►
From YouTube: IETF104-NETMOD-20190325-1350
Description
NETMOD meeting session at IETF104
2019/03/25 1350
https://datatracker.ietf.org/meeting/104/proceedings/
A
Our
last
session
was
a
little
bit
rushed
and
because
of
it
being
a
joint
session
with
net
confer
following
with
net
cough,
we
did
run
out
of
time
to
really
finish
the
discussion.
The
visit
we
mentioned
in
that
session
that
the
design
team,
the
versioning
design
team,
has
a
session
set
up
for,
has
a
session
set
up
for
tomorrow
and
Rob
Wilton
just
came
over
and
said
that
they'll
be
they'll,
be
sending
out
the
information
to
the
list.
So
if
you're
interested
in
participating
in
follow-on
discussions,
please
feel
free
to
go
there.
A
It's
not
a
formal
working
group
meeting,
but
it
is
with
the
members
of
the
design
team
that
are
here.
There
should
be
good
discussions
and
the
action
from
that
will
be
to
send
any
updates,
send
any
conclusions
from
there
to
the
list
for
for
follow-on
discussion.
So,
if
you're
interested
please
come
whoops.
A
So,
for
some
reason
it
feels
like
we're
late
in
the
week,
but
we're
really
not
we're
still
Monday.
So
some
may
not
have
seen
the
note
well
yeah.
This
is
the
standard
reminder
that
what
you
say
here
is
considered
an
IETF
contribution
and
that
it's
governed
by
our
rules
and
if
you're
interested
either
go
to
that
link
at
the
bottom
of
the
page
or
please
CBCP
79,
but
basically
anything
that
you
say
here
becomes
part
of
our
public
record.
A
A
A
So
we
have
two
RFC's
that
have
been
published
as
the
last
meeting.
The
first
one
I
think
is
worth
noting,
because
that
was
in
the
working
group
a
long
time.
So
having
that
published
is
a
really
nice
accomplishment
and
we
appreciate
all
those
who
contributed
to
it
and
scheme
amount
was
another
one
that
took
a
little
while
to
get
done
and
took
a
lot
of
hard
work,
and
we
appreciate
that
being
everyone
who
could
treated
get
us
there.
So
thank
you.
A
A
A
So
we
have
a
number
of
working
group
documents
on
the
agenda.
We're
going
to
talk
about
those
more
in
a
moment
not
on
the
agenda
are
that
these
two
documents
are
listed
basically,
a
few
weeks
ago,
the
authors
said
they
think
it's
ready
for
last
call
at
the
last
meeting
we
talked
about
it
being
close,
we
felt
it
was
a
little
too
close
to
come
to
try
to
do
the
last
call
right
before
the
IETF,
so
we're
planning
to
start
this
next
week.
Last
call
on
these
two
documents.
A
A
Thank
you.
Thank
you,
I
think
that's
our
intro
and
before
we
go
to
the
lodge,
we
wanted
to
have
Robert
speak
for
a
moment
on
a
change.
That's
happened
inside
the
general
IETF
and
specifically
related
to
yang
catalog.
So
Robert
is
the
project
manager
for
the
IETF
for
Yan
catalog.
So
you
just
give
us
update
what's
happening
there.
D
So
hopefully,
everyone
is
familiar
with
yang
catalog
set
of
tools.
It's
gathering
yang
modules
from
min
EST,
o--'s
and
implementation
communities
and
tools
to
develop
new
yang
modules.
The
operation
of
yang
catalog
org
is
moving
to
the
IETF
Secretariat.
We
have
a
project
underway
that
is
moving
the
the
machine
to
a
Secretariat,
managed
machine
and
adding
a
few
features
such
as
bringing
in
the
yang
validator,
dot-com
functionality
and
continuing
to
work
on
the
enhancements
and
bugs
that
the
community
identifies.
D
A
E
I'm
Bob
Sonya
from
erikson
speaking
about
your
instance
data
format.
This
has
been
an
ongoing
draft
for
some
time
and
we
had
some
new
changes,
so
the
concept
is
well
are
the
same.
We
need
yang
defined
data,
sometimes
offline,
not
just
from
rest
confronted
core
but
in
the
file
format
or
some
other
form
other
format.
E
So
what
changed?
But
we
have
actually
to
revision
changes
for
the
draft.
First
of
all,
we
have
the
concept
of
tARGIT
yang
modules,
so
these
are
the
yang
modules
for
which
we
are
defining
the
yang
data,
and
there
are
three
methods
now
to
define
which
modules
we
define
the
data
for
it's
not
just
the
modules.
Actually,
it's
the
modules
which
Reverend
date
of
that
module,
which
supported
features
and
which
deviations,
because
all
can
affect
whether
the
instance
data
is
usable
or
not.
E
One
method
would
be
to
put
it
in
line,
which
means
that
it's
in
the
same
XML
or
JSON
data
structure,
as
the
data
itself
contained
in
the
same
file.
Another
is,
if
you
think
about
the
use
case,
where
you
know
providing
diagnostic
information,
every
five
minutes
in
the
file.
You
don't
want
to
repeat
it
and
repeat
it
and
repeat
it
the
what
are
the
tortured
modules.
Then
you
just
use
a
URI
that
points
out
another
file
which
will
contain
this
information,
and
it
is
repetitive
files.
E
You
only
have
to
have
the
URI
and
in
some
cases
you
just
can
say
it's
externally
defined.
Whatever
is
the
context,
some
someone,
some
method,
you
know
what
files
you
are
using
or
what
yang
models
you
are
using.
This
was
redefined
to
yank
choice,
so
it's
easy
to
add
new
methods
if
we
want,
for
example,
package
based
method.
E
Connected
to
the
same,
where
work
it
was
raised
that
there
might
be
additional
data
connected
to
the
yang
library.
That
would
be
useful
for
identifying
the
target
target
modules.
For
example,
if
we
ever
introduced
December
version
or
sample
other
version
number
that
will
help
also
identifying
the
target
modules,
so
the
target
pointer,
which
is
actually
a
bit
of
a
data
structure
we
inside
the
choice,
allows
you
to
add.
Yes,
this
is
the
base
fire
that
will
tell
you
the
module
names,
revision,
etc,
but
I
also
support
an
augmenting
module.
E
E
Then
I
got
one
comment
that
it's
not
Oh,
IETF
yang
library
serves
our
purpose,
that
it
defines
module,
name,
revision
of
features
and
deviations,
but
maybe
I
would
like
to
use
some
other
yang
module
to
define
the
same
data
so
now
it's
possible
to
use
some
other
yang
module
for
that
also
earlier.
The
part
that
defined
the
target
yang
modules
was
mixed
up
in
the
content
data,
with
the
real
data
that
I
want
to
provide.
E
E
So
that's
for
that
reason.
Semantic
versioning,
for
example,
doesn't
apply
to
it
there.
There
are
a
few
considerations
about
backwards
compatibility.
So
if
you
documenting
some
list
data,
you
should
not
change
the
key
value.
If
there's
meaning
is
still
the
same
or
you
should
not
reuse
a
key
value
for
something
that
yet
has
has
not
changed,
and
this
is
just
an
example
how
this
looks
like
so
we
have
still
have
the
instance
data
set
name
for
it.
E
Some
other
additional
metadata,
which
is
not
not
everything,
is
listed
here,
and
then
we
have
this
inline
target
spec,
which
says
that
dot
yang
modules
are
defined
inside
the
same
data
set
inside
the
same
file,
and
then
we
have
the
target
module
separated
out
from
the
content
data.
To
tell
you
that
this
instance,
data
set
contains
data
above
about
ITF
net
of
ACM
provision
date,
which
is
not
mandatory
anymore.
In
some
cases,
it's
better
to
at
a
timestamp.
If,
let's
say
it's
a
file,
that's
produced
every
5
minutes.
E
On
the
other
hand,
if
this
is
instance,
data
file
is
something
that's
produced
by
human
design
activity.
Then
the
revision
date
is
still
or
the
data
is
still
the
good
thing
to
use,
and
then
this
would
be
the
content
part
just
couldn't
fit
it
on
the
same
page.
So
this
is
for
access
control,
common
example
that
provides
some
default
access,
control
rules.
E
E
F
G
E
Uh-Huh
you
raise
that
question
that
the
word
dodge
it
might
not
be
the
best
for
this
think
that
modules
for
which
we
are
defining
data
I,
would
really
appreciate
if
someone
could
come
up
with
a
better
word
for
it.
I
don't
want
the
full
sentence
there,
but
you
but
I
agree
with
you
that
it's
I
will
say
it's
not
bad,
but
this
is
not
the
best.
E
Personally,
I
think
it
would
be
acceptable
about
this.
Some
people
have
wanted
it.
Now:
it's
not
mandatory
to
use
the
yang
library.
What
is
mandatory,
that
the
first
module
that
you
use
in
these
tARGIT
yang
modules
must
be
able
to
specify
name,
revision
features
and
deviations,
because
these
four
are
clearly
needed.
Okay,.
B
So
first
Kent,
as
a
contributor
I,
did
some
email
this
morning
review
comments
on
this
draft,
which
we
could
take
potentially
as
last
call
comments
if
it
were
to
be
taken
into
last
call,
but
I
do
think
it
needs
some
work
still
naming
you
know
as
part
of
it
and
also
perhaps
some
structure
like
related
issues.
All.
B
A
Good
if
we
had
a
little
more
opportunity,
a
little
more
comments
on
the
list,
so
please
take
a
look
at
it
and
see
if
you,
if
you
like
the
version,
let
us
know
if
you
have
comments
on
it,
send
it
and
after
we
see
some
more
review
comments,
we
can
gauge
where
we
are
on
last
call.
