►
From YouTube: IETF96-ECRIT-20160721-1830
Description
ECRIT meeting session at IETF96
2016/07/21 1830
A
A
C
D
E
C
C
C
C
So
welcome
to
Egret
ietf
86
five.
Why
is
it?
The
slides
are
wrong.
It's
ITF
96
and
we
are
not
in
bonus
eras,
Roger.
If
you're
listening.
Thank
you
so
I'm
our
cleanser
co-chair,
Roger
Marshall
is
attempting
to
join
us
virtually
if
he
hasn't
already
doing.
This
is
the
note.
Well,
hopefully,
you've
read
this
and
you
understand
that
contributing
how
to
contribute
to
the
IDF
and
what
your
responsibilities
are
in
this
area.
C
Our
agenda
today
we're
going
to
talk
for
about
five
minutes
bashing
the
agenda.
So
if
you
see
anything
up
here,
you
don't
like
we're
going
to
hear
from
Brian
on
data.
Only
emergency
calls
validation
of
locations
around
playing
changes
and
extension
to
return
complete
and
similar
info.
Now
we
took
those
three
documents
and
actually
put
them
in
working
group
last
called
this
week,
but
we
thought
it
would
be
good
to
get
let
Brian
state.
What
he's
in
the
latest
changes
doesn't
believe.
You
just
updated
him
recently.
C
So
he's
going
to
talk
about
those
updates
and
then
you'll
have
comment
had
the
ability
to
comment
in
here
and
then,
of
course,
working
group
last
call
please,
you
know
review
them
intensely
in
comment
on
the
list.
What
I
mean
will
capture
your
comments
and
here
is
treat
image.
You
know
working
group
last
comment.
C
C
No
okay
document
status,
additional
data
related
emergency
calls
in
the
RC
editor
Q.
We
talked
about
next
generation
vehicle
initiated
emergency
calls
on
Pan
American.
We
believe
they're
real,
close
ready
to
go
to
the
iesg.
There's
been
a
lot
of
comments
on
the
lists
in
the
authors
and
in
those
commenters
are
working
those
issues
out
now
and
we'll
just
that
shortly,
and
these
are
the
three
documents
that
when
we
started
working
group
last
call
this
week
on
and
data
only
emergency
lost
extensions
in
validation
of
location.
So
milestones
we
move
those
out
to
November.
C
C
Right,
the
pan-american
II
call
in
the
in-vehicle
emergency
calls
are
both
listed
as
informational,
I,
believe
they're
in
the
drafts
as
standards-based.
We,
the
chairs,
believe
that
was
just
a
error
on
our
part.
We
fat
fingered
that
when
we
put
them
in
originally,
if
there's
anybody
has
causes
any
pain
for
anybody.
Please
let
us
know.
In
other
words,
if
you
thought
those
drafts
are
going
to
be
informational
and
now
we're
we've
announced
they're
going
to
be
standards-based.
If
that
cause,
you
any
concern,
please
speak
now
or
on
the
list
like
shortly.
F
E
C
F
So
data
only
is
a
way
to
send
a
message
that
contains
all
the
information
you
would
find
in
an
invite
and
use
this
when
there
is
no
real
time
media.
There
is
no
difference
between
sending
and
invited
some
emergency
call,
and
something
and
message
is
an
emergency
call,
except
that
you
wouldn't
have
SDP
and
set
up
a
session
with
with
with
media.
So
the
example
here
is
an
alarm
right.
An
alarm
goes
off.
F
You
send
the
message:
it's
got
all
the
same
data
that
would
be
an
invite,
gets
routed
like
a
any
other
emergency
calls
you're
in
service
SOS
and
the
and
the
request-uri,
and
are
out
that
you
get
from
a
lost
server
and
and
a
route
header,
and
it
does
exactly
the
same
thing,
but
it
doesn't
set
up
a
media.
There's
a
request
to
deal
with
a
statement
about
privacy,
mainly
that
privacy
for
emergency
calls
and
that
would
apply
to
data.
Only
emergency
calls
is
subject
to
local
regulation
and
they're
in
many
jurisdictions.
F
There
are
regulations
that
basically
reveal
or
tell
you
that
that
certain
information
that
normally
is
considered
private
is
at
least
shared
with
emergency
services,
people.
So
that
is
actually
a
meaningful
statement.
You
need
to
make
sure
that
that's
in
the
concern
in
the
consideration
so
I
put
that
in
there
was
reference
to
a
draft
that
I
put
in
that
lets.
You
do
an
update
to
additional
data.
We
haven't
worked
on
that
draft
for
a
while
and
I
decided
to
just
take
it
out.
That
was
just
the
easiest
thing
to
progress.
F
This
document
eventually
want
to
get
back
to
that,
because
some
data,
and
specifically
the
data
that
you're
using
a
message
alarm
data,
would
be
the
example
may
in
fact
change
over
time.
We
have
another
way
to
do
it.
You
can
send
it
by
reference
and
then
repeatedly
dereference
it.
So
we
have
a
mechanism
to
to
do
that.
But
a
push
mechanism
is
probably
better,
but
I
don't
need
to
reference.
