►
From YouTube: RATS Architecture Design Team, 2021-01-05
Description
RATS Architecture Design Team, 2021-01-05
C
So
should
I
see
a
changing
screen
share
right
now,
or
is
this
supposed
to
be
static.
A
You
should
I'm
just
taking
writing
down
the
attendance
okay
and
because
I
thought
we
might
have
another
person
or
two,
so
we
have
a
total
of
six
people,
but
I
think
we
have
quorum
of
editing
people
hi
peter
hi,
thomas
okay.
So
let
I
think
we
have
a
bunch
of
really
small
low-hanging
fruit
here
dave.
Did
you
get
a
chance
to
look
at
any
of
these
other
ones?
Yes,
you
did.
A
A
B
You
see
section
number.
A
B
And
I
agree
with
the
rationale
that
wavepen
had
put
in,
which
is,
if
you
expand
upwards,
it's
for
consistency
with
the
previous
section.
You
can
see
that
one
is
labeled
confidential
ml
model
protection
and
then
225
changes
to
protection,
and
you
can
see
227
says
this
is
a
generalization
of
the
ml
model
use
case
above
and
so
the
generalization
uses
the
same
word
as
the
more
specific
one.
B
A
So
okay,
so
where
is
this?
Is
the
fixes
231.
A
C
So
is
thomas's
refinement
of
my
proposal
reflected
already
in
dpr.
I
am
not
sure
thomas.
C
A
So
I
think
wade's
comet
here
is
pretty,
I
think
is
maybe
quite
pointed
I
don't
know.
Is
there,
can
you
explain
why
it
becomes
basic
to
abstract
thomas.
C
Yeah,
I
said
I
think
I
think
that
pros
and
cons.
I
think
very
nitty
comment.
Also
sorry
no
need
to
change.
The
comment
is
good.
That's
this
sorry!
I
already
I
missed
that
when
I
was
let's
so
so
the
so
the
I
think,
thomas
and
I
were
talking
about
unification
a
lot
during
that
phase
and
how
how
this
inter-opportunity
interoperability
works
and
therefore
it's
more
abstract
and
then
unifying.
C
But
yes,
ray
has
a
good
point
on
implementation
view
here
I
think
abstract
seems
unimplementable
and
so
yeah.
B
The
red
text
used
models
and
you
can
see
613
to
614
where
it
wraps
where
it
says
some
reference
models,
yeah,
and
so
I
kept
reference
models.
But
if
you
want
to
insert
the
word
interaction,
I'm
okay
with
that,
I
don't
I'm
not
asking
for
direction.
If
you
want
to
delete
that.
That's
okay
with
me
too,
but
I
left
it.
C
So
what
what
are
we
talking
about
here?
This
is
about
the
passport
and
no.
C
B
This
is
the
bridge
text
that
the
the
connecting
text
in
between
the
data
flow
diagram
and
the
two
models,
meaning
passport
and
background
check.
And
it's
saying.
G
B
B
B
Right,
okay
and
again,
that's
I
don't
know
if
it's
thomas's
or
somebody
else's
word
composition,
but
I
thought
that
was
fine.
There
in
6
14.
H
A
H
A
B
H
B
B
D
A
C
And
we
have
eric,
as
as
a
join.
C
A
Yeah,
if
you
do
it
wrong,
then
they
they
like
to
insert
their
own
meeting
link
because
they're
stupid.
A
Sorry
so
change
description
of
becoming
sections.
I
don't
know.
B
B
Okay,
I.
C
So
that's
that's
a
general
problem,
so
we
are
talking
about
model
model.
Is
a
relatively
over
loaded
term
patterns
is
relatively
unique
to
the
whole
text.
That's
the
final
editorial
pass
that
I
would
do
in
the
end
would
raise
the
question:
are
we
going
to
call
this
topological
models
or
patterns
in
the
end.
B
So
the
the
argument
that
I'll
make
in
favor
patterns,
which
is
why
I
said
it's
not
a
request
for
a
change
per
se
unless
you
guys
don't
like
it,
let's
read
613
describing
two
reference
models,
as
well
as
one
example:
composition
thereof,
there's
three
sections,
two
reference
models
and
one
composition.
If
you
say
a
pattern,
is
a
reference
model
or
a
composition.
B
B
That's
an
interesting
point.
I
I
hadn't
looked
at
608
before
top.
C
H
J
D
A
There
did
okay,
it
just
okay,
hey!
I.
D
G
D
B
B
B
F
A
C
A
A
A
Here
we
go
yeah
and
I,
when
I
upgraded
it,
I
I
neglected
to
buy
the
right
amd
processor.
