►
From YouTube: IETF100-SACM-20171115-0930
Description
SACM meeting session at IETF100
2017/11/15 0930
https://datatracker.ietf.org/meeting/100/proceedings/
A
A
D
B
C
C
E
F
F
G
I'm,
sorry,
clearly,
your
chairs
have
not
had
enough
caffeine
or
something
this
morning,
so
this
is
the
second
working
group
of
as
we've
already
indicated.
First
of
all,
we
are
three
chairs
today,
because
Adam
is
departing
and
Chris
is
coming
in.
So
first
of
all,
we
would
like
to
thank
Adam
for
his
leadership
over
the
years.
G
G
H
G
I
I
No
problem:
okay,
I'm
Steven
Bankart
with
NIST
I'm
gonna
be
providing
a
little
bit
of
a
cross
working
group
update
first
and
then
using
that
as
a
lead-in
to
the
sockem
draft
here
that
relates
to
that
so
I'm
gonna
be
talking
about
the
Rolie
girafft.
That's
in
Mile,
resource-oriented
lightweight
information
exchange.
This
is
just
a
really
brief
update
on
its
current
status,
so
that
everyone
here
is
aware
of
that
says
it
does
impact
a
couple
of
the
drafts
here
in
Sakon.
So
next
slide,
please
that's.
I
Okay,
so
the
Rolly
draft
in
Mile
is
currently
an
IAS
G
review.
We
were
navigating
a
few
remaining
issues
that
were
blocking
the
discusses
of
that
draft.
As
of
just
a
couple
days
ago,
we
have
resolved
those
primary
issues
and
we
are
now
fully
expecting
those
discusses
to
be
cleared
as
soon
as
we
post
a
new
version
of
the
draft
in
the
next
couple
of
days.
So
the
current
version
that
is
in
mile
is
version
of
13
once
we
post
14,
that
will
be
the
one
clearing
the
remaining
issues
and
hopefully
with
any
luck.
I
That
is
the
final
version.
We've
had
really
great
review
so
far
from
iesg
ad
last
call.
All
of
those
steps
we've
had
a
great
number
of
editorial
changes,
a
lot
of
great
review
and
a
small
number
of
minor
and
normative
changes
and
I'll
cover
some
of
those
briefly
I.
Think
on
my
next
slide.
Go
there
I
didn't
so
I'll
go
over
it
really
quick.
The
primary
change
that
we
did
to
clear
the
discusses
on
that
row.
Leak
or
draft
was
removed.
The
Ayana
registrations
for
the
dot,
well-known
URL
location.
I
Yes,
Thursday
afternoon
smile.
So
if
you're
interested
in
hearing
more
about
those
changes,
you
can
come
there
so
onto
this
draft,
the
back
one,
sorry
I
didn't
actually
make
it
to
that
slide.
I
went
on
a
tangent.
This
is
the
software
descriptor
extension.
This
is
the
draft
that
is
actually
here
in
sockem.
This
is
an
extension
to
Rolly
that
we've
been
working
on
here
that
covers
the
Rolly
capabilities
of
carrying
around
software
descriptor
information.
This
is
information
like
the
ISOs
wid
format,
although
it
is
not
confined
to
that
format
by
any
stretch
of
the
imagination.
I
So
this
draft
was
since
IETF
99
Prague
adopted
as
a
working
group
draft.
So
it's
official
title
is
now
draft
IETF
sockem
Roly
software,
descriptor,
0
0.
We've
continued
to
work
on
it
made
some
editorial
changes
to
the
text
but,
most
importantly,
we've
updated
the
extension
to
a
new
format.
This
is
a
standardized
format
that
we're
going
to
be
using
for
all
Rolly
extensions.
So
the
primary
change
in
this
version
is:
we've
changed
the
organization
of
the
document,
so
we'd
really
appreciate
it.
I
I
B
H
J
H
I
I
I
We're
looking
for
more
review
of
this
document
as
well
as
always
review
everything.
It's
really
helpful
for
the
for
the
authors
of
those
okay
next
slide
is
that
it
no
way
forward
right.
So,
just
as
we
were
talking
about
a
second
ago,
we
really
feel
like
software
descriptor
is
currently
in
a
pretty
good
place.
It's
a
technical
description
of
this
new
information
type.
It
covers
all
the
necessary
registrations
as
far
as
we
think
it
does,
and
it
really
kind
of
describes
the
extension.
I
So
at
this
point,
we
as
the
author's
feel
that
this
document
has
reached
kind
of
technical
maturity
and
we're
just
looking
for
review
so
that
we
can
really
kind
of
start
to
move
it
forward
and
find
out.
If
there's
anything,
we
need
to
change.
So,
if
Chris,
you
said,
you
found
some
knits.
That
would
be
great.
If
we
could,
you
know,
get
those
start
making
some
changes,
get
an
o1
version
out
and
and
continue
to
work
on
that
be
extremely
helpful.
