►
From YouTube: RATS Architecture Design Team, 2021-04-02
Description
RATS Architecture Design Team, 2021-04-02
A
Good
morning
hi
dave
I
had
hoped
to
have
more
time
to
do
some
foot
requests,
but
that
didn't
happen
so
yeah.
I.
A
B
Output,
okay
and
I
thought
we
could
walk
through
the
rest
of
hank's
tickets,
that
he's
opened.
A
I
asked
at
the
end
last
time,
given
that
I
haven't
done
the
work
yet
and
it
sounded
like
you
and
ned
actually
had
opposite
suggestions,
and
I
don't
have
a
suggestion
yet
until
I
actually
try
to
do
the
work,
because
I
I
may
find
that
that
one
of
them
is
more
correct
than
the
others.
When
I
actually
try
to
move
text,
but
okay,
but
the
head's
not
here
yet
so,
and
that
today
was
the
day.
I
wasn't
sure
if
hank
was
going
to
make
it
giving
us
good
friday.
B
He
said
he
would
try.
He
seemed
to
have
implied
that.
D
Hello,
my
name
is
yang
and
I'm
from
a
mobile
research
institute
and
now
I'm
starting
the
trusted
computing
and
the
remote
adaptation.
I
just
in
the
group
for
a
few
weeks.
I'm
still
keep
watching.
B
A
A
It
should
be
short,
it's
not
maybe
trivial,
but
okay,
so
one
thing
was
trivial,
which
is
deleting
the
yeah
yeah,
all
right,
okay
and
so
actually
number
three
is
closer
to
the
second
comment:
it's
a
duplication.
Second
number
two
is:
can
we
find
a
term
that
doesn't
say,
considerations
are
mitigated?
Is
it
you
know,
security
threats
or
security
issues.
C
B
Spent
spun
right
so
all
right,
so
let
me
commit
this
suggestion
so
that
we
yeah
fix
double
words
and
comma.
A
There
wasn't
a
comma,
but
that's
okay.
It
was
just.
A
A
A
B
I
would
say
that
threats
is
probably
right.
Okay,
that
looks
good
to.
A
B
Yeah,
okay,
okay
and
your
other
comment
is
security,
physical
security,
the
factory
part
of
an
implementation.
I
think
so
yeah.
A
A
A
B
I'm
thinking
like
we
want
to
say
both
code
and
infrastructure,
such
as
at
the
factory
yeah,
the.
A
B
A
Three
of
the
previous
comment
was
really
coming
on
the
word
proposal
in
1291,
yeah
and
so
yeah.
I
could
ask
the
same
question:
is
physical
security
of
a
factory
considered
part
of
a
specific
proposal.
B
A
So,
if
that's
the
intent
right
that
you're
trying
to
convey
you're
not
expecting
the
working
group
to
ever
see
a
specific
proposal
ever
right
because
nobody's
going
to
send
the
ietf
a
proposal
for
how
to
secure
physical
factory
right-
yes,
okay,
so
you're
not
trying
to
say
okay,
this
is
a
requirement
for
say,
internet
draft
that
describes
the
protocol
or
anything.
That's
not
what
you're
trying
to
say
here,
because
that's
only
a
tiny
subset
of
this
well.
B
So
you
know,
let's
just
pick
on
fido,
for
instance,
let's
say
fido
you
know
has
a
protocol
that
does
everything
that
for
fido
to
then
not
to
to.
I
would
expect
them
to
then
say-
and
this
relies
on
this-
some
kind
of
level
of
security
in
the
factory
and
that
if
you
fail
that,
then
the
protocol
breaks
and
they
should
say
that
in
their
considerations.
A
Well,
I
agreed
agreed
okay,
so
so
I
guess
the
text
here
is
not
wrong.
It's
just.
I
guess
the
question
is:
what
is
the
expectation,
because
one
reader
could
misread
this
by
saying?
Okay,
when
I
run
writing
internet
draft
or
photo
writing
another
standard
or
whatever
I
have
to
mitigate
all
of
these,
and
the
third
statement
could
be.
