►
From YouTube: NETMOD WG Interim Meeting, 2020-04-02
Description
NETMOD WG Interim Meeting, 2020-04-02
A
Appreciate
one
adapting
to
this
rapidly
changing
and
unique
in
our
lifetime
event
and
there's
not
idle,
not
sure
we
have
much
more
to
say
on
that.
There's
been
enough
discussion
on
the
list.
We
were
going
to
try
to
focus
on
our
agenda,
which
is
a
bit
it's
a
abbreviated
because
of
this
being
an
interim.
A
A
A
So
this
is
quite
different
than
one
of
our
normal
in-person
meetings,
but
it's
not
too
different
from
another
interim
or
other
interims,
although
we
do
expect
it
to
turn
into
something
that
might
be
a
bit
larger.
For
that
reason,
we're
asking
that
the
chat
window
on
WebEx
be
limited
to
queue
control.
So
if
you
want
to
enter
the
queue,
please
say,
plus
Q
and
then
wait
for
the
chairs
to
identify
you
and
say
you
can
speak
if
you
want
to
leave
without
being
called
on
just
to
Q.
A
A
If,
if
things
get
too
noisy
we'll
we'll
change,
how
we're
working
that,
but
right
now,
everyone
has
their
own
mute
control,
blue
sheets-
this
is
definitely
different
than
an
in-person
meeting.
Please
go
to
etherpad,
which
is
in
the
in
the
chat
window
and
go
to
click
on
that
link
and
add
your
name
to
the
bottom
of
the
ether
pad.
Please
be
careful
not
to
delete
other
people's
names
you
can
with.
We
do
have
history,
so
we
can
go
back
and
figure
out
who's
there.
A
So
if
you
see
someone
you
don't
like
being
there,
that
won't
really
help
the,
but
please
make
sure
to
add
your
name
at
the
bottom
and
once
you're
there
please
help
with
stay
there
and
please
help
take
some
notes.
We
are
on
jabber.
So
if
you
would
like
to
type
something
you
can
we'll
do
our
best
to
monitor
that
and
to
try
men
is
that
one
of
my
co-chairs.
B
Okay,
no
I
didn't
say
anything,
but
we're
gonna
go
two
minutes.
You
just
mentioned
minutes
if
everyone,
when
they
go
to
ether
pad
in
the
upper
right
hand,
corner
of
ether
pad
where
you
see
a
bunch
of
color
boxes,
you
know
select
that
and
type
your
name
against
your
color
you've
been
assigned,
and
in
that
way
we
can
sort
of
identify
who's
taking
the
minutes
in
the
in
the
ether
bed
after
them,
they're
actually.
A
A
Our
our
agenda
is
notably
without
times
we
we,
we
think
we
have
enough
time
to
make
it
through
all
these
topics,
but
they
we've
waited
the
agenda
with
the
the
key
working
group
items
at
the
front,
and
we
want
to
make
sure
that
they
have
sufficient
time.
We
do
think
that
we're
going
to
make
it
through
the
whole
agenda
and
if
we're
lucky,
even
less
than
the
two
hours
were
scheduled
for.
A
Computer
seems
to
be
having
a
little
trouble
here.
Okay,
as
many
probably
have
noticed,
we
have
adopted
a
nice
set
of
documents
from
the
yang
modeling
and
versioning
design
team,
and
this
is
really
an
important
piece
of
represents
an
important
piece
of
work.
It's
filling
a
really
important
gap
in
our
foundational
technologies
and
the
design
team
has
given
us
a
starting
point,
and
we
really
want
to
acknowledge
that
they
have
worked
hard
to
get
us
to
this
point.
That
doesn't
mean
that
we
don't
altogether
have
more
work
to
do.
A
We
absolutely
do,
and
you
know,
as
the
adoption
call
showed
there
were
some
there's.
Definitely
some
rough
patches
still
to
go
through,
but
we
really
want
to
acknowledge
every
person
who
contributed
to
that.
That
includes
the
formal
design
team
members
and
also
other
people
who
commented
either
publicly
or
privately.
So
we
really
appreciate
that
that
hard
work
and
want
to
acknowledge
the
whole
team
and.
A
A
A
A
D
A
A
A
B
A
A
E
E
The
agenda
here
is
I'm
gonna,
give
a
quick
solution.
Overview
I've
presented
this
before,
but
I
just
wanna
set
this
the
scene
for
everyone
asked
if
the
drafts
have
been
discussed
and
I'll
also
give
a
very,
very
one,
slight
update
on
the
requires,
update
and
also
quickly
discuss
or
the
sort
of
author
next
steps
are
here.
What's
planned,
that's
still
open
for
discussion
for
other
people.
E
So,
to
recap:
we've
adopted
I
think.
Is
it
a
seven
draft,
maybe
and
I
think
the
five
meaty
ones
here
are
the
ones
listed
here?
So
there's
a
module
revision
handling
and
that's
the
ones
discussing
effectively
notifying
BC
changes
between
module
revisions
and
allows
revision
labels
to
be
cited
with
revisions
and
allows
a
branch
revision
history,
if
that's
required,
doesn't
end
date
or
encourage
that
but
allows
it
on
top
of
that.
One
of
the
possible
revision
label
schemes
that
has
been
proposed
is
a
module
semantic
version
number
scheme.
E
If
that's
actually
I
think
the
modules
should
be
taken
out
of
that.
It's
just
a
yang
semantic
version
number
scheme,
because
proposal
is
that
same
scheme
could
also
be
used
for
packages
as
well,
and
this
is
a
scheme-
that's
based
on
December
2,
0
0,
but
extended
to
allow
fixes
in
all
the
revisions
effectively
and
is
one
in
fact.
The
recommended
revision
label
that
we're
suggesting
should
be
used
for
yang
modules
and
but
others
can
be
used.
E
Third,
one
is
yang
packages
and
that's
looking
at
the
schema
level
rather
than
for
individual
modules,
so
he's
providing
a
fashion.
That's
somewhat
similar
to
what's
in
yang
library,
but
it
allows
it's
defining
it
in
like
a
hierarchical
way.
So
packages
can
import
other
packages
to
build
up
yang,
schema
and
again
those
yang
schema
can
be
used
off
the
box
to
describe
describe
the
schema
is
that
the
device
is
using
ahead
of
time
as
well
as
being
out
to
download
that
at
the
time
as
well.
The
idea
here
again
is
that,
by
allowing
the.
E
The
fourth
draft
here
is
preparations
for
version
selection,
so
this
is
again
sort
of
building
on
young
packages
and
is
offering
a
solution
for
how
devices
can
support
backwards,
compatible
and
non
backwards.
Compatible
changes
to
the
schema
so
allows
devices
to
potentially
support
multiple,
different
versions
of
a
schema
and
to
allow
clients
to
choose
which
version
of
schema
that
they
want
to
use
and
the
the
idea.
E
This
draft
is
to
allow
quite
a
lot
of
flexibility
in
how
devices
decide
to
support
for
an
schema
selection,
so
they
may
only
support
a
single
schema
or
they
may
support
multiple
schemes
at
the
same
time,
and
then
the
last
draft
here
is
an
yank
schema
comparison
and
talling.
So
this
is
a
way
of
effectively
our
vehicle
and
finding
algorithms
to
compare
either
yank
schema
or
just
individual
yang
modules.
E
So,
even
in
the
sort
of
conversations
we've
had
about
young
version
number
schemes,
they
can't
cover
or
can't
you
can't
describe
all
changes
easily
in
a
single
version
number
and
it's
quite
a
crude
mechanism
to
describe
what
has
changed.
Eventually,
you
need
to
be
able
to
compare
the
schema
steamer
to
see
what's
actually
changed.
The
idea
here
is
to
actually
define
what
those
algorithms
are
and
to
provide
tools
that
actually
provide
that
comparison.
So
I
know
that
PN
provides
some
of
this
comparison.
E
I,
don't
know
it's
entirely
complete
or
whether
there
anything
needs
to
be
changed
there,
but
that's
the
sort
of
ideas
going
on.
So
on
top
of
these
five
drafts,
there's
also
requirements
draft
that
is
defining
what
the
requirements
are.
We're
meeting
and
a
solution
overview
draft
that
just
sort
of
ties
these
five
draft
together
and
finds
how
they
fit
together.
E
Next
slide,
please.