So
it's
really
important
to
to
comment
on
this
version,
because
that
will
help
guide
us
towards
whether
you
think
you
know
whether
we
move
towards
last
call
but
I
right
now.
I
J
K
L
Okay,
Marty
Marilyn,
presenting
the
data
extension
draft
it
actually
in
the
last
IETF
or
go
to
the
next
slide.
The
last
ITF
I
think
it
was
decided
to
an
analyst
to
move
to
a
new
keyword,
not
young
beta,
but
instead
structure.
So
the
module
is
now
called
ITF
young
structure,
extensions
and
the
statements
are
structure
and
augment
structure.
So
those
are
the
two
extensions
that
we
define
and
here's
just
a
simple
example
of
of
a
structure
that
could
could
be
defined
one
from
the
subscription
draft.
L
L
So
we
believe
that
this
draft
is
more
or
less
ready
for
working
with
plastic.
All
we
did
recognize
that
we
need
an
example
of
so
one
thing
with
the
structure
is
that
it
can
be
used
if
I
go
back
to
this
previous
slide,
a
structure,
we
find
something
that
that
is
not
you
know,
part
of
any
data
store
anything,
and
it
doesn't
say
where
it's
where
it's
used.
L
We
can
be
used
as
a
top-level
thing
in
an
SS
document
or
a
protocol
data
unit
or
as
in
this
example,
it
can
be
used
in
lining
something
else
like
in
that
conf
error
info.
You
can
you
can
you
could
put
anything
in
there,
for
example,
you
could
put
a
structure
in
there,
so
we
need
to
add
an
example
of
that,
probably
the
example
we
just
saw.
L
So
we
want
to
do
that
and
then
we
believe
that
the
draft
is
ready
for
working
class
call,
but
now
we're
also
working
on
this
yang
next
thing.
So
a
bigger
question
is:
should
we
continue
with
this
as
extensions
or
should
we
try
to
fold
it
into
the
yang
one
or
two
or
yang
next
work?
Instead,
yes,
that's
a
question
for
the
working
group,
so.
A
A
Years
from
now
and
before
before,
we
would
have
that
1.2
if
we're
lucky
18
months
and
I,
think
that's
really
aggressive.
So
the
question
is:
is:
do
you
want
to
wait
that
long
for
this
or
make
it
an
extension
and
gain
experience
with
it?
That
then
can
be
incorporated
into
1.2
and
so
I
think
it'd
be
useful
to
hear
from
the
authors
on
it
and
then
from
the
working
group
on
that.
L
Yeah,
so
this
I
mean
this
is
it's
based
on
the
experience
of
young
data
that
we
already
have.
So
this
is
like
the
second
iteration
of
this,
so
the
question
is
should
with
at
most
or
third
iteration
directly,
but
maybe
it's
I
mean
yeah,
it
might
be
okay,
I
mean
this.
Is
the
documents
on
is
almost
ready.
We
can
publish
it.
A
A
B
No
I
think
it's
more
of
a
chair
comment
is
braid
the
dependency
for
this
document.
I
think
zero
touch
document
is
no
longer
having
dependency
on
this.
It
moved
ahead
with
the
rest,
cough
yang
data
approach,
I
think
the
only
current
remaining
dependency
is
for
the
gang
push
messages.
Instance,
they
died.
L
E
M
L
M
Be
really
in
favor
of
doing
this
and
the
question
regarding
the
instance
data.
So
would
it
be
a
real
problem
if
my
lashes
draft
doesn't
use
this
at
all,
because
I
think
that
it's
just
a
separate
egg
module
so
there's
no,
no
interaction
with
other
and
defined
stuff,
so
I
think
and
anyway,
somebody
can
also
set
up
and
that
conserver
keeping
this
kind
of
data
like
yang,
catalogued
us
something
similar.
So
I
I
really
don't
see
any
reason
for
using
this
extension
in
that
particular
document.
It's
something
else
for
the
error.
L
M
B
L
M
L
N
E
O
A
B
Kent
as
a
contributor
and
I
mentioned
this
before
privately
as
author,
the
excess
right
now
is
underscore
ext
and
I'm.
Just
wondering
if
it's
important
to
maintain
that
the
structure
writes,
there's
structure
underscore
or
txt
and
I
mean
I,
guess
the
only
way
I
can
see
why
we
might
want
to
have
a
convention
for
extensions
always
having
you
know
in
the
module
name,
extension
or
ext,
but
it
also
is
a
little
bit
strange.
I,
don't
think
there
would
ever
be
another
module.
B
A
L
A
L
L
H
L
Yeah
that
we
shouldn't
do
this
at
all.
If
we,
if
we
were
to
believe
that
you
know,
if
you
use
plain
yang
to
to
mean
almost
anything
I
mean
so
currently,
if
you
define
something
in
yang,
you
define
schema
notes
for
a
datastore
conflict
or
coffee
falls.
That's
what
it
is,
and
now
we
we
would
change
that
definition.
So,
by
looking
at
you
a
young
module
with
normal
definitions,
you
don't
really
know
what
it's
supposed
to
be
used
for
right,
maybe.
H
H
L
H
Fine,
that's,
okay!
Do
they
want
to
implement
it?
They,
if
they're
importing
what
they
don't
understand,
then
that's
their
own
problem.
I!
Don't
think
that
someone
I
worry
about
it's!
The
other
comments
go
if
you
back
to
your
previous
slide,
or
you
said
that
now
you
have
a
container
effectively
the
top
level
and
you
always
have
to
have
that
extra
level
where
it's
used
I,
just
wonder
whether
there's
gonna
be
some
cases
like
arrow
strings
and
things
whether
that
might
actually
become
slightly
unwieldy.
To
have
that
way.
E
Bajrangi
Erickson
instance
data
actually
could
use
this
one
or
could
use
a
simple
yang
module,
but
I
think
that
it's,
the
fake
is
a
kind
of
a
special
usage.
You
don't
really
want
a
technical
data
in
the
naka
model
and
then
once
more
in
the
instance
data
model.
So
it
is
more
correct
for
to
use
this
structure
than
viewing
it
as
a
normal
module
can
be
done.
N
N
L
I
B
A
L
A
C
A
Yang
data,
so
clearly
we
need
to
do
address
those
before
our
request
calls
how
many
I'd
like
to
just
get
a
sense
of
the
room
of
who's,
read
it
and
how
ready
people
think
it
is
with
the
caveat
of
these
two
items
just
so
that
we
get
a
feel
of
whether
we
try
to
wrap
this
up
before
the
next
meeting
or
wait
one
more
meeting
cycles.
So
that's
really
the
objective
of
these
questions.
So
how
many
have
read
this
version?
A
A
There
are
additional
issues
that
those
who
have
read
the
document
think
that
need
to
be
addressed.
So
we've
definitely
heard
some
issues
and
we've
heard
some
rehashing.
Is
there
anything
else?
That's
out
there
to
me.
My
takeaway
is
we're
close,
but
we
definitely
need
more
reading
of
the
document.
So
maybe
a
little
bit
like
the
last
document.
Let's
see
if
we
can
get
the
update
out
there
get
some
more
comments
and
I
think
for
both
documents.
A
C
I
A
B
Hi
I'm
Kent,
I'm
gonna
present
Hanley
long
lines
for
in
inclusions
for
internet
drafts
and
RFC's.
This
is
presented
before
a
dream:
Ferrell
presented
a
103
I'm,
not
really
gonna,
go
through
this
slide
of
changes.
You
can
look
at
the
diffs
that
have
been
submitted
to
github
or
actually
using
data
tracker
tool.
There
are
three
issues
that
I'd
like
to
discuss
with
the
working
group
today.
B
B
This
idea
and
the
tool
chain
will
be
subsequently
easier
who
target
the
v3
output
only
so
he
really
objected
to
this
idea
of
using
the
word
artwork
and
so
I
changed
it.
So
now,
if
you
look
at
it
no
longer
says
our
work,
even
though
the
the
name
of
the
file,
the
name
of
the
draft,
is
still
having
artwork
folding.
B
In
the
name
which
actually
now
been
transitioned
over
to
using
the
word
inclusions,
so
you
can
see
in
bright,
yellow,
therefore,
I'm,
sorry
for
the
poke
folks
and
back
I
understand
it
might
be
difficult
to
see
the
black
background,
but
in
bright
yellows
handling
long
lines
in
inclusions
in
Internet
drafts
RFC's.
So
sometimes
it's
using
the
word
inclusions,
which
is
actually
from
the
ex
include
draft,
which
you
know
they
are
XML
to
RFC.
They
use.
X
includes
too,
that
the
use
of
the
word
inclusion
is
pretty
commonplace
within
the
tool
chain,
the
tools
team.
B
So
so
that's
that
nice
work
for
them
use,
but
it's
kind
of
not
the
word.
I
use
everywhere,
sometimes
I
just
say:
text
text
content,
it's
kind
of
in
the
nature
of
the
sentence,
so
anyway,
I
just
want
to.
Let
you
know
that
this
change
is
in
the
current
version
of
the
draft
I
view
it
as
an
editorial
change.
But
if
anyone
has
any
concerns
about
it,
please
let
say
so
now.
B
Okay,
so
we're
gonna
plan
to
move
forward
with
this
current
language
next
is:
should
we
also
have
a
single
backslash
option,
and
so
this
has
been
discussed
on
a
list,
and
but
there
was
no
clear
consensus
on
the
list
so
trying
to
get
the
consensus
here
today.
