►
From YouTube: RATS Architecture Design Team, 2020-03-27
Description
RATS Architecture Design Team, 2020-03-27
A
B
A
A
C
D
A
A
A
H
D
H
A
Right
so
I
I'm
I
I
have
some
doubts,
whether
at
14
meetings,
since
IETF
106
but
I,
know
don't
care
if
that
numbers,
a
precise
and
I
thought
about
highlighting
how
many
more
pull
requests
over
last
time.
But
whatever
the
point
is
to
give
people
notion
that
that
we
are
in
fact
not
sitting
around
doing
nothing
for
their
so
overview.
Presentation,
stick
something
relevant
here
in
the
right
hand
side.
So
that's
basically
what
I'm
I'm
talking
about
doing
I
wanted
to
bring
up
the
table
of
contents.
A
I
thought
about
sticking
some
of
the
issues
that
are
on
this
slide.
I
thought
about
sticking
them
on
top
of
this,
to
see
to
say
where
they
are,
but
I
I,
don't
know
whether
that's
just
too
busy
I
did
notice
that,
while
I
was
making
this
slide
that
the
numbers
changed
yesterday
yay
while
I
was
making
the
slide,
the
number
of
pull
requests
and
issues
changed
and
I
went.
A
Oh
that's
good,
so,
like
number
75
showed
up
from
you
Dave
yesterday
and
so
I
tucked
it
in
there
and
I
think
there
was
some
edits
to
the
defined
claim
that
happened
as
well.
So
I
just
stuck
this
there
and
I
think
the
point
was
to
my
point,
to
put
the
URLs
all
there
was
so
that
people
at
home
would
be
able
to
click
on
them
and
and
dive
deeper
into
it
and
see
if
there's
something
that
they
really
have
something
to
say
about
it's
entirely
possible.
H
A
H
F
A
H
Finished
you
have
a
slide
on
the
claim.
Stuffers,
that's
not
something
that
grant
spent
time
out
in
the
meeting
make
a
slide.
I
don't
have
a
slide
on
it.
You
should
least
mention
verbally,
whether
you
have
a
slide
on
it
or
whether
that's
gonna
go
to
the
list.
So
I
don't
know
that
you
need
a
slide
on
it.
The
question
is:
how
much
time
do
you
have,
and
you
have
too
many
slides
are
not
enough.
Slides,
okay,.
A
Well,
I'll
put
a
slide
at
the
end
and
we
can
decide
whether
or
not
we
want
to
have
a
slide
on
it.
Okay,
so
then
I
wanted
to
go
in
through
walkthrough
and
basically
wanted
to
walk
through
the
diagrams.
That's
was
my
goal.
Maybe
there's
a
diagram.
That's
missing
somewhere
I
would
love
to
have
throw
an
animation,
but
I
don't
have
that
much
clue
or
time,
and
so
I
wanted
to
emphasize
that
these
that
these
pieces
are
separate
but
are
not
subject
to
standardization.
Maybe
there's
another
way
to
do
that.
G
H
H
B
H
G
A
A
H
G
A
J
H
J
C
A
A
Was
waste
endorsements
would
get
to
the
verifier
by
out-of-band
security?
We
had
a
big
argument
about
that
at
some
point.
Yes,
but
I,
don't
think
it's
granularity
on
this
slide
to
talk
about
that.
Okay,
I
thought
if
I
didn't
say
that
that
someone
was
going
to
emphasise,
want
to
emphasize
that
they
have
to
be
signed
now
you
owe
but
does
the
appraisal
policy.
D
H
D
A
E
A
A
A
C
D
A
So
my
intention
was
to
so
conceptual
data
floats
so
that
there's
basically
I'm
in
the
PDF
there's
two
types
of
environments:
Sano's
that
contains
an
in
testing
environment,
the
target
environment.
Maybe
someone
would
feel
that
I
should
say
it
the
other
way
around,
but
maybe
that's
okay
and
then
some
people
ask
and
the
thing
we
trusted
to
measure
itself
and
our
conclusion
is:
there's
not
always
sense,
but
they
verify
knows
when
it
does.
Are
you,
okay
with
that
that
conversation
as
I
think
it
reflects
the
conversation
we
had.
