
in
the
information
model
and
the
other
half
would
be
in
the
eat
document
and
then
maybe
in
the
yang
document,
I,
don't
know
I,
don't
understand
the
vision
for
that.
But
that's
the
thing
that's
critical
for
me
is
that
the
definition
of
a
claim
in
a
coherent,
comprehensive
is
in
one
document.
So.
E
O
E
W
W
W
So
that's
an
information
model
and
we
can
point
to
other
instances
of
that,
but
it's
just
an
information
model
and
then,
when
we
go
say,
ok
I
want
to
define
a
claim
for
at
EEE.
This
is
this
is
now
a
document
model
that
uses
the
information
model
and
says:
ok,
there's
a
structure
I'm
going
to
define
what
an
attestation
token
looks
for
this
and
use
the
information
model
and
the
common
defined
elements
to
say
this
is
what
this
is.
W
E
So,
logistical
e
because
again
I
want
to
come
back
and
and
kind
of
sort
of,
but
really
be
black
and
white.
So
just
so,
you
guys
know
I'm
gonna
cut
the
line
at
Ming.
Ming
is
on
the
jabber.
Okay,
all
right
me,
deco,
Kevin,
confused
so
again
to
be
black
and
white.
My
initial
question,
and
preferably
yes
or
no
right,
is
that
you
and
Lawrence
since
you
are
agreeing
to
the
death
whatever.
That
means
right
is
what
I
was
trying
to
make.
E
O
E
E
K
I
heard
that
I'm
not
sure
what
Nancy
means
when
she
says
what
the
structure
looks
like
is
there
content
in
there
is
a
semantic
text
in
the
structure,
or
is
it
not,
it
can
have
semantics.
So
that
is
interesting
because
it
didn't
I
didn't
hear
that
from
the
text,
so
I
would
find
it
drastically
the
disastrous
if
we
split
information,
model
elements
all
across
the
data
model
documents
and
have
them
there
and
I.
Give
you
a
good
example,
because
then
we
would
define
nons
four
times,
why?
K
Why
not
define
it
at
a
central
point
and
just
refer
to
it
and
say
and
I'm
and
my
data
model
Ananse
is
represented
in
this
way:
data
moral,
wise.
It
makes
absolutely
no
sense
to
me
to
redefine
this
all
over
again,
and
that
is
why
I'm
sorry
fighting
against
you
back
to
back,
because
that
makes
no
sense
at
all
to
my
W.
W
Y
E
Y
Okay,
so
just
for
fun,
so
just
a
couple
of
comments
on
mines
on
the
information
model
where's
a
climber
first,
one
I
agree
with
different
wheeler
and
other
said
information.
Martin
to
me
said
worried
come
on
for
my
define,
just
structure
types
and
saw
the
claims
are
really
relevant
to
the
specific
domain
I.
We
might
have
have
some
common
claims
taken
locations
on
identifiers,
but
I
was,
it
should
be
extensible
of
different
workgroup
and
a
different
domain
defined
the
own
claims.
Y
So
what
you
you
mean
that
I
ain't
I
would
take
over
that
or
I
was
considering.
We
may
have
a
separate
document
for
separate
domain.
A
claims,
for
example,
team
may
have
a
say,
called
claims.
The
document
formally
in
the
same
information
model,
the
structure
types,
for
example,
like
a
tea
claim,
a
party
that
t8
terminology
not
appear
currently
80
and
Carnegie
information
model
trapped,
I,
see
right,
so
we'll
make
some
claim
what
immunity
is
Russian
vendor
or
saw
deformation.
So
that
will
become
a
claim
statement
in
some
document.
Y
So
attitude,
then
then
next
question
is:
how
do
we
know
all
claims?
You
said
model
right.
You
very
model
consists
on,
shall
we
have
a
registry
somewhere
in
this
workgroup
arts?
Is
that
or
go
to
IANA?
So
that's
a
open
question
I'd
like
to
ask,
but
if
a
claim
that,
in
my
comments
are
thinking
aloud,
which
is
extension,
it's
a
for
different
time
and
we
can't
define
its
own
set
of
a
clamps.
