►
From YouTube: IETF111-RTCWEB-20210729-2200
Description
RTCWEB meeting session at IETF111
2021/07/29 2200
https://datatracker.ietf.org/meeting/111/proceedings/
A
Okay,
our
intrepid
ad
has
gone
off
to
check
gather
town
for
the
other
working
group
chair,
but
in
the
meantime
we
can
go
ahead
and
get
started,
because
some
of
this
is
something
any
single
working
group
chair
should
be
able
to
handle.
Welcome
to
rtc
web
at
itf
111,
as
with
other
parts
of
the
ihf,
this
is
a
session,
is
being
recorded.
A
You
probably
know
how
to
join
the
meet
echo
since
most
of
you
are
here,
but
if
you
would
also
like
to
join
the
jabra,
that
is
also
possible
at
rtcweb
at
jabra.itf.org,
if
you'd
like
to
help
take
minutes
there
at
the
url
pictured
on
your
screen
or
at
the
tool.
Above
that's
the
note-taking
tool.
That
is
the
fourth
from
the
left
at
the
top
of
your
screen.
A
A
Here
is
our
agenda
for
today,
one
keep
smiling
we're
nearly
there.
In
fact,
we've
been
there,
we've
come
back
and
we're
about
to
be
there
again.
I
don't
think
we
need
a
jabber
scribe
per
se,
but
if
there
is
somebody
who
would
be
willing
to
relay
from
jabber
anybody
who
feels
that
they're
not
able
to
to
speak
to
them
directly
to
the
microphone,
for
example,
because
they
don't
wish
to
wake
up
their
household,
if
somebody
could
volunteer
to
serve
to
speak
for
them,
that
would
be
very
useful.
A
A
Jonathan
lennox
asks
in
chat,
I
was
hoping
he
was
volunteering.
Have
we
passed
the
rtc
web
10th
anniversary?
Yet?
Yes,
the
buff
was
in
march
of
2011,
so
we
are
now
past
the
10th
anniversary.
A
All
right,
well,
I
will
gene
provides
this
with
cake.
Thank
you,
gene.
I
will
watch
the
jabber
chat
as
best
I
can,
but
if
anybody
sees
anything
in
the
java
chat
that
needs
reflection
and
can
do
so,
I
would
appreciate
it
so.
A
The
status
of
the
working
group
recently
rechartered
hopeful
that
it
can
shut
down
again
soon.
The
only
thing
that's
on
our
agenda
for
today
is
the
update.
Justin
sent
the
slides,
probably
an
hour
or
two
before
the
meeting,
so
we
probably
not
too
many
people
have
reviewed
them
yet,
but
we'll
review
them
in
the
meeting
and
hopefully
make
some
tentative
decisions
for
a
confirmation
on
the
list.
A
A
B
An
irony
of
ironies,
the
camera's
not
working
for
me
here
in
meet
echo,
so
you
just
have
to
imagine
what
I
look,
like
apologies
for
being
late.
A
A
A
D
A
F
B
Okay-
let's
see
so
why
are
we
here.
B
The
recap,
from
the
last
time
and
bundle
both
had
their
own
thoughts
around
how
offers
and
answers
should
be
generated
regarding
bundling,
and
we
also
found
that
libwebrtc
differed
from
both
of
what
jason
recommended,
as
well
as
what
bundle
recommended
and
that
the
actual
impact
of
this
was
most
significant
when
using
the
max
bundle
bundle
policy
which
governs
how
bundling
is
actually
turned
into
an
offer.
B
B
B
We
also
said
we
were
going
to
update
jsep
to
you
know,
change
max
bundle,
since
max
bundle
resulted
in
behavior
that
we
were
no
longer.
You
know
wasn't
consistent
with
either
specification,
we're
going
to
basically
say
that's
deprecated
and
we'll
come
up
with
a
new
name
for
the
behavior
that
was
formally
described
by
by
max
bundle
christopher
suggested.
The
term
must
bundle.
I
think
that's
actually
pretty
good
to
the
point.