So
I
had
no
video
video
on
the
motherboard
when
I
upgraded
it
like
six
months
ago,
so
I
had
to
go,
buy
like
a
whatever.
It
was
a
16
core
two.
You
know
thickness
video
card
because
that's
the
one
I
could
order
in
overnight.
A
A
A
B
Well,
if
you
want
to
reword
it,
you
can
reword
it.
My
point
is
that
if
you
don't
make
a
choice,
that's
consistent
with
the
relying
party's
decision.
It's
not
going
to
work.
A
Yeah,
I
just
think
the
out
relying
party
is
being
embedded
as
part
of
the
application,
the
bigger
application
scope,
but
I
I
I
both
see
your
point
and
I
see
why
it's
the
wrong
words
yeah.
So
I
I.
B
I
I'm
not
trying
to
defend
the
original
word,
I'm
just
trying
to
say
the
original
word
is
better
than
this
one.
So
if
you
want
to
fix
it,
we
can
still
try
to
fix
it
or
approve
it
or
whatever.
But
we've
got
to
keep
that
in
mind.
As
I'm
saying.
F
I
mean,
I
think
I
think
the
original
is
is
not
correct,
so
keeping
it
there
is.
As
this
is
not
it's
not
the
right.
It's
not
so.
B
B
B
So
what.
B
By
application,
if
the
relying
party
is
a
web
server,
what's
the
application
is
the
web
is
the
application.
The
is
the
application
web
access,
which
includes
both
the
browser
and
the
web
server.
Would
you
say
the
web
server
is
the
application
of
the
browser?
Is
the
application?
Just
depends
on
what
you
mean
by
application
right,
so.
B
So
use
case,
then
that
application,
which
I
think
was
michael's
or
somebody
else's
point
use
case-
is
better
than
application.
A
B
G
B
H
So
because
I
I
like
patterns
better
because
when
you
talk
about
variations
on
the
models,
you
know
that
starts
to
pattern
starts
to
be
the
more
correct
word
right
and-
and
I
like
the
relying
parties
better,
because
it's
a
more
general
notion
and
more
tied
to
the
actors
in
attestations
where
applications
is
way
too
general
and
can
mean
anything
to.
B
Anybody,
so
I'm
fine
with
that,
although
I
think
the
point
about
using
different
patterns.
B
A
H
It
might
be
that
and
the
tester
is
in
different
use
cases
or
for
in
support
of
different
patterns,
expecting
different
kinds
of
evidence
and
the
choice
of
the
pattern
dictates.
What
evidence
is
there,
and
so
it's
the
same
with
tester,
but
it
has
the
variability
built
into
it
to
provide
different
kinds
of
evidence
depending
on
what
it's
doing
sure.
G
H
A
B
I
am
not
convinced
of
that.
I
think
I
agree
with
peter's
point.
I'm
not
sure
I
can't
prove
that
you're
wrong
so
that
you
might
be
right.
It
might
be
a
depends,
I
don't
know,
but
I
can
imagine
verifiers
that
have
no
clue
whether
you're
using
background
check
or
passport.
They
just
take
in
evidence
and
send
back
attestation
results
and
whether
it's
a
relying
party
or
they're
in
a
tester
sending
it
to
them.
Does
it
really
matter.
A
F
I
was
trying
to
make
my
placing
that
that
text
there
is
that
you
can
have
the
same
entity
that
does
some
kind
of
processing,
some
verifying
processing
or
some
are
testing
processing,
while
interacting
using
different
interfaces
or
a
background
check
one
or
a
passport
right.
But
these
are
interfaces.
It's
not
it's,
not
the
fact
that
the
business
logic
that
is
done
there
is
is
done
by
a
different
entity.
F
It's
the
same
entity
that
that
handles
the
payloads,
except
it
you
know
it
receives
and
and
and
and
and
forwards
stuff,
using
different
interfaces.
That's
it.
B
Okay
right
now,
I'm
convinced
that
I,
like
michael's,
suggestion
to
remove
the
parenthesis.
It's
a
it's
actually
not
worth
pointing
out
because
it
generates
more
confusion
than
it
helps.
B
I
I
don't
think
that
the
it
matters
you,
the
rob
the
verifier
might
know
what
kind
of
model
it
is.
I
think
that
keeping
it
open
is
better.