So
this
diagram
is
just
sort
of
showing
you
and
sort
of
dependencies
between
the
drafts
how
they
work
together.
You
can
see
that
the
module
revision
handling
is
the
base
draft.
On
top
of
that,
we've
got
the
optional
semantic
versioning
scheme
that
can
be
used
with
that,
and
the
packages
version
schema
draft
that
makes
use
of
the
module
revision
handling
so
those
three
sort
of
fit
together.
E
Next
slide,
please
so
I
mentioned
briefly
this
there's
a
requirements
of
solution
overview
draft.
The
requirements
draft
is
stable,
it
hasn't
changed.
There
are
questions
here
to
the
working
group
as
to
whether
we
should
try
and
take
this
to
informational
RFC
at
the
moment.
The
author's
opinion
I
think
is
that
probably
it
would
be
helpful
to
take
this
to
RFC
because
as
an
infertile
RFC,
because
the
solution
documents
reference
this
and
so
that's
I
can't
take.
But
we
want
to
do
to
raise
that
with
a
working
group
to
see
they
don't
have
any
opinions.
E
Similarly,
there's
a
solution
overview
draft
again,
which
would
be
informational
with
the
idea
that
it
just
sort
of
ties,
different
aspects
of
the
solution
together
and
explains
how
they
fit
together.
Again,
the
same
question
is:
do
we
aim
to
take
this
to
informational,
RFC
and
again,
the
preference
of
the
authors
is
that,
yes,
it's
helpful
to
do
so.
E
B
E
A
The
list
you
know
the
list
is
great
people
being
able
to
say
plus
one,
so
you
know
I
think
I
think
we
have
a
good
mechanism
for
it.
If
we
have
to
do
it
virtually
as
asking
a
question
to
you
as
co-author
and
also
to
the
other
co-authors,
do
you
have
a
feeling
on
timeline
of
when
you
would
want
to
proceed
process
these
documents,
this
event
to
the
I
I
issued
for
publication,
I.
E
E
The
solution
overview
draft
I
think
that
all
the
other
solution
draws
would
have
a
normative
reference
to
this
and,
as
such,
I
think
it
will
end
up
being
held
up
in
the
processing
queue
anyway.
So
again,
I
think
that
we
would
try
and
progress
that
at
the
same
time,
there's
some
of
the
solution
drafts,
but
that
does
raise
an
interesting
question
and
come
up
on
the
next
slide.
Okay,.
A
F
F
E
G
E
E
So
the
next
steps,
so
yes
as
as
Lou's
mention
the
design
team
is
officially
complete.
I
would
also
like
to
thank
everyone,
the
design
team,
the
for
all
that
work
in
this
effort.
The
next
steps
is
that
the
authors
would
like
to
meet
on
a
regular
basis
to
prepare
this
tip
of
grow
progress.
This
work
so
I
assume
that
or
hope
these
would
be
weekly
or
meetings
effectively.
The
meetings
will
be
open
to
all
everyone
can
join
them.
E
We
will
try
and
carry
substantive
discussions
and
things
on
the
lists
and
keep
the
work
fully
in
the
loop.
Those
the
current
time
suffers
weekly
meetings,
a
Tuesday
at
2:00
p.m.
UK
time
that
can
possibly
change.
If
L
people
are
interested
in
participating
and
and
effective,
what
to
keep
us
as
open
as
possible.
So
I
would
like
the
Lippitt.
It
should
be
an
alias
that
we
use
for
this
for
the
authors,
because
we'll
be
discussing
multiple
and
draft
at
the
same
time.
So
I
need
to
close
that
with
Lou
to
check
that
last
okay.
E
But
yes,
so
that's
the
sort
of
proposal.
In
addition
to
that,
we
would
like
to
try
and
face
the
document
delivery.
We
would
like
to
try
and
concentrate
on
the
module,
versioning
yanks
and
the
packages
as
a
first
block
and
then
half
of
the
version
selection
schema
comparison
to
follow,
and
the
aim
here
is
simply
to
try
and
get
all
the
members
of
the
working
group
able
to
focus
on
those
drafts
and
potentially
progress
those
more
easily
and
more
quickly
get
to
those
two,
maybe
to
a
working
group.
E
Last
call
stage
before
picking
up
more
detailed
discussion
on
the
version,
selection
and
schema
comparison,
so
I'm
not
trying
to
block
any
discussion
on
those
other
drafts
now
I'm
sure
we
will
have
those
but
I
think
we
would
quite
like
to
try
and
to
prioritize
those
three
drafts.
If
that
is
acceptable
to
the
working
group,
the
chairs,
that
makes
sense
any
questions.
A
It's
not
from
my
personal
perspective
as
long
as
things
don't
get
so
out
of
sync,
that
by
the
time
we
get
to
the
sort
of
second
set,
that
there
isn't
a
solution
for
the
second
set
that
can
sometimes
happen.
Is
you
end
up
with
a
base
set
that
doesn't
allow
for
support
of
some
other
capability?
So
we
don't
end
up.
In
that
case,
it's.
E
A
question
for
you:
Louise
is
an
option
here
to
take
the
first
three
to
work
last
call
but
then
effectively
to
hold
them
at
that
state
so
that
we
don't
take
them
any
further
until
we're
sure
that
the
the
other
two
drafts
are
have
solutions
effectively,
so
so
trying
to
try
and
focus
on
the
first
ones,
but
not
take
them
any
further.
And
then,
if
we
need
to
come
back
to
those.
If,
if
that's
required,
to
get
a
full
solution,
we
can
do
so
that.
A
B
I
So
this
is
a
recap
of
the
module
versioning
draft.
This
was
the
same
slide
was
presented
at
previous
ITF,
the
the
top
to
the
top
two
enhancements
which
are
mentioned
there.
Those
are
busy
the
crux,
the
main
stuff
provided
by
this
draft
that
were
first
allowing
nonlinear
module
development,
also
known
as
branching,
and
secondly,
it
allows
us
to
document
when
non
backwards.
Compatible
changes
have
occurred
in
revision
history.
I
What
we
basically
see
in
lots
of
the
dropps
revision
date
is
still
there,
but
we
allow
an
optional
text
label,
which
we
call
revision
label
to
be
associated
with
the
revision.
The
label
could
be
centered
sanbir
base
or
it
could
be
a
different
scheme
for
those
who
remember,
there
used
to
be
one
wrap
and
they
were
split
into
two
drafts.
The
sender
stuff
was
taken
to
a
separate
draft
which
will
be
presented
a
little
bit
a
little
bit
later.
I
There's
been
an
addition
of
the
import
by
a
revision
date
or
all
derived
to
allow
us
to
import
to
say
you
know,
I
want
something
from,
for
example,
2000
example:
we
define
what
backwards
compatible
and
non
backwards.
Compatible
changes
are
there's
some
clarifications
regarding
yang
status
handling
and
we
refresh
the
guidelines
for
a
big
in
yang
modules.
I
I
B
Can't
says,
sir
I
think
the
question
is
so
the
draft
is
adopted,
but
the
status
the
intended
status
currently
is
Pro
standard.
Potentially
it
could
get
moved
to
experimental.
Whether
or
not
that
makes
sense
is
something
we
need
to
determine.
Consensus
or
overt
I
mean
right
now,
I
guess.
The
question
is:
are
these
altering
opinions
in
the
rough
you
know
like
for
IETF
term
in
the
rough
or
is
there
actually
do
we
not
have
consensus,
so
we've
needed
to
determine
that
so
we'll
figure
that
out
as
we
go
forward,
I'll.
A
Say
we
don't
think
we
have
consensus
here
either.
You
know
from
my
perspective,
as
one
of
the
chairs
is
I,
think
it's
very
important
for
us
to
come
up
with
a
standards
track
solution
now,
certainly
if
the
working
group
thinks
it's
expect
not
mature
enough
to
be
a
proposed
standards,
experimental
is
better
than
nothing.
We
also.
We
have
a
few
people
in
the
queue
go.
F
First,
this
microphone
is
a
little
bit
better.
Okay,
I
I,
think
you're,
referring
to
your
email,
as
I
recall,
was
experimental
for
the
yang
silver
graph.
Not
this
the
yang
module
versioning,
but
I
think
this
issue
crosses
the
boundary
between
both
I'm
going
to
talk
a
little
bit
about
this
issue
as
well
from
a
different
perspective
in
my
presentation.
I
just
want
to
be
clear
on
what
we're
here.