We
intercept
a
talk.
Y
F
Right
so
so,
just
to
summarize
we're
really
driving
for
making
making
some
decisions.
What
we're
gonna
do
is
publish
these
on
the
list
so
that
we
can
encourage
continued
feedback
and
comment,
but
these
are
essentially
the
questions
that
we
want
to
want
to
ask
and
have
the
working
group
find
answers
for
which
is
in
terms
of
an
information
model.
Is
it
a?
Are
we
driving
toward
the
standalone
draft?
Are
we
driving
for
something
that's
merged
into
an
architecture
draft?
Is
it
something
that's
merged
into
the
eat
draft,
or
is
it
something
else
entirely?
F
We
had
some
discussion
about
whether
it
should
be
distributed
across
multiple
drafts
or
not.
These
are
all
really
important
questions.
I
think
what
I
also
heard
on
the
discussion
was
hey,
there's
something
like
domains
or
some
some
sort
of
context
in
which
we
think
that
there's
a
need
to
support
some
extensibility
in
terms
of
the
types
of
claims
that
can
be
expressed.
F
We
also
heard
that
there's
a
need
to
have
some
structure
and
pattern.
That's
common
across
everything.
That's
that
that
captures
some
of
these
other
concepts,
like
you
know
how
things
are
grouped,
how
they're
nested
that
sort
of
thing.
So
we
want
to
continue
this
discussion.
The
intent
for
today
was
to
just
open
the
discussion.
We
want
to
try
and
continue
this
until
we
and.
Z
E
F
E
AA
E
L
E
AA
The
work,
of
course,
is
described
in
the
draft,
but
there's
also
a
document
on
the
TCG
website,
which
gives
a
more
or
less
the
same
thing
in
a
in
a
different
format
with
like
colors
and
pictures
and
stuff.
If
that's
of
interest
to
anyone
and
that
document,
if
from
TCG's
point
of
view,
is
out
for
public
review
at
this
point,
okay,
so
what
I'll
do
is
talk
a
little
bit
about
what
what
we
see
as
at
a
station
at
TCG
and
then
we'll
go
into
some
of
the
documents.
AA
The
the
problem
statement
is
actually
kind
of
straightforward.
We
would
like
to
know
whether
the
software
running
on
a
particular
device
is
authentic
software
by
whoever's
definition
they
want,
whatever
software
was
put
there,
that
that's
what
is
there
and
it
hasn't
been
hacked
when
the
thing
boots.
We
can
do
this
using
a
measured
boot
process
and
we
use
the
TPM
a
way
to
store
the
results
and
report
the
results
and
sign
the
results.
AA
What
we
need
to
do
is
define
the
rest
of
the
workflow
that
goes
around
it.
The
the
ppm
part
is
has
been
done
and
stable
and
frozen
for
some
time,
but
the
protocol
part
has
not,
and
that's
the
part
we
want
to
work
on
here
and
of
course,
as
we
do
this,
the
protocol
work
should
be.
The
workflow
should
be
extensible
enough
to
fit
with
other
use
cases.
Other
applications
in
that.
Of
course,
what
the
rats
charter
calls
for
I
will
say.
The
last
bullet
here
is
an
important
one.
AA
The
the
way
a
DPM
works
is
defined,
and
we
need
to
retain
compatibility
with
that.
There
are
many
of
them
out
there
so
that
that
puts
some
constraints
on
what
we,
what
we
can
do
just
to
back
up
a
step
as
to
what
this
all
means.
Measured
boot
in
this
world
means
that
a
device
might
I'll
say
router
all
the
time,
because
I've
worked
for
a
router
company
device
would
have
an
operating
system,
software,
a
processor
of
some
sort
and
also
a
ppm
as
the
thing
boots.
AA
Each
stage
of
the
boot
measures
just
takes
a
hash.
Whatever
it's
going
to
run
next
and
stores
that
ash
in
the
TPM,
there
are
bazillion
details
that
go
with
that
Fudd,
but
essentially
that's
what
happens
if
he
PM's
job
is
to
receive
those
hashes
and
guard
them.