It
I,
don't
need.
F
We
don't
need
that
mechanism
to
make
this
useful
and
tying
the
to
the
dress
together
is
probably
not
a
good
idea,
so
I
just
took
it
out
just
deleted
the
text
and
then
I
needed
to
make
changes
to
conform
to
the
the
syntax
that
we
have
in
the
final
version
of
additional
data.
I
was
using
some
earlier
syntax
and
I
fix
those.
F
Those
are
the
changes
we
did
do
a
working
loose
last
call
on,
but
there
there
was
not
a
really
a
large
amount
of
review
and
I
didn't
make
changes,
so
the
chairs
agreed
well,
the
chairs
decided
that
they
would
just
rerun
another
last
call
which
is
fine
and
and
that's
what
we
that's
what
they
did
so
we
we
would
appreciate
a
few
reviews.
I'm
going
to
get
the
guys
to
refute,
to
to
do
some
reviews
we're
going
to
deploy
this
in
the
nena
architecture
of
x
generation
911.
F
So
this
is
going
to
have
many
interoperable
plantations
reasonably
soon.
I
think
and
I
would
appreciate
anybody
taking
a
look
at
it
and
and
give
me
any
feedback
questions
this
one's,
pretty
simple
and
and
and
fairly
mature.
It's
been
around
for
a
while
have
to
deal
with
a
lot
of
these
little
details,
making
sure
exactly
how
things
are
right,
but
it's
it's
actually
a
pretty
simple
mechanism.
F
F
So
similar
gives
you
a
way
to
return
location
information
on
a
fine
service
response,
so
this
is
used
in
two
circumstances:
one
if
you,
if
you
have
a
valid
location
but
you're
missing
some
information
that
can
occur
when
you
send
enough
information
in
the
location,
information
on
the
lost
request
that
we
can
uniquely
identify
at
the
location.
But
you
may
not
have
supplied
enough
information
that
that
we
really
like
you
to
have
in
the
location
that
you
send
when
you,
when
you
send
that
location,
an
emergency
call.
F
A
great
example
would
be
you
have
a
arm
you
have
a
street
or
for
which
there
really
is
only
one
such
street
with
the
right,
Street
house
number
and
everything
else.
Everything
else
is
the
same,
but
you've,
misspelled
or
or
in
some
way,
mangle
the
city
name.
Another
great
example
is
that
you,
you
you,
you
don't
include
the
county
or
the
a
two-armed
level
information,
the
the
911
authorities.
Are
they
I'm
sorry,
the
emergency
call.
Authorities
really
would
like
to
have
all
locations
have
that
a
to
information
and
but
you've.
F
Given
us
enough
information,
you
know,
there's
only
one
city
with
that
name
in
the
state
and
and
we
can
supply
that
information
back.
So
it's
complete.
It
gives
you
a
complete
all
the
fields
that
we
know
about
that
location
you
in
that
circumstance,
you
have
sent
a
valid
location
in
so
you're,
going
to
get
a
route
out,
but
there's
there's
information
that
we
know
and
we
want
to
be
able
to
return
it
to
you,
so
you
can
save
it
and
send
it
in
the
emergency
call.
F
The
the
the
second
use
is
when
you
send
an
invalid
location.
So
this
is
when
you
send
a
piece
of
information
in,
but
we
cannot
vote
uniquely
locate
you
in
them
in
the
lost
database.
You
don't
get
a
route
in
this
case
you
get
an
error
and
you
get
a.
F
There's
three
lists
of
valid
invalid
and
unchecked
and
you
you
know
the
individual
elements
of
location
information
are
put
in
one
of
those
three
thanks.
So
in
this
case,
what
we're
going
to
provide
with
is
a
valid
location.
That
is
a
maybe
you
meant
this.
So
this
is
a
piece
of
helpful,
hopefully
helpful
information
to
a
user
who's
entered
the
an
incorrect
location.
We
do
the
all
the
documents
for
all
of
this
say
that
it's
highly
desirable
to
validate
the
location
before
you
have
an
emergency
call
right.
F
You
want
to
make
sure
that
your
location
is
valid
before
you
have
to
go
and
get
a
route
for
it,
so
this
is
going
to
get
used
way
in
advance
when
you
first
get
your
service,
you
put
it
in
an
address.
If
you
put
in
an
address
that
we
can't
record,
we
can't
validate,
you
need
to
figure
out
what
the
valid
location
is
and
we're
going
to
give
you
helpful
information.
Give
you
a
a
valid
piece
of
information.
That
said,
maybe
you
meant
this.
It's
exactly
the
same
mechanism.
F
F
F
So
we
would
again
we'll
get
a
bunch
of
reviews
from
them,
but
anybody
in
the
room
who's
willing
to
review
it
this
one's
a
little
bit.
You
need
to
know,
lost
pretty
well
to
kind
of
understand
this
is
doing.
But
if
you,
if
you're
familiar
with
RC
5222,
then
then
there
the
draft
is
actually
fairly
short.
But
if
you
don't
understand
lost
you'll,
you
will
get
lost
reading
this
document.