The
proposal
is
for
to
also
support
both
headers,
so
current,
you
know
so
there'll
be
one
header,
bright,
yellow
with
note
with
a
single
backslash
line
wrapping
and
then,
whereby
you
know
you
know,
one
folding
only
occurs
on
the
maximum
column
and
always
continues
on
column.
B
So
the
the
two
headers
look
similar,
there's
only
the
one
character
difference
between
them,
but
that
one
character
difference
is
sufficient
for
human
readers
and
programs
to
understand
the
nature
of
what
they're
looking
at
and
how
to
unfold
it
or
to
read
it
so
so
the
question
is
I.
Guess
we
have
some
options.
One
is
a
leave.
The
draft
as
it
is,
which
is
to
say
we
only
support,
option
number,
two
or
B-
modify
the
draft
to
also
support
option
one
and
two
and
there's
also
a
sub
question
with
related
to
B.
B
If
we
actually
agree,
we
would
like
to
support
both,
which
is
how
do
we
support,
and
that
was
the
math
that
I
sent
out
the
other
day,
and
we
can
talk
about.
It
makes
sense
to
do
that
or
not
so.
First
I
just
like
to
get
it,
and
since
this
is
a
an
already
adopted
working
group
document,
I
can
ask
the
room
myself.
It's
not
really
a
chair
question
right.
No.
B
P
B
So
if
the
math
that
I
sent
out
the
other
day
became
like
one
in
200
was
the
probability
is
pretty
high
that
there'd
be
a
collision
uncomfortably
high.
So
so,
while
it
would
work
most
of
the
time,
it's
definitely
not
gonna
work,
sometimes-
and
it's
nice
to
have
that
fallback
solution
where
you
can
I
haven't.
L
Martin
Richland,
of
course
there
are
cases
where
the
pretty
thing
doesn't
work.
If
you
have,
if
you
need
to
fold
them,
you
have,
you
know,
lots
of
white
spaces
that
needs
to
be
part
of
the
or
even
just
one
weight
space.
Well,
then,
then,
if
you
didn't
you
shouldn't
fold
that
character
right,
so
you
need
to
be
careful
where
you
fold
smart.
B
L
Fairly
easy
also
for
a
program
to
to
know
fold
on
the
white
space
or
just
before,
just
after
the
white
space
I,
don't
think
that's
a
big
thing,
so
I
mean
the
math
is
I
have
to
admit
that
I
didn't
study
or
all
of
them,
but
I
mean
the
the
characters
are
not
just
random
characters
that
we
want
to
show
here
right.
We
talked
about
like
XML
JSON,
and
things
like
that.
So
I'm
not
sure
that
the
math
actually
works
in
those
cases.
Yeah.
B
C
B
All
have
like
pluses
and
equals
symbols,
and
things
like
that,
but
never
a
backslash
basics
before
is
one
of
the
things
that
typically
are
very
long
lines
and
wrapping
is
something
we
want
to
support
for
them,
but
like
XML
and
s,
strings,
which
are
also
something
typically
has
lone
lines.
Never
would
be
fine,
a
backslash
in
them,
I,
don't
think
not.
Even
in
the
the
end
of
the
element
character,
though,
there's
one
there
right,
yeah.
Q
B
C
P
Christian
hop
so
to
follow
up
on
that
I.
If
there's
a
choice
between
having
to
work
a
little
bit
harder
in
a
tool
to
use
a
single
solution,
I
think
it's
better
than
having
a
complex
standard
right,
I
mean
I.
Think
I
mean
we're
talking
about
automatic
folding.
If
it's
possible
to
always
do
it
with
one
choice:
I
think
we
should,
even
if
it's
a
little
bit
harder
I,
don't
know
that's
what
I
think
is
a
cost
benefit.
A
B
Think
there's
an
xkcd,
you
know
artwork
cartoon
if
it
Oakland
so
okay,
so
trying
to
get
some
sort
of
than
the
questions
I
have
on
the
screen
aren't
quite
right,
I
guess!
First,
we
need
to
ask,
if
you
know,
should
we
only
have
one
solution
without
even
talking
about
what
it
is
with?
One
is
versus
more
than
one
and
I
know.
That
sounds
a
little
bit
weird,
because
why
would
you
ever
want
more
than
one
but
again,
there's
complexity,
there's
a
little
bit
of
complexity
involved
and
the
one
that
we're
talking
about.
B
If
you
have
a
second
one,
it's
just
because
you
have,
it
doesn't
mean
you
have
to
use
it
and
it
could
it's.
Therefore,
the
fallback
right
so
for
the
I
think
I
put
them
female
for
the
programs
that
are
trying
to
do
the
automated
folding
they
do.
The
quick
test
you
know,
can
I
do
this
single
backsplash
approach
and
one
out
of
200
times
they'll
fail
and
it
says:
okay,
I'll
do
the
double
backslash
approach,
so
you
know
half
of
a
percent
of
the
time
it'll
have
to
use
the
backslash
approach.
B
H
B
B
Alright,
so
I've
heard
two
who
are
primarily
interested
in
just
having
the
one
solution.
One
I
guess
two:
if
you
count
myself
as
preferring
actually
having
two
solutions,
but
still
that's
not
enough
to
get
a
basis
for
the
sense
of
the
room,
so
I'll
ask
the
questions
again.
I
guess
a
would
be
leave
the
draft
as
it
is
meaning
we
only
have
a
single
back
double
backslash
approach
that
we
currently
have
or
B.
We
would
add
also
the
single
backslash
approach,
which
would
be
as
Rob
says.
A
B
You
only
do
one
one
one
double
I
yeah,
it's
one
delay
all
right,
yeah
so
I
mean
so
if
so
there's
so
there
would
be
a
case
where,
if
you
had
a
line
that
contained
nothing
but
spaces,
the
entire
line
like
that
went
past
max
care
column
right.
So
it's
a.
Why
would
you
ever
do
this?
I
don't
know,
but
that's
the
case.
B
Wait
a
second!
So
that's
the
one!
That's
unsolvable!
There's
other
cases
where
you
might
have
some
texts
and
then
some
like
and
you're
not
doing
the
smart
folding.
So
you
know
that
you're
not
writing
the
extra
19
lines
of
Python
code
you're,
just
using
the
1i
nook
where,
where
you
want
to
do
the
cut
on
the
max
call,
but
the
very
next
character
is
a
whitespace
whitespace
character,
so
that
would
then
fall
into
this
case.
It's
not
be
a
false
positive.
Now
we
could
write
a
smarter
folder,
but
if
you
don't
write
the
cellar.
K
B
B
Nobody
show
hands
for
those
who
think
that
we
should
support
both
the
double
backslash
approach
and
also
the
single
backslash
approach.
That
is
the
pretty
backslash
approach
and
I.
Think
you
want
to
know
whether
or
not
there's
gonna
be
a
third
option.
Okay,
so
there
will
be
a
third
option,
which
is
the
only
supporting
the
single
backslash?
Okay.
B
K
K
B
K
K
B
B
So
you
would
then
change
the
header
to
instead,
starting
with
the
word
note
to
start
with
the
word
start
and
then
you
would
also
have
a
footer
that
just
was
the
identical
thing,
but
ending
with
word
end
to
me
personally,
as
you
know,
at
the
motivation
for
wanting
to
have
a
footer
is
unclear.
Nor
is
there
a
request
by
the
RFC
editor
to
have
one,
but
you
know
what
would
the
motivation
be?
You
know.
Is
it
to
aid
extractions?
B
You
know
so
I
mean
I,
don't
there's
still
gonna,
be
lots
of
artwork
slash
source
code,
that's
not
folded
and
those
those
those
need
to
be
extracted.
Also
so
you
know
maybe
would
aid
some
instructions,
but
not
all
extractions,
so
it
doesn't
really
solve
the
bigger
problem
and,
besides
I
strongly
believe
that
we
should
move
away
from
trying
to
extract
from
plain
text.
Rfc's
xml
is
now
being
published,
I
don't
know
if
you've
seen
it,
but
now
in
the
link
where
you
know
we
click
on
you.
Your
choices
are
PDF
text
XML,
it's
now
available.
B
If
you
can
actually
download
the
XML
for
any
draft,
so
you
know
why
wouldn't
we
be
extracting
from
XML
instead,
so
I,
don't
see
it
really
aiding
instructions
and
is
it
aiding
readers?
So
you
know
again
what
about
all
the
unfolded
inclusions?
You
still
have
the
need
for
those
and
then,
lastly,
is
it
even
an
issue?
I,
don't
never
heard
anybody
having
an
issue
with
reading
and
mobiles.
B
H
People
are
going
to
use
this
against
text
RFC's
as
well,
unless
that,
unless
you
take
that
outside
the
scope
and
say
this
is
not
to
be
used
for
text
RFC's
and
the
other
column,
the
drafter
says
it
says
to
be
used
in
ITF
draft
and
can
also
be
used
in
other
cases.
I
think
there's
many
cases
where
your
document
you're
running
this
tooling.
You
don't
want
to
to
carry
on
through
the
entire
document
doing
this
unfolding,
where
stash
might
have
some
other
meaning
correct.
B
H
B
No,
so
how
it
got
extracted
is
a
separate.
You
know
like
a
precondition
to
some
other
piece
of
tool.
Tooling,
they
did
the
extraction,
whether
it
be
from
an
XML,
R,
star,
seed
document
or
some.
You
know,
third
party
s
do
somehow
someone
got
it
extracted,
it
that's
their
problem,
but
then
they
feed
it
into
the
unfold.
Er.
Okay,.