So
that's
what
I've
put
into
this
new
from
the
document
slide.
B
So
this
is
what
we
agreed
upon.
This
is
how
the
sdp
should
should
look
that
we'll
have.
This
is
balanced,
which
there
is
no
port
equals
zero
in
in
the
offer.
B
But
in
the
answer
you
see
the
use
of
a
shared
port
between
the
m
equals
audio
and
m
equals
video
online
and
there's
no
a
equals
max
or
a
bundle
only
anywhere
in
that
answer.
Stp
next
slide.
B
This
is
the
stp,
for
when
the
formerly
called
max
bundle
now
called
must
bundle
bundle
policy
is
used,
we
do
include
the
bundle
only
attribute
and
a
zero
port
in
amigos.
Video,
just
like
jason
had
recommended
before
and
and
in
the
answer.
There
is
no
zero
port.
There
is
no
bundle.
Only
again
this
was
the
the
json
policy.
The
bundle
has
now
yes,
bundle
now
basically
system
to
do
this
and
the
term
name,
the
terminology
or
the
name
of
the
monopoly
again
is
is,
must
pundle
rather
than
next
bundle.
B
So
here's
the
new
text
there's
basically
two
sections
and
I'll
just
kind
of
leave
this
up
here
for
a
bit
because
there's
a
bunch
of
a
bunch
of
words
here
for
folks
to
read,
but
this
basically
just
describes
you
know
what's
different
in
in
the
abyss
versus
you
know
the
88
29
I'll
sort
of
leave
this
here
for
a
second
for
folks
to
read
and
see
if
there's
any
questions.
D
This
looks
great,
I
think
one
more
sentence
might
be
helpful
here,
which
is
to
say
to
indicate
that
the
correct
behavior
is
the
behavior
that
formerly
was
max
bundle
just
to
like
avoid
confusion
right
assuming
I'm
not
mysterious
of
the
situation.
Yeah.
B
Sure
right
so
that
that
probably
be
good
to
put
into
the
intro.
I
think
on
the
next
slide.
When
I
talk
about
this,
the
specific
text
as
it
relates
to
must
bundle.
I
think
I
may
have
included
some
text
to
that
to
that
regard,
but
maybe
I
should
also
get
pulled
forward
into
the
intro
too.
B
B
Right
so
this
actually
talks
about
how
you
know
moss
bundle
is
the
you
know,
specified
text.
You
know
I
I
you
know
eric.
If
you
have
any
thoughts
on
on
what
specific
text
you'd
like
to
see
in
the
intro
section,
let
me
know
otherwise
I'll
I'll
take
a
shot
at
it.
B
Okay
sounds
good,
so
the
main
thing
is
just
making
sure
that
you
know
must
bundle
does
match
the
the
spec
behavior
or.
B
Anyway,
so
that
is
really
the
the
new
text.
There
is
one
question:
if
you
go
to
the
next
slide.
B
Roman
noted
this
in
the
email
discussion,
largely
you
know
summarized
that
you
know,
since
you
know,
subsequent
offers
do
not
include
port
zero.
They
have
a
shared
port
in
a
third-party
call
control
case
where
one
of
the
subsequent
offers
can
be
sent.
As
an
initial
offer
to
you
know
an
arbitrary
endpoint,
it's
possible
that
you
know
something
with
like
a
shared
port
could
get
sent
to
a
you
know,
something
that's
not
aware
of
bundle
and
then
we're
back
to
the
the
same
problem
that
we're
trying
to
avoid.
B
The
problem.
Is
it's
hard
to
avoid
this
this
issue
either
we
sort
of
move
when
we
have
port
zero
and
then
we
have
the
risk
to
existing
apps
that
we
sort
of
formally
talked
about.
You
know,
I
suppose
we
could
also
add
some
sort
of
knob
to
say:
hey
go
in
this
specific
case
generate
an
offer
that
uses
port
zero
and
bundle
only,
but
that
would
be
some
spec
work
as
well,
and
my
overall
feeling
here
is
that
you
know
bundle
has
been
extremely
successful.