Yes,
yes,
yes,.
F
F
C
Working
group
last
call
yeah
that
process
we're
asking
people
to
review
it
see
if
they
agree
with
the
mechanisms
defined
and
actually
look
for
editorial
nits.
If
you
don't
like,
you
know
that
stuff
for
your
keen
on
that
stuff,
we
do
need
people
to
review
these
things.
So
is
there
anybody
volunteering
to
do
help
us
with
a
working
group?
Last
call
Dale
hit.
F
And
finally,
we
have
planned
changes.
This
one
is
a
a
mechanism
to
deal
with
the
reality
that
the
data
in
the
lost
server
changes
slowly
over
time.
An
easy
example
is
they
build
a
new
road
in
a
town,
and
we've
got
to
add
that
that
piece
of
information,
another
example-
is
that
a
part
of
a
of
a
run
at
what
we
in
the
US,
it's
called
an
unincorporated
community,
it's
an
area
that
is
not
in
a
city
but
near
a
city,
and
the
city
expands
its
boundary
to
take
that
area
in
we
call
that
annexation.
F
There
are
other
terms
used
in
other
countries,
but
those
changes
are
an
actual
change
in
the
address
the
address
changes
because,
instead
of
having
the
name
of
the
community,
the
a
four-level,
be,
you
know
like
the
county
name
or
some
neighborhood
name.
It
changes
to
be
the
name
of
the
city,
so
what
was
valid
is
no
longer
valid.
That's
a
planned
change.
These
kinds
of
changes
happen
on
a
schedule.
They
say
at
midnight
on
this
date.
These
things
will
change.
Yes,.
F
Indeed,
it
did
we're
in
a
city
that
had
an
are
quite
a
number
of
planned
changes.
Thank
you
and
so
that
that
mechanism
is
so
that
that
happens
today.
The
only
way
that
we
have
to
to
have
to
communicate
issues
like
this
is
that
there's
this
statement
in
in
phone
bcp
that
says
you
should
validate
periodical.
You
should
revalidate
periodically
so
that
you
have
a
valid
location
every
once
in
a
while.
F
You
should
check
to
make
sure
that
it's
still
valid,
that's
a
really
lousy
way
to
handle
a
plan
change
the
the
way,
this
so
the
whole
mechanism.
What
that
revalidate
is
a
pull
mechanism.
What
this
draft
proposes
is
a
push
mechanism,
it's
a
way
to
say,
I'm,
going
to
tell
you
that
these
locations
will
become
invalid
at
a
certain
time
or
will
change
at
a
certain
time,
and
the
way
we
do
that
is
that
we,
we
save
a
URI
in
the
lost
server
that
we
poke.
F
If
you
if
a
changes
is
going
to
be
made.
So
when
you
do
a
validation
request,
you
can
supply
a
URI
that
you
or
I
can
be
saved
in
the
lost
database,
and
whenever
that
location
changes
that
you
are
I
that
you
provided
will
be
poked
to
tell
you
that
that
happens,
it's
all
optional
at
the
server
the
server
can
have
lots
of
control
or
what
you
are
eyes
it
will
or
won't
save.
You
are
not
guaranteed
that
your
eyes
you
give
it
will
be
saved.
You
are
not
guaranteed
that
you
can
give
them.
F
E
F
But
the
the
normal
case
is
that
there
will
only
be
one
two
or
three:
maybe
you
are
eyes
against
a
given
location
of
a
finite
size.
The
other
thing
it
does
is
it
provides
an
input
variable
called.
As
of
that
lets,
you
ask
the
server.
F
Will
this
location
that
I
gave
you
be
valid
as
of
this
date
and
that's
a
way
to
make
sure
that
you
can
create
that
when
this
plan
change
happens
on
a
known
schedule
that
you
can
validate
ahead
of
the
change
that
your
proposed
new
address
will
be
valid
as
of
the
date
of
the
change,
so
you
can
you
can
that
you'll
get
this
notification
that
says
this
is
going
to
become
invalid?
You
know
five
days
from
now.
You
can
three
days
from
that
location.
You
can
say
well
what,
if
I
changed
it
to
this?
F
F
We
did
adopt
that
as
a
working
group
draft
and
the
so
so
I
have
issued
a
0-0
of
drafty
crit,
it's
again
a
fairly
simple
mechanism,
but
if
you
don't
understand
that
lost
works
once
again,
it's
pretty
hard
do
basically
just
adds
this
uri
as
an
input
parameter
that
you
can
save,
gives
you
an
object
that
you
can
send
to
that.
You
are
I
to
tell
you
that
as
a
change
has
been
made
and
gives
you
this
as
a
thing
that
lets
you
say,
will
it
be
valid
as
of
a
specific
date?
F
This
is
a
real
problem
encountered
by
real
implementations
that
they
know
they're
going
to
have,
and
it
also
deals
with
a
fact
that
we
discovered
in
implementation,
which
is
that,
because
phone
bcp
says
periodically
revalidate,
the
number
of
transactions
on
the
lost
server
that
are
revalidation
is
two
or
three
orders
of
magnitude
more
than
the
number
of
other
operations
on
the
law
server.