I
K
I
K
K
L
Yes,
it's
mine,
yes,
okay
and
I'd
like
to
introduce
some
of
our
starting
work
about
the
network,
security
stages
or
poster
collection
of
the
some
networking.
First
in
network
infrastructure
devices
such
as
the
uterus,
which
I
think
this
is
maybe
maybe
a
natural
extension
of
current
a
second
walk
in
the
future.
So
we
proposed
here
and
hope
to
get
your
suggestions.
L
L
M
L
Our
an
idea-
and
we
try
to
do
something
here
so
here
we,
the
network
infrastructure
vices.
We
mean
that
you
know
the
uterus,
which
firewalls
those
devices-
okay
and
the
the
goal
is
to
collect
the
security
posture
to
do
the
assessment
of
the
devices
as
a
software
vulnerabilities
in
the
configuration
it's
totally.
M
L
L
So
the
the
overview
of
this
draft
is
first
today
we
want
to
mention
why
we
need
this
kind
of
mechanism
for
the
network
infrastructure
devices,
because
we
all
know
that
country
for
the
for
the
security
circumstances
of
the
network.
Devices
is
serious
because
there
are
a
lot
of
unsafe.
A
sex
channel
exists
here.
L
So
how
can
we
protect
them,
such
as
a
telnet
SNMP,
the
previous
version-
and
we
all
know
that
tcp/ip
is
a
very
open
network
protocol
and
it's
a
bring
us
a
quantity,
but
also
it
opens
a
lot
of
attack
in
attacking
the
face
for
network
devices.
So
can
we
protect
our
what
us
is
more
okay
and
then
the
network
and
the
device
is
becoming
more
and
more
complex.
L
So
that's
another
program
that
we
make
our
security
become
more
harder
and
the
another
key
point
is
that
the
Coverity,
the
traffic
processing
capability,
mismatch
between
the
date,
the
data
plan
and
the
control
end
of
the
device.
So,
based
on
these
readings,
we
see
that
if
we
have
some
kind
of
automatic
monitoring
and
posture
assessment
and
mechanism
for
the
network
devices,
that
will
be
good.
So
so
the
objective
of
the
mechanism
is
to
identify
the
stress
and
the
vulnerabilities
of
device,
such
as
unnecessary
services
and
the
instacart
in
insecure
or
configuration
and
abnormal.
L
M
N
L
Yes,
I
think
yeah,
that's
a
good
comment
and
I
think
that's
all
the
question
we
need
to
consider
later
together,
because
right
now
we
we
just
from
our
vendors
perspective
that
we
think
that
we
needed
those
information
to
be
collected
and
evaluated
to
to
access
our
network
device
security
but
today
and
how
to
do
it.
I
think
young
push
is
a
very
good
me
Canada
and
we
can
paste
on
that's
one
of
my
mind.
L
That's
one
point
and
the
secondary
I
think
I
also
agreeing
that
with
you
that
attestation
is
very
important
for
the
whole
work,
because
if
we,
if
we
just
provide
the
information-
but
we
can
not
count
in
that
information
is
integrated
and
it
has
a
trusted
base
so
that
we're
influenced
the
the
the
operators
charter
that
they,
this
information
is
useful.
So,
yes,
we
need
to
consider
all
these
program
as
a
whole
and.
L
The
data
sensitive
data
in
devices
and
how
to
do
the
privacy
protection
and
a
lot
of
singing
application
there
and
another
important
in
the
party
is
a
network
layer.
Okay,
so
this
layer
are
just
to
consider
how
to
protect
the
network
resources
and
their
service,
so
we
can
use
the
second
the
second
we
can
attend
to
collect
it
in
and
then
based
on
the
collective
attacked
we
can
hire
in
the
configuration
in
the
studio's
connection.
So
for
this
layer
we
have
three
controllers:
three
plan,
controller
and
Italy
and
the
manatee
plane.
L
They
all
cover
their
respective
features.
Security
related
features,
awared
use
later,
ok
and
the
last
one
is
the
infrastructure
layer.
This
is
more
about
the
device
itself
and
the
operating
system
running
on
the
network.
You
Isis,
so
we
also
care
about
their
security.
So
these
are
the
from
all
understanding
that
we
should
appear
about.
We
should
we
should
study
on
this
work.
Ok,
next
slide.
Ok!
L
So
so,
yes,
I,
think
this
I
think
is
this
page
adjuster
can
answer
some
of
the
questions
we
want
this
work
based
on
some
existing
technologies,
such
as
we
use
that
we
think
that
the
young
push
and
notifications
a
sub
pop
mechanism
is
very
useful
so
here
to
help
us
to
to
flexible
and
inhale
you
the
way
to
collect
those
information,
and
then
we
also
want
to
because
this
work
is
finally
used
by
the
Saku
Saku
architecture.
So.
M
L
To
matching
or
how
to
translate
the
network
poster
information
to
the
secona
information
motor,
so
we
need
a
broad,
I
think,
Hank,
as
tom
has
already
a
good
start
on
this
direction.
