►
From YouTube: IETF115-WISH-20221110-1700
Description
WISH meeting session at IETF115
2022/11/10 1700
https://datatracker.ietf.org/meeting/115/proceedings/
B
All
right
we
got
a
minute
I
guess
we
should
make
sure
that
Sergio's
here
since
he's
kind
of
the
show
there
he
is,
we
got
about
six
people
in
the
room,
some
of
the
usual
suspects.
So
that's
good.
B
B
Right,
let's
move
on
to
the
good
stuff:
this
is
some
meeting
tips.
Please
use
the
medeco
tool
to
be
able
to
log
in
that's
how
we're
going
to
do
attendance
so
next
time
we
can
make
sure
we
get
the
biggest
possible
room
that
we
can
get.
B
There
are
about
seven
or
eight
people
in
here,
so
we
won't
get
the
biggest
room,
but
that's
also
how
we're
going
to
use
the
queue
to
manage
the
queue
to
get
people
in
and
out
to
make
sure
that
the
remote
presenters
are
and
participants
are
actually
equally
supported,
all
right.
So
next
we
got
the
ITF
nope.
Well,
it's
Thursday,
hopefully,
you've
seen
this
before.
B
This
is
primarily,
if
you
know
about
IPR,
say
something
if
you,
if
you
can't
say
anything,
that's
all
right,
just
don't
say
anything
and
some
other
things
about
anti-harassment
code
of
conduct
Etc,
but
we'll
go
into
that
a
little
bit
more
now.
This
is
kind
of
the
code
of
conduct
guidelines.
This
work
group
has
not
had
any
of
these
problems,
which
is
nice
to
basically
cheat
everybody
with
respect.
You
know
kind
of
use.
Your
head
find
the
best
solution
for
the
whole,
the
internet.
You
know,
I'll,
let
you
read
this.
B
So
the
next
slide
I
think
we'll
show
it
is
the
mass
policy
everyone
in
here
is
massed
up
if
you're
remote
do
what
you
want,
but
when
you're
here
you
need
to
have
a
mask
on
unless
you're
you
do
actually
get
up
here
to
talk
and
you're
six
feet,
awake
no
need,
but
I
think
we're
good
all
right.
The
agenda
pretty
straightforward.
The
only
thing
that
we're
really
going
to
pause
here
for
is.
We
need
a
note
taker.
E
B
B
We're
basically
just
going
to
note
the
actual
action
items.
Basically,
we
got
one
sorry.
B
We
can,
we
can
figure
it
out
between
two
of
us.
I.
Don't
think
it's
going
to
be
that
much
discussion.
Hopefully
we
really
got
one
thing
on
tap,
which
is
our
working
group
draft,
which
is
wish
whip
we're
gonna,
let
Sergio
go
through
I.
B
Think,
like
one
issue,
we
still
got
about
10
minutes
to
sign
for
that,
and
then
we
have
20
minutes
to
talk
about
WEP
and
the
rest
of
the
time
is
kind
of
to
talk
about
the
future
and
the
way
forward
and
to
note,
if
you
were
a
dispatch,
you've
probably
seen
some
of
the
slides
of
Sergio's
going
to
talk
about,
but
you
know
again
we're
we're
here
doing
our
thing,
so
we're
kind
of
kind
of
maybe
doing
some
of
the
dispatch
process.
B
B
If
you
do
that,
because
they're
loaded
up,
you
ought
to
be
able
to
load
them,
I
don't
know
meals.
If
there's
any
way,
you
can
come
to
the
rescue
here,
but
I'm,
unfortunately
not
able
to
connect.
D
E
E
So
the
change
between
the
title
is
wrong
should
be
the
changes
from
zero
four
to
zero.
Five,
mainly
we'll
say
about
a
some
minor
field
that
I
have
received,
and
some
of
the
some
additional
changes,
one
some
of
them
that
are
maybe
Brown.
E
E
E
We
have
also
added
a
note
about
the
calling
order
of
said
local
description
and
and
the
on
the
HTTP
boss,
the
request
to
avoid
ice
restarts,
and
also
we
have
added
that
they
prefer
options
to
to
configure
Ice
is
to
use
the
information
return
in
a
2001
response
and
not
to
rely
on
the
options
on
making
an
option
request
today
to
the
to
the
web
server
if
it
is
not
really
mandatory
for
the
client
for
the
client,
because
it
does
not
support
to
updated
configuration,
the
ISO
reconfiguration
one
say
after
the
third
local,
this
return
has
been.