H
D
A
A
D
E
E
A
So
my
goal
here
is
to
just
basically
stop
the
stop
the
war,
because
we
could
have
some
people
that
won't
have
this
in
their
space
and
there's
some
clearly
some
people
where
they're
like
well,
that
makes
no
sense
and
I
and
I
get
both
points
of
view,
and
so
I
just
want
to
capture
that
tussle
here,
right
that
that
there's
a
that,
we
don't
have
to
agree
on
this.
We
just
have
to
know
that
we're
not
confused
but
the
wrong
we're
not
applying
the
wrong
policy.
That's.
H
So
since
it
doesn't
make
sense
according
to
many
security
interpretations
or
whatever,
then
I
don't
think
you're,
the
architecture
should
go
out
of
its
way
to
allow
it
I
think
the
point
is
that
without
changing
the
architecture,
you
can
use
the
architecture
to
do
it.
Anyway,
we
have
a
sentence
in
there
that
actually
is
pretty
close
to
this.
H
E
The
question
of
whether
or
not
the
is
to
be
trusted
to
measure
itself:
it's
not
the
fact
that
it's
measure
net,
so
there's
always
a
place
for
the
verifier
to
ask:
can
it
trust
the
measurements
period
wherever
the
measures
were?
Was
it
something
that
happened
during
boot?
Was
it
happened
from
some
external
entity
that
was
measuring
and
watching
from
afar,
and
so
whether
it's
you
know
where
the
measure
is
and
is
the
measure
trusted,
is
a
meta
question
that
the
verifier
always
has
to
be
aware
of
and
and
I
think.
E
A
A
A
A
A
Okay,
so
next
one
I
didn't
come
up
with
any
bubbles
for
this
one
I
beliee
are
two
examples:
are
line
cards
and
a
chassis
aggregate
of
similar
systems,
or
he
had
many
routers
working
together
as
a
single
unit
and
I
thought.
That
was
a
good
way
to
write.
This
smart
phones
seem
to
be
composite
devices
where
the
we
collect.
The
one
CPU
collects
data
about
the
state
of
the
other
pieces,
whether
they're,
broadband
or
Wi-Fi
chips,
or
something
and
I
don't
know
if
we
had
a
I
couldn't
find
a
third
example.
A
H
E
So
you
could
consider
different
kinds
of
software
environments
so,
for
instance,
a
a
measurement
agent
that
was
doing
something
from
another
virtual
machine
in
a
virtualized
environment
or
even
in
the
smartphone
case,
something
running
in
trust
owns
looking
outside
of
trust
own
to
do
measurement.
Those
can
be
looked
at
as
composite
at
the
stations
you.
H
A
H
Realized
a
question
that
I
would
ask
if
I
would
see
this
for
the
first
time
and
I,
don't
know
if
it's
just
early
and
I
can't
think
right
now
or
if
we've
actually
discussed
it
before.
If
you
have
evidence
of
a
composite
device,
what
do
you
get
back?
Do
you
get
back
and
attach
station
or
results
of
a
composite
device?
You
get
back.
H
A
A
E
D
A
H
So
the
question
I'm
asking
is:
can
a
test
strip
be
a
tester
C
dot
I,
don't
know,
maybe
line
current
on
a
chest
or
something
like
that
and
all
of
those
BCE
things
have
their
own
network
communication
by
network
communication.
They
have
our
IP
address
and
able
to
come
Kate
with
relying
parties
themselves.
No.
A
A
H
So,
for
example,
let's
say
that
the
leta
tester
a
is
the
main
CPU
and
a
server
and
a
tester
B
is
a
specific
NIC
with
a
processor
in
it
yeah,
and
so
then
a
tester
B
could
do
its
own
communication
or
do
some
you
know,
management
link
or
whatever
it
is
independent
of
what
the
management
connection
sorry
independent
of
whatever
the
Leda
tester
a
is
doing
right
because
you
can
offload
stuff.
It
may
do
stuff
inside
of
its
processor
there.
E
H
So
I
think
the
answer
is
that
the
assumption
is
that
BC
do
not
have
any
network.
Is
the
only
communication
between
say
a
dystrophy
and
the
verifier
is
via
the
leta
tester
a
in
which
case
the
only
evidence
that
exists
and
the
only
communication
of
the
verifier
from
the
whole
composite
device
comes
to
leta
tester,
a
I
think
that's
what
this
diagram
is
trying
to
show?
Yes,.