H
H
A
So
I
think
that
the
point
trying
to
expand
here
this
is
a
combination
describing
combinations,
and
I
think
that
the
point
I
think
the
point
is
supposed
to
be-
that
you
don't
have
to
inter
invent
new
protocols
and
new
ways
of
conveying
evidence,
because
you've
decided
to
have
a
different
pattern.
B
B
B
My
point
earlier
is
that
you
have
to
have
one
that's
consistent
with
other
line
party
things,
so
it
doesn't
work
right
and
the
united
party,
I
might
argue,
is,
is
unique
to
that,
although
maybe
that
would
be
true
for
the
tester
too,
I
don't
know
how
to
think
about
it,
but
here,
if
we
focus
on
who
can
use
stuff,
that's
not
relying
parties.
A
I
think
that
the
relying
party
mostly
gets
to
specify
the
topology
or
the
pattern
that
he
expects
so
he's
in
charge.
From
that
point
of
view,
at
least
from
a
business
point
of
view
right,
he
says
this
is
the
part.
This
is
the
interaction
that
I
expect
and
use
it,
but
the
the
thing
that
I
I
thought
the
sentence
was
trying
to
get
at
is
that,
despite
having
different
patterns,
there
are
still
commonalities.
B
So
if
we
keep
relying
parties,
then
we
probably
have
to
change
pattern
back
to
mod
model.
This
is
partly
all
why
I
think
it's
actually
easier
with
the
sentence,
but-
and
that's
because
we
already
said
example
of
a
pattern-
is
the
hybrid
where
the
relying
party
is
using
background
check
and
the
tester
is
using
passport,
and
you
can
still
make
it
work
right
as
long
as
as
long
as
you're
doing
it
right,
then
you
can
construct
a
pad.
B
Whether
lying
party
does
not
know
that,
there's
that
there's
a
a
passport
model
on
the
other
side,
he
doesn't
actually
know
that,
and
so
in
that
sense,
it's
a
choice
of
model
that
the
online
party
makes
not
a
choice
of
pattern
because
he
doesn't
actually
get
to
know
what
the
texture
is
doing.
I
H
But
it
needs
it
needs
to
be
both
right,
and
so
I
think
the
core
concept
is:
there
needs
to
be
agreement
of
what
it
is
and
and
to
say
that
the
relying
party
is
the
superior
party
that
gets
to
decide
is
really
taking
away.
The
ability
of
an
attester
to
have
its
own
policy
decide
whether
it
should
participate
or
not,
and
so
yeah.
B
I
can
decide
whether
to
participate,
but
if
it's
gonna
fix,
but
it
has
to
send
messages
that
the
relying
party
will
accept.
H
Exactly
right,
so
so
there
has
to
be
an
agreement
that
says
once
we
engage
in
this
communication,
this
application
that
we're
going
to
do
that.
Both
sides
have
to
know
exactly
what
to
produce
and
what
to
expect
right
and
and
how
they
have
it.
G
H
Make
it
clear
that
where
you
don't
want
to
prescribe
that
in
this
model,
this
is
what
has
to
happen
right,
there's
a
variability
in
terms
of
what
the
exact
pattern
is
that
you
might
want
to
engage
in
and
and
then
once
the
pattern
is
chosen.
What
evidence
would
be
presented
using
that
pattern,
so
so
having
the
some
of
the
words
in
there
that
suggest
that
that
variability
is
possible
on
either
end
and
that
there
needs
to
be
agreement
is
really.
I
think,
the
point
that's
important.
A
H
Is
not
in
the
context
of
this
work
here,
but
this
is
why
we've
concluded
that
there
really
needs
to
be
a
negotiation
facility
to
have
both
parties
involved
and
have
more
complex
things
possible,
and
so
that's
outside
of
what
we're
talking
here.
But
there
largely
needs
to
be
that
and
whether
it's
in
a
given
use
case
prescribed
to
say
this
is
what
model
we're
going
to
use
or
whether
or
not
it's
discovered
during
an
active
attempt
for
a
resource
request
either
way
is,
is
okay.
A
D
B
B
B
The
yeah,
the
bottom
of
the
screen
in
white,
the
728
through
7,
now
731,
so
no
just
the
first
paragraph
yeah
that
part
right
there.
So
a
lot
of
your
point
is
covered
in
there
and
the
question
is
out
of
all
the
stuff
you
talked
about,
which
I
agree
with
a
lot
of
what
you
talked
about
is
covered
by
those
four
lines.
But
maybe
there's
a
point:
that's
not
in
there
and
if
there's
a
point,
that's
not
in
there,
then
that's
what
the
sentence
should
say.