F
Now,
I
don't
bring
up
the
intended
status
yet,
but
I
bring
up
again,
should
I
eat
EF
modules
or
must
IETF
modules
have
a
revision
label
that
is
a
Yankton
I.
Think
that's
the
crux
of
the
question
to
act
to
answer
and
then
that
could
help
dictate
the
module.
Intendant's.
Sorry
that
that
intended
sets.
That's.
I
The
other
issue
was
regarding
tax
in
Section
three
one
one.
Regarding
ordering
for
our
PC
input,
I
thought
was
Martin,
who
made
the
point
that
it's
not
just
our
PC
to
action
and
how
to
refer
to
not
just
input
and
output.
We
have
to
update
the
draft
it
side.
Note
on
this
was
rob,
was
asking
questions
about
what
happens
with
all
grants
and
looks
like
there's
potential
existing
issues
with
altman,
since
the
order
is
not
predictable,
but
that's
not
something
which
is
being
brought
because
of
this
class
is
just
a
something
which
came
up.
I
I
E
And
so
as
a
contributor,
one
of
my
thoughts
here
is
I.
Think
when
we
looked
at
this
before
one
issue
with
having
an
extra
statement
is,
is
then
be
able
to
define
the
set
of
different
versioning
schemes.
Initially,
you
might
think
using
identities
with
that
office
doesn't
work
in
the
yang
language
as
opposed
to,
as
instance,
data.
So
one
option
may
be
for
each
versioning
scheme
to
define
its
own
extension
saying
you're
using
this
versioning
scheme.
A
J
You
guys
hear
me:
yes,
sorry
is
on
the
previous
one,
I
guess:
I'm
I
think
we
some
of
the
authors
or
were
we're
good
with
the
concept
of
different
versioning
schemes,
but
I
I
think
it
would
be
fairly
useful
if
we
find
a
way
to
kind
of
have
a
a
preferred
scheme.
If
we
can,
if
we
can
get
ourselves
there
and
and
what
this
point
about
is
kind
of
saying.
Look
if
there's
a
if
there's
a
certain
pattern
for
numbers
like
major
minor
dot,
dot.
J
Editorial
a
3,
a
3
digit
number
scheme.
It
would
be
very
helpful,
I
think
for
the
industry
that
there's
a
standard,
well-known
interpretation
of
of
that
pattern
in
a
way,
I
guess
what
I
would
I
recommend
is
is
if
we
can
find
a
way
that
we
reserved
that
pattern
for
a
well-known
scheme.
I
think
it
would
be
useful
for
everyone.
J
Yeah
I,
just
we
would
I
think
what
we
probably
want
aim
for
as
a
group
is
that
we
reserve
1.2.3
pattern
as
meaning
something
specific
and
no
other
scheme
could
use
exactly
that
pattern.
They'd
have
to
be
penned
it
with
a
with
a
prefix
or
have
some
other
way
of
making
it
very
intuitively
obvious
to
a
human
that
it's
a
different
scheme.
H
A
From
my
standpoint,
I
think
the
documents
continue,
as
they
are
up
until
the
point
that
the
the
working
group
decides,
that
we're
ready
for
last
call,
and
then
we
can
revisit
whether
they
should
move
from
proposed
standard
to
experimental.
That's
generally,
what
happens
with
documents
and
end
up
experimental?
Is
it's
not
until
the
working
group
decides
that
it's
done,
that
there's
agreement
that
what's
being
proposed,
is
not
sufficiently
mature
or
well
understood
for
it
to
be
proposed
to
entered
so
I.
A
A
K
This
is
Tim
Carey
I.
Think
I
would
like
to
just
make
sure
that
everyone
knows
you
know
from
from
the
standpoint
of
the
versioning
and
I
kind
of
agree
with
Jason
in
the
sense
that
I
know
for
the
open-source
community
as
well
as
that
that
stands
for
at
least
from
the
broadband
forum
community
work.
We
would
prefer
to
see
a
pattern
there
so
that
we
can
put
our
tooling
around
this
and
in
order
to
do
some
comparisons
of
modules
and
and
make
some
decisions
around
it.
So
the
pattern
would
be
important
so
that
we
can.
L
I
I
I
So
the
draft
you
know
we
use
the
add
provision
label
which
replaces,
at
date
and
marketing,
made
the
comment
that
tools
which
look
for
the
date
you
know
would
break
because
of
this.
The
reason
we
did
this
was
not
to
break
the
tools,
but
basically,
if,
if
a
module
is
using
a
revision
label,
you
know
it
should
be
that
module
should
be
searchable
and
portable
by
that
revision
label.
I
I
The
new
extension
status
description,
so
the
comment
was
made
that
this
is
not
necessary,
that
we
can
kind
of
update
what
we
can
just
update
the
description
of
the
of
the
data
note.
Instead,
the
reason
this
was
done
with
understand
here
that
to
work
around
for
description,
the
status
which
is
not
allowed.
So,
yes,
you
can
change
the
description,
but
whenever
you
change
the
description
you
it
helps.
You
figure
out
whether
it's
a
user
or
tooling,
whether
the
change
is
beaten,
backwards,
compatible
or
not
backwards
compatible.
I
We
may
need
to
annotate
the
description
change
with
you
know,
to
identify
that
is
backwards,
compatible
and
stuff
like
that,
so
it's
it's
I
mean
I,
don't
think
it's
the
most
important
question
right
now
and
I
think
even
within
the
now-defunct
is
19.
There
is
no
consensus
as
to
whether
this
is
needed
or
not,
but
this
just
explains
why
that
was
added.
I
So
the
new
these
nodes,
deprecated
and
obsolete
use,
an
T
the
comment
was
made
would
be
better
to
be
using
boolean.
Instead,
the
authors
are
okay,
with
using
boolean
as
long
as
the
don't
change
the
one
that
Noah
doesn't
exist,
we
don't
change
the
existing
implementations
that
don't
supply
those
big
nodes,
but
from
what
I
recall,
there
was
no
pushback
on
changing
this
to
bull
from
a
key.
F
You
can't
so
Rashad
is
kind
of
touched
on
a
number
of
the
issues
that
we're
going
to
touch
on
here,
but
nonetheless,
let's
recap:
what
what
yang
symbol
means
what
we
mean
by
that?
Not
much
has
changed
between
the
previous
draft
and
we'll
jump
right
into
the
issues.
So
next
slide,
please
so
as
a
recap,
yang
silver
is
designed
or
defined
to
be
the
full
silver
20.0
spec
that
you
would
see
if
you
need
to
subvert
or
with
some
addition.
So
this
is
a
superset.
F
If
you
were
to
choose
to
use
the
Yang's
there
sorry
December
2000,
it
would
be
completely
compatible
with
this,
but
as
the
disbanded
design
team
talked
about
just
incorporating
silver
early
on,
we
ran
into
the
issue
that
what
and
Rob
describes
is
very
well
that
if
you've
got
a
relatively
stable
branch,
let's
say
you're
one
dot,
o
branch
and
you're
iterating
that
1.1
1.1
1.1
1.1
at
2.
So
on
so
forth
and
a
customer
comes
and
says
you
know.
F
Minor
change
has
been
added,
like
a
new
feature
has
been
added,
and
the
uppercase
M
signifies
that
a
non
backwards
compatible
change.
Major
type
revision
change
has
been
added
to
this
branch
and
at
that
point,
all
bets
are
kind
of
off
that
that
branch
has
been
signified
as
being
poisoned.
We
do
not
think,
however,
that
that
would
be
a
common
that
common
of
an
occurrence
and,
as
we'll
talk
about
when
we
get
to
the
issues,
the
intent
is
that
the
it's
currently
documented.
F
This
way
that
idea,
we
incorporate
the
use
of
Yang
Cimber
as
a
revision
label
scheme,
we'd
already
kind
of
talked
about
some
of
those
issues
there.
In
addition,
this
draft
talks
about
this
idea
of
gang
artifacts
and
not
just
modules.
It
used
to
talked
about
just
modules,
but
we
decided
when
packages
kind
of
evolved,
that
of
the
design
teamwork.
We
wanted
a
kind
of
a
single
noun
that
would
represent
being
able
to
tie
a
game
similar
to
a
yang
module
as
well
as
package,
so
that
if
you
read
that
that's
what
that
notation
needs.
F
So
this
is,
we
evolved.