B
So
there's
two
options
and
there's
no
need
for
a
third
option,
so
le
the
drafters
is
meaning
that
we're
only
supporting
the
header
and
be
modified.
That
draft
to
support
both
a
header
and
a
footer
so
can
I
get
a
raise
of
hands
for
those
who
think
that
we
should
leave
the
draft
as
it
is
only
supporting
the
header,
raise
your
hands,
please,
okay
and
then,
or
those
who
think
we
should
modify.
The
draft
support
both
I
had
her
and
a
footer.
B
Next,
all
right,
so
there
has
been
some
three
virtual
interims
one
on
February,
6
xx
and
then
March
2
xx.
The
focus
for
all
three
was
on
scoring
the
game.
Next
issues.
There
was
a
total
of
70
issues
that
were
entered
over
the
course
of
three
years.
The
the
gang
next
repo
was
created
on
March
11
2016.
This
presentation
is
reviewing
all
the
this
preacher.
This
preacher
table
reviews
the
results
of
those
meetings
and
there
will
be
a
deep
dive
meeting
on
this
Wednesday
in
carlin
3
between
3
&
5.
B
Next,
please,
okay,
so
first
I
wrote
a
Python
script
that
would
reach
out
to
github
and
extract
all
the
issues
and
run
it
through
Matt
plot,
and
so
this
is
a
3d
view.
It's
impossible
to
see
cuz,
it's
not
interactive
with
a
mouse
that
you
can
rotate
and
it
looks
very
nice,
but
okay,
but
I
did
then
next
slide.
Please
I
did
a
couple
on
axes
views
so
I
actually
didn't
mean
to
do
these
because
they're
not
terribly
important,
but
on
this,
on
the
left
hand
side,
you
see
complexity
times,
importance
and
I.
B
Think
the
the
bottom
line
to
take
away
is
that
most
of
the
important
issues
have
low
complexity.
That's
how
I'm
reading
this
particular
view
and
then,
on
the
right
hand,
side.
You
have
complexity
times
backward
compatibility,
and
the
takeaway
here
is
that
most
of
the
backward
compatible
issues
have
low
complexity.
B
B
Seems
like
I'm
missing
something
in
the
background,
but
anyway
I
think
it's
just
this
one
picture
anyway.
So
so
this
slide
is,
is
showing
backwards-compatibility
times
importance.
So
it's
sort
of
an
on
axis
view,
but
these
are
the
two
axes
that
we
really
care
about
when
we
think
about
what's
important
right
we
want
to.
We
want
to
do
what's
important,
but
the
backward
compatibility.
If
it
says
it's
a
high
backward
compatibility,
it
means
that
it
could
be
a
yang
1.2,
we're
not
kind
of
break
compatibility
with
the
existing
game.
It
says
low
backwards
compatibility.
B
B
This
is
the
only
slide
where
there's
kemp's,
so,
if
you're
interested
in
looking
at
the
how
many
of
the
issues
fall
into
the
various
levels,
this
is
the
slide
to
be
focusing
on
to
see
that,
but
you
can
see
there's
only
you
know
for
the
low
backward
compatibility,
there's
only
three
issues
that
have
been
scored
as
actually
being
low
backward
compatibility,
but
there's
actually
eight
issues
that
have
unknown
backwards
compatibility
and
any
of
those
eight
issues
could
potentially
be
backwards.
Compatibility
breakers
on
next
slide,
please.
B
So
this
light
is
showing
the
same
view
as
before,
but
because
the
unknowns
are
kind
of
hard
to
interpret,
I
took
the
value
of
the
unknowns
and
sprinkled
them
evenly
into
the
low
medium
high.
So,
for
instance,
if
I
said
there
was
two
unknowns:
I
added
two
to
the
low
and
I
added
two
to
the
medium
and
I
added
two
to
the
high
and
pretended
this
view,
I
just
really
unfortunate
that
the
text
is
kind
of
occluding
the
graphics
here,
but
the
size
of
the
the
circles
are
representing
the
number
of
issues.
B
B
So
this
is
the
same
as
the
previous
slide,
but
the
only
difference
here
is
that
now
I've
multiplied.
The
size
of
the
circle
is
representing
the
complexity
so
now
you're
having
a
sense
of
how
much
efforts
it's
going
to
take
us
to
to
work
on
this.
If
we
wanted
to
tackle
the
medium
importance
and
high
backwards
compatibility,
that's
that's
almost
all
the
work
you
70%
of
the
work
is
right
there
that
big
light
green
circle
in
the
background
next
slide,
please,
okay!
So
where
do
we
want
to
focus?
B
B
We
want
that
topmost
row
and
you
know
mate,
but
if
we're
gonna
break
backward
compatibility,
if
we're
going
in
with
the
intention
of
breaking
backward
compatibility,
then
we'd
also
probably
go
ahead
and
grab
the
medium
and
the
low
issues
that
we
know
we're
going
to
break
backwards,
compatibility
go
ahead
and
do
them
now,
because
we
don't
want
to
have
to
break
backwards
compatibility
again
any
time
soon.
So
let's
just
go
ahead
and
take
care
of
all
those
ones
that
we
think
that
have
the
potential
for
breaking
backwards.
B
Pride
of
work,
even
the
unknowns
down
there,
that
we're
not
sure
if
they
might
break
backward
compatibility,
would
so,
on
the
left
hand,
side
of
a
focus
on
importance.
That's
probably
what
we
would
do
I
mean
subject
to
discussion.
Of
course,
on
the
right
hand,
side,
if
the
focus
is
on
backward
compatibility,
we
want
that
rightmost
column.
Those
are
all
going
to
be
giving
us
the
backwards
compatibility,
but
then
there's
also
those
unknowns.
We
don't
really
know
yet.
B
If
they're,
you
know
how
highly
backwards
compatible,
they
might
be,
and
then
you
see
that
little
medium
with
a
circle.
It's
just
one
issue,
but
we
probably
go
ahead
and
grab
that
one,
because
it's
it's
a
you
know
medium
would
when
we
say
medium
backwards,
compatibility
is
probably
a
way
to
at
least
do
some
of
it
in
a
backwards
compatible
way.
So
if
we
were
to
do
that,
that's
what
we
don't
end
up
with.
So
that's
actually
the
end
of
the
presentation
and
where
I
want
to
stop
I'm.
B
G
Young-Shin
vada
I
think
there
are
two
other
dimensions
that
are
important.
One
dimension
is:
do
we
actually
know
how
to
solve
it?
So
sometimes
you
have
questions
where
people
say
yeah,
that's
how
important
what
we
don't
really
know
how
to
do
this
and
the
other
question
is
typical
cases.
You
have
something
where
there's
a
solution
or
even
multiple
solutions,
but
people
don't
agree
on
that.
G
R
B
J
R
B
R
Will
say:
I
don't
know
about
adoption,
I
mean
even
going
from
yang
102
one
one.
You
know
it's
it's
taken
time
for
the
industry
to
to
pull
into
that
right.
You
know
so
putting
out
a
yang
1.2.
You
would
wonder
you
know
just
just
what
that
adoption
time
frame
would
be
that
you
know
people
are
barely
on
one
one
right
now.
If.
A
A
A
So
you
know
it's
a
little
hard
in
the
abstract
to
say
you
care
a
lot
about
a
particular
one,
and
you
see,
if
want
to
see
an
extension
immediately,
but
the
feedback
focus
on
things
that
are
highly
compatible,
so
we
can
do
it
in
a
modular
way
is
something
that
would
be
useful
feedback
now,
if
that's,
if
that's
what
the
room
feels
but
I'll
take
I'll
interpret.
What
you're
saying
is
be
very
careful
about
pushing
hard
for
a
anything
that,
beyond
yang
1.2,
I'm,
sorry
beyond
1.1
yeah.
R
Cuz
they,
you
know
the
industry's,
not
even
there
yet.
So
it's
a
question
to
me
of
you
know
it:
should
we
you
know,
is
it
the
next
one,
1.2
or
2.0,
and
and
then
it's
really
based
upon
when
the
industry
adoption
will
curb,
will
happen
on
it
right
right?
But
if
there's
but
then
you
said,
there's
you
know
two-thirds
of
the
features
out
there
are
already
backward
compatible
and
and
if
you
can
box
them
up
and
do
a
release,
you
know
I'm
just
twice
right.
Just
information.
B
A
B
We
did
have
a
dimension.
Complexity
was
the
third
dimension,
so
these
were.
That
was
the
3d
view.
I
showed
you
the
beginning
and
and
actually
think
you
go
to
the
next
slide.
You'll
see
we're
kind
of
know.
One
more
slide.
Please
yeah
that
one.
So
this
one
tries
to
factor
in
the
complexity
as
well,
it's
kind
of
hard
to
read,
but
you
got
a
sense
for
it,
the
most
the
effort.
It
would
be
doing
that
the
that
one
thing,
but
your
can
also
made
the
other
point,
which
was
when
would
what's
the
timeline?
B
What's
the
timeframe?
Do
we
do?
We,
as
a
working
group,
have
a
sense
I
mean
where
we
open-ended
about
this
I
mean
Lou
made
a
point.
The
beginning
would
be
at
least
two
years,
I.
Think
last
time,
I
was
actually
three
years
before
when
we
started
to
finish
you
know,
is
there
a
hard
deadline
that
were
what
we
need
to
work
towards
or
should
work
towards.
H
H
So
I
think
it'd
be
useful.
I
hope
we
announce
you've
done
so
far
is
useful,
but
actually
I
think
with
the
list
of
issues.