So
they
can't
be
hacked
later
a
verifier
and
come
along
and
say
to
the
TPM
hey.
AA
AA
Sense
we'll
look
at
this
again,
but
there's
a
source
for
software
which
might
be
the
device
manufacturer,
although
it
could
be
someone
else,
there's
the
device
that
we
are
at
testing.
There's
the
verifier.
The
information
on
from
the
device
manufacturer
should
go
to
the
verifier,
so
it
knows
what
to
look
for
then
the
verifier
asks
the
question
to
say:
who
are
you
and
what
do
you
have
and
the
device
returns?
The
responses
I
should
say
in
ppm
speak.
A
quote.
AA
AA
Why
is
this
hard?
Maybe
it
isn't,
but
one
part
of
it
at
least
for
what
we
do
in
the
router
business
is
telling
whether
you're
running
the
right
software
doesn't
matter
much.
Unless
you
know
you're
asking
the
right,
router
you're
not
going
to
put
your
hands
on
the
thing
it's
you're
off
in
the
ether
somewhere,
there's
some
device
out
there.
There
have
been
a
number
of
embarrassing
circumstances
that
got
summarized
through
a
thing
called
the
a
Zygon
attack,
oh
I'm,
supposed
to
stand
on
the
red
X.
Yes,
oh
Wow,
okay,
yes,
so
we're!
AA
AA
AA
AA
We
need
to
first
identify
verify
the
identity
in
this
work,
we're
using
we're
leveraging
the
I,
Triple,
E,
dev,
ID,
stuff
and
use
the
key,
which
is
in
the
TPM,
to
verify
the
TLS
connection
and,
having
verified
the
connection
that
we
have
the
right
device,
then
we
can
go
ahead
with
the
next
stage
of
asking
for
quotes
so
that
we
can
then
verify
that
whatever
is
on
the
box
is
what
we
thought
was
going
to
be
there.
We
also,
of
course,
need
the
reference
measurements.
AA
AA
So
those
are
the
those
are
the
flows
that
we
want
to
be
able
to
do
with
this
to
support
within
TCG
we've
collected
this
together
under
the
name
of
remote
integrity,
verification
informally,
it's
been
a
remote
attestation.
For
years
and
years
we
started
using
the
Rif,
which
I'm
told
should
be
changed
to
River,
but
we're
using
this
acronym
to
describe
this
particular
collection
of
bits
and
pieces
which
can
be
put
together
in
a
certain
way
to
do
this
function
right
so
within
within
TCG.
AA
AA
This
one
is
focused
on
network
equipment
because
I
work
for
a
router
company
I
think
it
extends
to
embedded
systems
pretty
easily
of
more
general
sort
and
probably
two
other
things.
It's
not
meant
to
close
off
by
just
that
at
the
moment.
This
is
what
we're
focused
on
and,
of
course,
the
point
of
being
here
is
we
want
to
coordinate
this
work,
so
there
are
a
lot
of
pieces
that
go
into
this.
AA
The
important
parts
here
I
think
it's
obvious,
that
all
of
the
documents
that
relate
to
how
a
TPM
works
are
and
should
remain
in
TCG,
but
if
it
involves
protocols
and
models
and
communication
and
so
on,
then
it
probably
belongs
here.
So
we're
trying
to
figure
out
how
to
get
that.
That
wants
to
happen
the
the
stack
that
we're
working
with
at
the
moment.
This
is
kind
of
a
draft
of
how
we
can
put
all
these
pieces
together
has
both
TCG
documents.
AA
AA
He
did
did
a
mid
4
for
this.
The
same
query
response
to
quote
stuff
and
it
was
a
tough
project
right,
so
I
think
if
we
can
focus
on
this
stuff
in
IETF.
That
would
be
great
there's.
Another
half
of
this,
which
I
think
is
out
of
scope
for
rats
right
now,
which
is
how
you
handle
the
reference,
measurements
and
I-
think
we'll
probably
discuss
that
more
later.
But
that's
not.
T
AA
AA
There
are
a
number
of
documents
on
this
work
at
TCG
and
the
draft
and
the
and
the
version
on
the
TCG
website
names.