A
I
think
so,
div,
and,
and
so
so
I
think
that
the
reason
why
we
have
internal
links
are
network
connections
is
so,
if
you
consider
a
tester
B,
yes
to
be
some
big
routing
platform
where
100
gig
links
great,
but
the
path
to
the
verifier
is
the
management
connector,
which
is
on
it.
Elita
tester,
a
right,
and
so,
although
it
could
do
Giga
bits
of
data
elsewhere
and
maybe
is
connected
to
a
tester
a
via
a
hundred
gig
back
plane,
which
is
IP
addressed,
it's
still
not
actually
externally
visible.
We're.
A
H
B
D
E
H
A
I
wanted
to
basically
yeah
I
wanted
to
match
that
I'm
instead
put
my
boxes
on
top
of
the
ASCII
text,
so
that
people
would
connect
directly
to
it.
There
was
no
like.
Oh
did
you
change
the
diagram.
Actually
I
changed
one
diagram
so
that
my
boxes
would
line
up
better
I
didn't
think
that
mattered,
like
I,
changed
it
in
the
document,
but.
D
H
A
A
A
H
A
H
A
E
H
It's
still
true
in
here,
because
this
is
a
tester,
a
B
and
C,
and
it
just
so
happens
and
the
evidence
of
composite
device
a
tester,
a
B
and
C's
evidence
are
combined
together.
They
compile
the
evidence
for
composite
device
is
not
in
his
diagram,
so
each
individual
evidence
is
specific.
With
the
evidence
of
device
you're
right,
it
may
include
multiple
evidences,
each
of
which
are
indeed
different
format.
E
That's
important
because
you
can
get
away
from
all
that
complexity
by
talking
about
that
that
agreement,
whatever
you've
agreed
to
whether
it's
negotiated
or
predefined,
will
specify
which
pieces
are
signed
and
by
whom
and
what
that
needs
to
look
like,
and
so
both
sides
need
to.
You
know
the
producer
of
that
evidence
needs
to
make
sure
that
all
the
cases
are
there
are
all
the
format's
are
supported
and
the
verifier
needs
to
be
able
to
interpret
them
all
so
that
you
know
we've
discussed
in
our
work
questions
about.
E
H
I
agree,
though,
I
think
what
this
diagram
is
showing
is
what
happens
at
the
end
of
that
which
is
eventually,
you
settle
on
a
particular
thing
that
you're
going
to
send
right
after
you've
done
pre-negotiation
after
you've
done
the
configuration.
However,
it
isn't
you
learned
it
in
many
cases,
the
the
device
on
the
left
that
sends
the
evidence
may
only
be
capable
of
generating
one.
E
The
point
where
the
way
we've
talked
about
that
is,
that
they're
still
at
the
negotiation
phase,
but
it
just
defaults
to
no
negotiation.
That's
what
pre-selected!
You
know
there
is
no,
no
actual
negotiation
necessary
because
you're
gonna
select
only
the
thing.
That's
there.
So
you
know
and
I
and
the
terminology
that
we've
used
I
know
it
makes
sense.
Is
that
there's
the
negotiation
phase
which
leads
to
a
selection
and
that
selection
there's
is
what
may
be
predefined
and.
H
E
H
E
H
B
H
I
H
J
A
Think
I
think
the
architectural
point
here
is:
is
there
a
negotiation
phase
and
if
so,
where
does
it
fit
into
the
architecture?
Okay,
is
it
part
of
the
architecture
that
the
architecture
is
defined?
There's
a
negotiation
here,
the
protocol
blah
blah
blah
well
I,
don't
think
we're
going
that
way,
or
it
says
it's
part
of
the
use
cases.
A
H
E
Is
not
that
there
needs
to
be
the
possibility
of
one
and,
if
there's
not
I
mean
even
in
the
cases
no
issue,
there
needs
to
be
some
notion
of
what
domain
of
interpretation
we're
talking
about,
so
that
when
you
express
it's
going
to
be
in
this
format
that
that
format
means
something
to
both
the
tester
an
appraiser.
It
doesn't
need
to
be
universal,
but
there
needs
to
be
some
way
of
saying
that
I
know
what
they
are
on.
Both
sides
is.
D
E
I
think
an
important
point
and
whether
you
put
it
in
the
slide
or
not,
is
this
idea
of
a
domain
of
interpretation,
because
if
there
isn't
one
there,
it
has
to
be
universal,
so
you
need
to
be
able
to
define
between
you
know
when
I
a
new
use
case
is
created
to
be
able
to
say
within
the
context
of
that
use
case.