E
It
has
been
created
and
also
there
has
been
fixes
in
the
example
to
make
it.
So
it
was
quite
in
another
document
and
also
we
have
made
the
data
an
opaque
value
So
to
avoid
a
misinterpretation
by
implementers
and
thinking
that
it
has
to
have
some
value
when
it
is
not
meant
to
to
be
to
have
any
sense.
E
B
B
I
might
give
it
a
little
extra
longer
than
two
week
working
group
last
call
because
of
some
major
American
holidays,
which
most
people
will
be
Mia
on,
but
definitely
I
would
like
to
try
to
if
you,
if
you
can
literally,
if
you
can
promise
me
that
you'll
get
it
done,
we
can
get.
The
working
group
last
call
started
next
week
and.
B
Of
course
can't
see
the
job
room
or
anything,
because
my
network
connection
is
shot
so
unless
we
see
anything,
we're
gonna
go
okay.
Let's,
let's
proceed
with
that
that
path,
Sergio
and
okay,
we
can.
We
can
switch
over
to
web.
B
And
again
so
we're
we're
kind
of
begging,
begging,
forgiveness
from
the
working
group,
because
technically
this
is
an
in
scope.
But
since
we
have
you
here-
and
we
have
been
talking
about
this-
a
little
bit-
we're
hoping
you're
willing
to
indulge
the
overstepping
of
the
bounds
of
the
charter.
While
we
have
this
discussion.
B
E
B
E
Okay
so
I
want
to
well
I
wanted
to
do
on
last
meeting,
but
we
didn't
have
time
for
it
so
I'm,
just
mostly
going
to
present
the
same
slides
that
I
did
let's
say,
I
presented
for
last
last
meeting,
but
with
minute
updates.
E
First
I
know
that
the
grace
is
explicitly
out
of
its
scope
of
the
working
group,
but
also
right
now.
The
equivalent
graph
of
a
web
is
a
reuse
most
of
the
mechanics
that
has
been
put
in
place
for
Whip
and
the
first
drop
was
mostly
mainly
yes,
replacing
whip
by
web,
and
that's
all
that
I
I
did
for
work
with
the
first
wrap,
and
so
why?
Why
do
we
need
web?
E
Why
do
we
need
to
handle
the
the
play
by
side
or
the
of
the
streaming
with
the
webrtc
first,
because
we
are
seeing
that,
at
least
in
my
side,
I
seen
that
quite
a
lot
of
friends
lately
to
be
able
to
be
able
to
have
internet
ability
between
webrtc
services
and
product
and
mostly
I,
see
also
a
trend
that
is
mostly
about
having
backup
solutions
that,
for
example,
being
able
to
switch
from
webrtc
provider
to
another.
E
If-
and
there
are,
there
is
any
downtime
in
any
of
them
and
also
to
avoid
the
service
login,
so
you
can
have
multiple
or
you
can
rely
on
multiple
webrtc
services
or
providers
without
having
to
be
logged
in
so
this
is
say
well,
this
was
not
the
the
one
of
the
the
initial
goals
for
the
when
I
pressed
when
I
brought
a
web
since
I
have
that
graphene
I
see
a
lot
of
attraction
in
in
that
regular.
E
Also,
there
are
other
benefits
like,
for
example,
we
can
reduce
the
player
server,
the
name
that
can
be
integrated
easily
I
mean
there
is,
for
example,
with
them,
epic,
that
sound
I
feel
like
that.
You
don't
have
to
write
your
own
player,
you
just
use
whatever
player
is
available
and
there
is
a
quite
a
lot
of
my
versions:
either
open
source
of
of
closures
that
allows
to
have
a
much
better,
a
user
experience
that,
if
you
build
your
own
and
also
this
protocol,
will
allow
integrating,
with
the
impacted
us.
E
For
example,
this
has
been
a
report
and
when
we
there
has
been
a
joint
effort
or
what
they
went
in
the
name
impact
as
if
to
be
able
to
have
like
a
common
player
that
could
use
both
mpegdas
and
webrtc.