We
could
probably
go
through
that
list
and
then
classify
them
into
ones
that
we
think
are
definitely
worth
doing,
because
either
the
scope
isn't
too
hard.
We
understand
the
solution
etc,
and
we
agree
it's
worth
doing
so.
That's
one
set
another
set
of
issues
is
ones.
We
definitely
don't
do
either
it's
not
important
enough
or
it's
too
complex.
H
Whatever
we
just
push
those
officer
look,
we
know,
we
don't
discuss
those
and
then
there's
a
set
in
the
middle
of
these
ones
require
further
discussion
is
unclear
either
way,
but
I.
Think
then,
with
that
saw
discussion,
we
could
maybe
end
up
with
a
set
I'm,
more
bounded
scope
of
what
we
think
the
real
problem
is
bring
the
back
to
the
working
group
and
say
look.
This
is
what
we
think
the
scope
would
be.
It'd
be
two
feet
to
fix
these
issues.
We
know
the
answers
to
these,
or
these
ones
are
still
open.
B
S
Hello
could
have
impaired
one.
This
is
Hugh
from
Galway
and
I'm
here
to
discuss
is
authority
for
setting.
This
is
not
the
first
time
to
introduce
this
chapter
and
actually
recapture
a
little
bit
for
this
province
pace.
Actually,
you
know
current
Nanticoke
for
the
doesn't
allow
you
to
reset
the
wrong
in
datastore
for
Canada,
so
you
also
cannot
reset
without
changing
their
wronging
and
what
dynamic
you
also
cannot
have
reset
to
the
initial
state,
so
the
same
apply
to
the
rest,
convo,
so
the
problem.
So
the
solution
we
proposed.
S
Actually,
we
defined
a
method
that
you
allow
you
to
reset
the
datastore
to
the
initial
state
to
the
faculty
for
contents.
Actually,
this
can
be
used
during
the
zero
touch
to
certain
stage
or
when
the
existing
convolution
actually
has
some
major
errors,
and
so
in
this
chapter
we
define
when
a
piece
a
were
called
a
reset
datastore
a
PC.
In
addition,
we
actually
define
a
new
factory
default
text
or
it
is
read-only.
S
It
only
can
be
set
by
the
server
and
it
cannot
be
changed
by
the
imagine
of
operation
like
a
at
config
etc,
and
this
is
a
status
update.
Actually,
this
chapter
the
history:
actually
we
moved
a
chart
from
the
netcode
to
the
Nano
mode
and
we
have
several
discussion
in
master
IT
meeting.
We
discussed
this
chapter
had
a
lot
of
support
and
recently
the
change
actually
based
on
the
discussion.
S
We
actually
have
a
two
minor
to
closer
to
minor
opening
issues
that
discuss
in
the
last
meeting
and
we
actually
put
back
the
copy
configure
based
on
robust
comments
and
also
we
have
a
few
reference,
a
potat.
This
is
all
the
changes
were
made,
and
so
we
like
to
you,
know,
request
the
walking
boot
up
top
is
work
I.
A
B
B
B
Good
number
yeah
how
many
people
think
the
workgroup
should
not
adopt
this
version
jobsworth
basis
for
working
through
talking
crazy,
it's
no
one,
okay,
great,
so
I
think
that's
will
confirm
that
on
the
list.
Thank
you
and
actual
dis
confirm
it.
As
a
last
furthers
any
objections
on
the
list
otherwise
will
be
adopted
by
default.
Thank
you.
T
S
This
actually
a
very
short
shaft
and
motivation
to
do
this,
because
we
already
have
never
called
Bates
event
that
defines
the
ups,
a
foot
for
TC,
obviously,
a
60
64
70
actually-
and
this
can
be
used
to
monitor
the
knack
of
searching
and
to
repose
some
common
system
event.
But
when
we
introduce
the
young
empty,
maybe
additional
event
need
to
be
incorporated
can
delay.
The
problem
is
when
we
introduced
like
intend
pedestal
and
operation
pedestal.
They
may
involve.
Some
interval
process
mean
you
need
to.
S
You
know,
apply
it
some
competitor
from
intent
with
operational,
so
how
to
keep
shadows
at
some
of
the
event
like
you
summer
obviously
cannot
apply
this,
so
we
can
allow
the
user
to
you
know
to
reporter
the
the
fare
of
the
object,
also
to
to
notify
some
reason
why
it
cannot
apply.
So
we
supposed
to
define
one
an
empty,
a
specific
notification
and
history.
S
The
tow
packages
of
history
actually
Xetra
also
moved
from
the
network
over
to
the
net
motor
and
last
meeting
we
received.
Some
discussion
represent
this
chapter
and
two
comments.
Actually,
on
a
many
nice,
the
one
is
about
clarification
for
the
committee
committee
RPC
and
this
new
notification
define
in
this
chapter.
Actually,
we
think
for
for
how
do
you
identify
the
management
setting?
Actually,
we
can
potentially
use
the
username,
but
it
is
optional.
We
also
can
use
some
other
parameter
composed
use
a
multi
attack
to
do
that.
S
Another
one
actually
is
to
discuss
the
the
you
know
for
in
Kenda
and
running
really,
when
you
config
apply
the
configure
stated
to
the
operational.
This
a
potato
will
usually
actually
happen.
You
know
usually
immediately
so
changes
since
master
version.
Actually
we
add,
as
a
mode
attack
actually
to
as
a
another
parameter
to
identify
the
management
sessions
in
some.
In
some
cases
the
same
user
may
have
a
different
application,
so
we
can
use
model
tag
to
indicate
that
the
second
a
is
we
actually
remove
two
redundant,
the
notification.
S
Why
is
apply
in
tennis
star,
the
other
one
is
applied
in
ten
and
four
in
tennis
star?
Actually,
we
sink
because,
as
a
clarify,
it
also
happen
at
the
same
time
as
the
commit
operation,
so
it
hadn't
ended
also
for
the
intend
any
actually
we
can
use
the
an
MTA
apply
the
updater
to
in
the
indicated.
So
this
is
something
seems
not
an
idiot,
so
we
remove
it
plays
to
next.
S
This
is
the
updated
for
the
model
structure.
Actually,
you
can
see
we
actually
added
the
application
package,
reuse,
the
module
tag
to
do
that,
identify
the
management
session.
In
addition,
we
actually,
we
can-
you
know,
use
this
model
together
with
MMD
ativ
and
to
use
if
to
compare
the
difference
between
the
antenna,
pedestal
and
Operation
Thetis
thought
you
identify.
What
is
the
missing,
obviously,
which
object
is
not
applied
next,
so
this
is
a
the
the
notification
with
the
final.
It
will
be
used
at
a
stage.
S
B
A
I
B
P
T
P
Oh,
that
bar
oh,
so
why
do
we
need
this?
So
many
applications
have
a
need
for
a
geolocation.
You
might
want
to
map
your
data
center
router,
firewall
files,
server
location.
You
might
also
want
to
map
your
fiber
endpoints
failures
like
fiber
cuts,
where
your
customers
are
weather
stations,
cameras
so
forth.
So
it's
good
to
have
a
common
grouping
that
you
know
can
be
reused
throughout
all
the
different
models
that
might
need
a
geolocation.
P
P
So
there's
a
lot
of
standards
around
this.
Actually,
as
you
start
digging,
and
digging
open
open
it
up
where
I
think
the
can
of
worms,
there's
a
lot
to
actually
learn
and
there's
a
lot
of
work.
That's
done
on
this,
primarily
in
the
area
of
the
standardized
reference
coordinate
reference
systems,
the
one
that
we're
all
using
with
our
phones
is
the
GPS
one
is
wgs84
and
it's
sort
of
it.
P
There's
there's
lots
of
like
fun
things
to
talk
about,
but
originally
when
that,
when
they
sent
the
satellites
up,
the
geode
is
off
by
the
jig
maps,
where
they
sort
of
think
mean
sea
level.
Is.
It
was
off
by
at
least
a
hundred
meters
in
certain
places,
so
they
they
send
up
another
satellite
that
maps
the
gravity
right,
because
you
have
a
different
iron
and
you
know
different
concentrations
of
things
in
the
earth
and
it
actually
changes
where
at
sea
level
is
so
they've,
slowly
gotten
better
and
better
right
and
gotten
closer
and
closer.
P
So
so
all
these
things
you
have
to
learn
about
and
you
have
to
pay
attention
to
them.
There's
other
standards
as
well.
W3C
has
a
standard
object,
GML,
KML
or
markup
markups
XML.
That
include
objects
as
well
as
IDF.
We
have
one
in
a
URI
there's
also
things
you
might
need
to
consider
like
continental
drift
right.
Your
location
is
going
to
change
over
time,
so
if
you're,
if
you're
referring
to
a
router
and
a
building,
it's
actually
gonna
move
a
few
centimeters
or
so
per
year.
P
He
also
might
be
concerned
with
other
planets
or
asteroids
have
routers
on
the
moon
and
then
finally,
you
might
actually
want
to
map,
have
a
database
of
geo
locations
in
a
virtual
reality
or
you
might
want
to
have
an
alternate
mapping
right
where
you've
modified
reality.
You're
still
reality
you're
talking
about,
but
somehow
you
want
to
look
at
it
in
a
different
way.
So
this
is
the
grouping,
and
this
is
the
grouping
that
I
came
up
with.
P
It
tries
to
cover
all
of
us
a
lot
of
it's
optional.
So
when
you
just
want
that
long
and
maybe
height,
that's
all
you
have
to
put
in
there.
But
if
you
need
the
other
things
it's
in
here.
So
if
you
look
the
alternate
system,
this
is
sort
of
a
catch-all
if
you
want
to
sort
of
go
way
out
there.