I
have
to
implement
all
of
these,
and
both
of
those
would
be
incorrect
readings
of
your
intent
so.
B
A
A
B
I
don't
know
I'm
trying
to
wing
something
here:
some
threats
apply
to
protocols
and
some
threats
are
against
physical
infrastructure.
Well,
there's
such
as
factories
there's.
A
And
an
example
of
one
that
applies
to
code
is
the
text
that
talks
about
protecting
stuff
inside
of
it
protecting
code
inside
of
a
tee.
We
actually
have
some
code
about
that.
Sorry,
some
text
about
that
and
then
a
comma
after
your
code,
parentheses.
B
A
A
Well
hold
on
hold
on
before
you
do
that
you
have
two
sentences
there
using
implementations
in
both
sentences
and
implementations
has
a
different
meaning
in
those
two
sentences.
B
A
Yes,
yeah,
your
red
spell:
checker
must
be
set
to
uk.
B
Yeah
yeah,
okay
yeah
that
looks
okay
to
me.
B
You
go
yeah,
so
I
was
hoping
to
ask
hank
about
this.
I
didn't
couldn't
figure
out
his
connection
and
if
anyone
else
did
figure
that
out
leave
it
on
the
screen
here,
let
me
read.
B
A
Okay,
I
understand
what
the
original
text
meant.
I
understand
your
comment.
I
don't
understand
hank's
thing,
so
I
think
I'm
with
you
right
now,
yeah,
okay,.
B
Most
of
these
seem
to
be
suggestions
for
improvement,
rather
than
knowing
what
we're
gonna
do.
Yep,
and
you
had
your
question
that
you
wanted
to
go
back
to
about
your
well.
A
I
wanted
you
to
be
able
to
read
ned's
comment
and
see
if
you
have
any
comments
on
ned's
comment,
so.
C
Yeah,
so
basically
we
last
time
we
discussed
the
possibility
of
writing
a
section
8.0,
which
is
kind
of
an
introduction
to
what
we
mean
by
a
conceptual
message.
C
C
You
know
a
particular
message.
In
other
words,
in
the
context
of
describing
a
message
we
threw
in
some
explanation
of
hey.
This
is
a
conceptual
message,
but
without
really
saying
those
words,
so
what
I
was
trying
to
do
is
say,
take
trim
those
you
know
two
or
three
sentences
from
the
intro
put
it
into
section
8.0
and
then,
and
then
substitute.
C
You
know
the
the
specific
message
name
with
the
with
you
know
the
text
conceptual
message
so
that
a
prize
applies
broadly
across
all
of
the
messages
and
then
what's
left
over
is
it
is
sort
of
a
well?
You
know
one
way
or
the
other,
so
I
opted
for
putting
all
of
the
conceptual
message:
definitions
together
in
the
section
below,
because
the
the
you
know
overview
explanation,
sort
of
in.
B
B
So,
whereas
I
had
argued
for
removing
endorsements
from
here
and
moving
it
up
to
here,
you're
arguing
for
are
you
are.
Would
you
rather.
C
A
Yeah
yeah,
so
right
so
in
ned's
proposal
we
would
move
three
one
and
three
two
down
to
like
eight
four
and
eight
five
or
some
other
intelligent
reordering,
as
you
point
out,
if
there's
a
better
way
to
reorder
them
but
yeah,
I
would
be
fine
with
that.
A
D
B
Hank
is
joining
us.
I
I
guess
I
don't
feel
strongly
one
way
or
the
other,
but
I
see
that
the
current
state
is
poor.
I
I
feel
strongly
about
getting.
A
I
I
have
no
preference
on
which
way
to
do
it,
but
I
do
have
a
preference
if
somebody
else
is
going
to
do
it
that
that
suddenly
gives
me
a
preference.
A
C
A
There
he
is,
there's
hank,
you
know
I'll
probably
want
to
wordsmith
some
stuff
afterwards,
but
I
can't
tell
you
ahead
of
time
what
I'm
going
to
do,
but
in
terms
of
the
moving
the
text
around
massaging
it
and
taking
first
cut.
I
appreciate
you,
volunteering.