When
you
will
use
webrtc
for
the
live
version
of
the
of
the
stream,
and
you
can
use
it,
that's
version
to
use
the
timesheet
version
of
the
of
the
of
the
screen.
E
So
you
can
start
watching
the
one
event
in
in
life,
but
you
can
go
back
in
time
and
connect
to
the
to
the
impact
Das
stream
and
that
it
is
has
when
you
can
just
select
from
what
time
you
want
to
watch
the
the
screen
and
with
seamless
in
the
same
player
and
also-
and
there
is
also
one
very
important
case-
that
it
is
a
while
in
traditional
in
webrtc
Services.
We
are
running
the
all
the
services
in
either
in
a
native
application
or
in
a
full
browsers.
E
E
By
just
having
to
provide
to
be
able
to
play
a
web
streams
and
not
have
to
provide
the
full
flavor
apis
that
it
is
much
difficult
to
maintain
and
and
and
there's
some
megabyte
levels
in
their
apis
so
now
the
last
thing
is
that
a
weapon
web
could
be
used
together
for
service
in
the
operating.
So
you
can
have
you
have
a
web
service
and
a
web
service.
You
can
try
to
to
just
create
a
gate
with
that
post.
Do
the
interoperative
it
in
both.
E
So
the
question
is,
is
if
we
should
restart
this
working
groups
to
to
include
also
egress,
as
a
scene
has
said,
I
presented
it
in
this
part,
so
Monday
and
I
think
that
the
construction
was
that
we
should,
or
the
work
in
the
with
working
groups
should
be
able
to
decide
if
we
want
to
recharge
the
the
the
working
group
or
not-
and
we
don't
have
to
go
back
to
this
patch
and
to
try
to
find
a
way.
E
So
whatever
we
decide,
we
are
free
to
to
decide
if
we
want
to
the
to
change
the
scope
of
the
of
the
charter
and
out
of
the
the
web.
The
protocol
as
a
working
group
item.
B
So
so
Adam,
let
me
just
jump
in
here,
real
quick,
so
I
guess
I
would
I
just
want
to
say
that
I
I
completely
agree
with
the
way
Sergio
characterized.
We
went
to
dispatch
and
they
were
like.
Why
are
you
coming
to
us
you're
already
a
working
group?
You
can
just
recharge
her,
but
they
I
think
in
other
instances.
They
might
have
said.
Why
didn't?
Why
didn't
you
come
to
us
if
you
so
we
kind
of
got
a
free
pass.
B
F
B
All
right:
well,
maybe
we'll
try
to
loop
back
to
Adam
and
Jonathan's
in
the
room
and
so
is
Harold.
So
maybe
we
can
give
him
a
little
bit
of
time.
D
Should
I
go
ahead?
Okay,
can
you
hear
me
now?
Yes
yeah,
we
can
great
all
right,
I
think
it
did
not
survive
a
device
earlier
in
the
conference.
Sorry
about
that,
so
I'm
gonna
do
a
full
180
from
what
I
said.
Last
time
we
had
this
discussion
and
and
say
that
I
now
support,
although
somewhat
reluctantly
the
notion
of
retartering
to
do
the
web
work.
D
D
Did
this
first
time
around
to
help
move
the
work
along
and
it
the
volume
of
discussion
on
the
list
is
still
rather
slow,
but
I
am
definitely
seeing
a
lot
of
interest
in
the
industry
for
this
and
people
are
moving
forward
with
the
protocol
as
specified
I
see
like
major
service
providers
already
deploying
this
in
like
beta
trials
and
I,
think
it
would
be
far
far
better
to
have
you
know
a
written.
D
D
C
Yeah
Jonathan
Lennox
I
mean
I,
think
your
analogy
was
Dash
as
somewhat
convinced
me.
I
was
sort
of
dubious
before,
but
I
think
sort
of
the
idea
that
in
browsers
this
is
just
implemented
in
JavaScript,
though
there
could
be
libraries
for
it,
but
for
embedded
things
it
needs
to
be,
it
can
be
implemented
for
real,
which
is
easier
than
letting
them
run
arbitrary
JavaScript.
That
seems
reasonable.
My
concern
I
mean
and
I
think
the
cert,
the
your
regular
expression
at
the
second
bullet
is
certainly
a
very
easy
thing
to
do.