The
whole
versioning
thing
to
do.
Rashaad
just
talked
about
previously
using
the
revision
history
as
a
lineage,
but
for
the
purpose
of
a
quick
human,
readable,
human,
understandable
notation.
We
introduced
this
concept
of
a
revision
label
and,
as
Roush
I
mentioned,
we
wanted
something
that
kind
of
resonated
with
people
and
we
landed
on
silver,
which
open
config.
Also
uses,
and
because
of
this
desire
to
do
this
multi
branching
and
the
realization
that
not
everything
is
perfect.
F
We
arrived
on
this
notion
of
this
lowercase
and
uppercase
M,
so
that
gives
us
somewhere
as
it
is
today,
and
the
one
bullet
I
didn't
really
mention
is
like
full
silver
2.0.
We
also
recognize
that
you
can
have
a
zero
major
version
and
that
doesn't
follow
the
backwards
compatible
non
backwards,
compatible
rules,
so
that
December
as
it
is
as
it's
documented
in
the
draft
today
next
slide,
please
back
in
time.
Okay,
so
this
is
just
I
walk
through
this.
This
is
just
a
visual
example
of
how
this
works.
F
Where
you
have
the
beta
0
version,
then
you
have
an
initial
release.
We
talked
about
that
recently
on
the
list.
The
idea
would
be
the
yang
module
kind
of
RFC
press
release
would
be
1.0,
and
then
you
can
see
how
the
branches
might
evolve
and
where
those
and
lowercase
uppercase
M
notation
could
come
into
play
because
of
that
back,
porting
of
minor
or
major
backwards
compatible
non
backwards,
compatible
fixes,
and
you
can
see
otherwise
how
the
version,
the
revision,
label,
hierarchy
or
lineage
would
progress
again
following
for
the
most
part,
the
silver
Toyota
rulz.
F
F
The
two
revisions
are
the
same
module
and
you
can
that
a
module
perspective
between
revision
now
and
revision
previous
and
say
yes,
this
is
either
backwards
compatible
or
non
backwards
compatible,
meaning
a
human
can
easily
do
that.
This
is
this
is
something
that
is
convenient
for
that
operator,
that
consumer
of
the
yang
module
to
look
at
and
know
that
there
are
either
non
backwards,
compatible
changes
and
they
can
dig
into
that
more
deeply
or
there
aren't
any
non
backwards,
compatible
changes.
F
So
my
to
link
my
consumption
of
that
yang
module
can
continue
to
proceed
normally,
and
so
the
proposal
of
the
authors
at
this
point
is,
yes,
is
define
a
revision
label
scheme
and
we
feel
that
there
should
be
a
some
revision
label
that
we
go
forward
with
as
something
that
is
documented
and
defined,
not
to
say
that
other
producers
of
yang
modules
can't
define
their
own
revision.
Labels
pause
there
because
I
know
the
mic
line
is
building
Carson.
M
Carson
I'm
you!
Yes,
can
you
hear
me
yeah
right,
so
what
my
general
experience
in
life
is
that
it's
hard
to
have
two
things
that
diferencia
antics
and
and
only
differ
in
their
case
being
used
by
humans.
So
if
you
think
about
the
number
of
people
who
talk
about
Millie
Hertz
when
they
mean
megahertz,
maybe
it's
not
a
good
idea
to
use
capital
m
and
lowercase
M
for
two
different
semantics.
So
maybe
you
should
look
out
for
a
different
character,
different
letter
to
use
there,
instead
of
using
the
same
letter
twice.
F
F
That
kind
of
we
picked
M
because
a
major
minor,
that's
why
and
that's
why
the
the
lowercase
for
minor
the
uppercase
for
major,
but
it's
something
that
I
think.
Ultimately,
you
get
the
point.
We
want
to
signify
that
this
branch
has
something
not
just
quite
right
with
it
based
on
the
pure
symbols,
so
it's
certainly
worth
considering
whether
or
not
that's
lowercase
and
uppercase
M
or
some
other
or
less
ambiguous,
notation
right.
Thank
you.
F
I
think
that's
the
mic
line
so
next
slide.
Please
I
would
say:
they've
got
something
worth
discussing
and
bringing
up
on
a
list,
so
we
could
have
our
kind
of
master
record
of
that,
because
I
don't
believe
that
proposal
or
that
point
has
been
raised
before
so
thanks.
Kirsten
second
issue
raised
during
adoption
and
Martin
has
now
talked
about
this
on.
This
call
and
Kent
has
also
raised
this,
so
the
document
currently
mandates
that
the
ietf
modules
will
use
yang
Cimber
in
the
revision
label,
and
do
we
want
to
do
this?
F
Martin
mentioned
that
the
the
intended
status
might
dictate
the
answer
to
this
and
I
took
the
tack
the
other
direction.
If
we
answer
this,
then,
if
we
answer
this
in
the
positive,
then
we
kind
of
know
what
we
want
to
do
with
the
intended
status.
If
we
answer
this
and
the
negative,
then
we
can
have
that
discussion
of
what
the
intended
status
should
be.
F
Thinking
is,
though,
that
yang,
ietf
or
IT
of
yang
modules
typically
not
introduce
non
backwards,
compatible
changes
that
could
and
almost
certainly
will
introduce
backwards,
compatible
changes,
so
they
should
conform
to
the
pure
silver
2.0
rules.
So
the
complexity,
the
the
uppercase
lowercase
M,
goes
away.
You
probably
would
just
have
a
one
dot
dot,
one
dot,
X
stream
of
modules
and
the
module
history
revision.
History
will
capture
that
lineage.
F
That
said,
the
author's
feel
that
we
should
say
that
we
mandate
this
for
ietf
modules.
For
the
point
I
raised
on
the
previous
slide
that
the
the
convenience
factor
for
the
the
consumer
of
the
yang
modules
can
quickly
see
that
module
food
now
is
got
a
minor
version
a
bump,
so
there
are
new
features
that
might
be
worth
exploring
and
seeing
if
there's
something
that
I
could
consume
or
use
differently.
So
we
feel
that
this
is
still
something
that's
useful
to
have
for
ietf
yang
modules.
N
I've
brought
this
up
that
this
is
a
point
where
I
think
that
I'm
not
sure
we're
gonna
get
modules
in
and
some
of
the
bgp
features
which
won't
have
non
backward,
compatible
features,
but
I'll
take
that
offline
I
just
want
to
point.
This
is
the
I
mentioned
it
on
the
mail
list
during
adoption.
I
just
want
to
link
to
this
is
the
the
place
where
I
that's
reached
and
I
made.
That
comment.
F
A
Next,
I'm
not
sure
this
is
counter
to
your
to
sue,
but
I
definitely
agree
with
sue
that
non
backward
audible
changes
are
gonna
happen
and,
in
fact,
I.
Think
introducing
wolf
is
the
work
that
we're
doing
here
will
make
it
more
common,
and
we
also
have
an
existence
proof
that
the
IETF
will
do
backward
incompatible
changes.
If
you
remember,
we
had
an
instance
of
that
already.
L
Thank
if
we
can
use
the
same
schema
both
in
built
for
standards
and
for
vendor
our
own
products,
Ericsson
products.
That
would
be
very,
very
useful.
This
something
that's
meant
a
number
of
vendors
would
need.
Also
I
will
have
later
a
presentation
about
3gpp
and
the
AG.
Some
other
standard
organizations
don't
take
backwards,
compatibility
so
strictly,
and
they
will
need
this
as
well.
Probably
Thanks.
G
Thank
you,
I
think,
yes,
I.
Think
I
also
would
like
to
have
this
Hema
in
idea,
because
I
see
some
cases.
For
example,
even
if
everything
is
backward
compatible
and
linear,
you
may
have
a
for
example
day.
One
zero
zero
in
is
G
and
A
one
one
zero
in
the
working
group
processor.
So
then,
if
you
don't
have
this
rule
I'm
a
little
bit
confused?
How
you
can
have
this
thing
together.
F
Is
is
that
human
late
to
the
tag
that
is
consumable
the
by
the
are
usable
by
the
consumer,
the
yang
module?
The
thing
that
the
drafted
Rashaad
presented
on
yang
module
versioning
describes
the
canonical
way
one
would
use
to
to
use
the
lineage
of
the
yang
module
and
bc
NB
c--
tags
to
figure
out
the
the
true
nature
of
where
they're
at,
but
that
is
a
human,
is
kind
of
convoluted.