So
the
law
servers
has
to
be
scaled.
F
Much
much
much
larger
just
because
of
that
periodic
revalidation
we'd
like
to
turn
that
pull
notion
into
a
push
mechanism
so
that
we
can
reduce
the
load
on
the
law.
Sir,
it's
really
the
validation
part
of
the
sir.
So
that's
the
deployment
experience
tells
us.
We
made
a
mistake:
periodic
revalidation
is
causing
a
lot
of
load
that
we
didn't
anticipate
and
we
would
like
to
turn
that
mechanism
around
make
it
push.
F
F
H
F
F
H
F
There's
you
should
look
at
the
text.
The
security,
the
security
considerations
is
an
important
part
of
this
particular
document
and
and
I've
tried
to
think
through
make
sure
that
we
can't
you
can't
get
a
malicious.
You
know
it
helps
that
you
can
only
post
right,
there's
no,
no
data
you're
getting
back,
but
nonetheless
you
could
get
a
you
know.
You
could
get
a
an
error.
You
know
it
could
be.
No.
We.
F
Could
cause
some
innocent
party
to
be
sending
out
sure?
That's
why
they're
you
limit
the
number
right
so
that
you
can't
do
it's?
Not
it's
not
a
way
to
do
an
application
attack.
You
will,
you
will
poke.
You
know
you
will
try
to
push
a
piece
of
document
to
some
things.
Yes,
the
the
you
you
would
like
to
know
who
who
is
requesting
this
you
have
there
occur
you
would
like
to
have.
You
should
have
a
policy
and
I
pretty
sure
I
talked
about
this.
H
F
H
F
Yeah
I
mean,
but
that's
again
in
the
in
the
scale
of
databases.
That's
that's
pretty
small
as
long
as
there's
really
only
three
of
them
right.
If
there
were
5,000
of
them
per
record,
then
maybe
it
would
be
a
concern
and
and-
and
there
is
definitely
advise
about
you-
know
bluemont-
how
many
of
these
you
handle
any
other
questions
or
comments.
I
think
I'm
done.
Reviews.
C
C
This
one
first
Randy
you're
next
is
this
one
first,
oh,
you
can't
see
that
kidney.
Maybe
if
I
do
this
and.
H
H
H
H
So
that's
and
there
are
a
lot
of
editorial
changes
and
I
can
go
on
to
the
next
slide,
which
helps
about
where
I
got
comments
from
yes.
So
we
get
a
lot
of
comments
from
the
number
of
people
I
want
to
thank
all
of
them.
We
had
a
lot
of
editorial
changes
resulting
from
that
and
then
the
big
technical
changes
that
I
mentioned
before
I
want
to
thank
Christopher,
especially
for
spending
a
lot
of
time.
I
think
we've
got
things
in
a
really
good
shape.
H
Now
next
slide
yeah,
so
there's
really
just
to
open
issues
remaining
one
is
a
related
to
where
I
mentioned.
Just
a
second
ago
about
moving
the
stuff
that
wasn't
directly
needed
for
eCall
from
the
Eco
document
to
the
car
crash
and
one
of
those
is
a
capabilities
element
that
gets
sent
from
the
vehicle
to
the
piece
app
to
let
the
pset
know
what
the
vehicles
capabilities
are.
So
the
pset
will
not
request
anything
that
the
vehicle
does
not
support.
H
H
The
other
open
issue
is
about
the
use
of
about
the
info
package
registration
right
now.
The
e
call
document
says
that
the
e
call
outside
the
info
package
I
can
be
used
to
send
a
body
part
a
mime
type,
that's
listed
in
the
emergency
call
data
set,
and
so
there's
been
some
debate
about
whether
that's
legal
/,
the
info
document.
If
it
is
legal,
everything's
cool,
if
it's
not
legal,
then
we
have
a
few
options.
One
option
is
that.
H
So
one
option
would
be
that
we
require
a
new
info
package
registration
that
would
include
everything
in
the
earlier
ones,
for
every
new
mime
type
that
might
get
deployed
in
a
different
region.
Another
option
which
is
kludgy
ER
but
avoids
the
legalistic
argument
would
be
that
we
don't
use
info.
We
use
reinvite
or
options
or
something
like
that,
but
some
other
sit
message
that
doesn't
have
the
legalistic
restrictions.
H
A
third
option
is
that
we
define-
and
this
is
all
so
ugly
and
occlude,
but
perfectly
legal
I-
think
you
define
a
a
mime
type,
that
that
says:
multi-part
emergency
services,
data
and
then
put
a
parameter
in
there
and
what
you
do
is
you
just
only
for
info.
You
just
take
your
actual
mind
type
and
cram
it
into
this
multi-part
as
an
extra
wrapper
around
it,
so
that
the
info
package
only
has
the
one
mime
type
it's
horribly
ugly,
but
it
would
get
around
the
wording
of
the
info
package
restriction.