C
My
worry
is
how
much
more
do
we
need
in
particular,
if
we
need
to
be,
you
know,
Feature
Feature,
equivalent
to
Dash
and
basically
have
let
you
know
that
people
say
well
in
dash
I
if
I'm
doing
it
over
Dash
I
can
do
you
know,
subtitles
I
need
to
be
able
to
do
subtitles
in
web.
Well,
that's
just
one
example
off
the
top
of
my
head:
I!
Don't
know
why
everything
that's
in
dash!
That's
my
worry.
C
G
E
Yeah,
the
in
fact
this
is
in
one
of
my
last,
my
latest
slides
about
what
is
the
the
Delta
that
we
have
to
to
add
from
from
web
to
week,
because,
as
you
have
said,
there
are
some
some
functionalities
that
are
specific
to
to
playback
that
we
have
not
worried
yet
in
in
in
just
size.
So
this
is
already
covered
also
in
the
slides
later
on.
H
H
This
seems
to
be
a
simple
specification
for
doing
a
simple
job.
We
have
other
groups
like
mock,
who
are
trying
to
build
the
the
grand
total
solution
of
changing
everything,
so
that
so
that
we
can
achieve
things.
We
cannot
do
but
a
simple
tool
for
a
simple
job.
We
should
simple
tools
need
specifications.
We
should
specify
it
go
ahead.
Okay,.
G
Yeah
10
pounds
in
here
yeah
I'm
in
favor
of
us
doing
this
I
think
particularly
the
embedded
stuff,
makes,
makes
sense.
I
have
a
slight
preference
for
us
doing
it
after
we've
got
got
whip
out
the
way
I
kind
of
can
we
is
there
a
is
a
problem
with
sequencing
it
so
that
we
don't
change
the
charter
until
we
get
whip
out
the
door.
B
G
B
B
We're
going
to
run
a
working
group
last
call
when
that
working
group
last
call
completes
Niels
and
I
are
going
to
put
our
head
together
and
flip
a
coin
and
figure
out
who's
going
to
be
the
shepherd
and
we're
going
to
push
it
forward
out
of
the
working
group,
and
at
that
point
we
can
start
to
do
Charter
revisions
to
look
at
that
because
it'll
take
weeks
for
that
to
happen
as
well.
So
if
that
is
kind
of
the,
if
we
do
it
at
that
kind
of
pace,
is
that
kind
of
work.
F
Sean
Sean
I
I
have
to
I've
done
the
or
started
the
shepherding
for
for.
F
B
I
mean
I'm
I'm,
thinking
like
in
the
new
year,
we'll
we
will
start
we'll,
put
up
a
GitHub
repo,
we'll
put
in
the
original
charter,
and
then
people
can
start
walking
away
at
making
changes
so
I'm,
hoping
that
the
changes
that
we
would
get
via
the
director
reviews
and
through
the
isg
wouldn't
won't
be
earth-shattering
to
this
so
I'm,
hoping
that
we
won't
have
a
cycle
problem,
but
I
I
do
acknowledge
that
we
don't
want
to
do
too
many
things
at
once
here,
because
we're
just
running
out
of
Cycles
from
people.
B
So,
let's
try
to
what
we'll
do
is
we'll
start
the
retarding
process
after
we
get
the
document
out
the
door,
but
nothing
to
say
that
Sergio
can't.
We
can't
keep
talking
about
this
actually
on
the
list.
If
people
do
have
Cycles
Pierce.
A
Yeah,
it
seems
like
the
integration
with
Dash
can
potentially
get
to
fairly
complicated,
as
mentioned
some
of
the
issues
around
it's
all
timestamps
in
media
and
how
you
do
subtitles
and
everything.
In
some
ways
it
seems
as
potentially
a
more
a
nasty
sort
of
part
to
it,
whether
it
be
pulled
apart
and
at
it
as
a
separate
thing.
I
don't
know,
but
not
quite
sure
how
we
separate
out,
but
it
might
get
because
I
would
we're
involved
in
that
Dash.
A
If
report
that
started
to
try
to
tackle
some
of
these
issues,
but
then
it
seemed
like
there
was
quite
a
few
things
to
deal
with
I.
Don't
know
what
you
think,
Sergio.