O
O
Branch
right,
we
would
decide
not
to
do
that.
We
wouldn't
make
use
of
that
kind
of
added
thing
that
will
put
for
vendors
to
help
with
their
use
cases,
so
we
would
make
use
if
we
had
to
we'd
make
a
non
backward
compatible
change,
but
then
we
would
bump
the
major
version
rather
than
putting
an
M
capital
a
is
that
am
I
understanding
correctly.
O
F
J
Hi
I
guess
I'm,
not
I'm,
not
sure
I'm
I'm
fully
in
line
Joe,
I,
think
I
think.
Typically.
Yes,
there
were.
There
would
be
a
pretty
strong
attempt
to
not
use
the
m-80s
modules,
but
as
Surat
up
there
may
be
examples
of
branching
in
standards
modules,
and
that
does
mean
there
is
a
possibility
than
in
a
older
branch.
It's
possible.
There
could
be
a
decision
in
the
future
to
make
a
non
backwards,
compatible
change
in
that
older
version.
P
P
E
A
contributor
I
just
had
a
thought
on
I
I.
Tell
those
comments
about
draft
being
worked
in
the
working
going
to
ice
is
G
in
terms
of
version
numbers
I
understand
how
that
works
quite
well
for
a
module.
That's
before
1
0-0,
that's
fine!
When
we
are
doing
new
revisions.
That
could
be
an
issue
but
I
do
know.
The
semver
also
allows
descriptions
to
be
added
after
the
main
version
numbers,
so
I
think
we
could
potentially
look
as
to
whether
there
could
be
some
appropriate
strings
to
use
of
December.
E
H
Thanks
yeah
I
just
wanted
to
to
point
out
again
that
the
the
mechanism
that
makes
all
these
things
that
people
are
asking
for
possible
is
the
other
draft,
the
marginal
versioning
draft.
So
when
you
say
that
we
need
no
backwards-compatible,
even
though
we
don't
encourage
it,
it's
the
module
working
draft
that
allows
that
this
draft
is
just
a
symbolic
string
for
the
version.
F
Some
for
the
the
comments
keep
going.
Thank
you
next
slide.
So
now
we're
into
our
github
issues
and
Rishabh
posted
the
link.
That's
he
filtered
on
module
versioning,
but
all
of
the
issues
are
there.
We
have
tags
for
various
ones.
Number
48
happens
to
be
that
and
rashad
touched
on
this
a
little
bit
as
well
right
now.
F
The
pattern
defined
for
revision
label
for
gangstas
is
very
specific
and
matches
on
these
2.0
2.oh
rules,
including
the
extra
pre-release
and
bill
Nam
and
clay
sure
that
Rob
just
mentioned,
plus
our
mm
lower
case
M
uppercase
M
notation.
So
the
draft
states
that
anything
that
matches
that
must
be
interpreted
using
the
rules
that
we
described
already
done
on
the
second
slot,
and
then
it's
already
been
raised.
Well,
what?
If
I
want
to
use
a
something
that
looks
like
a
silver
but
I
want
to
use
different
rules
that
the
vendor
might
have
a
release.
F
Version
looks
very
much
like
a
silver,
but
they
don't
follow
the
strict
major
means
non
backwards
compatible,
so
on
and
so
forth.
Martin
I
believe
brought
up
that
this
was
kind
of
a
layer
violation
on
the
list
and
to
Authority
been
discussed.
We
could
have
additional
metadata.
That
indicates
the
scheme
being
used
or
we
could
have
required
that
different
revision
label
schemes
defined
different
extension
points
to
again
unambiguously
stay,
that
this
is
the
scheme
being
used
for
label
and
the
and
then
their
pointers
to
how
the
label
should
be
should
be
processed.
F
Ultimately,
the
author's
want
something
that
is
unambiguous
and
Tim
on.
This
call
brought
up
that
the
pattern
makes
sense.
You
can
look
at
that
and
say:
okay,
it
matches
this.
So
these
are
the
rules
we're
going
to
to
use,
but
we
want
to
make
sure
that
if
we
are
going
to
say
that
the
revision
label
matches
that,
if
their
original
label
matches
his
pattern,
this
is
in
addition
to
mandating
that
ietf
modules
use
this.
That's
going
to
go
towards
that
final
document
status.
So
the
author
still
feel
that
this
makes
sense.
F
Chuckles,
that
was
the
last
of
the
major
issues.
So
the
again,
the
the
big
driving
issue
here
is
before
we
get
to
intended
status.
The
the
authors
wanted
to
answer
of
the
kind
of
a
normative
language.
Do
we
mandate
that
IETF
modules
use
silver
as
yank
silver
as
a
revision
as
a
versioning
scheme,
and
then
how
do
we
interpret?
That
is
it?
Is
it
the
pattern
or
do
we
need
something
else,
and
that's
the
discussion
and
the
question
they
need
to
answer
here
and
then
we
can
move
on
to
based
on
that.
B
P
P
First,
I'm
going
to
give
a
quick
overview
and
Rob
presented
in
versioning
solution
overview.
Young
packages
is
a
versioning
at
the
much
higher
level
than
the
individual
module.
So
the
goal
of
the
young
package
definition
is
to
provide
Americans,
concise
and
more
concise
way
for
for
clients
to
confirm,
with
the
young
schema
on
the
device.
P
For
example,
the
set
of
young
modules
to
implement
near
to
VPN
on
the
device
could
be
defined
as
a
young
package
and,
in
addition
to
the
running
server
solution
of
flying
discs
is
also
considered
so
that
the
reason
why
young
package
instance
fire
I
defined
so
then,
what's
the
main
definition
of
young
package
I
think
the
first
one
is
a
packaged
version.
A
packaged
version
is
a
mandatory
and
it
could
be
used
by
their
revision
label
and
also
revision
date
and.
P
The
just
discussion,
some
version-
can
also
be
used,
so
that
means-
and
another
characteristic
is
package
hierarchical.
That
means
type
each
can
include
other
packages,
and
the
last
definition
should
here
is
the
checksum.
The
checksum
definition
in
the
young
package
is
for
the
integrity
checks,
which
is
to
solve
integrity,
problems
and
downloading
and
package
from
any
and
URL.
P
P
And
then
provide
the
guidance
to
this.
The
third
planet
related
to
the
first
one
Saints,
the
package
court
package,
definition
on
the
server
and
offline
young
business,
I
unified,
so
the
core
part
of
redefined
and
the
last
one
is
similar
with
the
I
just
mentioned
before
it
could
be
a
revision
label
and
for
a
take.
So
this
is
some
main
update.
So
next,
please
Jason.
J
Yeah
it
was,
it
was
actually
to
address,
lose
point
about
time
constraints.
I
know
we
started
off
the
meeting.
Thinking
we're
gonna
have
lots
of
time,
but
we
are
chewing
through
it
quickly.
At
this
point,
our
our
X
design
team
has
taken
an
hour
and
we
still
have
several
drafts
to
go
so
I
was
gonna.
I
was
going
to
propose
one
of
two
options
and
see
what
people
think
we
had
already
discussed
this
amongst
the
authors
that
we
definitely
wanted
to
get
through
these.
J
These
first
three
drafts
and
detail
up
to
including
packages
based
on
the
priority
of
how
we
want
to
work
in
the
working
group
will
be
the
the
next
two
ones,
my
presentation
and
Rashad's.
We
could
either
extremely
abbreviate,
as
in
maybe
do
the
five-minute
version
touch
on
to
a
couple
of
slides
or
even
push
to
the
end
of
the
session.
If
there's
time,
I
don't
know
what
the
feeling
is
from
other
people,
but
I'd
suggest.
Maybe
we
do.
B
P
Okay,
so
this
is
the
the
young
package
model
structure.
This
is
the
new
structure.
The
left
one
is
the
actually
taken
out
from
the
young
package
instance,
it's
a
similar
one,
but
the
point
here
is
that
now
this
is
the
common
package
definition
for
both
on
the
young
instance
data
file
and
also
the
device.
So
you
can
see
on
the
left
on
the
right.
The
young
packages
container
just
uses
grouping
directly.
So
in
that
way
the
young
instance
looks
very
the
unified
version
and
they
also
change
the
package
Association
to
the
data
star
schema.
P
As
for
the
young
library,
augmentation
improve,
there's
a
lot
of
reusing.