E
Which
is
fine,
quick
question
to
the
remaining
sorry
by
the
way,
I'm
late,
because
I
did
not
find
the
invite
did
not
find
it
by
looking
for
addressing
for
links,
so
I'm
very
sorry
doesn't
seem
to
be
my
inbox.
I
don't
know
why
the
request
that
is
still
open
did
we
abandon
that?
Did
we
push
that
back
to
lawrence?
There
was
a
reply
recently.
E
Yes,
and
so
my
question
is,
is
what's
our
open
action
there.
A
C
A
E
Oh
yeah,
that
might
be
also
the
case
so
so,
but
maybe.
E
B
Yeah,
well,
we
I
we
we
worked.
We
we
dealt
with
an
introduction
that
I
added
to
the
security
considerations,
saying
they're,
not
really
we
they
need
to
be
evaluated
against
a
real
protocol
or
a
real
implementation.
E
B
A
A
Have
an
opinion
on,
I
think
my
opinion
on
hank's
question
is
yes
and
opinion
on
lawrence's
suggestion
is
no
in
terms
of
taking
it
out,
but
yes
requires
an
introduction
or
reference
or
something
the
by
analogy.
The
type
architecture
document
does
use
the
same
term
and
I
think,
has
more
of
context
introduction
type
stuff.
A
But
the
other
point
is
that
that
term
is
not
only
in
that
section,
there's
other
sections
that
also
use
it
so
like
the
verifier
section
and
trust
model
also
uses
that
same
term,
and
so
it's
wider
than
just
a
section.
This
issue
is
wider
than
this
section.
It
just
so
happens
that
the
sections
that
hank
points
to
is
the
first
occurrence,
and
so
that's
why
it's
fine
to
file
it
that
way,
and
so,
if
you
want
to
assign
this
one
to
me,
that's
okay,
because
I
know
what
we
did
in
teep.
A
I
can't
use
the
same
approach
here
and
keep
there's
a
much
more
natural
approach,
because
there's
a
whole
section
in
the
teep
architecture
on
who
has
what
keys,
okay,
and
so
the
trust
anchor
store
would
be
used
in
context
by
referencing.
That
particular
discussion,
which
we
don't
have
a
similar
discussion
in
rats
architecture,
so
I
had
to
come
up
with
some
other
way
to
do
it.
But
given
that
I
did
the
text
in
the
keep
architecture
document.
B
A
Does
where
does
trust
anchor
store
show
up
in
this
section?
You
search
for
the
word.
Trust
anchor
store
it's
in
the
second
paragraph
of
the
verifier
subsection,
which
I
think
is
point
four,
all
right.
That
is
twice
in
the
in
the
section
that
hank
pointed
to
at
least
twice
there
at
that
testation.
A
A
B
A
A
A
B
And
I
remember
we
said:
oh,
this
is
such
a
pig
in
a
blanket
or
I
don't
know
what
the
word
is
so.
A
It's
defined
in
context
as
a
component
that
is
intended
to
provide
trustworthy
device,
identity
and
protected
storage
for
measurements.
That's
our
informal
definition
of
root
of
trust
right,
but
it's
only
under
typically
right.
So
it's
it's
not
formal
right,
so
I
might
be
able
to
do
something
similar,
but
I
think
that
there's
already
definitions
elsewhere
in
other
documents
about
trust
anchor
store,
I
might
be
able
to
just
reference
so
well.
A
It's
used
a
lot
in
this
paragraph
yeah,
but
it's
a
different
kind,
because
root
of
trust
is
like
the
piece
of
hardware,
for
example,
whereas
trust
anchor
store
is
the
place
that
you
stick
your
root
keys,
for
example,
so
yeah.
E
Or
your
certificate
paths.
A
B
So
all
right,
so
let's
look
at
some
other
issues
here
this
one
I
clicked
on
the
tab
by
random.
So
here
we
go
a
few
words
on.
Why
is
that
that
is
the
case
are
needed.
I
don't
even
understand.
A
By
the
way
the
text
quoted
is
old
text.