Otherwise
we
have
the
astronomical
body
which
would
default
to
earth
and
that
geodetic
system
now
the
geodetic
system
is
what's
actually
gives
meaning
to
your
coordinates.
Right
so
tell
you:
what
height
is
relative?
P
What
your
lips,
oil
values,
aren't
like
Latin
long
and
then
so
as
a
part
of
that,
the
datum
value,
that's
what
that
would
be
like
wgs84,
in
other
words,
a
GPS
one.
There
are
others
and
we'll
talk
a
little
bit
about
that
later,
but
we
also
might
want
to
specify
our
coordinate
accuracy
and
our
hike
accuracy.
Sometimes
those
are
not
the
same,
so
we
can
split
them
and
then,
as
far
as
the
coordinates
we
have,
we
cover
ellipsoidal,
which
are
lat/long
as
well
as
cartesian,
which
would
be
like
a
vector
from
a
center
of
mass.
P
The
velocity
is
there
primarily
for
stable
movement.
This
isn't
meant
to
track,
you
know
orbital
dynamics
and
stuff,
it's
it's
to
track,
control
drift
or
something
has
very
stable
move,
and
we
also
need
a
timestamp.
When
did
you
take
the
measurement
so
in
so
I
define
the
grouping
and
then
I
need
to
make
sure
that
we're
actually
compatible
with
what
other
people
have
done
out
there?
We're
not
just
inventing
something
new.
That's
not
gonna
work,
so
I
went
through
the
different
of
the
various
different
other
standardized
dopings.
P
The
IETF
URI
defines
lat/long
an
optional
height,
a
nuptial
uncertainty
and
an
optional
CRS
or
geodetic
datum.
It
does
not
include
time
soon,
so
the
this
is
fully
expressible
in
the
Eng
I'll
note
this
on
each
one.
We
use
decimal
64
values
in
the
yang,
so
different
different
other
markups.
We
use
different
other
models,
we'll
use
different
values.
So
in
this
case
it's
strings
right,
see
the
sort
of
infinite
precision.
P
W3C
we've
got
lat
long,
optional,
height,
optional
accuracy,
optional
height
accuracy,
and
it
includes
heading
in
speed.
I
didn't
see
that
there's
any
way
to
actually
change
the
CRS,
so
it's
just
defaulting
and
always
wgs84
the
GPS.
In
this
case,
they
the
types
are
doubles.
So
the
only
difference
you
get
here
is
we.
We
have
decimal
use
vessel
64.
We
actually
map
a
touch.
That's
able
to
have
a
huge
accuracy,
but
doubles
it's
a
very,
very
extremes
that
doubles
may
be
a
little
bit
better,
but
for
all
intents
and
purposes
they
map
back
and
forth.
P
P
It's
not
clear
whether
they
standardized
their
CRS
but
there's
a
place.
You
know
you,
can
you
specify
what
the
CRS
is?
I
didn't
see
if
they
had
a
registry,
though
so
it
could
just
be.
Writing
up
strings
well,
not
random,
but
it
also
includes
an
observation
which
would
include
the
time
stamp,
but
it
also
the
observation.
Object
includes
things
like
the
target
camera
you
know
or
result
of
a
measurement.
P
P
This
is
super
Sun
at
w3c,
but
also
modifies
and
gives
you
more
control
over
what
the
height
is
and
they
include
clamp
to
ground,
which
means
we're
not
going
to
specify
the
height.
It's
just
the
ground
relative
to
ground
and
absolute
absolute
would
be
like
the
geoid.
So
how
far
away
from
the
giri
you
are
and
the
geoid
is
generally
mean
sea
level.
P
It
also
has
a
clamp
to
seafloor
and
relative
to
sea
floor
and
and
then
it
includes
other
things
like
the
camera
direction.
It's
used
like
for
Google
Earth,
so
this
is
also
fully
accessible.
Although
you'd
have
to
map
the
seafloor
values,
we
don't
have
a
way
to
do
that.
We
don't
have
that
a
seafloor
in
our
urban.
P
So
some
examples.
This
is
a
restaurant
in
New
York.
It
shows
just
a
Latin.
Long.
Locatable
item
is
not
part
of
the
grouping
I'm
just
using
that
because
that's
where
the
grouping
is
used,
you
could
imagine
so
a
more
complex
example:
the
pump
they
arts
in
Paris.
This
includes
a
timestamp
lat/long
and
a
height
35
meters
above
mean
sea
level.
P
An
on
earth
example
the
Apollo
11
landing
site.
You
can
see
that
here.
This
is
interesting
because
we
have
we
had
to
specify
a
new
reference
frame.
The
ask
for
among
the
astronomical
bodies,
the
moon.
There
is
a
standard,
geodetic
datum.
In
this
case.
The
string
is
me
in
the
registry
that
actually
is
defines
what
me
is
and
we'll
get
to
the
registry
and
whether
we
need
one
and
then
a
few
slides
and
then
the
lat
and
longitude
on
the
moon
of
the
landing.
The
times
is
actually
when
the
Eagle
landed.
P
I
also
heard
that
one
of
our
esteemed
chairs
once
did
a
virtual
reality
project
and
design
date,
spotted
cow,
so
we
have
here
Kent's,
Spotted
Cow,
and
it's
in
the
alternate
system
of
Kent
space.
This.
P
So
the
open
issues
should
we
have
so
height
height,
there's
open
issue
of
height.
Should
we
have
more
relativity
here
instead
of
just
to
the
geoid?
Should
we
had
a
options
like
kml?
Has
we
had
liked
to
ground
verse
to
be
wide
and,
and
or
you
know,
maybe
just
have
a
generic
string
that
allows
the
local
user
to
you
know,
put
a
value
there
to
indicate
what
the
height
is
relative
to
the
other
issue.
That's
probably
gonna
cause.
Some
commotion
is
the
registry.
P
We
already
have
a
registry
in
the
ietf
for
that
uri
and
so
it'd
be
nice
to
reuse
it.
One
issue
that
I
see
with
using
the
URI
registry
is
that
its
specification
required
and
iest
or
approval,
so
it's
very
hard
to
get
things
in
this
registry,
which
is
fine
if
you're
just
going
for
very
standard
values.
But
you
know
we
include
things
like
alternate
system
in
virtual
realities.
P
The
registry
I
specify
is
literally
just
to
stop
having
duplicate
values
right,
so
I
had
specified
it
as
first-come
first-serve
that
might
be
too
loose,
in
which
case
we
upgrade
it
to
designated
expert
and
say
direction
to
the
designated
expert
of
no
duplicates
right.
You
do
not
allowed
to
have
opinions,
you
know
put
the
value
in
as
long
it's
not
duplicating
the
value.
P
P
Right
so
the
other,
so
the
other
option
is:
if
we
keep
with
a
with
our
own
register,
can
we
modify
the
URI
registry
to
be
less
strict?
Would
that
people
you
know
with
the
IETF
be
open
to
that
I'm,
not
sure
or,
and
also
should
we
use
the?
U.
Rn
is
used
right
now
to
to
standardize
the
values
in
the
URI
registry.
So
you
have
like.
P
You
are
n
o
GC
for
the
open
geographical
consortium
standards,
it's
a
very
interesting
that
they
use
AES
PG,
so
the
world
uses
values
that
are
done
by
what's
now
called
the
International
Association
of
oil
and
gas
producers,
but
that
E
is
for
European.
It's
fun
all
the
different
people
involved
in
this.
P
So
it
should
be
just
reserve
earned
prefix
to
point
at
that
registry
and
no
one's
allowed
to
reserve
that
Prix
anything
with
that
prefix
re,
and
then
we
can
just
use
that
very
standard
worldwide
registry
for
most
of
the
values.
And
then
we
leave
open
the
ability
to
add
things
for
people's
virtual
realities
or
ultimate
systems,
so
that
that's
sort
of
an
open
issue
was
the
new
work
that
no
one's
had
a
chance
to
see
it.
B
P
A
good
question
I,
you
know,
I
just
published
this
and
I've
started
getting
a
little
bit
of
feedback
on
it.
The
Cartesian
is
actually
what
like
wgs84
uses
and
then
they
map
to
the
ellipsoidal
so
that
we
can
actually
make
sense
of
it.
Okay,
so
you
know
it
seems
worthwhile
to
have
it
in
there,
because
it's
how
it's
stored.
Okay,.
B
U
P
B
P
B
B
At
you,
okay,
and
also
the
one
thing
I
know
from
my
history
of
mapping
back
and
forth
between
wgs84
and
Cartesian-
is
that
the
ability
to
the
resolution
get
the
math
lose
the
resolution
very
quickly.
I
think
we
can
support,
maybe
at
a
30
kilometer
with
kilometers,
so
you
would
actually
have
to
have.
Maybe
that's
sufficient
for
your
use
case.
But
it's.
P
Much
better
now,
so
the
the
it
had
there's
a
couple
different
things
in
play
with
that
it
has
to
do
with
where
they
have
a
table
because
they
measured
the
the
the
gravitational
changes
are
on
the
surface
right
because,
like
you
know,
the
earth
is
not
perfectly
solid
iron,
and
so
they
have
these
mapping
tables
that
adjust
because
the
the
GPS
geoid
is
a
mathematical.
You
know
ellipse
right
and
then
they.
U
P
Adjust
that
based
on
measurements,
they've
taken
so
the
value
you're
talking
about
the
accuracy,
is
how
accurate
those
tables
are
right,
so
that
used
to
be
like
200
kilometers
or
something
like
that
right.