J
Christer
and
Krista
Homburg,
yes,
first
of
all,
I
also
like
to
thank
you
for
we've,
given
a
lot
of
comments
and
things
which
came
out
from
tree
TPP
other
comments
that
you
had
and
things
are
going
very
smoothly,
but,
as
you
say,
there
are
there,
there
are
few
open
issues.
I
think,
there's
one
more
issue
than
this
also,
which
was
raised
by
my
colleague,
which
you're
discussing
offline,
was
this
usage
of
Content,
ID
and
and
and
and
and
call
in
forward,
but
I.
Think.
J
J
You
didn't
really
know
what
people
supported
and
what
they
were
going
to
do
when
they
receive
is
so
so
we
define
the
info
package,
which
has
all
the
semantics
and,
and
one
thing
which
just
info
package
does
say,
is
that
you
define
which
mime
types
are
associated
with
which
this
info
package
now
the
way
it's
done
here
in
equal
is
a
different.
Instead
of
listening
these
mine
times,
it
says
that
you
can
use
any
mind
x,
mm
I'm,
pipe
which
exists
in
this
specific
I
on
our
registry.
J
So
whenever
people
want
to
use
something
new,
they
don't
need
to
define
an
info
package.
They
just
add
stuff
into
this
I
in
our
registry,
and
that
is
my
main
issue,
because
I
think,
first
of
all
it
cydia
mean
we
don't
know
in
the
long
run
what
people
are
going
to
put
there
as
long
as
they
can
justify
it
somehow
being
in
you
know,
emergency
related
and
another
example
or
I
also
think
it
gives
a
bad
precedence.
J
I've,
given
example,
I
will
give
it
I,
don't
know
it's
the
best
one,
but
assume
that
someone
registers
a
nympho
package
for
transporting
you
know
game
online
game
information
and
they
create
iono
registry
for
online
game
mime
types
we
start
getting
all
these
games,
10
games
hundred
games,
200
games,
they
all
define
their
own
mime
type
and
they
all
put
this
into
the
saints
registry
and
use
the
same
info
package.
So
I
think
it's
it's
it's
a
slippery
slope
and
a
dangerous
way
to
go.
J
So
my
suggestion
is
really-
and
I
think
that
this
original
intention
of
info
package,
when
you
need
something
new-
you
define
a
new
info
package.
I
said
here
earlier,
you
mentioned
RFC,
but
that's
not
true.
You
don't
need
an
RFC
to
define
a
new
in
the
package,
its
specification
required
and
an
expert
review,
so
you
can
define
it
whatever
you
want.
Maybe
any
you
know
the
Chinese
Americans
whatever
they
they
find
is
it
can
be
done
and
so
an
info
packages
rrrr
Chi.
In
that
sense,
so
I
really
think
that
should
be
done.
J
We
list
explicitly
the
info
packages
and
these
are
my
plans.
You
use,
and
nothing
else,
also
what's
important
to
think
that
this
is
not
only
about
the
mine
finds.
Another
thing
which
we
have
in
the
info
package
is
a
section
where
you
need
to
either
define
or
reference
a
specification
which
describes
how
this
info
package
is
used,
and
that
of
course
includes
how
the
mime
types
are
used.
J
So
now,
assuming
that
you're
adding
a
new
mime
time
to
this
registry
which
coming
into
this
info
package,
you
would
also
have
to
update
this
add
a
reference
to
the
specification
which
now
describes
the
usage
of
this
mime
type
and
so
on,
and
maybe
it
could
be
doing
something
else.
Maybe
it
could
affect
some
of
the
other
things
which
are
in
the
in
the
info
package
rate.
J
Whatever
we
have
I
mean
we
don't
know
so,
then
you
would
have
to
update
the
info
package
and
there
is
really
no
procedure
for
updating
the
info
package,
the
actual
template
that
has
been
provided
to
Anna.
There
is
no
no
mechanism
for
that,
so
my
suggestion
and
I
have
said
offline
I
am
willing
to
help
whatever
I
can
do
to
do.
That
is,
we
define
this
info
package.
We
call
for
ikura
cars
crash,
sorry,
they
we
define
our
another
waiting
for
package.
We
can
use
similar
name,
no
problem.
J
We
can
even
it's
a
lot
of
copy
paste.
We
can
do
and
then,
for
you
know,
Asia
whoever
needs
they
do
in
their
info
box.
I
think
that's
a
clean
way
is
the
way
you
know
exactly
what
it's
there.
You
don't
need
to
go
and
look
in
this
rigid
tree
where
there's
something
new
stuff
there,
even
though,
and
so
on
and
I-
think
that's
a
generic
way
to
do
it.
I
know
that
there
is
this
capabilities.
J
In
this
specific
case,
there
make
an
instant
that
we
have
done
so
we
can
know
what
supported
but
I,
don't
think
that's
way.
We
should
go
that
individual
info
packet
is
start
defining
their
own
inclination,
a
mechanism
that
not
what
we
should
do.
There
is
a
common
negotiation
mechanism
which
you,
when
you
send
your
invite.
You
include
this
receive
info
header
field,
which
name
stain
for
package.
If
I
support
this
in
package
and
I
know
which
mine
pies
are
associated
I,
everything
is
good.