We've
updated
that
if
you
remember
to
talk
about
the
confidentiality
is
not
necessary
for
the
public
key.
So
there's
an
if
here.
E
And
the,
if
is
the
problem,
if
it's
totally
it's
just
a
claim
and
and
it
is
phrased
so
abstract
that
first
of
all
I
was
like.
Why
am
I
reading
this
and
then
okay,
probably
there's
a
reason,
but
I
can't
see
the
reason.
So
my
question
is
in
the
issue
now.
C
So
they
were
saying
that
there's
other
text
that
hasn't
been
merged
yet
or.
B
C
A
E
Exactly
and
if
it's
so
low,
it's
hard
for
reach
for
me,
because
the
webex
elements
are
over,
it's
zoomed
in,
thank
you
so
yeah.
It
seems
to
be
the
same
thing.
So,
basically
the
weird
sense.
If
public
key
cryptography
is
used,
then
only
integrity
is
necessary,
which
I
can't
pause
seriously.
What
was
that.
C
B
B
A
Okay,
so
maybe
we
can
replace
both
of
those
last
two
sentences
with
a
with
a
sentence
about
the
opposite
parity,
which
I
might
even
merge
it
with
the
previous
paragraph.
So
let
me
just
say
it
out
loud
starting
okay.
No,
I
can't
say
it
out
loud,
okay,
so
I
was.
A
Yeah,
go
ahead
all
right,
so
I'm
going
to
start
with
the
previous
paragraph,
I'm
going
to
try
to
read
it,
read
the
two
sentences
and
then
add
a
third
sentence,
that's
kind
of
what
I'm
thinking.
So
it
could,
because
I'm
proposing
to
merge
the
two
paragraphs
so
in
general,
a
combination
of
some
physical
security
measures
and
some
cryptographic
measures
is
used
to
establish
confidentiality
protection.
A
E
A
Yeah
yeah:
okay,
if
the
validation
key
is
uses
public,
private,
key
cryptography,
meaning
you
got
to
keep
pair
right,
then
then
confidentiality
protection
applies
to
the
private
key.
D
A
B
Dave,
I
think
I
would
go
on
and
and
and
add
another
point,
which
is
that
there
may
be
privacy
reasons
to
keep
the
the
public
key
confidential.
But
there
are
no.
C
C
A
D
E
B
E
B
I
I've
heard
what
dave
said
if
and
I
I
believe
I
can
remember
it
long
enough
to
put
make
a
pull
request
so
I'll
take
that
on.
Where
did
that?
Go.
E
D
B
See
so
that's
why
I
have
the
problem
with
not
there
okay,
so
I
will
sign
it
to
me
or
go
on
all
right.
I
think
this
is
self-explanatory.
We
just
need
somebody
to
write
it.
C
B
A
B
B
B
E
Yeah,
I
was
surprised
how
not
appropriate
the
example
was.
It
was
like
I
need
to
even
read
this
yeah,
for
example,
so
for
the
last
paragraph,
an
entity
that
both
connects
to
a
wide
area
network
and
to
a
system
bus
is
talking
on
both
the
test
and
verifiables
yeah,
whatever
that
exactly
means,
so
an
system,
bus
connectivity,
and
so
so.
E
This
is
basically
a
entity
can
take
on
so
take
on
both
at
the
same
time,
the
role
of
a
test
and
one
layer
of
rats
and
then
the
layer
of
the
role
of
verify
another
layer
of
reds.
These
are
actually
not
collapsed
roles
in
a
single
layer,
which
I
think
what
this
is
about
right.
E
A
So
here,
why
not
just
delete
this
paragraph?
If
now
that
I
say
that,
partly
because
another
example
was
added
to
this
section
recently,
and
so
two
paragraphs
above
was
the
text
that
used
to
be
in
the
reference
values
section,
and
we
said
we
should
move
down
to
here,
because
it
was
combining
of
roles.
So
this
was
text
that
was
an
example
elsewhere
that
moved
into
the
section.
So
we've
already
got
a
new
example
in
this
section,
so
I'm
thinking
it's
actually
easier
to
just
delete
the
paragraph
it
doesn't
like
so,
which
makes
sense
to
me.