This
is
what
we
mean
when
we
say
this.
A
A
E
It's
another
idea
documents,
but
the
general
idea
is:
it
says
that
it
specifies
what
the
words
we're
using
in
this
communication
mean
right.
So,
where
I
know
it's
there
for
four
factors
in
the
discussion
of
crypto
algorithms,
when
you
say
I'm
going
to
use
this
crypto
algorithm
I
need
to
define
what
domain
we're
talking
about
for
them.
So
I
said
camp
is
a
good
example
yeah.
So.
A
Is
a
camera
he
came
up
with
DOI
and
largely
regarded
as
one
of
the
biggest
mistakes
we
made
in
it
because
it
it
added
an
layer
of
indirection
in
the
document.
What
made
it
very
hard
to
reason
about,
and
no
one
ever
came
up
with
a
second
do
I.
So
that's
why
I'm
fearful
of
the
words
because
I
feel
it
has
a
bad
rap,
I
think.
H
H
E
E
The
question
comes
up,
as
do
you
believe
that
the
tester
and
the
relying
party
always
know
each
other,
are
they
even
in
the
same
administrative
domain?
If
that's
the
case,
then
you
have
a
lot
more
freedom
to
do.
What
you're
talking
about,
if
you're
gonna
say
to
strangers,
are
going
to
maybe
need
to
do
an
attestation.
You
need
to,
in
a
sense,
settle
on
a
vocabulary
that
you're
going
to
use
so.
I
I
We
are
having
a
problem
with
the
fact
that
this
conversation
is
so
stratospheric
that,
if
I
were
in
the
room,
I
would
have
no
clue
as
to
what
this
means
in
terms
of
how
I
implement
what
is
it.
You
know
it's
so
far
from
code
that
an
example
is
not
just
warranted
but
really
mandatory,
and
that's
part
of
the
reason
why
I
think
we
can
spend
what
people
will
spend
so
much
time
on
this
slide.
I
When
we
talk
about
you
know,
anytime,
we
see
a
tested,
proteases,
a
specific
format,
relying
party
it
demands
a
specific
format,
we're
at
the
ITF
right.
What
is
the
format
right?
What
are
the
format's?
We
know
how
to
negotiate
right
I
mean
if
it's
an
HTTP,
you
can
use,
accept,
headers
and
other
things
like
that,
but
we
are
so
far
from
that.
We,
you
know
there
needs
to
be
a
little
bit
of
grounding
here.
I
A
H
This
slide
is
about
the
sorry,
the
lines
on
here
from
evidence
attestation
results
or
what
the
data
flow
is
not
the
protocols.
The
protocols
are
on
the
slides
that
have
the
passport
model
and
background
check
model,
and
so
on.
So
you
write,
you
have
to
look
at
those
to
talk
about
protocols.
You
have
to
look
at
one
of
the
other
diagrams
to
talk
about
protocols.
I
I
A
A
I
E
Right
I
think
would
be
okay
here
to
to
say
that
this
is
an
area
where
still
discussing
and
that
you
know
that
this
some
room
for
for
change
in
further
definition
here,
because
I
think
we're
spending
a
lot
of
time
discussing
something
which
is
fairly
controversial.
That's
not
really
getting
you
toward
what
you
need
to
do
on
Tuesday.
A
H
The
words
don't
match
what
we
talked
about
it
so
I
understand
entus
right.
So
let's
say
that
there's
a
negotiation
protocol
that
happens
at
runtime
right.
Then
it's
not
true
that
the
document
specifies
which
format
it's
specified,
how
to
determine
which
format
right
in
which
you
already
say
in
the
bottom
left
bullet
of
the
slide
and
so
I
I
II
thurr
want
it
to
be
removed
or
I
want
to
be
consistent
with
the
bottom
left
bullet.
H
H
H
A
H
A
H
H
E
Think
you're
gonna.
If
you
try
to
make
a
slide
here,
it's
going
to
maybe
cause
more
problems
than
it
solves,
because
there's
going
to
be
a
lot
of
questions
about
variability
and
how
you
represent
that
on
the
one
slide,
and
so
please
find
something
like
that.
Have
that
picture
in
your
document,
I
would
hold
off
on
that.