J
B
J
F
Prime
rozum,
the
when
you
add
a
new
block
of
data,
all
we're
doing
is
adding
a
piece
of
data
right.
It's
a
hunk
of
XML
that
we're
sending
between
the
entities
right.
We
need
to
decide
what
level
overview
is
appropriate
to
add
a
piece
of
data
using
the
mechanism
for
providing.
If
we
decide
that
specification
review
is
the
right
level,
a
specification
requires
the
right
level
of
documentation
to
add
a
new
thing,
then
adding
to
the
registry
should
have
specification
required
if
we
decide
that
that's
not
an
appropriate
level
of
documentation.
F
It's
too
much
overhead
for
the
for
the
kind
of
information
that
we
have.
Then,
maybe
we
go
back
to
just
an
expert
review
by
requiring
that
you
were
doing
a
new
info
package.
You
are
demanding
that
you
do
specification
required.
That
might
be
what
we
want,
but
it
isn't
a
decision
that
we're
making
it's
just.
You
got
to
do
it
that
way,
because
that's
what
nation,
what
when
m.phil
says
it
is
low
overhead.
We
can
build
a
template
that
says
to
do
a
nympho
package.
You
know
filling
these
blanks
done,
so
it's
not
a
huge.
F
J
Krista
I'm
not
talking
about
standard
required.
What
maybe
I
remember
the
definition.
Maybe
I
remember
the
definition
of
specification
required
wrong,
but
my
what
I,
what
what
what
is
needed
is
there
needs
to
be
a
publicly
available
document
which
takes
per
trigger
account
because
the
expert
reviewer
needs
to
you
know,
look
at
this.
It
doesn't
have
to
be
an
RFC.
It
doesn't
have
to
be
an
opposite
exactly
so
yeah
classification.
E
F
So,
but
but
all
I'm,
Sam
and
just
for
reinforcing
the
same
time,
who's
going
to
repeat
it
right.
We
need
to
decide
what
the
level
of
control
that
we
want,
/,
adding
another
hunk
of
data.
If
we
decide
that
specification
required
is
appropriate
great.
Let's
do
that
by
pushing
my
requiring
info.
We
don't
get
to
make
that
we
get
to
make
it
stronger
right.
We
can
say
and
adding
something
that
a
registry
takes
more
than
specification
required.
F
I,
don't
think
we're
going
to
do
that,
but
we
could,
but
we're
stuck
at
the
minimum
level
of
specification
report
which
may
or
may
not
be
appropriate.
I
happen
to
think
it's
not
appropriate,
but
we
can
we
can.
We
can
make
that
decision
independently.
However,
I
am
not
in
favor
of
playing
games
with
the
the
sip
info
package,
RFC
right.
If,
if
the
SIP
guys
say
no,
you
can't
do
that.
He's
right.
You
got
to
specify
what
mimetypes
going
the
package.
That's
the
end
of
the
story,
we're
going
to
do
it.
F
D
G
Good
yeah
and
Ryland
yeah
I
have
a
lot
of
sympathy
for
krista's
points
here.
I
think
the
reason
we
did
all
this
work
to
come
up
with
with
the
new
version
of
info
with
the
info
packages
rules
to
prevent
the
kind
of
proprietary
explosion
of
the
use
of
info
which
info
wichman
didn't
interoperate
and
I,
feared.
If
we
add
these
ad
mimetypes
our
outsides
of
the
control
of
the
actual
package
that
effectively
you've
got
different
versions
of
the
same
package
that
aren't
going
to
interoperate
as
Brian
said,
I
think
we
can
make
this
a
template.
G
H
So
if
I
can
actually
just
play
devil's
advocate
for
a
moment,
I
think
I
very
well
understand
the
concern
that
you
guys
have
about
the
reason.
There's
info
package
and
the
reason
requirements
and
how
allowing
us
to
do
this
could
potentially
create
a
slippery
slope
that
allows
you
know
the
game,
people
to
to
just
go
hog-wild
with
completely
different
types
of
data,
with
completely
different
semantics
and
and
so
forth,
and
that
would
be
a
mess
but
just
to
play
devil's
advocate.
If
I
was
a
game
designer
trying
to
do
that
sort
of
thing.
H
L
This
is
on
this
is
hummus
yeah.
I
think
the
case
that
Krista
are
provided
here.
Sort
of
de
bourrée
seems
a
little
bit
constructed
I'm,
like
the
fact
that
you
come
up
with
gaming.
The
gaming
community
is
is,
is
no
no,
it's
like
it's
like.
Is
this
a
real
concern,
like
literally
the
gaming
community,
of
writing,
sip
applications
and
then
putting
stuff
in
for
emergency
service
to
must
be
kidding
right
now,
he's
not
saying
they.
L
J
Define
dated
would
define
their
own
iono
registry
for
gaming
mimetypes
and
they
even
then
they
would
one
have
one
single
info
package
which
points
to
that
registry,
and
then
everyone
starts
putting
these
these
mime
types
into
the
same
registry
and
it
all
gets
associated
with
one
single
info
package.