E
Yeah
I
mean
I
would
like
to
have
the
like:
the
basic
media
flows
established
with
web
and
at
least
I,
probably
on
the
first
draft,
just
because
mainly
we
have
already
make
most
of
the
working
whip.
So,
yes,
a
happy
meeting
webpage
seems
like
a
very
easy
leaf.
Then
a
is
bet
when
we
are
able
to
have
like
a
first
version
of
web.
E
So
we
can
reduce
some
of
them,
so
I
don't
know
yet,
but
this
is
the
best
of
them,
but
I
think
that
if
we
try
to,
we
get
some
momentum
having
the
the
web
protocol
in
in
the
in
draft,
they
are
specified.
It
will
be
much
easier
to
to
to
get
Advanced
on
those
topics,
so
I
will
try
to
or
what
my
proposal
would
be,
not
to
the
same
thing
that
we
have
done
with
beep.
E
Don't
make
the
these
other
features
be
a
blocker
about
having
a
first
version
of
web
or,
for
example,
early
drafts
and
try
to
see
if
we
can,
for
example,
add
that
add
them
later
on
with
the
essentials
either.
If
it
is
a
HTTP
res
interface
or
it
is
by
adding
data
challenge.
Support
for
web.
A
Okay,
yeah
yeah
I
think
it
sounds
reasonable,
I
guess,
because
you've
got
sort
of
media
type
description,
so
you've
got
STP
trying
to
map
that
to
the
media
descriptions
within
Dash
and
also
then
you've
got
the
time
frame
aspects
around
the
dash
MPD
that
doesn't
really
have
a
kind
of
an
analogy
in
in
the
webrtc
world.
I
think
in
terms
of
description,
okay,
yeah.
B
All
right,
I
think
it's
probably
not
a
good
use
of
our
time
to
actually
slap
up
the
charter
and
start
to
edit.
It
I
think
that's
kind
of
maybe
kind
of
a
complete
waste
of
time,
so
I
think
the
we
should
set
everyone
free
here,
give
them
30
minutes
back
of
their
lives
unless
there's
something
else
to
discuss.
B
C
A
E
So
basically,
the
the
web
protocol
is
basically
the
same
thing
that
as
whip,
so
we
have
the
the
web
players
and
an
HTTP
post
with
sap
offer.
Then
the
endpoint
return
on
a
DP
answer
with
the
the
200
or
201
response,
and
then
everything
the
the
request
and
the
ice
and
the
details
is
established
with
the
media
server
and
the
rtcb
float
start
to
go
from
the
media
server
to
the
web
player.
E
In
the
there
has
been
some
discussion
that
I
don't
know
if
this
is
a
the
best
thing
to
do
or
not,
because
we
don't
want
to
have
two
different
ways
of
operating
in
the
in
the
draft.
So
this
and
the
the
current
graph
also
covers
it.
But
I
don't
know
if
we
want
to
remove
it
or
not
that
they
will
say
that
there
was
a
also
like
an
interest
of
having.
E
The
service
to
be
able
to
provide
the
sdp
offer
so
and
then
the
player
does
not
need
to
know.
For
example,
you
want
to
establish
an
audio
video
session,
or
only
audio
and
also-
and
there
was
some
claim
that
some
implementation
don't
support
great
offers
so
and
also
this
will
allow
weapon
weapon
to
probability.
E
But
there
is
a
big
con,
at
least
from
my
side
that
it
is.
The
media.
Server
does
not
actually
know
what
codecs
may
be,
maybe
used
by
the
by
the
stream
when
it
is
established,
because,
for
example,
at
least
in
in
miligas,
we
allow
the
players
to
connect
before
the
Apple
stream
is
started
and
the
publication
can
choose
whatever
codecs
they
use.
So
we
are
able
to
switch
between
dynamically
between
the
Codex.
E
So
if
the,
if
the
it
is
difficult
for
us
at
least,
to
provide
an
sdp
offer
with
the
to
the
player,
because
we
will
just
have
to
to
provide
a
list
of
all
the
potential
corrected
and
that
may
that
assist
in
fact,
so
this
is
one
of
them.
One
point
that
if
we
that
we
will
need
to
to
decide
if
we
want
to
support
it
or
not,
just
none
so
a
in
this
one,
no
yep.
E
So
what's
missing
us
Jonathan
I
think
it
was
younger
than
the
one
that
said
that
so,
as
Jonathan
said,
that
web
has
more
requirements.