A
number
of
the
documents
TCG
process
is
a
little
different
from
here,
so
some
of
them
are
already
published
and
freely
available.
Some
of
them
are
members
only
still
I
know
there
will
be
a
number
of
companies
in
this
room
that
are
members
of
TCG.
So
for
those
people
you
can
just
go,
look
them
up
right
and
some
of
the
ones
that
are
in
process
are
coming
pretty
close
to
public
review.
AA
AB
AA
Tcg
process,
the
document
is
developed
inside
the
organization
until
its
you
know,
fairly
cooked,
and
then
it's
put
out
for
public
review
and
then
published
openly.
So
this
this,
the
document
referenced
at
the
very
beginning
of
this
talk
was
unusual
in
that
it's
been
published
early
in
the
cycle,
so
that
we
could
do
this
work.
AA
What
do
we
do
next?
We
would
of
course,
appreciate
help
in
clarifying
the
workflow.
There
are
a
number
of
assumptions
that
people
who
have
been
plugging
away
at
this
at
gcg,
for
so
long
can't
see
anymore.
So
helping
us
understand
what
it
is
that
we
need
to
explain
to
get
the
picture
right
to
feed
into
you
know,
through
the
use
cases
into
whatever
it
is
at
rats
produces
to
ensure
that
we
can
support
this
case
would
be
helpful.
I
think
we
Justin
I
did
not
write
the
security
considerations,
section
and
I.
AA
Think
a
discussion
of
the
trust
model
probably
fits
in
there
pretty
well,
and
of
course,
what
we
want
to
ensure
is
that
the
is
that
what
comes
out
of
rats
works
with
this?
With
this
model,
there
are
a
lot
of
PPM's
out
there
and
if
we
can
start
using
them
actually
for
attestation,
that
would
be
really
cool.
I
think
I'm
done
questions
hey.
Q
Dave
Taylor
thanks
for
presenting
this
I,
have
a
clarifying
question
you're
both
of
your
diagrams.
It
was
the
figure
to
figure,
for
they
both
showed
a
tester
and
verifier,
but
they
did
not
show
a
relying
party.
And
so
my
question
is:
is
the
relying
party
the
same
as
the
verifier,
or
is
this
just
talking
about
the
evidence
flow?
This
is
our
fog,
'''l
what'd
you
do
to
the
providing
party
which
one
did
you
mean
it.
Those
two
from.
Q
E
AC
AC
AC
Q
E
O
O
F
AA
E
M
M
AA
AA
E
A
Q
A
pest
ie
be
member
and
the
person
here,
oath,
I,
be
wiki,
page
technically
Bernard,
authored
the
text
and
then
I
put
the
site
on
the
page
in
order
to
get
a
liaison
person
liaison
manager,
then
the
IB
has
a
specific
set
of
questions
that
they
like
to
answer
and
among
those
and
I
can't
remember
all
of
them.
I
think
they're
on
the
public
wiki.
Maybe
Roman
has
that
path.
Yeah
yeah
yeah
that
right!
So
it's
a
like
what
do
you?
What
would
the?
Q
A
Q
Y
Q
AA
E
E
E
E
Okay,
so
the
reason
why
I'm
saying
this
is
I
guess
I
should
have
started,
but
the
first
question
is:
we
talked
about
teep
dependencies
on
brats
and,
having
that
use
case,
there's
a
long
discussion
so
which
will
be
ongoing.
It
wouldn't
be
a
bad
idea.
I'm
looking
at
the
tea
co-chair
to
see
if
he
cannot
of
having
a
joint
virtual
intern
interim
of
I'm
thinking
an
hour.
So.
Q
E
So
let
me
ask
the
question
value
in
the
joint
tea
and
rats.
If
you
object
just
raise
your
hand
going
once
twice
okay,
so
we
will
put
a
doodle
poll
to
put
that
one
together
and
I'll
caucus
with
that.
Okay,
so
for
the
rats
working
group
items,
given
that
there
was
a
lot
of
discussion
and
we
still
have
a
ways
to
go
visa
via
the
information
model,
would
it
be
worthwhile
to
have
one
versus
two?