The
model
says
definition
to
relate
to
to
associate
to
the
packages,
but
right
now
this
is
the
only
left.
It
part
is
to
associate
the
package
to
the
Davis
does
schema
and
the
previous
version,
this
package
to
the
schema
mapping
is
one-to-one
in
this
version.
It's
one
two
and
nothing
in
that
way.
It
very
I
think
it
in
that
way
it
will
be
more
compatible
with
the
young
signals,
like
selection,
so
next
I
think.
P
P
Allow
me
da
already
defines
different
data
star
and
each
race
the
different
associated
data
stars
image
and
our
discussion.
We
think
there
could
be
many
ways
that
datastore
schema
could
be
represented
using
packages,
so
we
give
some
recommendation.
Hot
packages
could
be
used
to
define
theta
star
schemas
things.
Packages
are
mostly
used
for
package
conformance
between
client
and
device.
P
So
with
this
mapping,
then
that
would
be
more
efficient.
So
we
give
the
to
two
major
suggestions
on
the
on
the
package.
Definition
to
the
data
store
wine,
statist
all
have
more
packages
for
the
operation
of
data
stores.
That
schema
must
include
all
packages
for
all
the
configuration
data
stores
and
any
data
store
that
exactly
have
the
same.
Data
store
must
use
the.
Q
P
P
P
P
P
For
example,
when
it
matches
module
is
no
revision
date
how
to
define
that
behavior
and
another
related
question
is
also:
is
it
okay
for
namespace
to
be
optional
right
now?
It's
mentoring
around
England
young
library,
but
in
package
names
it
space
is
optional.
So
here's
our
proposal
we
retain
package
require
every
module,
has
a
revision
date
but
tape
the
XML
namespace
as
optional.
P
A
P
P
P
The
arguments
is
that
packages
could
potentially
update
more
quickly
than
modules
like
when
they
were
automatically
Bute
at
the
output,
and
perhaps
revision
date
is
sufficient,
like
we
can
like
package
version,
can
use
revision
date
as
a
person,
so
identify
a
pack.
So
we,
our
proposal,
is
retained
times
time
to
want
any
potential
confusion
with
the
revision
date
use
as
a
package
version.
So
if
there
are
any
feedback
on
this
we'd
like
to
hear
from
the
watching
next,
please.
P
This
is
quite
hard
one,
but
we,
if
the
there
any
feedback
on
this,
we
will
welcome
them
feedback
from
for
now,
the
young
package
definition.
We
think
that
and
a
registration
for
young
package
definition
is
quite
useful,
but
we
think
should
this
I
just
cover
it,
have
standardized
young
packages
or
all
young
packages
or
does
I
end
up
managing
the
package
revision,
or
should
we
try
and
just
use
something
like
J
top
with
some
X
expert
review
or
release
process?
So
we
with
we're
not
sure
about
this,
and
we
are
seeking
suggestions
from
the
achievement.
B
J
So
just
a
reminder
of
what
this
presentation
is
about:
schema
selection,
so
this
is
a
functionality.
It's
mainly
server
based
functionality
with
some
minimal
extensions
on
the
client
side.
Basically,
servers
can
and
they
can
support
multiple
kind
of
versions
of
a
schema
or
multiple
schemas
and
and
advertise
those,
and
then
the
client
can
decide
which
one
they
want
to
use.
So,
as
example,
here
you
have
vendor
modules
version
6,
0
1,
you.
J
Modules
version
5
to
3
the
the
server
the
router
could
advertise
these
and
the
client.
If
it
had
previously
integrated
with
the
old
version,
5
could
request
and
select
and
say,
hey
I
want
to
keep
talking
to
this.
This
upgraded
server
but
I
want
to
keep
on
using
the
old
versions
5
to
3
and
the
clients
can
can
ask
that
there
is
a
default
schema
set
if
a
client
doesn't
request
anything
so
for
clients
that
don't
know
about
this
whole.
This
will
scheme
a
selection
process.
J
They
could
just
connect
and
they
would
be,
they
would
use
a
default
schema,
which
is
configurable
on
the
server
side
if
desired
and
skeena
sets
are
basically
built
up
of
yang
packages
and
I'll
show
that
if
you
skip
to
slide
6
now
that
kind
of
looks
so.
This
is
an
illustration
of
of
what
the
basic
unit
of
selection
is,
so
a
client
selects
a
schema
set
and
on
the
right.
This
illustrates
what
a
schema
set
looks
like
it's.
J
Basically,
each
data
store
has
a
list
of
packages,
so
packages
are
associated
with
a
data
store
and
you
kind
of
define
which
packages
are
part
of
each
data
store
and
that
makes
up
a
schema
set.
You'll
notice
here.
In
the
example
at
the
top
I
have
my
running
and
candidate
you'll
notice.
The
same
packages
are
defined
for
the
running
and
candidate,
which
hopefully
makes
a
lot
of
sense
and
then,
similarly,
both
of
those
packages
vendor
config,
3,
1,
2
and
2
5
2
you'll
notice.
J
So
hopefully
it
gives
a
bit
of
an
idea
of
what
what
a
schema
set
means.
You
define
the
packages
for
each
data,
store
that
you
support,
ok
and
and
to
make
this
very
brief,
I'm
going
to
skip
the
slide
13
and
highlight
the
really
the
one
key
issue
will
have
to
follow
up
on
a
list
with
other
issues,
but
the
one
key
issue,
I
think
is,
is
where
should
this
document
live
net
mod
versus
net
comp?
J
So
the
this
draft
does
define
some
updates
to
net
comp.
There
they're
not
huge,
but
it's
during
the
kind
of
the
capabilities
exchange
flow.
The
hello
has
a
new
advertisement
and
the
the
client
can
for
some
semantics
there
for
a
client
to
select
optionally
select
which
schema
set.
It
wants
to
use
similarly
there's
some
extensions
for
rest
comps,
where
the
rest,
coffee,
URI
kind
of
encodes,
the
the
desired
schema
set.
J
J
The
possible
split
would
be
to
put
protocol
changes
and
we
focus
really
narrowly
on
exactly
just
the
Netcom
changes
and
what
document
just
the
rest.
Kant's
protocol
changes
another
document
and
then
a
third
document
that
has
the
stuff:
that's
not
really
protocol
specific
that
talks
about
how
schema
sets
are
built
and
configured
etc.
So
our
proposal
from
the
X
design
team,
the
authors-
is,
it's
probably
most
simple-
to
keep
it
all
together
as
one
document
but
address
it
in
net
comps,
but
we're
open
to
discussion
about
this.
P
B
J
B
A
I
I
Okay,
so
quickly
on
the
white.
So
basically,
this
defines
optional
extension
to
annotate
changes
which
could
be
I'm
ambiguous
to
improve
the
accuracy
of
comparison.
For,
let's
say
what
this
drop
is
that
they
seek
to
provide,
describes
algorithms
to
compare
young
yang
modules
to
be
able
to
figure
out
whether
you
know
changes
they're
backwards
compatible
or
not.
I
Here
you
can
see
so
I
mean
description,
changes
can
be
backwards
compatible
or
not,
depending
on
whether
you're
changing
the
actual
semantics.
So
in
this
case,
it's
just
no
stuff
is
written
with
1f
in
the
initial
version-
and
you
see
here,
the
second
half
was
added
and
we
have
the
annotation
to
say
that
this
is
an
editorial
change.
I
This
drop
is
probably
the
greenest
of
all
all
the
graphs
in
that
in
that
series
so
really
like
to
hear
from
the
working
group,
especially
people
who
worked
on
Khan
tooling,
you
know
question:
should
we
move
all
the
extensions
to
the
module
versioning,
there's
pros
and
cons
the
extension
4nbc?
Are
they
useful?
We,
the
exe
19,
but
everything
is
useful.
I
Anyway,
so
the
whole
point
of
this
to
summarize
is
to
help
with
tooling
to
compare
between
versions.
One
other
thing
which
I
think
Rob
has
brought
up
last
ITF
is
whether
we
have
to
settle
on
a
bigger
tool.
I,
don't
know
how
the
working
group
feels
about
that
or
whether
we
go
with
all
our
various
tools.
We
have
today
and
hope
that
you
know
one
or
more
do
support
this
functionality,
and
this
is
it
for
me.
B
Q
Q
Yes,
so
this
job
has
been
around
for
a
while.