I'll
eat
I
didn't
say
that
they
were
going
to
use
the
emergency
iono
registry,
but
they
were
going
to
dosometing
sim
similar
with
their
own
owner
registry.
J
Basically
allowing
you
to
have
one
info
packager,
then
people
can
start
adding,
adding
adding
mimetypes
internet.
That's
what
I
want
to
avoid,
and
and
of
course
you
know,
anyone
can
define
an
XML
type
which
is
very
flexible
on
it.
We
can't
prevent
that,
but
I
mean
but
but
I
mean
that's
not
a
reason
for
us
to
do.
You
know
the
wrong
thing
just
by
saying
yeah,
but
someone
else
could
do
this,
but
you
know
we
do.
We
should
do
what's
right
and-
and
you
know
that's
over
sure,
but
but
see
if.
H
The
reason
to
make
it's
the
reason
for
us
to
do
something
in
a
more
difficult
way,
which
isn't
the
semantics
that
we
want
is
to
prevent
us
from
being
a
precedent
for
somebody
else
to
do
the
wrong
thing,
even
without
us
being
that
precedent.
People
who
want
to
do
the
wrong
thing
can
do
the
wrong
thing.
Chris.
J
I,
don't
think
it's
easier
because
even
in
even
if
we
would
move
forward
this
way,
there
will
be
a
lot
of
text
needed.
We
need
text
either
an
update
to
RFC
68
is
86
or
something
describing
how
this
is
done.
Yes,
you
can
create
these
registers
and
all
these
things
and
and
we
need
rules,
how
it's
done.
How
do
you
update
the
info
package?
J
C
I
I
M
I
H
G
And
rollin
well,
I
would
hope
that
the
expert
review
of
the
info
package
would
at
least
catch
some
of
those
if
it
looks
like
you're
just
basic
regni
generic
container
to
stuff
in
whatever
you
like
means
gear
up
said
in
20
PP
we
could
create
a
nympho
package
called
free
to
BP
application
and
just
like
for
everything
in
the
future
and
I
have
to
come
back
to
ITF
again.
This
is
not
what
we
want
to
encourage
so
and
I.
Don't
think
it's
really
that
hard
pretty
much.
G
H
But
the
issue
isn't
with
what
we
do
for
eco
and
car
crash.
The
issue
is:
what
does
the
next
region
that
might
need
that
good?
So
ms
t
is
one
set
of
data
in
North
America
Ruiz
beds,
which
is
a
pretty
flexible
set
of
data.
But
what
happens
if
China
or
Australia
or
Saudi,
Arabia
or
some
other
region
or
country
decides
they
want
something
that
isn't
quite
beds?
They
want
their
own
data
set.
So
what
the
RFC
with
the
documents
say
now
is
well,
you
define
a
mime
type
of
you.
H
G
Just
the
document,
that's
all
it
is
publicly
available
document
and
following
that
template,
so
we
provide
them
a
template
and
we
identify
okay
here,
fill
out
your
own
stuff
fill
out
your
own
stuff.
It
doesn't
seem
that
it's
a
huge
piece
of
documentation
and
basically
we
want
good
documentation.
So
we
can
implement
that
thing
without
having
to
go
for
all
kinds
of
hoops
and
hurdles
in
order
to
find
out
what
it
is.
J
J
It
sound
wearing
some
documents
right,
so
so,
no
matter
how
it's
done
there,
jit
they're,
just
not
gonna,
send
an
email
to
iono
and
hey
here
is
our
mime
type
and
then
it's
not
occupant
anywhere
I
mean
it
and
I,
don't
think
even
I,
don't
know
what
the
ITF
procedures
for
registering
new
mime
types
are,
but
I
would
assume
that
that's
not
gonna
pass
either.
Just
someone
saying
please
please
register
is
my
time
without
any
indication.
What's
in
it,
so
they
will
have
to
define
any
info
package
day.
J
J
J
H
H
Actually,
there's
this
really
very
little
say
on
this:
you
know
it's
just
a
mirror
of
it.
I
mean
the
stuff
that
came
out
of.
He
call
was
moved.
The
econ
metadata
control
object
was
moved
in
here.
There
are
a
lot
of
editorial
changes
and
I
think
that's
it.
So
once
we
decide
whether
we're
moving
the
the
capabilities
object,
whether
we're
leaving
it
here
or
moving
it
back
into
eCall,
so
it
forms
like
part
of
the
base
pack
for
extension.
Here,
that's
really
the
only
open
issue
that
affects
this
document.
F
H
G
H
Right,
absolutely
not,
and
particularly
because
what
we're
really
talking
about
here
is
as
various
not
just
public
safety,
we're
talking
about
vehicle,
initiated
emergency
calls
and
we're
talking
about
region.
Specific
data
sets
for
them.
Car
manufacturers
would
probably
want
to
implement
this
all
of
the
regions
in
the
same
vehicle
so
that
the
vehicle
just
knows
what
to
send
based
based
on
where
it
is.
You
wouldn't
couldn't
do
that
in
the
vendor
tree.
You
wouldn't
want
to
Brian.