B
A
B
Yeah,
so
those
three
we're
not
talking
about
deleting
that
the
question
is
what
you
we
got,
those
three
points
in
there:
the
evidence
for
different
line
parties,
evidence
for
different
use
cases
and
both
models
by
the
same
device.
Those
three
points
are
in
there
is
there
a
fourth
point
and
if
so,
that's
what
the
sentence
should
say
is
whatever
the
fourth
point
would
be,
because
otherwise
we
can
delete
726
and
just
replace
it
with
what
is
it?
That's
in
addition
to
those
three
points.
B
Right
now
it
looks
like
it's
redundant
with
7
328
to
731.,
meaning
the
the
green
text.
There
looks
to
me,
like
it's
trying
to
say
the
same
thing
as
we
already
say
in
the
next
paragraph.
B
A
B
A
G
A
B
A
A
Okay,
all
right,
so
I'm
going
to
commit
the
suggestion.
B
I
was
the
person
who
put
in
the
opcwa
stuff.
I
if
I
understand
the
updated
text
it,
meaning
the
proposed
new
text,
then
I
am
fine
with
this
other
than
you'll
see.
I
made
I
make
one
other
minor
suggestion
in
the
next,
but
if
the
point
is
you
can
make
the
same
point
without
referring
to
opc
ua
and
just
referring
to
tls
in
general.
If
that's
the
main
thing,
I'm
fine
with
that
approach.
C
You
you
removed
the
the
encoding
references
in
any
case
and
and
flattened
it
even
more.
Was
that
correct
dave?
Am
I
remembering
this
correctly.
C
With
my
yeah,
but
but
but
with
the
with
my
suggestion,
I
think
you
you,
you
wordsmith
it
again
and
you
removed
some
other
specific
ass
that
I
introduced,
and
that
is
in
the
green
text.
Now
right.
B
F
B
B
But
I
think
all
I'm
saying
is
I:
I
think
that
the
green
text
is
fine
other
than
1001
we
might
drop
opc
ua,
so
their
fc
editor
doesn't
ask
us
to
add
the
reference
back
in.
B
A
C
Okay,
yeah,
okay,
I
I'm
I'm
not
rooting
for
any
kind
of
removal,
but
I'm
confused
by
it.
It
did
now
now.
As
a
reader,
I
read
it.
I
was
like
okay.
I
have
to
look
that
up
now.
Everything
else
is
like
off-the-shelf
items.
I
know
opc
ua,
I
might
have
to
now
look
up.
I
actually
know
opc
ua
better
than
I
know
I
know
I
know
I.
H
B
C
Okay,
so
that
okay
yeah
you're
right
this
is
level
layer
specific.
Apparently
I
I'm
fine
so.
A
H
D
A
The
place
and
that's
why
I
like
the
opc
way
on
I
I
dave
I'm
sure
you
know
it
better
than
but
I
I
was
introduced
to
it
in
2019
and
actually
I
think,
they're
going
to
go
move
to
tls,
but
they
aren't
there.
B
A
B
A
Would
I
would
yeah
so
it
would,
I
don't
know
if
we
would
opc
ua
like
that's
what
I
would
do,
then
we
have
to
fix
the
then.
C
Tlscwt,
which
I
think
is
fine,
but
that's
also
a
little
bit.
This
is
like
the
experimental
realm,
so
it's
an
example
for
where
what
was
emerging
here
right
now
right.
So
so,
if
that's
intent,
I
am
fine
with
that
was
another
intent
here
for
including
the
certificate
type
in
the
the
text
of
two
four
one.
E
C
A
tls
certificate
type
that
is
very
specific,
I'm
not
complaining,
but
I'm
I'm
asking
why
actually.
B
C
B
C
Reason
to
keep
it
somewhere
could
be
the
tls
certificate
type
list
in
the
bottom,
because
this
is
how
you
actually
do
it
and
now
imply
that
maybe
somehow
there
is
something
but
having
a
explicit
reference.
So
that
would
be
a
good
reasoning
I
think
for
including
the
tscwt
id
here.
Otherwise,
it's
an
id,
and
so
I
think
it
has
this
emerging
experimental.
This
will
be
in
other
things,
also
flavor.
If
that's
the
intent.
That's
fine.
B
Everybody
uses
tls,
there's
tls,
there's
dtls
and
there's
stuff
that
uses
neither
and
each
of
those
could
be
using
one
or
more,
depending
on
the
on
what
supports
of
those
of
those
different
formats.
C
Yeah,
but
but
highlighting
above
that,
cwts
and
x549
pub
key
sweats,
for
example,
are
somehow
equivalent.