The
egm
earth's
gravitational
model
2008,
which
is
the
most
reason
one
gets
it
down
to
it's
a
small
number.
It's
like
a
hundred
meters.
P
Because
it's
just
from
you
right,
so
it's
from
your
location,
that's
where
it's
you're,
basically
you're
heading
in
speed
right!
So
originally
I
had
the
value
as
heading
in
speed,
so
it
was,
but
so
for
continental
drift.
The
continents
not
just
they
don't
just
move
on
the
plane
or
what
they
big
move
up
and
down
what.
B
P
So
you
so
the
so
the
the
Cartesian
or
the
ellipsoidal
gives
you
your
location,
the
velocity,
it's
a
velocity
vector
right.
So
it's
it
is
where
you're
at
so
you've
identified,
where
you're
at
whether
Cartesian
or
ellipse
I
do
now
you
have
a
location.
Now
the
question
is:
what
direction
are
you
going
and
how
fast
and
that's
that's
what
that
vector?
Is
it?
P
Doesn't
it
doesn't
map
into
like
an
ellipse
ideal
case
if
it
was
heading
in
speed
that
would
be
in
Alex
Roy
deal
because
he
would
have
a
360
degree
heading
and
then
the
speed
along
that
this
is
a
three
three
XYZ
right.
So
it's
just
saying
where
you
know
how
fast
am
I
going
in
the
XYZ
direction,
the
magnitude
of
the
vector
and
where
the
vector
is
I'm.
B
P
P
B
P
P
P
He
actually
does
I,
don't
remember
what
telecom
he
works
for,
but
he
he's
been
over
those
draftkings
and
specifically,
he
said
oh
yeah,
you
know,
and
he
actually
helped
me
a
little
bit
right
with
making
sure
that
I
hadn't
made
any
mistakes,
so
we
do
need
to
get
people
who
are
doing
it.
Who
would
be
using
it
yeah
yeah,
oh
yeah,
let's
get
the
right
technical
review.
I
have
to
review
the
rest
of
it
yeah.
So
maybe
I'll,
try
to
you
know,
sign
people
up
to
send
something
in
mailing
lists.
Saying
I've
done
this
mapping.
B
So
that
actually
brings
in
an
interesting
question
about
you
know:
polling
the
room
for
interest.
If
we
don't
have
the
you
know,
people
room
if
actually
it's
a
separate
group
of
community
that
would
need
to
contribute
to
that
discussion,
but
nonetheless,
I
think,
would
be
interesting
to
get
some
sense
from
the
room.
I
realized
that
this
is
a
zero
zero.
You.
B
You
know
first
time
being
presented,
but
still
it's
always
interesting
to
get
a
sense
from
the
room
at
level
of
interest,
so
how
many
people
in
the
room
have
actually
read
the
draft.
Please
raise
your
hand
wow
so
few
and
and
there's
only
one
version
of
drafts.
I
don't
need
to
ask
any
other
question
there.
Cuz.
B
Gave
me
one
of
three
like
any
things,
for
the
velocity
all
right
and
and
then
how
many
people
think
this
is
an
interesting
work.
Please
raise
your
hand
actually
more
more
than
read
the
drafts.
That's
good
and
I
think
it's
probably
premature
to
actually
ask
for
anybody.
You
know
to
do
an
adoption
to
sell
or
actually
I.
P
Don't
know
if
it's
premature,
because
there
I
think
more
to
do
so,
so
no
I
I
mean
it
is
the
first
time
I'm
presenting
I,
wouldn't
mind
asking
on
the
list,
maybe
if
or
at
polling
the
room.
If
people
think
we
could
ask
soon
about
adoption,
because
it
has
been
on
the
list
before
right,
the
people
have
talked
about
the
need
of
for
a
geolocation
object
on
the
list.
So
it's
it's
work,
I
think
we've
already
kind
of
sort
of
agreed
that
we
might
want
to
take
on.
R
Adoption
so
this
tip
Karen,
okay,
I,
was
just
gonna.
Make
that
same
statement
I
mean
you
know
to
be
honest
with
you
long
lat
height,
that's
80%
or
more
of
what
what
we
need
right,
and
so
you
know
to
try
to
do
the
you
got
a
long
tail,
hair,
alright,
and
so
you
looks
like
you
got
everything
that
you
basically
need
and
I
know
that
in
fact
last
week
this
actually
came
up
and
said:
hey,
look.
We
need
to
get
some
geolocation
stuff
for
some
metadata
for
some
network
elements.
R
B
P
Before
you
ask
that
question
I'd
like
to
defend
it,
okay,
this
came
up
from
people
the
the
reality
is.
We
do
have
networks
right
out
there
we
put
we
put
computers
on
them
right.
We
we
send,
we
send
spaceships
to
asteroids
to
land
on
them.
There's
no
reason
to
limit
ourselves
here
and
I'm,
not
I'm,
not
gonna,
break
anything
by
putting
it
here
and
I
did
my
research
right.
P
So
when
I,
when
I,
when
we're
putting
when
I
show
the
Apollo
11
landing
site
right,
it's
actually
referring
to
a
worldwide
standard
for
the
geolocation
on
the
moon
right,
it's
not!
It's
not
like
I'm
just
playing
around
throwing
things
in
there
for
games.
U
P
Don't
I,
don't
know
anybody
lasers
to
the
OGC
or
anything
but
I'd
be
open
to
hearing
at.
V
Bob
so
I
think
it's
interesting
now,
if
you're
asking
a
question
who
has
expertise
here
to
review
the
location
aspect
of
it,
not
the
young
part
right,
and
if
there
is
nobody,
we
have
to
ask
who
would
be
the
right
experts
outside
of
this
working
group
outside
of
the
ITF
wherever,
but
if
you've
got
the
feedback
of
those
guys
selling?
This
is
good
from
a
geolocation
point
of
view.
Then
we're
fine
and.
B
V
H
But
it
does
also
look
like
you've
already
done
a
lot
of
research
here
in
existing
other
other
models
that
modeling
the
same
data.
So
it's
not
like
this
a
blank
slate
where
you
just
sort
of
come
up
with
this
new
scheme.
It's
like
I've
looked
at
these
ones,
I've
cholestacare
into.
I
think
it's
the
right
one,
so
I
think
you've
done
a
lot
of
the
legwork
already
here.
Yeah.
P
The
comparison
that
was
the
real
point
behind
the
comparison
right
was
that
I
found
three
under
three
other
standards,
right
that
it
didn't
standardize.
They
have
stos
behind
them
who
have
all
looked
at.
It
came
to
very
similar.
You
know
very
similar
outcomes
right
and
then
I
and
I'm
in
the
draft
showing
that
I'm
we're
not
doing
we're
not
deviating
from
these
right
and
that
we
can
map
back
and
forth
to
them.
I
do
think
that
does
grant
some
legitimacy
to
the
work.
H
There
already
and
then
my
other
companies
I,
think
that
I
think
ITF
takes
a
long
long
time
to
produce
young
models.
I'm
gonna,
showing
that
again
again
again,
whereas
I
think
they
would
be
really
really
useful,
so
the
world
in
general
to
have
more
standard
yeah
molds
coming
out
of
ITF
the
people
can
use
and
either
they're,
not
hundreds,
imperfect,
I
think
we'll
get
to
the
stage
where
we
can
fix
those
mistakes
and
we
get
some
that's
a
ninety.
Nine
percent
is
good
enough.
There's
a
starting
point:
okay,
if.
B
You
can
do
it
quicker,
I
think,
actually,
I'm
I'm
convinced
it
seems
like
there's
a
lot
of
interest
in
the
room
and
you've
done
the
homework.
So
let's
go
ahead
and
move
with
a
poll
for
adoption.
How
many
people
in
the
room
so
I've
already
asked
how
many
people
read
the
draft
and
have
interest
in
the
drafts
now
we're
just
gonna.
Do
the
poll
for
adoption
any
people
in
the
room
would
like
to
see
this
draft
adopters
working
group
as
the
basis
for
a
working
group
document.
B
Please
raise
your
hand,
as
many
people
have
raised
her
hand
and
how
many
people
would
not
like
to
see
this
document
used
as
the
basis
for
a
working
group
document.
Please
raise
your
hand,
that's
none!
Okay,
so
we
will
confirm
that
on
the
list
and
like
the
last
time,
if
no
one
controls
you
know,
will
be
a
doctor
by
default.
T
G
Young-Shin
Veda
I'm,
going
to
talk
about
an
update
to
the
young
datatypes,
the
common
yang
data
types.
Almost
every
young
module
is
importing
and
I
list
of
things
to
add
on
my
computer
and
at
some
point
in
time,
I
thought
I'd
just
start
putting
that
into
a
new
document,
so
people
can
actually
look
at
what
might
be
in
there
next
slide,
please.
So
what
I've
added
is
a
type
for
a
date
and
the
time
for
time,
because
we
only
had
date
and
time,
but
sometimes
you
only
need
one
of
them.
G
I
added
types
for
time
intervals
like
MIT
in
minutes
seconds.
What
have
you
you
can
read
the
whole
list
really
basic
stuff,
but
popular
stuff,
I
added
something.
That's
called
a
revision
identifier
for
the
current
version
of
yang
kind
of
telling
you,
which
version
you
are
talking
about
and
I
copied
in
a
version
of
the
node
instance
identified
at
prototype
from
neck,
because
it
seems
to
be
useful
outside
of
the
neck
and
context,
and
people
might
not
want
to
have
a
dependency
on
a
cam
if
they
just
want
to
use
a
node
instance.