Internal
functionality
than
with,
for
example,
within
the
and
I,
think
that
the
whole
thing
is
that
we
have
a
good.
E
We
can
check
what
are
the
the
misinfusion
ideas
by
taking
what
does
provide,
because
we
should
try
to
be
like
a
a
future.
We
should
have
should
be
able
to
provide
the
same
functionality,
as
does
provide.
That,
for
example,
would
be
would
be
about
having
multi
languages
in
in
audio
to
be
able
to
do
remote,
PESA
and
mutant
avoid
having
Western
resources
if
they
pay
the
the.
E
If,
if
the,
if
the
player
is
supposed
to
support,
subtitles
and
legations
supporting
metadata,
specifically,
for
example,
the
for
time,
stamping
and
and
having
a
like
the
smtpa
and
the
timestamps,
and
also
how
to
do
client-side
resolution
and
quality
selection,
for
example,
to
be
able
to
have,
even
if
the
service
is
doing
a
layer,
switching
between
supercars
or
SBC,
to
be
able
to
allow
the
client
to
say
to
specify
to
to
know
which
layers
are
available
and,
for
example,
to
yes
and
say
that
and
send
me
a
lower
resolution.
E
Even
if
there
is
a
if
I
am
able
to
do
a
a
Higher
One.
So
the
I,
don't
I,
think
that
this
can
be
implemented
as
add-ons
today
to
the
to
the
web
specifications
and
they
don't
change
the
what
we
are
going
to
to
implement.
So
we
can
add
it
on
top
and
probably
as
a
optional
and
feature
that
we
can
negotiate.
C
Yeah
Dartmouth
I
think
sorry,
I
should
have
put
myself
in
the
queue
I
apologize,
I,
I
think
now
the
I
think
sort
of
the
big
picture.
Consequence
of
this
is
a
lot
of
these
things.
Possibly
everything
but
subtitles
is
going
to
require
fresh
offer
answers,
so
we're
going
to
need
General
fresh
offer
answers
in
a
bunch
of
places,
not
just
offer
answers
for
ice
restart.
So
that's
going
to
be.
C
You
know
a
fairly
notable
change
to
the
protocol.
I
think
that's
something!
That's
something
we're
going
to
need
to
think
about
and
I
guess.
The
question
is:
are
you
going
to
support
are?
Can
we
do
this
entirely
with
client
initiated
offer
answer,
or
does
it
have
to
be,
or
is
there
some
place
where
they're
you
know
where
the
server
decides
hey?
We
need
a
fresh
offer
answer
because
that's
going
to
be
even
more
complicated,
does
he
need
to
have
some
sort
of
polling
to
notify
of
that.
C
Yeah
and
then
I
think
that's
probably
for
subtitles.
You'd,
probably
want
somebody
something
over
data
Channel,
but
I,
don't
know
the
details
and
we
need
to
let
somebody's
on
the
chat
commented
that
there's
you
know
the
MP4
container.
Does
you
know
everything
you
ever
amazing
quad
number
of
different
things?
The
question
is
which
one
of
those
things
how
many
of
those
things
do
we
actually
need
so.
F
Individual
just
clarification
on
the
on
the
slide,
where
you
said
like
whether
we
should
support
reversing
the
the
offer
answer
roles.
If
you
connect
the
players
first,
don't
you
run
into
the
problem
that,
if
the,
if
the,
if
the
whip
client
then
later
connects,
you
might
still
run
into
the
problem
that,
like
you,
have
codec
in
this
match
or
you're
like?
How
are
you
going
to
enforce
from
the
server
side
that
basically,
the
one
codec
which
would
get
set.
E
I
mean,
but
if
the,
if
the
player
sends
the
zp
offer
and
data,
for
example,
the
unit
super
ab1
I
mean
I
cannot
do
much.
At
least
you
know,
I
mean
I
can
I
can
try
to
even
do
trust
coding
in
the
server
side,
but
I
mean
the
server?
Will
the
player
will
send
me
all
that
it
supports
is
so
it
is
if
I
run.
F
Into
this
yeah,
let's
say
like
if
you,
if
you
connect
to
the
web
players,
and
you
have
negotiated
something
with
them,
there's
no
guarantee
that
whatever
you
have
negotiated
with
them
will
match
with
the
with
the
which
player
what
if
it
connects
later,
basically
or
yeah,.