E
They're,
like
me,
I
also
I
seem
to
have
done
this.
E
B
I've
been
put
trying
to
drag
myself
away
through
the
catch-22
sequel:
it's
not
as
good
as
you'd
think.
B
Interested
sorry,
is
it
just
an
aside,
there's
a
there's,
a
there's,
a
a
case
where
the
character
randomly
he's
charged
with
censoring
letters
from
the
you
know,
troops
back
home
and
he
just
randomly
censors
letters.
Words
from
the
you
know.
Like
every
second
letter,
he
censors
for
fun
right
see
if
the
letter
still
makes
sense.
Sorry.
E
E
Is
fine
but
a
real
world
example
where
trust
is
handed
off
from
the
I
don't
know,
fires
to
the
bootloader
to
the
kernel
or
something
like.
Oh,
I
see
now
so
now,
because
that's
your
chain
of
trust
and
jen
of
trust
was
also,
I
think,
mentioned
already
in
a
comment
somewhere
else.
So
that
would
be,
I
think,
something
to
do
here,
but
it
was
so
invasive
that.
B
Is
there
is
there,
is
there
a
diagram
and
or
text
that
we
could
lift
and
quote
from
some
other
document
ccc
or
I
don't
know
what
something
I
don't.
A
Think
you
need
a
diagram
because
I
think
the
existing
diagram,
because
the
whole
point
is
there's
already
a
a.
B
A
You
just
need
a
list,
you
know
a
b
and
c
so
it
goes
from.
You
know,
rom
to
the
first
stage.
Bootloader
to
you
know,
if
you're
talking
about
the
normal
world,
it
would
be
to
something
like
you
know:
goofy
or
you
know,
grub
and
linux
and
so
on
or
in
a
te.
You
might
have
you
know
going
through
the
trust
you
know
rmtf
a
and
to
opt
into
ata.
B
So
I
would
change
this
paragraph
continuing
with
this
example,
or
it
put
it
after
this
thing,
a
real
world
example.
It
would
be.
We
already
talked
about
bios,
but
you
know.
B
I
I
think
that
what
you're
saying
is
that
it
while
it
tries
to
be
it,
has
some
specifics,
but
at
the
same
time
it's
trying
to
be
abstract.
So
it
doesn't
mention
yuffie
and
it
doesn't.
It
says
something
about
the
kernel,
but
you
know
it
doesn't
say
who's
colonel.
So.
B
D
A
A
read-only
bios
and
because
we
re-label
these
right
so
then
b
was
was
b.
The
kernel
I'm
trying
to
read
this
sentence
here
because
after
we
because
remember
it
was
originally
different
and
we
said
well
that
was
wrong
and
then
we
updated
it
so
that
a
is
now
the
read-only
bios
b
is
let's
see
c
is
the
kernel
b
is
what
the
firmware.
E
B
This
is
inconsistent
now,
actually,
because
b
is
the
bootloader
and
c
is
the
kernel.
A
E
If
you
just
keep
at
least
for
one
second,
that's
because
could
you
scroll
up
a
little
bit
again,
because
I
do
not
see
this
introduces
me
an
example.
That
was
my
problem.
I
think
it's
it
starts.
There
is
an
illustration,
and
then
it
starts
with
the
details.
It
doesn't
say
why
we
are
not
talking
about
illustration
that
if
you
scroll
below
the
illustration,
it
starts
integrating
the
details
of
the
illustration
for
two
very
long
sentence,
paragraphs
for
three
actually
and
then
it
says
this
example
could
be
extended
and
contributed.
E
A
Okay,
I'm
gonna
fix
this
right
now.
So
give
me
two
minutes.
E
A
A
B
C
Dot
yeah
and
I
think
at
one
point
we
had
had
more
than
one
intermediate
layer,
but
then
we
removed
it
because
the
the
diagram
was
too
big.
A
Okay,
so
I
I
okay,
I'm
telling
you
what
I'm
thinking,
I'm
thinking
that
the
that
it's
easier
just
to
walk
through
a
specific
example
and
then
end
with
a
statement
that
says
in
general.