Actually
the
the
version
five
actually
was
presented
in
a
lobster,
Singapore
meeting
and
actually
a
great
I
thought
these
were
looking
good
jobs,
but
the
chair,
such
as
we
should
to
be
independent
from
management
plan
protocol
like
a
man
called
red
caboose.
So
you
know
so
we
made
a
new
revision
regime
fixed.
Q
Expression,
so
it's
very
complicated
so
based
on
amenities
discussing
actually
Andy
it's
better
than
be
actually
if
the
chair,
maybe
we
can
use
to
expose
something
or
experts
expression
for
this
logical
expression,
and
we
also
discussed
a
lot
about
how
to
you
say
whether
we
should
fully
represent
is
a
model
or
or
not
so
in
latest
version
we
made.
Actually
we
actually
harmonized
with
eager
job,
and
we
add
additional
attribute,
like
a
qualitative
variable
function,
call
and
also
a
piece,
a
50
fetcher.
Actually,
so
we
also
absolutely
tried
to
actually
define
clearly
about
the.
Q
What
is
the
event
within
the
condition?
What
is
action,
and
we
introduced
up
the
animal
like
a
policy
variable
and
also
we
decouple?
This-
is
a
model
from
nanoparticle,
make
it
more
generic
to
be
applicable
to
many
other
emerging
protocols.
So,
for
this
is
a
multiple
people
who
don't
know,
what
do
you
say?
A
model
is
doing.
Actually
this
is
a
model.
Actually
I
think
is
a
very
important
comment
and
actually
for
the
nano
automation.
It
really
can.
Q
Q
So
there's
a
lot
of
a
lot
of
issuing.
We
need
to
address
actually,
so
people
are
single
big
challenges
about
a
logical
expression.
So,
especially
as
we
mentioned
when
you
decompose
is
the
logical
expression
to
a
set
of
the
young
is
so
it's
very
difficult
to
do
so,
what
what
are
they
suggest
actually
is
use
X
path
for
this
logic,
expression,
and
we
also
find
actually
a
positive
variable-
is
very,
very
useful,
especially
when
you
maintain
some
state
we
can
use
we.
Q
So
probably
we
can
actually
expire
to
reference
to
the
products
available
in
the
expression
based
on
the
Andes
suggests,
and
he
also
think
that
we
should
take
a
look
at
our
model
that
already
be
pops
up
see
and
in
our
model
they
also
have
notifications
events.
So
probably
what
we
should
overuse
is
kind
of
notification
when
or
we
defined
for
a
new
event.
So
this
is
something
we
need
to
figure
out.
Q
We
also
have
an
edge
up
women
on
this,
but
a
single,
probably
we
need
to
take
a
look
at
is
also
for
for
server
actually,
especially
when
server
suppose
you
say:
if
a
cable
server
need
to,
you
know,
support
sufficient
expensive
approaching.
Also,
they
may
have
needles
about
the
general
public,
scripting
environment,
such
as
path
and
how
other
scripts,
so
so
the
proposal
from
our
that
we
think
we
should
you
know,
provide
a
universal.
C
Yeah
so
Joel
Joel
yeagley,
one
of
the
chairs
I,
think
one
of
the
challenges
that
we
we
experienced
in
the
in
the
adoption
call
is
that
there
definitely
is
some
some
opposition
to
adopting
it
as
it's
written
today
and
I.
Think
that
I
think
that's
a
relatively
serious
challenge
to
to
the
question
of
whether
we
can
advance
it
in
those
in
this
working
group.
This
point.
C
C
Q
Thank
chosen
for
the
comments
actually
I,
think
eager.
We
take
our
ego
and
us
and
I
took
a
lot
of
time
to
discuss
with
Andy
I
think
based
on
amenity
to
discussing
where
is
rich.
A
lot
of
women
and
Andy
also
suggests
a
lot
of
import
actually
Stephanie.
We
need
to
think
about
that.
So
our
plan
is
to
report
really
vocally
the
job,
and
we
also
also
actually
will
worker
with
hang
up
because
and
what
are
you
suggesting?
C
Like
pulling
for
adoption
again
as
it
stands
now,
I
don't
think.
Oh
six
is
something
that
can
advance.
B
Q
Q
It
is
another
new
job
that
we
already
put
in
is
the
idea
in
last
idea
meeting
and
one
of
the
issue
weighted
cars
into
how
these
relate
to
the
another
meeting
capability
chapters.
So
we,
you
know
the
genuine
made
actually
which
are
going
on
with
that
job,
and
we
clarified
the
use
cases
which
define
the
telemetry
data
tagging
parameter
at
a
lot
of
attention
on
Purnima,
try
to
support
a
multi-dimensional
dimensional,
operational
data
classification
and
in
which
a
situation
and
a
model
we
add
a
lot
of
new
elements.
Next,.
Q
Q
So
when
all
that
interesting,
theater
export
to
the
anime
and
MSO
connector
within
30
days
can
provide
a
multi-port
dimension
of
inevitability
realities,
so
the
bigger
benefit
is
we
can
identify
the
province
of
memory,
the
data
for
services
or
assurance
application.
In
addition,
we
can
provide
net
availability
to
several
categories.
Data
such
as
such
as
many
one
know
the
traffic
flow
capacity
next.
Q
So
what
can
imagine
data
tagging?
Look
like
how
do
you
classify
the
antenna
metadata
tag?
We
actually,
you
know
lister
for
an
important
context.
Telemetry
data
tag
after
his
oppose
the
important
tacitly
we
call
the
OPM
tagging.
The
OPM
actually
is
a
relation
of
the
object,
type
of
property
and
metrics,
and
this
OPN
tagger
can
help
to
capture
the
problems
in
memory
related
data
to
represent
the
object
ID
for
attribute
and
perform
the
Med
sugar
relate
to
the
young,
didn't
order
or
a
set
of
young
people
note.
In
addition,
we
can
provide
the
fine
granularity.
Q
Opm
targets
such
as
OPM
type
tag,
actually
use
these
tagger
you.
If
you
could
connect
the
performance
metric
of
value.
You
can
indicate
all
this
is
a
metric
of
value
for
don't
freedom
for
latency.
It's
a
maximize
the
value
or
minimize
the
value.
Also,
we
can
no
indicator
the
metric
or
scale
or
magical
preceding
season.
The
second
category.
The
tacit
way
we
provide
is
civil
caca.
Actually,
we
can
use
three
taggers,
you
know
for
for
some
young
people.
They
don't
know
that
they
can
be
used
for
multiple
public
purpose.
Q
So
we
can
use
this
service
tacit
to
describe
each
other's
food
number,
air,
civilian
service
or
tunnels
of
it
and
the
third
category.
We
call
the
status
also
at
type
we
can
identify
different
data
types
that
relate
to
the
connectivity
result.
Hardware
and
the
last
packet
we
provided,
we
call
the
parent
improving.
This
is
something
related
to
the
MA
TESOL
segregation.
Q
So
how
do
we
use
this
telemetry
data
tag,
so
typical
in
this
case
is
actually
web.
We
actually
provided
how
to
use
the
OPM
tag.
Actually,
with
the
OPM
tackle
you,
you
can
actually
take
a
hardware
model
alarm
model,
a
similar
interface
model
to
capture
the
performance
of
memory
related
her
to
classify
them
into
optical
type
property
magical.
Q
So
we
can
provide
a
multi-dimensional
telemetry
data
classification,
so
this
OPM
target
and
the
version
space
in
all
the
objects
in
the
device
model,
and
we
can
you
know
advertise
these-
can
measure
data
tag
to
the
connector
or
MMS,
so
I,
never
so
can
know.
So
what
kind
of
ten
imagine
a
that
they
really
want
to
subscribe?
All
connect
so
indicate
we
can
provide
a
multi-dimensional,
can
eventually
dialysis,
so
we
gave
this
example.
I
Q
The
second
case
is
about
a
mother
sauce
aggregation.
Actually,
when
you
connect
opponent,
basically
like
the
latency
from
Lankenau,
one
unkind
to
maybe
in
the
same
metal
element
a
so,
you
really
want
to
know
which
the
latency
from
which
in
Bangkok
and
Gruber
together.
So
you
can
use
the
parent
and
good
in
fact,
to
do
that
next.
Q
Q
So
we
think
that
the
K
value
for
for
this
telemetry
data
tag
is
that
we
can
use
them
to
identify
the
phone's
memory
really
later,
but
I
think
the
most
important
things
that
you
can.