He
has
a
already
has
a
draft
on
edge,
so
we
can
work
together
with
his
draft
to
to
deal
with
this
program
so
and
another
very
important
thing
that
we
need
to
do
is
that,
because
these
are
all
related
to
the
network
devices
information.
M
L
L
L
So
in
general,
there
were
good
news
and
the
bad
news
for
us
know
for
Socko,
okay,
so
possibly
for
the
date
plane
hit,
play
and
young
mother,
young,
okay,
Ted
play
and
security
posture,
young
mother
and
with
such
the
other
existing
your
mother
any
place.
We
don't
find
the
existing
and
about
this
part.
So
so
right
now,
in
order
after
we
leave
the
some
awful
point
that
we
think
is
important
for
the
networking
devices
security.
L
O
L
Avoid
flooding
or
the
DDoS
attack
to
the
device
and
how
to
detector
the
MM,
fragment
or
Martha
on
the
packets
and
to
just
that,
if
cut
it
in
and
that's
just
from
a
very
initial
idea
of
wrong
or
perspective,
I
think
that
we
think
that
is
not
complete,
and
maybe
it's
some
it's
just
very
beginning
stage.
So
we
hope
that
later
we
can
refine
them
and
we
can
simplify
or
add
more
useful
information
here
next
page,
okay
and
about
to
the
date
plan.
L
We
also
write
our
also
write
our
control
and
story
control
in
security
post
bit
align
your
model,
but
for
this
drafter
we
ran
way
after
our
search
of
the
current
existing
chapter,
we
find
that
most
of
them
are
already
covered
in
the
in
several
charts.
For
example,
the
PGP
is
already
covered,
OSPF
is
MPs
kitchen
DTS
and
all
this
are
covered
in
existing
draft,
but
we
also
identify
some
some
some
missing
content
in
those
draft,
so
I
think
for
us.
Our
strategy
is
that
we
can
contribute
to
to
those
exceed
those
existing
jobs.
L
So
that's
our
current
idea,
okay,
so
the
next
steps
we
counted.
We
want
to,
we
think
we
can
to
further
work
to
simplify
the
current
data
motor,
because
a
lot
of
details
may
be
not
needed
for
it,
and
also
we
can.
We
should
consider
how
to
combine,
where,
with
current
young
push
and
the
notification
mechanism
and
how
to
work
together
with
us,
working
in
the
scope
of
sockem
so
based
on
the
stock
confirmation,
modal,
and
maybe
we
think
that
is
incomplete
information.
L
I
Hi
Steven
Bank
are
just
a
quick
question,
so
I
I
have
read
the
draft.
I
was
just
looking
over
it
again
and
I
just
wanted
to
ask
a
quick
clarifying
question
about.
Maybe
if
you
could
state,
but
what
exactly
are
the
goals
of
this
draft?
Are
you
attempting
to
just
register
new
yang
modules,
or
is
this
trying
to
do
something
more
than
that
I
mean?
What's
the
what's
the
goal
of
this?
Yes,.
L
Okay,
I
yeah,
that's
a
good
question.
I
think
our
ultimate
goal
is
that
we
want
to
based
on
the
current
Sock'em
architecture,
to
collector
the
security
posture
of
the
network
devices,
because
from
our
long-standing
current
ii,
work
only
covers
how
to
collectors
enterprise
devices
security
posture,
but
how
to
extend
their.
They
can
fit
the
existing
work
to
the
network
scope.
So
that's
what
we
want
to
do,
because
this
is
just
a
zero
zero
version
of
charge.
L
Oh,
we
can
only
say
that
Calgary,
we
provided
some
raw
materials
here,
and
we
think
that
this
information
is
useful
to
help
us
to
assessment
at
the
network
status
of
the
security
studios
of
network
device,
but
how
to
use
them
and
how
to
collect
and
how
to
evaluate
it.
Then
how
to
do
the
smart
filtering
of
them
I
think
there
are
a
lot
of
work
need
to
do.
M
I
L
M
I
M
Yeah
I
think
the
term
baseline
is
just
a
little
bit
confusing
at
this
point
here.
So
this
is.
This
is
about
security
posture
either.
This
is
the
attend.
I
get
from
this
presentation.
That
is
not
about
baseline
security.
It
is
about
the
crying
specific
security
posture
from
network
devices
that
you
know,
yeah
right.
L
L
M
The
yang
modules
are
supposed
to
be
the
interface
for
this
additional
information
that
is
probably
or
maybe
not
yet
covered
in
the
information
model
of
second
today
or
unfortunately,
as
you
put
it,
this
is
the
bad
side
is
already
covered
and
other
yang
modules
have
to
strip
it
out.
So
next
steps
are,
is
that
is
making
basically
order
in
the
draft
and
find
out
what
is
redundant
and
safe
from
that
and
find
all
viruses
distinctly
new,
and
if
that
is
missing
from
the
information
model,
make
sure
I
met
you
there.
L
B
Cristinaw
co-chair
hat
off,
so
looking
at
the
draft,
you
had
a
lot
of
yang
data
elements
listed
in
your
draft
and
it
was
more
type
information
than
it
necessarily
included
what
I
would
consider
kind
of
a
baseline.