B
As
noted
in
the
previous
presentation
at
itf-110,
you
know
99.9
percent
of
all
sessions
where
bundles
offered
are
actually
accepted
and
the
cases
where
it's
not
accepted
are
almost
always
just
kind
of
legacy,
audio
only
devices
where
bundling
doesn't
matter
and,
finally,
the
fact
that
in
a
third
party
call
control
case,
bundle
would
have
already
been
negotiated
and
you
would
have
a
single
transport
over
which
multiple
streams
were
being
multiplexed.
B
That
then,
like
the
remote
side,
would
either
have
to
accept
or
just
get
a
single
stream,
because
there's
no
there's
no
option
to
really
unbundle
at
that
point
in
time,
because
you
haven't
proposed
new
candidates
or
anything
else
like
that.
I
think
we'd
probably
just
sweep
this
under
the
rug.
B
F
So
again,
I
think
this
is
fine.
I
just
want
to
probably
just
have
like
a
note
saying
that.
Well,
this
is
not
like
any
subsequent
offers
generated
are
not
compliant,
cannot
be
used
as
initial
offers
bundle
and
do
whatever
you
want
for
that.
Something
on
those
lines,
just
note
saying
that
this
is
somewhere.
B
Okay,
do
you
think
that
should
go
into?
I
guess.
Bundle
already
has
some
text
about
this
point.
Yeah.
F
See
I
just
again,
I
just
I
need
the
warnings
for
somebody
so
that
they
know
that
this
thing
is
there
but
like
if
they
really
need
it.
As
we
discussed
on
the
list,
they
can
just
again
they
can
launch
sdp
or
but
again,
like
the
chances
of
anybody
needing
it
are
small,
so
that
doesn't
doesn't
need
to
be
part
of
the
spec.
Just
need
to
be
needed.
People
just
need
to
be
made
aware
of
that.
B
Okay,
move
on
to
the
final
slide.
B
Is
this
thing,
okay
yeah?
So
I
I
think
we
may
be
mercifully
at
the
end
of
the
j7,
with
it
rtc
web
journey,
unless
there
are
any
other
sort
of
questions
or
concerns.
C
A
little
quick
justin,
I
appreciate
that
it's
like
so
I
was
literally
like
is
this.
This
is
one
of
the
things
where
we're
done
so
quickly,
it's
great
to
be
great,
because
there
was
agreed
and
they've
all
read.
It.
F
Yeah
one
other
thing
is:
there
was
a
question
on
the
music
for
to
review
the
bundle.
So
if
people
who
are
participating
in
rtc
web
can
just
chime
in
and
said
that
we
looked
at,
the
updated
bundle
looks
good
to
their
tcp
rtc
web
participants.
That
would
help
probably
getting
this
thing
published
sooner
rather
than
later.
So
that's
the
only
comment
I
have.
G
Yes,
murray:
go
ahead,
hey
it's
still
so
weird
to
see
a.d
next
to
my
name,
even
though
it's
been
a
year
and
a
half,
we
have
a
document
that
was
left
over
from
the
original
work,
the
examples
document
it
was
never
published
it's
sitting
out
there
kind
of
in
a
dead
state.
I
don't
know
how
important
it
might
be
to
reconsider.
Do
we
want
to
do
it
here?
While
we
have
everybody's
attention?
Is
it
not
worth
doing?
I
have
no
idea,
I'm
not
invested
in
it.
I
just
came
up
in
conversation.
H
C
Because
I
remember
it,
we
decided
that
we
were
going
to
proceed
without
it
and
I
was
going
to
sit
on
the
side
and
then
it
was
going
to
be
ad-sponsored,
because
there
was
enough
information
in
there
to
provide
examples.
And
that
was
it.
So
that's
where
you
kind
of
decided
to
move
forward,
but
that
was
that
was
the
approach.
G
And
then
I
guess,
because
it
got
dropped
out
of
the
cluster
and
then
lost
like
it,
received
no
further
attention
after
that,
so
it's
kind
of
in
limbo.