There
could
be
an
arbitrary
number
of
layers
and
other
implementations
right,
but
walk
through
one
example
first
and
then
say
how
to
generalize
it
rather
than
trying
to
come
up
with
a
picture.
That's
a
generalizable
one.
I
don't
know
that
there's
much
value
in
having
a
general
picture
and
a
specific
picture
both.
A
B
B
C
A
Yeah,
I
I
think
that
there's
already
text
that
does
that
at
the
bottom
of
the
section
here
I
can
read
it
to
see
if
it
needs
any
tweaks
or
whatever,
but
I
think
that
the
last
two
paragraphs
in
this
section
already
do
that.
B
A
B
I
think
we
can
trust
you
on
the
one
sentence
to
get
right,
so
I
think
it's
better
to
look
at
it
all
as
one
okay.
In
that
case.
A
A
B
B
A
B
Somebody
we
didn't
assign
it
to
somebody,
okay,
okay,
I
will
try
to
do
that.
I
think
I
remember
what
we
discussed.
B
D
E
I
have
to
also
investigate
that
a
little
bit
trust
model.
B
E
A
The
last
sentence
that
was
quoted
was
not
that
one
was
about
a
different
point:
the
sentence
before
that.
I
think
yes
ned,
but
I
don't
remember
so.
There's
two
sentences.
A
Already
the
last
sentence
there,
I
don't
think
it's
covered
any
place.
Sells
I
mean,
maybe
you
can
find
something,
but
the
last
sentence
there
is
talking
about
the
case
of
okay.
I
use
a
relying
party,
and
so
I
want
a
tester
right,
so
I
learned
some
things
about
their
lying
party
by
using
it
as
a
tester
myself
right,
I
learned
what
type
of
what
what
the
claims
need
to
be
in
order
for
him
to
grant
my
request.
Rather
than
rejecting
my
request.
A
Okay,
now
that
I've
done
that
now
I
can
watch
the
traffic
and
see
what
whether
other
requests
from
other
people
are
getting
accepted
or
rejected
if
they're
accepted,
then
I
know
that
they
must
have
sent
claims
that
were
similar
to
the
ones
that
I
sent
and
there
I
can.
Therefore
I
can
infer
things
like.
A
B
C
We're
we're
just
saying:
hey,
there's
a
potential
for
privacy
consideration
because
using
using
you
know
various
forms
of
observation,
you
can
infer
what
the
what
the
appraisal
policy
is.
A
And
therefore,
from
the
appraisal
policy
you
can
infer
the
value
of
specific
claims.
Is
this
last
sentence
there
right
so.
B
B
Well,
if
I
go
get
my
passport
validated
and
the
claims
are
put
in
my
passport,
then
that
could
be
from
a
timing
point
of
view
completely
disconnected
for
when
I
I
I
communicate
to
get
my
traffic,
I
mean
if
you're
watching
the
internet
and
I'm
watching
a
movie,
then
you're
sure
about
it
right.
But
no.
But
but
I'm
saying
if.
A
C
A
To
observe
the
traffic
to
the
relying
party
right,
just
the
resource
request
that
you're
using
the
attestation
to
grant
or
deny
right
saying
if
you
know
that
certain
values
were
accepted
by
the
relying
party
right,
you
can
tell
that
they
were
by
whether
the
request
was
accepted.
So
you
know:
did
the
http
get
return
actual
content
or
did
it
return
an
authorization
denied
right
or
I'm
just
giving
you
an
analogy
or.
B
Or-
or
you
observe
you
don't
observe
any
of
that,
but
you
just
observe
this:
the
the
quantity
of
traffic
results
afterwards,
it's
presumably
larger
when
they
get
the
movie
in
lower
if
they
don't
yeah.
So
so
what
what
I'm
trying
to
say
is
that
there
are
would
could
be
mitigations
where
you
download
the
whole
movie.
B
So
so
that
defeats
the
traffic
analysis
pattern
and
there's
other
ways
of
doing
that
in
passport
type.