F
H
C
C
The
document
status,
so
we
have,
we
have
no
individual
drafts
hanging
out
there
targeted
at
egret
right,
so
nobody
has
submitted
anything
for
a
while.
We
have
these
five
drafts
that
are
on
the
screen
that
are
there
three
of
them.
We
called
working
group
last
call
this
week.
Second
working
group
last
call
in
a
couple
of
those
okay,
so
we're
getting
to
the
point
where
our
work
is
winding
down
and
when
we
submitted
the
request
to
have
a
meeting
at
IETF
96,
there
was
25
meeting
requests
more
than
they
had
space.
C
For
so
my
point,
I'm
getting
to
is
we're
almost
if
we
don't
start
wrapping
up
some
of
this
and
winding
down
and
keep
asking
for
meeting
slots,
we're
going
to
start
doing
the
IETF
a
disservice
by
you
taking
a
meeting
so
awkward
and
there's
a
work
group
that
really
does
need
to
meet
so
I'm
couraging
people
to
review
the
working
group
last
call
documents,
so
we
don't
have
basically
don't
have
to
meet.
You
know
in
Seoul
I
mean
that
a
alyssa
is
going
to.
C
You
know
overpower
me
in
any
way
shape
or
form,
but
we
got
to
get
through
these
issues
that
between
and
the
car
crash
and
II
call
I
think
those
are
not
insurmountable
issues,
I
think.
If
we
pay
attention
to
those
in
work
through
it,
we
can
get
those
documents
done
prior
to
to
Seoul,
so
I
mean
that's
that's
from
a
chair
position.
That's
what
I
would
like
people
to
to
start
working
on.
F
It
depends
on
reviews,
it
depends
on
reviews,
I'm
going
to
work
hard
to
get
reviews
so
that
we
can
do
those
things,
but
I
do
believe
that
we
that
there
is
more
work
to
do
and
that
more
work
is
a
result
of
deployment
experience
that
works.
We're
having
I
know
at
least
two
things
that
we
want
to
do
so
I.
You
know
they
can
be
done
as
individual
drafts.
We
can
do
it.
You
know
we
can
leave
the
mailing
list
open,
I,
understand
all
those
kinds
of
things,
but
I
think
it
would
be
good.
F
C
E
So
that
would
that
would
be.
My
recommendation
is
after
we
get
through
the
rest
of
these,
that
we
go
and
go
into
dormant
state,
and
then
people
can
always
bring
the
list
with
new
things
as
they
come
up
and
if
you
know
a
year
goes
by
or
something
and
we
change
our
minds.
We
can.
We
can
close
the
group
or
just
keep
the
list
or
we
have
a
lot
of
options.
Sorry,
that's
a
nice
addition!
I.
C
C
G
Andrew
I'm,
just
one
comment
on
the
slide:
mark
preach
previous
working
requirements
for
like
service
I
think
it
is
the
same
service,
eco,
sorry
for
the
ecole,
just
a
comment
on
the
wording
of
the
slide.
Ok,
it
really
is
there
isn't
the
IETF
eco
and
if
they
disappoint
the
chief
mate
and
the
free
to
be
p,
equal,
there's
a
single.
E
I
think
his
point
is
just
that
we
basically
did
this
for
that
for
the
service
being
developed
in
3gpp.
We
wrote
the
draft
here
for
that.
So
it's
just
one
service
but
I
mean
I,
guess
the
time
but
I'm
assuming
the
timing
on
that
is
like
we
basically
know
everything
like.
If
we
resolve
these
last
couple
open
issues,
we
can
go
to
a
group
last
call
and
go
to
ITF
last
column.
That
will
be
okay
for
a
3gpp.
Is
that
correct.
G
E
G
C
J
J
Think
to
work
is
gonna
is
already
pretty
stable
as
far
as
a
3d
BB
work
is
concerned.
Obviously
they
reference
and
we
have
all
this
open
issues
and
I
know
they
have
a
meeting
next
week,
so
they're
gonna
be
they're
gonna
be,
but
I
think
it's
I
said
it's
it's
there,
no
main
issues
I
mean
we.
We
found
issues
that
previous
meeting
but
they'll
have
those
have
all
been
taken
care
of,
say
things
very
much
went
much
smoother
than
it
thought
so.
C
Okay,
so,
but
you
you
remember
the
original
idea,
we
were
going
to
run
it
through
IETF
last
column,
before
Alyssa
hit
the
pub
request
button,
we're
going
to
check
practice,
see
if
there's
anything
else
we
needed
we
hadn't
covered
from
3gpp
perspective,
so
I
mean
we
can
still
follow
that.
But
if
you
know
something
that
says,
we
don't
have
to
deal
with
that
any
more
than
you
know.
Please
it
please.
Let
us
know:
okay,
I.
G
Wouldn't
say
we're
quite
that
right
now,
but
I
guess
we
could
also
initiate
as
well
right.
We
could.
We
could
VA's
to
ITF
decreed
from
free
to
be
p.m.
so
we
think
this
is
good.
Now
so
don't
hold
it
anymore
or
if
you're
waiting
one
out
since
then
sir
communication
and
we'll
we'll
get
back
to
you.
Okay,.