I
don't
I,
I
don't
think
it
needs
to
be
published
if
nobody
thinks
it's
important,
I'm
just
I'm
sort
of
cleaning
up
loose
ends.
G
We
and
I
don't
know
if
suas
is
here,
so
I
shouldn't
speak
for
him,
but
I
we
had
decided
that
there
was
a
ton
of
information
that
was
useful
here
for
people
who
are
actually
trying
to
implement
because
of
annotations
for
what
all
the
things
you
need
to
implement
and
we
were
going
to
go
ahead
and
publish,
but
I
don't
know
what
we
were
waiting
for
on
that
we
had
agreed
to
get
the
other
stuff
out
first
and
then
it
probably
just
the
ball
got
dropped.
G
If
the
document
is
in
state
to
publish
my
preference
would
still
be
to
publish
it,
but
you
know
that
that's
what
I
thought
was
happening
like
that's.
That
was
my
understanding
of
where
we
were
so.
We
can
change
that,
of
course.
G
So
if
if
we
were
to
do
that,
is
it
reasonable
to
bring
it
through
the
working
group
for
one
last
review,
one
will
ask
we're
getting
your
last
call
and
just
do
it
that
way
or
like
sponsoring.
It
is
the
other
option.
But
since
I
have.
C
You
all
here,
assuming
this
working
group,
is
going
to
be
open
for
the
next
three
four
months.
I
think
that's
that's
probably
fine.
Okay
and
we
can
we
can.
We
can
run
it
through
this
process,
so
we
can
press
the
buttons
on
the
data
tracker
to
re-accept
it
in
the
working
group
and
move
on
forward
and
give
it
get
right
in.
G
C
We'll
close
the
loop
with
suas
to
make
sure
that
everything's
is
ready
to
go
so
the
document
that
we
last
call
is
the
one
to
go,
but
it's
really
it's
really
great
that
the
json
this
like
that's
we're
done
right
like
that's
it
we're
not
gonna,
have
to
meet
again
in
december.
That's
our
november!
This
is
fantastic.
A
The
other
thing
I
wanted
to
point
out
on
on
the
examples
is
it:
the
review
of
examples
is
quite
difficult
right
because
it
is
a
whole
bunch
of
very
detailed
sdp
review,
and
so
it
would
not
surprise
me
if
the
reason
it
got
stalled
is
if
we
sent
it
out
for
review
and
people
stalled
on
the
reviews,
and
it
didn't
come
back.
H
A
It
it
is
just
a
it's
a
whack
of
different
stp
arcana
to
get
through,
and
it's
great
to
have
all
that
sdp
arcana
collected,
but
it
it's
also
quite
difficult
to
find
the
right
people
to
to
review
it.
So
I
suggest,
if
we
are
bringing
it
back
in
in
into
the
process
that
we
solicit
reviews
today
and
say
anybody
who's
willing
to
review
it.
A
Please
send
a
note
to
the
chairs
today
before
you
leave
and
we
will
also
go
through
and
ask
the
same
of
m
music
as
soon
as
we
get
a
confirmation
from
sukhas
that
he
thinks
that
the
revived
draft
is
ready
for
review.
But
it
is.
It
is
a
non-trivial
review
task
honestly
because
it
is
a
a
significantly
more
sdp
review
than
textual
review,
and
so
you
really
need
to
to
sit
down
and
swat
up
so
to
speak.
G
I
mean
I
need
to
run
in
a
second,
but
just
to
follow
up
on
that.
If,
if
you
bring
it
into
the
working
group,
and
then
you
decide
there
isn't
enough
energy
or
just
no,
you
just
don't
want
to
bother
abandoning
it
is
just
fine
too
I
just
I
just
don't
want
to
leave
it
dangling
if
there's
a
chance
that
it
could
be
useful
to
somebody.
So,
however,
you
want
to
handle
it's
fine,
I'll
work,
the
chartering
side
of
it
to
make
sure
all
that's
legit,
and
then
we
can
proceed
from
there.