Things
so
probably
support
that
better
than
background
checks,
because
if,
as
you
just
said,
if,
if
I
go
in
to
walk
up
to
the
gate-
and
the
first
thing
they
do
is
reach
out
to
an
obvious
background
checker,
then
I
know
that
that's
happening
right.
That
point
or
the
observer.
D
B
E
E
E
C
E
B
You
can't
go
forward
all
right,
so,
let's
see
here,
do
we
have
any?
This
issue
is
unowned.
E
Yeah
brother,
this
is,
I
remember,
I
don't
know
different
timeouts.
Apparently
I
I
don't
know
the
exact
problem,
but
I
really
remember
that
there
are
two
items
in
here
and
they
are
basically
the
same.
B
B
Secured
secured
locked.
E
The
limit
is
to
passionate
is
security
trains,
so
you're
just
specializing
limiting
and
you're
special
you're
abstracting
locked.
So
I
it's
sorry.
No
yeah.
A
You
prefer
combining
those
two
into
one
use
case,
so
just
involving
no
encryption
at
all
comma,
such
as
locking
doors
limiting
personnel
that
I
can
enter
the
facility
period.
E
I
would,
I
would
add
this,
I
would
say
like
like
locked
doors,
low
security
and
a
a
increase
as
a
dedicated
vetting
of
every
person
entering
the
residence.
But.
B
Low
security
case,
I
think
the
high
security
cases
is
much
more
obvious
and
maybe
was
covered
in
a
previous
paragraph
too,
and
so
I
think
we
just
need
to
say
well,
there's
a
bot,
the
bottom.
The
bottom
could
be
quite
low.
C
B
Want
a
range
of
of
of
scenarios
here
within
the
the
in
well.
I
think
that
we're
trying,
I
think
the
point
we're
trying
to
say
is
that
that
there's
not
a
one
secure
one
security
fits
all
that
there's
a
risk
benefit
analysis
that
you
have
to
make
and
that
for
some
devices
the
the
risk
is
low
and
the
benefit
is
low.
So
don't
do
a
lot.
C
B
C
B
A
Think
the
rule
that
it
was
being
used
in
the
document
right
now,
I'm
not
trying
to
defend
it
or
pick
on
it.
But
I
thought
that
the
rule
was
hyphenate
when
using
as
an
adjective
and
space
when
using
is
a
noun.
A
Okay,
well
all
right!
Well,
although.
A
B
B
A
E
Yeah,
but
not
at
the
first,
whatever
yeah,
so
that
is
that
is
some
pure
editorial
stuff
yeah.
So
I
I
don't
think.
B
C
I
think
it's
global.
B
B
C
C
The
thing
you
were
just
highlighting
an
example
of
it
being
referenced
in
text.
B
No,
it's
not
that
it's
it's
missing
the
reference
it
doesn't
have
a
it
should
have
like.
B
C
D
B
I
could
figure
it
out
yeah,
but
this
is
this
is
the
example.
So
then,
when
you
click
on
that,
you
go
down
to
this
thing
and
then
you
get
the
actual
okay
piece.
I
don't
I've
never
found
that
really
great.
I
always
want
to
go
directly
to
the
place,
but
anyway
that's
not
how
it
works.
Okay,
so
I'm
going
to
put
you
as
the
guy
here
and
then
I
think,
we'll
finish
our
call
and.
A
You
don't
need
a
reference
for
t-e-e,
because
I
think
we
copy
the
definition
into
this
document
that
was
agreed
on
by
the
t
working
group.
There's
different
definitions
out
there
and
the
one
that
we
use
is
the
one
that
the
ietf
defined
and
it's
consistent
between
ietf
and
ccc,
and
maybe
other
people.
But
it's
a
little
bit
different
from
say.
Global
platforms.
A
Definition
work
level
platform
defines
te
is
something
conforming
to
their
spec,
which
is
which
is
a
much
more
narrow
than
what
we're
talking
about
is
because
there
is
a
global
platform
tee
api
and
when
they
just
use
the
term
unqualified,
they
mean.
B
B
I
I
would
put
it
here
in
this
use
case
yeah
the
reference
to
t.