So
you
didn't
give
guidance
on
any
of
those
elements
to
be
set
in
a
range
you
know
turned
off
turned
on,
etc.
L
No,
no
I
I'm,
not
sure
right
now,
because
I
think
the
most
important
work
ago
for
this
work
is
that
we
just
want
to
collector
or
want
to
specify
what
information
we
need,
but
it's
very
hard
to
define.
You
know
the
what
value
we
anticipated,
something
maybe
some
some
part
of
them.
We
can.
We
can
provide.
So
if
we
open
or
if
it's
turn
on
off
in
certain
off
but
there's
some
some
some
some
counter
or
something
cannot
be,
give
our
only
value
to
evaluate
right.
So
I
think
it's
is
half
of
your
requirement.
L
O
Hello
Jessica
from
Huawei
I'm,
going
to
present
as
a
security
baseline
for
network
device
management
plain-
and
this
is
a
part
of
the
previous
presidential
security
baseline
walk,
as
we
all
know
that
Sakuni
neck
network
devices
is
a
Chinese
a
and
a
crucial
work
for
both
organizations
and
the
operators.
So
our
purpose
of
this
world
craze
true
defines
security
baseline
for
network
devices,
and
we
want
this.
O
Work
can
be
fostered
used
in
the
security
posture
assessment
and
this
document
will
focus
on
the
multiplying
security
and
we
provide
a
data
model
one
and
we
want
to
include
in
incorporate
the
best
practices
from
the
industry
and
that
you
provide
a
minimal
set
of
Security's
features
for
naturalize
management
plain
and,
as
we
all
know,
that
they
are
already
many
work
about
llamado
see
in
IDF.
So
we
will
reduce
the
exists
in
your
models
and
provide
additional
models
under
groupings
for
the
missing
parts.
Next
slide.
Please.
O
So
this
picture
is
the
overall
structure
of
our
of
this
document.
Change
can
be
divided
into
four
parts:
user
authentication
and
authorization
system,
management,
security,
lock,
security
and
foul
security,
and
at
the
bottom
of
this
page
will
list
the
existing
Llamados.
There
are
two
RFC's
that
has
already
published
it
about
the
security
met
about
system
management,
and
we
can
you
reduce
the
security
parts
of
these
young
models
and
the
zerah.
Also
several
drops
led
working
on
TOS
ssh
net
account
syslog
anja
is
a
el
llamado
next
slide.
Please.
O
O
O
The
second
section
is
about
authentication
and
authorization
for
users
away,
meaning
folk
sounds
configuration
of
readers
and
taxes
servers
and
as
a
secure
communication
between
client
and
the
server,
and
the
last
part
is
the
user
profile
it
meaning
about
the
user.
Credential
configuration
and
the
password
complexity
check
next
slide.
Please.
O
For
lock,
security
and
the
file
security
were
many
folks
out
to
a
specters
of
the
security
on
the
log
and
the
files
that
stored
locally
and
another
is
the
remote
transportation.
So
so
for
lock,
security
remaining
focus
sounds
the
local,
the
local
log
access
control
and
is
a
remote,
remote
security.
We
focus
on
the
syslog
security.
This
pad
has
been
already
defined,
so
adjuster
reduce
each
and
full
file
security.
O
O
This
is
a
whole
picture
as
this
document
chanda.
Our
next
step
is
to
try
to
refine
the
data
model
and
we
want
to
incorporate
use
of
faster
practices
from
the
industry.
So
please
review
and
give
us
comments
about
this,
and
now
we
will
also
try
to
combine
this
work
with
second
frame
work
and
to
to
to
meet
our
goals
to
access
the
network
security
posture,
and
we
also
try
to
combine
wins.
The
existing
young
push
push
and
the
sub
pop
mechanisms.
M
I
think,
probably
you
will
be
there.
I
don't
know
if
not
because
I
can
give
you
the
point
in
time
you
can
see
if
this
works
out
and
yeah,
we
have
to
make
sure,
maybe
just
create
one
reporting
yang
module
that
can
be
used
for
subscribe
notifications
and
use
them
both
workgroups,
probably
because
seriously
some
of
the
stomach's
identical.
So
that
would
be
a
good
way
to
move
forward.
I
think
at
least
that
men
and.
N
I,
second,
that
one
other
thing
it's
really
good
to
see
is
a
couple.
Attempts
ago
we
talked
about
the
use
of
yang,
push
and
I'm
glad
that
sacrum
is
listening
and
and
adding
it
we're
trying
to
take
yang
push
to
towards
working
group.
Last
call
and
a
lot
of
the
impetus
for
this
is
going
to
be
the
users
of
this,
whether
it
be
sockem
I
to
NSF
DTN,
and
so,
if
there's
anybody
who
would
like
to
see
and
usually
I
can
push
and
get
it
to.
The
last
call
to
be
great.