D
Prairie
I
would
like
to
I'd
like
to
propose
that
we
do
not
attempt
to
afford
advance
this
document.
I
have
two
reasons.
One
is
the
reason
that
ted
just
indicated
nameless
a
lot
of
work
and
the
second.
I
got
three
reasons.
D
The
second
is
no
doubt
there
are
defects
in
this
document
which
will
not
detect
review
and
will
then
leave
people
to
be
confused,
and
if
there
are
interesting,
rfc
it'll
be
very
hard
to
change
those
and
will
not
do
so,
and
then
it
will
be
wrong,
whereas
if
it's
in
a
draft
you
can
simply
publish
a
new
internet
draft.
Third,
it
seems
to
me
the
world
has
gotten
along
quite
well
with
this
being
an
rfc
and
so
like
expending
itf
effort
to
make
it
an
rfc
seems
like
a
really
low
value
proposition.
D
So
I
think
it'd
be
fine.
If
it
stayed
in
id,
you
know
it's
not
like
the
ids
go
anywhere.
We
can
perfectly
all
find
them,
and
you
know
so.
I
recommend
we
get
back
to
not
to
attempt
to
advance
at
darcy.
I
I
only
wanted
to
mention
that
I
did
an
attempt
of
reviewing
the
samples
draft
and
it
is
enormous
and
I
tried
to
do
a
careful
review
in
the
beginning
and
then
ran
out
of
steam
kind
of
like
midway
and
only
skipped
over
the
last
pieces.
So
yeah
I
mean
timothy
penn
is
suggesting
to
divide
and
conquer.
I
guess
so
if
at
all,
that
might
be
a
feasible
thing
because
for
one
person
it's
basically
I
don't
know
you
would
have
to
set.
C
Done
in
the
past
right,
it's
unfortunate.
This
is
this.
This
would
have
to
be
this
document,
but
there
there
are
times
where
they
be
split
up
the
review,
for
you
know
you
take
the
first
50.
The
next
person
takes
the
next
50..
So
if
we
can
get
a
couple
people
we
can
maybe
people
that
are
willing
to
review
it.
People
can
split
it
up
and
do
that
to
help
save
the
you
know
the
weight
of
the
review.
G
G
Do
that,
let's,
let's
just
review
the
stuff
we
didn't
get
the
reviews
for
on
the
last
time
was,
but
I
know
magnus
and
nels
sent
a
ton
of
really
good
comments
on
that
stuff
have
been
been
resolved,
but
that's
you
know
I
I
I
will
flag
that
when
you
ask
a
set
of
volunteers
to
do
an
incredible
amount
of
work
for
you
and
then
are
like
screw
you.
We
can't
even
be
bothered
to
publish
it.
It
does
not
encourage
volunteers
to
keep
doing
that.
For
you.
B
I
I've
spent
a
bunch
of
time
reviewing
examples
as
part
of
of
jsep
and
I'll
be
willing
to
do
some
review
here.
The
one
thing
I
would
say,
though,
is
my
experience
with
with
jsep,
where
we
actually
went
to
machine
generation
of
examples
was
that
even
with
review,
it
was
almost
impossible.
It
was
literally
impossible
to
find
all
the
issues
that
you
know.
The
machine
generation
actually
found
a
number
of
issues
that
we
thought
in
in
in
sap
that
we
thought
was
correct.
B
Otherwise,
and
so
I
I
sort
of
feel
that
unless
these
examples
are
machine
generated
or
machine
checked,
there
will
be
errors,
and
you
know
regardless
how
diligent
we
are.
So
I
I
don't
know
if
that's
something
you
know,
sue
house
has
spent
any
time
or
has
actually
generated
these
like
from
machine
or
anything
like
that,
but
I
I
just
have
some
concerns
that
you
know.
We
just
need
to
be
aware
that
everything's
gonna
slip
through.
D
So
yeah,
I
think
I
feel
the
need
to
repeat
myself
who
is
advocating
for
this
to
go
to
rfc.