E
G
Yeah
Tim
Hanson
couple
of
things:
I,
don't
like
the
switching
of
the
answer.
I,
don't
see
that
I,
don't
believe
that
purported
benefits
like
you're,
never
going
to
be
able
to
connect
a
whip
to
a
whip
without
something
in
the
middle
anyway,
and
so
saying
that
it
makes
it
easier.
I
think
is
kind
of
artificial
to
be
honest,
I
think
it
complicates
the
the
things
somewhat
unnecessarily
so
I'm
I'm
kind
of
not
I,
don't
buy
that
and
it
adds
a
round
trip.
Of
course,
so
I
don't
buy.
G
That
and
I
think.
The
second
thing
here
that
kind
of
the
additional
features
I
think
those
channels
are
the
way
to
go,
and
even,
if
you
end
up
needing
to
renegotiate,
you
can
renegotiate
over
the
day's
channel
so
and
it's
worth
I
mean
we've
done
a
ton
of
stuff
where
we
use
the
label
of
the
data
channel
to
indicate
what
the
purpose
of
that
channel
is.
So
you
don't
have
to
like
you
know,
it's
sort
of
you
don't
have
to
agree
it
in
the
STP.
G
You
know:
web
subtitle
Dash
en
it
might
be
the
title
of
the
the
label
of
the
channel,
and
then
you
know
what
it
is
so
I
think
we
can
I
I
think
we
can
get
a
long
way
towards
these
sorts
of
features
without
needing
to
do
another
round
of
offer
answer,
and
even
if
we
have
to
do
another
round
of
offer
answer
to
add
a
track
or
something
you
can
do
that
negotiation
over
a
spare
and
negotiating
data
channel.
G
So
you
can
create
today's
channel
for
doing
offer
answer
over
and
you
do
offer
answer
over
that
and
and
keep
your
keep
your
your
HTTP
server
out
of
the
loop
because
it
doesn't
need
to
be
in
it.
G
H
That
will
just
lead
to
well
kissing
contest,
but
there
are
a
number
of
things
that
we
could
write
specs
for
that
would
be
applicable
in
any
webrtc
context.
Subtitles
especially.
E
E
Yeah
yeah
I
agree
with
that.
For
example,
a
I
already
have
been
working,
for
example,
in
a
solution
for
subtitles,
a
for
whip
that
obviously-
and
this
will
have
to
be
done
somewhere
else-
probably
maybe
Decor
or
a
music-
and
then
we
can-
we
could
reuse.
Then
they
were
here.
So
that's
what
that
was
also
my
goal.
So
I
think
that
having
a
full
set
of
functionality
is
like
a
like
the
goal
for
me
for
my
service,
but
it
they
don't
need
to
be
all
done
in
web.
E
C
John
Athletics
again
yeah
I
think
the
sort
of
subtlety
here
is
there's
I
mean
we
have.
You
know
a
number
of
layers
of
indirection
here.
Right
I
mean
you
know:
web
users,
webrtc
webrtc,
uses
sdp,
offer
answer,
STP
offer
answer,
uses
sdp
and
you
know
and
I
think
it'll
for
a
lot
of
these
things,
Yes
somewhere
in
the
30-year
history
of
sdp.
There
have
been
solutions
to
find
for
these.
Are
they
good
I
don't
know?
Are
they
implemented
in
browsers?
No,
could
they
build
under
browsers?
C
Probably
will
they
be
ask
Harold,
but
I
mean?
If,
if
you
say,
if
you
say
hey,
if
we
add
this
to
webrtc,
so
it
might
just
be
a
matter
of
not
even
defining
new
Solutions,
but
finding
like
Adam
mentioned,
there
is
an
RTP
payload
for
ttml
I,
don't
know
if
it's
quite
solves
our
problem,
but
probably
roughly
does
so.
C
Maybe
if
you
can
maybe
just
say,
hey
browsers,
add
this
ttml
to
your
webrtc
stacks
and
you
know,
and
that
will
and
there's
an
actual
use
case
for
this,
because
we
want
to
be
able
to
do
this,
for
you
know,
live
Dash
and
I.
Think
that's
a
use
case.