H
So
Dave
altameyer
so
I'm
definitely
not
an
expert
in
yang
models,
or
you
know
what's
what's
out
there,
but
in
reading
this
the
one
thing
that
I
think
is
missing.
That
may
be
useful
to
inform
some
of
this
conflict
resolution.
Discussion
is
is
around
you
know.
It
seems
like
this
is
motivated
by
the
fact
that
there's
some
missing
stuff
and
and
what's
currently
out
there
as
far
as
the
a
models
and
and
so
I,
would
like
to
see
more
analysis
around.
H
You
know,
what's
what's
missing,
and
you
know
why
there's
a
need
for
more
work
in
this
in
this
space.
I
think
that
would
actually
be
helpful
to
inform
and
maybe
engage
other
other
folks
within
the
yang
community
on
you
know
what
well
what's
needed
to
really
fill
the
gaps
that
you're
you're
trying
to
it
to
address
so
maybe
would
be
worth
writing
a
draft
that
instead
focuses
on
instead
of
focusing
on
new
work
would
focus
on.
O
L
Just
add
some
clarification
to
that
question.
I
think
this
manager,
plan
and
Java
is
the
Aruban
motion.
Actually
we
noticed
this
program
that
we
country
what
we
are
doing.
It's
already
exist
in
some
some
young
related
attracts.
So
we
do
the
brief
analysis,
just
like
Choshu
said
and
then
I
think
we
should
the
two
mode
of
this
work
to
give
an
overall
picture
of
what
is
missing
in
current
existing
young
model.
That's
that
part
of
the
work,
but
that's
not
all,
of
the
work,
I
think
for
the
core
of
our
work.
L
We
want
to
do
this
working
second,
because
that
we
think
that
stuck
on
is
a
very
good
place
to
to
the
the
end-to-end
and
network
security
post,
your
collection
and
assessment,
so
that
doesn't
mean
that
the
doesn't
mean
only
the
content.
We
what
we've
content
we
need
so,
but
also
covers
the
mechanism
how
to
how
to
process
this
message,
this
information,
so
how
to
collect
and
how
to
evaluate
then
I
think
there
are
still
a
lot
of
work
related
to
to
these
Network
Devices
how
to
how
to
manage
isn't
so.
L
G
M
I'm
Hank,
Adam
and
I
were
the
champions
of
the
second,
the
combined
that
come
second
hackathon
project
of
the
ITF
100
and
last
time
for
a
quick
recap:
we
for
the
first
time
combined
the
yang
push
technology
with
the
XMPP
quit
broker
draft
technology
that
is
incubated
in
Maya,
and
so
what
we
did.
There
was
a
static
combination.
We
had
two
participants
of
vendors.
M
It
was
for
y
from
the
People's
Republic
of
China
and
Cisco
from
the
States
of
America
and
both
provided
how
we
are
in
their
own
locations,
so
we
were
basically
live
connecting
across
the
globe
and
aggregating
that
information.
We
are
subscribed,
notifications
here,
they're
in
Prague
and
fed
that
into
the
second
domain
for
the
post
process,
man
that
is
defined
by
the
at
the
moment
expired
architecture.
M
This
time
we
addressed
the
new
items
that
are
defined
for
the
we
chartered
for
the
new
charter
of
second,
which
are,
if
I
remember
correctly,
messaging
variation
and
collection,
and
so
most
of
it
is
about
collection
orchestration.
But
this,
which
is
the
messaging
all
of
them,
include
the
term
orchestration,
which
is
like
a
meta.
All
of
that.
So
this
time
we
are
focusing
on
discovery
of
target
endpoints.
M
This
is
done
assuming
that
we
are
addressing
network
equipment
primarily
and
that
network
commitment
is
supposed
to
have
yang
interfaces
active
and
those
are
then
again
yang
push
capable.
So
the
way
forward
was
to
have
a
route
network
device
and
then
crawl
through
the
network
by
discovering
the
capability
that
a
network
device
can
find
its
own
neighbors
in
the
layout
who
broadcast
domain,
we
are
lldp
and
CDP
or
CDP
and
has
a
corresponding
module
yang
module
capability.
M
We
then
query
that
so
subscribe
to
that
we
have
new
other
slides
now,
I'm
surprised,
okay,
I
mean
knowing
that
and
and
yeah.
Therefore,
we
can
define
new
management
plane
interfaces
for
other
devices.
Again
we
have
a
capability
given
and
subscribe
to
appropriate
modules
and
therefore
discover
the
complete
technology
if
it
is
enabled
with
CD
Pol
ADP.
That
is
one
step.
The
other
step
is
that
we
orchestrate
the
collection,
so
we
provided
or
created
our
first
second
repository,
the
second
phone,
which
is
the
imperative
guidance
repository
in
this
database.
M
Basic
key
value
store,
we
find
device
type
device
model
and
potentially
even
the
IP
address,
which
is
easiest
thing
to
do
at
first
and
then
find
out
by
a
manually
configured
value
which
capabilities
are
interesting
or
defaults
to
subscribe
to
in
this
device.