E
I
think
so,
because
someone's
everyone
who
looked
at
is
gonna
see
what
they
want
to
see
in
it
that
doesn't
fit
what
they
have
in
mind
and
what
you're
really
trying
to
do
is
put
one
slide
together.
The
deals
with
their
ability
of
a
lots
of
these
cases
that
they're,
extensible
and
I
think
that
I
agree
with
the
idea
that
we
need
to
address
that.
But
if
you
try
to
do
this
in
a
hurry
to
fit
into
a
briefing
on
Tuesday,
you
may
hurt
the
cause
more
than
help
it.
E
H
H
My
question
is
to
Michael
when
you
say
how
this
all
fits
together.
I.
Think,
oh,
my
interpretation,
is
you
added
the
slide
in
response
to
Eliot?
Is
that
correct?
If
so,
then
I
agree
with
the
comments
that
depending
goes
on
here,
it
might
okay,
how
the
architecture
all
fits
together.
I,
don't
know
what
would
be
out
of
slide
when
you
do
that.
I,
don't
know.
H
I
E
L
E
I
A
H
I
think
the
danger
is
not
putting
in
too
much
detail
well
still
being
able
to
talk
about
how
to
apply
the
architecture
and
fit
the
other
things
together.
Keep
one
of
those
examples,
but
if
you
don't
use
the
word,
zero,
ERP,
I'm
thinking,
I
think
it
is
a
yes
without
necessarily
having
to
tie
it
to
teep
I,
think
you
could
generalize
it
slay
me
or
you
can
use
specifically.
Teep
is
the
point
to
tie
it.
You
have
to
tie
it
specifically
easier
to
either
background
check
or
passport
or
something
else
and
for
simplicity.
H
I
would
tie
it
to
something
that
is
just
background
check
or
something
that's
just
passport
model.
Teep
combines
the
two:
that's
why
I
think
that
you
get
some
at?
This
is
just
an
example.
Right
I
would
pick
an
example
that
should
pick
say
a
passport
with
I'm.
Just
gonna
make
something
up
here
right,
because
you
could
either
have
a
negotiation
or
no
negotiation,
so
I
might
pick
I'm
gonna
arbitrarily
pick
out
of
the
matrix
passport
with
negotiation
and
then
a
background
example
with
no
negotiation,
fine
and
just
say
yep.
These
are
just
two
arbitrary
example.
I
I
It
just
just
to
show
that
you
know:
okay,
we've
added
this
compose
out
of
these
components.
You
know
into
the
architecture.
You
can
see
how
the
architecture
is
laid
out
right
now,
here's
sort
of
a
real
way.
You
would
put
the
stuff
together,
one
example:
if
it's
just
unreal
real
right
and
then-
and
you
say
you
know,
here's
it-
here's
how
it,
because
that
will
help
people
relate
to
the
componentry
that
they
have
to
build
in
this
and
that
they
and
understand
what
they
you
know
how
they
can
use
this.
A
D
H
A
H
This
is
for
one
of
the
presentation,
so
we
had
a
bunch
of
discussion
and
lists
and
things
about
the
timing.
Stop
I
gather
we're,
not
gonna
have
a
chance
to
talk
about
that
before
the
actual
meeting,
and
so
we
might
be
in
our
internal
disagreements
during
the
actual
meeting
that
might
be
the
best
we
can
do.
The
timing,
chemo.
A
M
H
A
L
M
You
can
see
the
slide.
What
I
have
right
here
is
a
what
I
think
intersected
would
Dave
and
I
had
talked
about
thread
and,
during
a
brief
call
on
potential
times
of
interest,
as
events
at
times
of
particular
events
or
specific
events
that
occurred,
what
you
see
in
purple
are
ones
that
one
or
the
other
of
us
think
it's
important
ones
in
there
in
white.
We
both
think
are
important
and
the
open
questions
for
people
on
the
call.
M
H
H
Is
in
the
pluralist,
with
an
explanation
of
why
that
one
is
important
for
security
purposes,
the
AAA
HD,
RP,
RR
and
ya
are
the
ones
that
are
not
in
the
poor
requests,
because
I
removed
them,
because
I
didn't
have
any
text
that
actually
relied
on
them,
and
so
I
said,
if
there's
no
security
check
that
so
many
was
gonna.
Do
that
used
that
event
and
the
event
was
not
interesting
in
the
security
part?