I
hear
a
lot
of
people
sort
of
vaguely
offering
to
do
some
reading
if
it
has
to
happen,
but
who
is
strongly
advocating
for
that?
I
heard
colin
advocating
some
people
might
be
put
out.
Who
are
those
people?
Would
they
please
step
forward
because,
like
the
the
fact
that
we
sunk
a
lot
a
lot,
some
cost
a
lot
of
effort
into
doing
this
before?
D
I
think,
there's
not
like
a
strong
argument
again,
especially
for
the
reasons
just
indicated
so
like,
like.
I
understand
like
that,
you
know
people
may
feel
guilty.
I
feel
like
they
ought
to
do
a
lot
of
people
wanting
to
volunteer
but,
like
we
all
have
important
things
to
do
so.
I
must
so
like.
I
would
hear
someone
like
strong
advocate
so
tell
me
why.
E
To
second
justin's
concerns,
I
think,
having
looked
over
a
lot
of
the
examples
in
the
document,
I
doubt
if,
if
they're
automated
with
it,
you'll
find
a
large
number
of
errors,
probably
because
if
anything,
the
stp
has
also
changed
over
the
period
right,
we've
moved
to
new
unified
plan,
implementations
and
so
forth.
So
I
think,
there's
a
it's
very
likely
that
an
enormous
amount
of
work
will
be
required
here
and
that
just
the
auto
effort
to
automate
the
whole
thing
is
going
to
be
a
lot
of
work.
E
B
That
you
know
we
did
build
a
python
based
generator
for
jsab
and
if
that,
if
we
decided
we
wanted
to
proceed,
and
we
did
want
to
try
to
assure
ourselves
that
that
would
be
examples
to
be
correct.
I
I
bet
we
could.
Probably
you
know,
adapt
that
to
generate
a
lot
of
the
things
that
are
in
this
document,
without
a
ton
of
effort,
and
then
maybe
we
don't
need.
Quite,
as
you
know,
many
reviews,
because
we
can
be.
You
know
if
we
feel
the
structure
is
right.
B
You
know
the
generator
like
something
that
we
got
everything
right
for
jason,
which
I
I
feel
like.
We
spend
a
lot
of
time
on.
You
know
we
don't
even
look
at
every
single
online
as
do
the
ips
match
and
all
that
sort
of
thing,
so
I'm
I'm
offering
to
help
there.
If
we
decide
that
we
do
want
to
publish
this
document,
and
you
know
we
do
want
to
find
a
way
to
generate
these
things.
I
One
word
of
caution
regarding
the
automatic
generating
last
time
I
read
the
draft.
It
had
lots
of
lots
of
the
further
sections
basically
said,
like
oh
we're,
abbreviating
the
stp
here
for
the
sense
of
shortness.
Since
you
have
read
the
previous
examples
already,
this
is
leaving
out.
I
don't
know
like
the
bundle
parts
and
like
these
bits
and
pieces,
and
so
on
so
like
automating.
That
would
basically
change
these
samples
like
drastically
to
be.
H
Yeah,
I
think
I
sort
of
agree
with
ecker,
also
because
it's
not
quite
clear
to
me
who
exactly
the
audience
to
consume.
This
is
because
I
think
very
few
users
of
webrtc
apis
are
actually
hopefully
constructing
raw
sdp.
I
feel
like
that
would
be
you
know,
not
a
good
outcome.
You
know
for
random
api
users
to
be
doing
that
as
opposed
to
just
treating
it
as
an
opaque
blob,
so
or
else
using
some
library
written
by
somebody
who
actually
understands
sdp
to
do
the
munching
of
it.
So
it's
I
kind
of
agree.
G
I've
already
said
the
main
thing
I
had
to
say,
but
I
mean
just
on
the
the
implementers
like
people
who
don't
understand
sdp,
I
mean
there's
you're
right,
they're,
not
creating
stp,
but
every
time
I
go
to
debug
anything
with
webrtc.