Type
device
model
combination,
so
probably
you're
interested
in
something
else
when
it
is
a
just
a
purely
a
to
switch,
then
if
it's
a
full-fledged
carrier-grade
router
so
again
that
this
is
guidance
what's
to
collect
from
a
specific
type,
mother
and
yeah.
These
are
our
two
basic
accomplishments.
M
I
think
we
focus
on
orchestration
or
on
static
configuration,
and
we
created
offers
repository
for
our
guidance
in
this
case,
imperative
guidance
and
yeah
big
I
think
we
want
a
tailored
price
that
is
targeted
at
CrossFit
group
collaboration
and
considering
that
we
now
create
the
days
and
our
third
binding
to
yang.
That
is
not
net
current
and
not
restaurants.
We
hope
that
we
can
even
show
feasibility
for
the
comai
binding
of
yang
crush
even
next
year
in
London,
I.
P
M
The
terminology
please
and
now
I-
will
answer
the
question
and
imperative
guides
is
telling
you
what
to
do.
It
is,
for
example,
it
can
be.
A
subset
is
typically
configuration
on
the
management
plane.
It
is
how
a
device
also
behave.
It
is
making
a
device
change
its
behavior.
In
this
case
it
makes
the
collector
do
specific
collection
operations
on
specific
device
types
and
models.
M
The
contrast
to
that
is
imperative.
Sorry
declarative
guidance.
This
is
guidance
that
tells
you
how
it
should
be.
So
if
the
temperature,
for
example,
is
as
a
threshold
of
50
and
it's
always
at
60
you're
violating
declarative
Gardens.
So
therefore
it
could
be
that,
for
example,
the
term
policy
which
is
like
overloaded
and
burnt
term
here
and
the
ITF
is
declarative
guidance.
Also,
you
can
define
the
courage
of
guidance
via
description
logic,
languages
and,
if
John
stressed
nervou
would
not
be
here.
M
He
would
be
basically
at
the
mic
and
tell
us
how
this
is
done,
but
there
are
existing
solutions
to
do
very
sophisticated
ways
of
defining
imperative
guidance,
declarative
guidance
and
intend.
It
is
a
third
category,
but
here
starts
just
stops.
My
expert
knowledge
as
a
John
says,
never
be
the
one
to
ask
at
this
point.
H
M
We
read
I'm
actually
not
sure
how
much
of
that
went
into
drafts
already
again.
I
would
contact
John
stressed
not
offline
I.
Try
to
send
him
to
you
and
Stephen
yesterday,
I'm
not
sure
that
of
successful,
because
he
was
writing
slides.
All
the
time
is
that
he
has
was
working
on
actual
solutions
in
all
three
areas.
So
I
guess
we
should
go
and
ask
him.
I
mean.
H
M
Maybe
there
is,
there
is
a
draft
called
something-something
capabilities
in
the
I
to
NSF
the
ECA.
The
event
condition
action
model
is
described
in
detail
there,
and
it
includes
some
of
the
variations
of
the
intent
and
declarative
guidance.
Language.
So
I
would
read
that
draft
first
and
order
to
get
a
little
bit
familiar
with.
The
ECA
model
is
used
to
aggregate
hierarchies,
offer
event,
refinement
or
notification,
refinement
and
I
to
NSF
yeah.
H
These
drafts
yeah
I
think
it
would
also
be
useful
for
those
of
us
that
are
not
I
to
NSF
experts
to
understand
how
that
more
closely
relates
to
the
work
here
in
sockem
would
be.
Maybe
a
couple
of
ideas
on
how
could
happen
would
be
posting
a
quick
summary
to
the
list
talking
about
how
the
work
relates,
or
you
know,
or
writing
a
draft
talking
about
how
okay.
M
Yeah
I
will
not
have
the
time
to
write
a
draft
for
that,
but
I
would
recommend
everybody
to
join
us
at
Butterworth
at
5:00
p.m.
today.
Although
it
is
a
small
room
and
maybe
I
should
not
invite
anymore
I
reckon
he
said
it's
a
small
room.
We
are
already
a
lot
of
people
there,
but
I
would
again
start
reading
existing
drafts
and
I
to
NSF,
which
are
always
so.
H
J
M
M
H
Q
One
edition
action
top
of
that
there's
actually
discussion
in
MRG
currently
actually
is
also
similar
to
that
busy
regarding
how
how
intent
of
policy,
how
service
models,
how
all
those
things
actually
relate.
There
was
like
a
big
discussion
on
Monday
regarding
this
and
there's
also
actually
a
draft
on
this.
Perhaps
the
imperative
and
declarative
the
guidance
are
some
things
to
do
also
after
the
discussion
month
anyway.
M
The
the
problem
here
is
I
think
to
build
a
general
solution
without
a
very
specific
problem.
Domain
failed
twice
here
in
the
ITF
and
now
nobody
is
doing
that
anymore.