The
brothers
are
actually
interested
in
so
but
yeah
I
think
this
might
just
be
a
matter
of
scouring
the
web,
the
RFC
archive,
rather
than
defining
new
work.
A
I
can't
quite
think
what
I
was
going
to
say:
yeah
the
data
Channel
thing,
so
I,
just
sort
of
webrtc,
whether
whether
you,
whether
you
want
to
sort
of
stay
with
just
using
the
webrtc
data,
Channel,
whether
you
want
to
be
kind
of
open
to
using
just
websockets,
so
in
some
senses,
maybe
a
higher
level
abstraction.
When
one
talks
about
using
it
again,
it's
something
that
kind
of
came
out
of
the
out
of
the
work
that.
E
A
Dash
if
thing
I
mean
I
guess
because
well
I
guess,
with
the
sort
of
these
types
of
approaches
we're
using
Dash.
Typically,
the
data
Channel
approach
doesn't
necessarily
isn't
necessarily
kind
of
needed,
because
it's
often
not
a
peer-to-peer
type
of
communication,
but
I
guess
it
doesn't
have
to
be,
but
just
a
thought
really
as
to
whether
or
wants
to
sort
of
allow
that
flexibility,
as
opposed
to
just
stipulating
just
use
the
data
channel
in
web
ITC.
E
Yeah
I
mean
I,
don't
have
a
I.
I
have
not
started
to
work
on
a
solution,
and
so
then
I
don't
have
to
check
how
this
implemented
with
them
and
back
does
but
I
can.
There
is
no
limitation
about
producing.
Where
was
socket
over
data
challenge
at
the
moment.
Probably
when
we
start
to
check
it,
we
will
see
that
it
makes
more
sense
to
use
one
on
another
or
maybe
probably
with
web.
E
G
So,
coming
back
on
the
today's
Channel
thing,
yeah
we
turns
out
the
this
channel.
Api
looks
very
like
the
websocket
API
in
the
browser.
So
actually
we've
done
a
few
things
where
you
consume
one.
On
top
of
the
other
and
the
the
so
JavaScript
author
actually
doesn't
know
which
it
is
and
doesn't
care
because
semantics
are
pretty
similar.
The
open
semantics
are
different,
but
once
it's
open
it's
pretty
indistinguishable.
G
To
be
honest,
so
you
can
you
can
fake
one
with
the
other,
so
you
might
not
actually
need
to
kind
of
enforce
that
at
the
client
side.
Obviously,
at
the
server
side
it's
completely
different
and
I
think
that's
where
you're
going
to
come
up
against.
It's
like.
Where
does
this
metadata
arrive
get
generated?
Does
it
get
generated
kind
of
close
to
the
media
server
where
your
media
is
coming
from?
Is
it
tied
to
that,
in
which
case
data
channels
is
the
right
path?
G
Whereas
if
it's
coming
from
you
know
where
your
HTTP
request
went,
then
websockets
is
the
right
path,
and
so
I
think
that
you're
going
to
fight,
because
the
media
server
is
probably
a
long
way
back
behind
some
that,
whereas
your
your
web
server
is
is
out
the
front
so
I
think
like
which
you
end
up,
choosing
is
going
to
be
almost
an
architectural
decision
rather
than
anything
else.
But
yeah
I
mean
our
experiences
that
from
the
client
side,
you
almost
can't
tell
the
difference
between
them.
F
All
right
do
we
want
to
ask
now
for
I
think
we
had
on
the
agenda
the
the
question
whether
regarding
the
rechartering.
B
Is
there
anyone
that
would
like
to
not
not
does
not
want
us
to
try
to
recharter
that
I
would
be
interested
in
in?
Why.
B
B
We
will
throw
in
the
charter
and
we'll
start
editing
it
and
we'll
invite
everyone
to
provide
issues
and
pull
requests
on
the
charter,
and
then
we
will
proceed
forward
and
with
that
I
think
we'll
draw
this
meeting
to
a
close
unless
there's
anybody
else
that
has
any
other
business
that
they
would
like
to
add
at
the
last
second
here
going
once
going
twice
sold
as
the
Techno
music
starts
to
pound.
In
the
background,
everyone
have
a
good
night.
Thank
you
very
much.
B
And
thanks
Niels
for
helping
me
out
with
my
network
connection
problems.