Actually
you
have
large
amount
of
device
model.
You
have
many
the
object.
You
are
interested
in.
You
use
these
dynamics
and
metadata
tagging.
You
can
know
how
to
you
know,
correlated
data
object
in
different
a
model
together,
so
they
can
provide
a
good
another
way
inside
to
to
these
telemetry
data.
So
we
think
it's
a
very
useful
walker.
A
Q
A
L
This
is
not
a
proposal
for
work.
This
is
just
the
background.
Information
that's
what's
happening
out
in
the
world,
I'm
working,
both
in
IETF
and
also
in
3gpp,
which
defines
the
mobile
networks
and
they
went
there
from
web
or
mobile
networks.
3Gpp
has
recently
in
the
last
year
adopted
Yang
and
that
Gulf
in
a
big
way.
Next,
please
next
slide,
please
bit
of
background
how
we
GPB
is
working.
They
define
everything
game,
so
called
stage.
1
2,
3
solutions,
stage,
1,
&
2,
basically
stays
what
how
the
function
t
should
work
a
stage.
L
3
there's
gives
you
a
precise
definition
that
you
can
be
really
code
against.
Currently,
each
3
there
are
yang
Netcom
for
JSON
schema
and
an
XML
schema
solution
sets
out
being
defined
for
5g
networks.
Next,
please
I
was
very
heavily
ok
how
they
are
defining
this.
They
use
UML,
which
is
object,
oriented
and
as
many
different
rules.
They
have
restricted
UML.
L
L
I
have
been
heavily
involved
in
how
to
map
these
object.
Oriented
our
models
in
the
yang
there
have
been
actually
multiple
airports
around
this
ours
might
be
one
of
the
bigger
ones.
We
produced
one
specification
that
defines
how
the
mapping
is
done.
L
Some
parts
of
it
is
rather
straightforward,
so
an
object
becomes
a
grouping
inside
lists.
An
attribute
becomes
Alif,
but
some
parts
are
quite
complicated
like
in
how
do
you
model
inheritance,
inheriting
attributes
easy,
but
inheriting
containment
is
very,
very
tricky.
Also
attributes
what
they
call
attributes
can
be
very
still
a
big
structure.
It
can
be
a
simple
string,
but
it
can
be
a
structural,
including
a
structuring
through
the
structure
of
of
strings
and
integers
and
whatever
next
slide
please.
L
This
is
just
an
example
of
how
the
mapping
is
done.
We
have
here
bosses
an
abstract
class
that
only
has
a
grouping
defining
a
few
leaves.
We
have
a
managed
element
that
will
be
concrete
class
I,
think
it
inherits
from
the
abstract
class,
and
then
it
is
really
instantiated
in
this
list
of
manager,
and
then
it
also
has
contained
class.
It's
the
simple
case,
X
I,
please.
L
So
containment
can
be
very
tricky.
They
have
four
first
four
records
if
containment,
which
can
be
mapped
to
something
like
list
in
IETF
interface,
what
is
much
more
troublesome
is
containment
by
inheritance
when
an
abstract
class
it
contains
other
classes
and
then
suddenly
and
a
lot
of
clutter
costs
inherits
from
the
abstract
class,
and
suddenly
you
say
that
ok
I
modify
the
abstract
class,
so
anyone
who
has
inherited
from
the
abstract
class
needs
to
modify
their
containment
structure.
This
is
like
if
we
could
modify
or
augment
the
grouping
which
we
don't
allow.
L
Addressing
is
are
different
and
we
had
to
have.
We
have
to
come
to
compromise
on
this
day
using
something
called
distinguished
names
which
held
up
also
users.
They
have
used
it
for
a
long
time,
so
they
will
not
change
for
yanks
sake,
though,
and
also
distinguished
names
are
a
network
that
will
hierarchy.
So
it's
not
just
the
yang
server
or
the
napkin
server
the
addressing,
but
they
can
address
which
net
conserver,
which
network
system
they
are
using
with.
L
L
They
have
default
values,
but
they
do
really
mean
something
different
one.
What
we
have
there
invariant
variables,
which
cannot
be
changed
once
they
have
been
really
created
and
they
have
system
created
objects,
so
read-only
objects,
read-only
lists
that
can
contain
configurable
leaves-
and
yes,
we
don't
like
this.
We
decide
not
to
do
this,
but
this
system
also
works.
It
has
been
working
for
20
years
exploit.
Please.
L
L
They
are
handling
all
these
yang
models
in
Microsoft
Word
documents,
which
is
a
terrible,
terrible
idea,
but
it's
not
the
3gpp,
the
only
one
who
has
this
crazy
idea.
One
interesting
point
is
that
they
really
have
a
full
release
every
18
months
and
the
way
or-
and
they
don't
consider
backwards,
compatibility
so
important,
which
is
very
strange
for
me,
but
that's
what
they
do
excite
these.
L
They
also
accepted
the
usage
of
that
core
protocol,
but
with
yank
push
we
have
great
problems
we
might
get
it
accepted,
but
as
we
it
was
quite
late,
I
mean
it
might
not
be
accepted
after
all.
Next
slide,
please,
okay,
that's
that's
it
if
anyone's
interested
I'm,
one
of
the
main
yang
guys
there.
So
please
contact
me
yung
lean
that
helped
me
some
as
well.
Thank
you.
M
M
So,
basically
the
idea
is
that
in
the
constraint
space
we
like
binary
information,
so
we
are
presenting
in
C
bar.
You
know,
I
could
short
identifiers,
so
we
have
a
way
to
map
yang
item
identifiers
into
numbers
which
we
call
SIDS
and
we
are
using
Co
app
instead
of
HTTP.
But
apart
from
that,
it
looks
a
lot
like
rest
country.
So
this
is
a
whole
set
of
documents
that
works
together,
which
is
why
why
this
took
a
while
to
complete,
but
parts
of
the
whole
suite
may
be
useful
for
other
applications.
M
M
So
the
people
you
want
to
talk
to
probably
are
the
authors
which
are
down
at
the
bottom
of
the
slide.
I
happen
to
be
either
document
Shepard,
because
I
was
a
working
up
chair
at
the
time
when,
when
we
did
the
working
last
call
car
is
now
chaired
by
by
Jaime,
reminisce
and
Marco
de
Luca,
and
the
responsible
ad
for
the
Cobra
group
is
very
Libra
I'm.
Just
mentioning
this
in
case,
you
need
to
talk
to
people
as
opposed
to
an
anonymous
mailing
list.
M
Next
slide,
there
are
four
documents:
Yangtze
bar
is
pretty
much
fully
cooked,
so
it
doesn't
hurt
if
more
people
look
at
that,
but
I
do
not
expect
big
problems.
Turning
up
there
anymore
SIPP
is
a
pretty
aggressive
way
to
encode
young
item
identifier.
So
that's
interesting
and-
and
please
do
have
a
look
at
the
application
process
we
came
up
with
and
then
we
have
something
that's
pretty
called
specific,
which
is
the
mapping
to
coop,
but
of
course,
that
still
needs
to
be
net
compatible
and
in
some
way
or
restaurant
from
that
word.
M
So
we
have
completed
the
working
group
last
call.
We
have
a
number
of
open
issues.
This
slide
lists,
maybe
the
five
more
important
ones,
but
I
won't
really
have
time
to
go
into
that
next
slide
and
last
slide.
So,
given
the
working
last
call
is
now
done,
the
next
steps
are
likely
to
be
that
the
authors
process.
M
These
comments,
generate
new
versions
of
the
four
drafts
and
that's
for
the
chairs
to
decide,
but
I
would
expect
that
we
will
have
a
second
person
last
call
and
if
you
missed
the
first
one,
please
do
have
a
look
at
the
second
word.
Noob
last
call
and
we
do
plan
to
ship
to
the
isg
at
the
end
of
this
month.
So
this
will
not
be
an
endless
story.
This
will
go
to
the
is
tree,
so
I'm
done.
B
Okay,
great
thank
you.
Everyone
we've
made
just
in
time,
and
so
we
will
be
posting
a
follow-up
recording
of
the
session
and
the
minutes
and
and
also
the
data
tracker,
all
the
slides
we
posted
there
and
if
there's
any
remaining
conversation
and
was
taken
to
list,
let's
please
do
that
and
stay
safe.
Thank
you.