So
every
working
group
no
starts
its
own
version
of
it,
and
now
people
find
out
hey
this
somehow
always
sucks
in
looks
similar,
which
is
a
conundrum,
so
basically
really
know
this,
and
finding
out
that
our
problems
are
similar,
helps
us
to
create
an
actual
lightweight
combined
solution.
M
Q
B
J
E
I
just
like
to
sort
of
get
some
guidance
or
suggestions
as
how
we
should
move
forward
in
terms
of
you
know,
second,
technology,
whether
you
know
we
should
combine
or
mean
to
me
at
the
moment,
I
think
it
still
makes
sense
to
have
a
separate
terminology
draft
for
a
second
and
we
can
revisit
that
decision
later.
If
you
know,
if
we
want
to
move
it,
you
know
merge
it
with
some
other
drafts.
We,
you
know
that
doesn't
preclude
that
and
I
think.
E
M
M
Not
expired,
this
is
due
to
as
we
are
working
on
refactoring
a
version
of
the
architecture,
but
it
is
simply
accumulating
too
much
work
on
my
side.
I
cannot
were
also
the
hackathon
was
more
important
to
find
out
and
through
a
future
feasibility
from
okay.
This
makes
absolutely
no
sense,
and
that
really
helped
follow
us
through
iterations
now,
and
it
will
have
a
very
good
effect
on
I
think
both
the
x-pyr
drafts
and
the
terminology.
R
Hi,
this
is
Nancy,
so
I
just
want
to
say
the
terminal
terminology
draft
is
important.
I,
don't
have
any
strong
opinions
about
convergence
merging,
but
just
as
an
FYI
as
I
am
working
with
now
a
new
co-author
on
the
XMPP
grid,
which,
admittedly
is
a
mile.
We
found
it
helpful
to
reference
the
terminology
that
is
currently
in
sockem
to
help
clarify
the
intent
that
what
we're
trying
to
do.
G
R
M
S
M
Caron
was
pointing
at
Adam,
who
looked
very,
very
happy,
so,
yes,
it
is
moving
forward.
It
is
just
not
moving
as
fast
as
it
did
due
to
the
reasons
Jared
just
said,
it
is
not
intended
to
stall.
All
is
basically
effectively
not
finished
yet
and
I
think
that
the
alignment
work
in
the
data
models
that
we
are
now
aiming
at
will
also
consolidated
the
terminology.
More
also,
it
helps
that
John,
Cessna
and
I
are
off
of
the
iPhone
is
F
terminology.
M
M
And
I
think
it's
I
don't
know
the
extra
charm
out
here
there
are
non-essential
drafts,
like
I
don't
know
like
a
problem
statement
or
so
that
it
should
never
be
pushed
to
is
gb
because
I
forgot
the
term.
I'm
sorry
they're
supported
supportive
draft
support
rafts
and
those
should
never
be
pushed
to
the
is
she
because
it
is
a
gateway
function?
They
are
overloaded
by
that
and
just
would
see
the
context
and
combine
drafts,
and
I
try
to
understand
and
read
the
definition
of
support
raft
and
it
seems
the
terminology
is
not
covered
by
it.
K
So
Kathleen
weary
one
clarifying
point.
The
word
never
was
not
part
of
the
statement.
It
is
up
to
the
working
group
to
justify
why
something
might
need
to
be
published
and
there
are
split
views
amongst
the
iesg
and,
of
course,
with
a
turnover
in
March
I.
Don't
know
where
that
split
will
move
to
right.
It's
been
on
our
trends
of
disliking
support
documents,
more
and
more,
it
could
swing
back
or
it
could
swing
even
further
to
disliking
them.
K
It's
just
encouraged
to
have
them
the
wiki
based
so
terminology.
We
don't
see
these
too
often
coming
through
it's
fine.
If
the
working
group
wants
to
publish
it,
you
know
I
will
happily
support
that
and
and
move
that
through,
but
it's
up
to
the
the
working
group
we
often
see
them
combined
with
other
documents,
so.
G
K
H
K
So
you
might
use
your
sockem
wiki
to
point
to
expired
drafts
or,
if
you
want
to,
if
you
think
it's
important
to
have
them
published,
so
that
the
broader
community
using
these
specifications
coming
forward,
really
need
them,
then
just
make
that
clear
and
myself
or
subsequent
ad
will
fight
for
that
right
because
we
are.
We
are
here
to
support
the
working
group.
All.
G
K
G
Are
not
stable,
so
what
I
had
asked
at
the
last
meeting
was
that
we
at
least
keep
at
the
last
meeting.
There
was
a
conversation
around
you
know.
The
terms
that
were
in
the
terminology
document
were
not.
You
know,
weave.
Those
terms
have
evolved
somewhat
in
the
conversations
with
the
I
to
NSF
and
this
group
that
were
working
on
terminology,
but
that
wasn't
getting
rolled
back
into
the
draft
and.
K
G
M
G
M
M
G
M
G
R
Means
again,
yeah
I
mean,
as
the
author
of
the
architecture
draft
pitted
despair.