So
there's
all
kinds
of
events
like
when
was
the
IP
header
added.
When
was
a
claim
a
done
versus
claim
being?
M
That
the
thought
were
interesting
above
had
been
placed
into
the
pull
request.
My
initial
attempt
at
timings
included
timings
that
heard
on
this
call,
but
perhaps
people
who
are
on
this
call
can
defend
them.
A
good
example
would
be
HD,
which
could
be
interesting
to
Hank
because
of
to
DES
and
the
possibility
of
having
some
central
nuts
distribution
time,
something
that
was
interesting
and
the
question
of
do
we
put
that
into
base
architecture
document
or
not,
is
something
it's
worth
going
here.
Additionally,
the
attest.
Our
awareness
was
something
that
I
thought.
H
Why
and
I
would
also
assert
that
until
somebody
else
generates
text
that
uses
one
of
them
in
an
equation.
The
other
ones
are
not
that
one's
a
weak
assertion,
because
I
can't
prove
the
negative
right.
It's
a
request
for
somebody
else.
True,
the
positive
and
I
will
accept
it,
the
other
for
if
and
only
if,
somebody
has
a
text
with
an
equation
that
actually
uses
that
value
in
some
security
check.
M
And
I
have
no
problem
with
that.
The
weakest
one
on
the
list,
I
think,
is
RP
RP
for
results.push,
it's
possible
to
generate
results
and
push
them
several
times,
because
you
might
different,
send
them
to
different
different
places
and
that
might
or
might
not
be
needed.
The
results
relayed.
That
is
an
interesting
one
where,
if
you
create
a
set
of
information
which
goes
from
a
ground
check
to
verify,
ER
and
then
relay
the
results
somewhere
else.
M
H
H
I,
don't
think,
there's
a
personal
opinion
by
us
here.
I,
don't
think
in
our
question.
Mark
should
be
a
question
anymore,
because
that
there
has
been
a
case
explained
in
pork
was
next,
unlike
the
other
ones,
so
I
think
they
should
be
either
white
or
in
no
question
mark
or
color
coded
differently
than
the
purple
ones,
for
which
that's
not
the
case.
H
My
point
is
whether
you
can
think
of
it
or
not,
is
not
at
the
point
of
the
car
coding.
The
color
coding
is
whether
text
has
been
generated
for
review
and
text
has
been
generated
for
a
review
with
a
diagram
and
explanation
for
the
NR,
but
not
for
the
other.
Four,
and
that's
why
I
said
it's
a
different
color
coding,
because
one
is
out
for
review
and
the
other
ones
are.
H
I
agree
agree
that
this
only
make
this
whole
slide
only
makes
sense
if
there's
a
slide
before
it
or
something
that
defines
the
concepts
that
are
sort
of
in
section
nine
right
now,
which
is
particularly
I,
mean
what
are
the
security
checks
you
want
to
do
right.
You
want
to
know
whether
certain
things
are
fresh.
You
want
to
know
if
you're
a
testing
result
is
fresher.
You
want
to
know
if
your
evidence
is
fresh,
you
want
to
know
particularly
is
fresh.
H
You
have
to
have
that
as
context
first
to
say
what
it's
up
to
security
checks.
You
want
to
do
right.
What's
the
threat
you
trying
to
mitigate
then
after
then
that
you
could
say
what
events
are
necessary,
because
how
would
I
do
that
check?
Well,
that
would
mean
such
and
such
information
to
do
that
and
here's
the
information
you
need
so.
H
H
What's
not
obvious
on
the
slide
here
is
how
does
the
how
these
times
use
so,
for
example,
a
relying
party
right
if
you
I'm
gonna,
put
something
here:
I,
don't
remember
if
this
matches
what's
in
the
text
that
I
wrote,
but
you
get
the
point,
it
says
the
relying
party
checks
that
time
er
time.
Ns
is
less
than
a
threshold,
for
example,.
B
M
M
M
D
That
has
the
the
purple
values
after
slide
four,
and
then
we
can
just
say
that
hey,
there's
some
other
ones,
there's
some
other
possible
ones
that
we're
considering
and
we
need,
you
know,
use
cases
to
motivate
them,
and
so,
but
but
so
I
think
it's
the
case
that
slides
three
and
four
aren't
using
any
of
the
values
that
are
perfect
purple
currently.
Is
that
correct?
That's
correct,
yeah!