That
is
still
some
something
for
people
to
digest
in
the
itf
and
maybe
in
the
world
outside
properly.
I
guess.
G
C
That
there
is
something
emerging
there
that
is
converging
here.
That
will
be
it's
all
the
same
in
the
end.
That
is,
that
is
something
that
is
not
accepted
by
isg
effectively
today,
so
they
will.
I
am
relatively
sure
about
that,
but
it
will
take
one
or
two
years
so
so
going
with
an
idea
is
better
than
going
with
an
idea
here.
I
think.
B
Michael
the
sentence
that
was
in
between
what's
now
1001
or
1003.
do
we
need
something
to
connect
in
your
blank
line
there
there
was
a
sentence
in
between
and
the
red
text
up
above.
D
B
All
right
so
do
we
want
a
sentence
or
not?
I
I
don't
know
I
was
trying
to
figure
it
out.
So
you
notice
you
have
a
blank
line
there
that
looks
like
it
needs
some
connector.
So
the
suggestion
takes
up.
G
B
B
D
C
I
don't
it's
hard
for
me
to
parse
this
edit
to
the
edit
actually.
A
B
Other
people-
I
don't
know
if
you've
reviewed
the
green
text
starting
at
1003,
I
did,
but
I
don't
know
if
other
people
have.
B
C
B
E
A
A
Okay,
so
we
have
the
reference
here
to
opc
ua
tcg
dice
down
here,
so
we
kept
this
text.
C
C
Yeah
we
will
get
the
mustang
shot
feedback
right.
So
there's
that.
B
So
just
a
question
for
somebody
who's
an
expert
on
the
tls
handshake
in
line
1020,
I'm
looking
at
the
word
may
and
in
1020.-
and
I
just
want
to
understand-
is
there
exactly
two
ways
to
do
it?
Extensions
and
certificate
types?
Is
that
the
meaning
here
yeah?
You
have
to
choose
one
or
the
other,
but
there's
only
two
ways
to
do
it.
It's
not
those
aren't
four
examples.
The
four
examples
are
within
those
two
things,
but
there's
only
two
branches
for
for
the
2s
handshake
extension
is
that
right.
B
D
B
C
Cheers
but
the
the
topic
was
drop
opcua
and.
G
F
We
have,
we
have
itf
specs.
That
you
know
can
illustrate
the
purpose
of
this,
whereas
we're
referencing,
something
that
is
you
know
outside
outside
the
itf
world
and
it's
behind
a
paywall,
and
you
know
you
know
making
it
a
more.
It's
not
hard
to
pay.
B
B
Yeah,
it's
it's
not.
I
verified
it
when
I
did
the
reference,
so
you
know
all
the
teachers,
all
the
specs
for
the
published
ones
are
publicly
available.
Okay,
cool,
but
the
rest
of
your
point
is
correct.
Right.
I
agree.
That's
what
I
said.
It's
improvement.
Thank
you.
Okay,
cheers.
A
All
right,
so
I
stared
at
this
two
problems
that
I
had
had
self-assigned
a
month
ago
and
then
last
night
again,
I
think
I've
looked
at
this
three
or
four
times,
and
I
don't
know
what
to
do
with
this.
Ned
has
helpfully
or
unhelpfully
explained
the
term,
but
not
in
a
way
that
I
have
any
clue
how
I
would
decide
to
put
this
in
right.
A
So
I
mean
I
wasn't
vague
on
the
term
in
the
first
place,
it
was
guys
comment,
so
I
don't
know
what
to
do
with
this
one
and
the
other
one.
B
A
All
right
well,
perhaps
you'll
we
can.
A
At
least
I
can
tell
you
which
ones
I
had
problems
with,
that
did
the
same
thing
on
this
one.
So
here
he
wants
to
know.
Is
this
a
description
of
mutual
authentication?
The
point
is
that
the
tester
may
first
wanted
to
authenticate
or
have
attestation
from
the
the
verifying
party
as
to
its.
B
Is
this
a
description?
Mutual
authentication
answer?
No,
no,
I
don't
think
so
either
so
because
authentication
and
attestation
are
dif
have
different
meanings,
meaning
one
may
imply
the
other,
but
they're
not
synonymous.
B
A
C
A
A
Will
prepare
a
poll
request
doing
that
and
then
we'll
have
something
to
argue
about:
okay,
fine,
fair
enough,
all
right!
So
we're
going
to
continue
with
this.
Until
you
know,
new
year's,
as
by
the
during
the
year
of
the
rat,
and
so
next
week
same
time,
all
right
all.