I
have
to
look
at
the
sdp
and
I
have
to
be
able
to
try
and
understand
it,
and
if
our
answer
to
them
is
go
read
every
stp
spec,
of
course,
that's
impossible.
They
can
look
at
something
that
helps
them
walk
through
what
they're
trying
to
do
the
examples
and
some
sort
of
creative.
G
You
know
something
that
guides
them
through.
What
they're
trying
to
do
that
it
does
help
find
problems
and
sort
things
out
at
times.
I'm
not
saying
this
document
solves
that
or
doesn't
solve
it,
but
look.
It
was
a
document
that
was
constructed
to
try
and
show
what's
useful
about
the
sdp.
Not
not
be
complete,
worked
examples
of
sdp
right,
so
you
know
whatever
that
that
that
is
what
it
is,
but
I
I
do
think
everybody
is
trying
to
debug
a
webrtc
pro
unfortunately
needs
to
look
at
the.
J
I
don't
I
kind
of
disagree
with
carl
and
I
think
you
can
debug
webrtc
within
a
browser
using
the
browser
tools-
they're
quite
rich
now,
but
I
think,
if
you're
implementing
a
stack
from
scratch,
then
examples
would
be
super
useful.
That's
a
very
small
audience,
though,.
C
I
guess
what
I'm,
what
I'm
kind
of
hearing
is
that,
without
you
know,
taking
any
kind
of
sense
of
a
formal
goal
said
like
we're
willing
to
have
people
review
this,
but
we
don't
really
want
to
keep
the
working
group
open
to
get
through
this.
This
process
so
murray
as
long
as
you're.
Okay.
With
that,
we
can,
you
know,
go
through
the
process
of
making
sure
that
we
try
to
solicit
reviews
and
get
other
input,
and
then
you
know
basically
hand
it
back
to
you
whether
you
want
to
progress
it
forward.
A
A
So
let
me
suggest
that
one
of
the
early
things
we
should
do
is
first
check
in
with
subhas
and
see
how
much
more
energy
he
wants
to
put
into
this,
because
if
sukhos
at
this
point
is
not
advocating
for
publication
and
doesn't
have
the
energy
to
put
into
revising
the
the
document
post
review,
then
I
think
that
that's
very
telling-
and
we
should
take
that
signal
once
we've
talked
to
suhas-
then
sean
and
I
think
you
we
should
take
the
the
action
item
to
to
then
circle
back
with
murray
and
confirm
with
him.
A
That
he's
willing
to
go
forward
with
this.
If
we
solicit
the
reviews
as
ad
sponsored-
and
I
think
if
that's
the
case,
then
it
continues
with
perhaps
a
little
bit
more
review
energy
than
it
had
before,
but
actually
on
the
same
path.
It
was
on
when
we
closed
the
working
group,
the
first
time
because
we
closed
the
working
group
for
the
first
time
with
this
not
complete
in
part,
because
the
working
group
had
at
that
point
lost.
A
Obviously
we
have
more
people
here
today
than
we
had
at
that
point,
because
we
really
had
gone
down
to
just
very
little
energy
at
that
point,
but
I
think
the
chances
that
we'll
maintain
the
energy
through
the
process
of
building
an
automated
checker
or
doing
any
of
the
other
things
that
people
have
have
suggested
seems
pretty
low.
So
I
think
it's
really
does
have
the
the
energy
to
revise
this
in
the
face
of
reviews.
A
A
It
doesn't
look
like
we
have
any
strong
objections
and
ecker
who
raised
the
issue
of
advocacy
for
the
rfc
is
okay
with
that
way
forward.
So
I
think
let's
call
that
the
plan
of
record
and
sean
and
I
will
first
follow
up
with
suhas
and
then
with
murray.
A
That
brings
us,
I
think,
to
the
end
of
both
the
agenda
and
the
additional
issues
that
were
raised
at
the
end
of
the
agenda.
So
I'll
ask
the
question:
is
there
any
other
topic
for
the
working
group
at
this
time?.