G
G
So
the
issues
you
think
a
date
is
simple:
a
trip
is
simple,
but
then
you
think
about
what
does
it
mean
to
have
a
canonical
format
for
a
date?
You
think
it
should
be
simple,
you're
looking
to
access
the
specification,
they
have
a
data
type
of
day.
They
have
text
about
canonicalization
of
that
and
I
didn't
understand
it,
and
that's
why
it's
here.
It's
probably
a
reminder
for
myself
to
keep
looking
into
this.
G
Unless
there's
someone
in
the
audience
who
understands
date,
kind
of
neck-
realization,
if
not
I'll,
just
keep
it
as
as
a
reminder
for
myself.
The
next
one
is
probably
something
where
more
people
have
an
opinion
about,
and
as
is
the
number
of
bits
that
you
need
to
represent
time,
a
different
resolution.
So
if
you
have
a
nanoseconds,
timer
and
probably
people
would
say,
yeah
64
bits,
it's
not
a
bad
idea,
because
that
takes
pretty
fast.
If
you
have
a
time
that
you
express
some
minutes,
then
probably
32
bits
are
fine.
M
Know
what
I
want
to
propose
to
modify
the
regex
pattern
for
the
domain
name,
because
in
fact
it
doesn't
do
what
the
description
promises
it
says
that
can
be
use
in
DNS
a
and
aaaa
records.
This
is
not
really
the
case
because
in
those
because
you
can
also
have
first
wildcards,
which
means
asterisks
and
second
slash.
That's
used
for
the
reverse
domain.
M
Registration,
basically,
like
I,
don't
know
something
/
64
for
for
cider
style
domain
delegation
delegation,
so
I
actually
wanted
to
propose
it
in
the
mailing
list.
I
wanted
to
to
propose
some
pattern
for
that,
but
I
haven't
had
time
to
test
it
so,
but
I
think
it
would
be
a
good
addition,
because
otherwise
it
really
doesn't
do
what
what
promises.
On
the
other
hand,
there
is
this
host
type
and,
in
this
case,
I
think
that
it
it
should
be
strict
just
to
follow
the
stricter
rules
for
for
host
names.
M
In
this
case,
is
it
just
it's
a
union
of
domain
name
and
IP
addresses,
and
then
in
this
case
it
it's,
probably
not
good
to
permit
just
a
single
dot
which
which,
which
is
permitted
in
domain
name.
So
I
think
it
would
be
useful
to
extend
the
domain
name
type
and
restrict
the
host
type.
Probably
so,
please
send
an
email,
because.
G
P
P
Don't
know
if
this
would
be
our
NBC
case,
but
it
would
be
really
nice
if
we
had
more
precision
in
the
canonical
format.
Maybe
we
have
it
to
have
a
new.
You
know
that
might
be
a
non
backward
compatible
but
yeah
and
as
far
as
the
counters
yeah
I
I,
don't
know.
If
I
would
say
both
I
would
just
say
64,
you
know
it's,
because
a
lot
of
representations
are
string.
P
P
B
Can't
as
a
contributor
so
I,
my
first
thought
was:
maybe
we
should
have
some
more
types
in
addition
to
ours.
We
could
have
days
weeks
months
years.
Perhaps
who
knows
how
these
models
might
be
is?
But
my
second
thought
was:
are
you
defining
these?
Because
we
talked
about
the
missing
duration
and
I
mean?
Are
these
trying
to
fill
in
that
gap
for
a
duration.
G
Well,
there's
there's
a
definition,
for
we
don't
have
a
definition
for
duration,
but
if
you
look
at
real
data
models,
if
people
model
a
time
out,
they
usually
have
something
that
is
a
millisecond
or
a
second.
So
it's
very
commonly
used
to
which
we
have
these
unit
specific
types
or
not
like
an
axis.
The
duration,
which
is
a
relatively
complex
construct,
could
do
the
same,
and
it's
commonly
what
not
people
implement
I'm.
B
Just
thinking
about
a
model
where
my
duration
was
something
I
know
a
couple
weeks
or
a
month,
it's
like
this,
it's
a
it
was
a
larger
and
it
would
have
been
convenient
to
represent
it
in
terms
of
that
like
if
you
had
larger
granularities
days
and
weeks,
it
makes
it
more
natural
to
express
these
kinds
of
things.
Yes,.
G
I
mean
I
mean
that
I
believe
is
even
an
ISO
standard
format
for
durations.
That's
out
there.
We
could
add
something
to
that,
but
I
don't
think
it's
a
replacement,
because
if
you,
if
you
really
just
model
a
time
out-
and
you
know
it's
going
to
be
in
milliseconds-
you
would
just
want
to
have
a
number
rather
than
something
more
complex.
B
G
B
B
All
right,
so
how
many
people
would
like
to
see
this
version?
The
document
used
as
a
basis
for
a
working
group
document.
Please
raise
your
hand
quite
a
number
and
and
how
many
would
not
like
to
see
this
document
adopted
as
the
basis
for
a
working
document,
none,
okay,
so
I
guess
we
could
confirm
an
adoption
poll
on
the
list
and
I
mean
whether
or
not
it
ever
gets.
You
know
to
last
call
we'll
see.
P
So
this
is
this
is
really.
This
is
very
interesting
case
right
because,
basically,
it's
a
standard
module
for
defining
types
that
we're
gonna
use
everywhere
and
all
we're
basically
saying
is:
let's
keep
doing
that
right.
Let's
just
keep
rubbing
that
thing
and
adding
more
common
types,
that
kind
of
makes
sense,
sure.
A
Drops
that
last
point
I
was
really
thinking.
Does
it
make
sense
to
Rev
this
as
abyss
or
to
just
have
a
new
module
that
defines
these
new
types
and
I?
Think
it's
a
good
experiment
going
back
to
Chris's
comment,
let's
see
how
fast
we
can
do
it,
let's,
let's
just
use
the
model
of
keep
revving
this
okay.
H
I'm
all
the
question
is
how
hard
it'd
be
for
us
to
agree
on
what
goes
in
there
and
put
on
github
and
be
done
so
and
that's
the
sort
of
thing
is
the
the
the
publication
delay
is.
Does
that
help
us
here
at
all,
or
is
this
somebody
pushing
it
through
that
publication
queue
a
lot
quicker,
but
otherwise
it's
sort
of
we
just
delayed
by
an
extra
year.
H
G
A
H
H
A
Maybe
okay
so
I
its
the
ITF
and
we're
in
the
business
of
bruising
RFC's.
So
let's
try
it,
but
it's
an
interesting
point.
Let's
think
about
it.
Maybe
talk
to
Melissa,
take
the
talk
outside
and
we
have
a
few
minutes
left
for
the
next
one
and
by
the
way
I'll
note
that
we
only
made
it
through
like
a
third
of
your
slides.
So
that
was
effective
and
please
take
a
look
at
the
the
rest.
If
you're
interested
in
the
topic,
I.
X
Yes,
this
document
basically
introduced
a
John
Daly
model
for
ESA,
a
business
or
policy
management,
so
it
can
provide
ability
to
network
management
function
and
you
can
Joana
configure
the
monetary
state,
a
change,
our
natural
element
and
take
a
simple
action
when
triggers
a
condition
on
the
system
estate-
and
this
is
document
at
first
introduced
in
Lhasa,
a
Vida
meeting
and
the
warden
who
was
just
asked
to
coordinated
ways.
Another
document.
X
So
we
had
to
some
update,
forcibly,
polished
the
text
and
clarify
the
difference
between
the
booing
trigger
condition
and
who
this
recommendation
and
synchronize
some
terms.
Always
one
already
published
a
talk,
RFC
its
introduced,
participative
measurement,
a
fragment,
Fred,
Fredman,
Ranma,
walk
and
the
way
he
split
eight.
X
He
went
management
that
yeah
I'm
not
doing
to
treat
their
trigger
young
and
another
is
event
young
makers
a
trigger
when
make
it
a
even
trigger
loops
bow
and
the
NASA
document
areas
is
grouping
and
defines
there's
a
condition,
and
we
also
change
the
two
leaves
the
name
to
help
carry.
There
is
a
understand,
oxidize.
X
A
A
H
H
A
Y
I
love
mentioned
some
other
drafts,
so
two
eight
years
ago
we
presented
a
very
similar
concepts
of
the
models
and
we
didn't
cut.
There
is
no
feedback,
we
didn't
know
if
it's
interesting
that
enough
for
the
working
group,
so
we
haven't
continued
to
march
on
that
side.
So
from
what
I
see
here,
this
is
totally
to
placate
to
that
pretty
much
same
concept.
Y
A
A
Okay,
I
actually
did
have
sort
of
that
feeling
that
there
was
some
overlap
here
with
Netcom
one
of
the,
so
one
of
the
actions
I
think
will
be
forward
to
the
neck
mod
and
neck
cuff
chairs.
To
talk
to
each
other,
talk
to
yourself
and
see
where
we
end
up.
It
would
be
good
to
not
try
to
reinvent
something
that
is
already
been
good
work
done
on,
so
you
try
to
bring
those
together.
That
would
be
really
good.
I
I
think
from
this.
X
A
Z
You
mentioned
previous
after
the
first
session
that
you
wanted
to
discuss
briefly
the
timing
experiment
just
to
want
to
do
that.
You
think
you
have
time
I
think
we
are
over
okay.
If
you
want
to
make
one
comment
on
it,
that
would
be
great.
So
one
comment:
please
provide
the
feedback.
You
will
receive
a
survey
after
the
meeting
and
please
provide
the
feedback
about
what
you
think
about
this
experiment
wave
the
unstructured
time.