True
I
actually
stopped
working
on
it,
because
I
sensed
and
not
sense
I
got
feedback
that
questioning
whether
it
was
useful
or
agreement.
So,
if
there's
belief
that
there's
use
to
it
I'm
more
than
happy
to
bring
it
back
and
move
it
forward,.
R
G
B
B
C
M
B
M
Was
my
form
of
getting
in
soon
again?
We
had
a
lot
of
things
on
the
plate
in
the
last
iteration,
Dave
and
I,
and
we
soon
spiced
so
I
think
we
can
be
better
than
this
and
this.
This
is
not
the
problem
that
the
content
is
actually
quite
clear
and
quite
understandable
and
going
forward,
and
it
is
about
just
doing
it
now.
I
mean
if
you
compare
the
editors
version
with
the
submitted
version
on
the
github,
and
you
see
the
significant
additions
that
are
already
in
I.
G
H
De
vuelta
mire
so
yeah
I'd
like
to
get
this
curse.
Word
document
done
me
as
far
as
milestones.
It
would
be
great
if
we
could
aim
for
a
working
group
last
call
on
the
draft
by
the
next
meeting,
because
we've
been
working
on
it
for
awhile
and
I.
Think
we're
getting
to
the
point
where
you
might
be
addressing
new
use
cases
around
curse
web,
but
we
can
do
those
and
an
additional
drafts
I
think
it's
important
to
pull
the
trigger
and
get
the
core
done
as
soon
as
possible.
So.
M
G
M
M
We
use
a
lot
of
attributes
that
align
with
the
ISIL
standard
and
some
of
those
are
not
as
well
defined
or
useful
at
some
points
make
sense
as
it
should
be
making.
So
my
current
thinking
is,
we
will
diverge
from
some
of
these
definitions
and
make
them
better
I'm,
actually
more
useful,
so
but
this
stuff
that
I
have
to
line
with
Dave,
who
was
also
involved
in
creating
the
ISO
standard,
and
we
have
to
prune
everything.
M
My
biggest
problem
was
that
the
current
of
Weber
schema
definition
from
ISO
makes
absolutely
no
sense.
It
is
broken
in
so
many
ways
that
you
can
define
switch
that
are
basically
no
sweats,
and
so
we
want
to
concise,
with
definition
that
doesn't
have
that
problem
and
I
think
also.
There
are
some
implicit
definitions
somewhere
in
the
text
or
somewhere
in
the
XD,
that
I
never
explicitly
defined
and
I
think
gave
us
our
best
expert
on-site
in
this
domain.
M
M
F
Hey
I
can
talk
about
that
a
little
bit
so
UCP,
I
think,
is
ready
to
be
updated.
We're
going
to.
I
think
the
authors
are
debating
what
kind
of
diagram
they
want
to
include
in
it
if
I
understand
correctly.
So
it's
something
that
can
happen
on
lists,
which
is
why
it
wasn't
a
presentation
about
it
today
that.
I
G
I
So
I
don't
really
have
anything
extra
beyond.
Basically
what
I
said
at
the
end
of
the
thing
just
waiting
for
review
for
people
to
look
at
it,
I'll
probably
go
over
it
with
a
microscope
and
see
if
I
can
get
some
editorial
changes
to
make
before
the
next
IETF
posted
zero
one.
So
if
things
continue
as
they
do
now,
I'll
probably
ask
for
a
last
call
at
London
at.
H
G
G
I
G
G
B
B
G
So
all
of
the
so
the
Charter
has
gone
to
the
iesg.
We've
gotten
some
comments
back.
The
some
of
them
have
to
do
with.
We
aren't
quite
clear
enough
in
some
of
our
language,
so
Chris
and
I
worked
for
a
while
yesterday
on
editing
a
little
bits
of
it.
This
is
not
making
any
substantive
technical
changes
to
it.
It's
just
trying
to
be
a
little
bit
clearer
in
the
language.
G
The
one
thing
that
we
did
get
feedback
on
was
that
they
would
like
to
see
some
milestones.
The
milestones
are
not
actually
part
of
the
Charter
they're
actually
separate,
but
we
do
need
a
concise
set
of
milestones,
so
I
guess
what
I
would
like
to
see
is
a
discussion
about
what
our
milestones
are,
because
the
list
of
things
that
we
work
on
well,
primarily,
the
feedback
was
boy
that
list
of
things
that
you're
going
to
work
on
is
pretty
abstract,
and
wouldn't
it
be
nice
if
it
were
more
concrete,
so
there's
a
couple
different
ways.
G
We
could
do
this.
One
way
we
could
do.
It
would
be
to
go
through
that
list
of
work.
Items
will
include
from
the
Charter
and
discuss
which
specific
documents
might
map
to
those
that
people
are
planning
on
or
are
currently
in
the
process,
I'm
open
to
suggestions
on
how
to
do
this
without
getting
into
a
lengthy
architecture,
not
not
specific,
not
sockem
architecture,
but
a
lengthy
conversation
around
it.
So
thoughts.
G
H
Well,
you
were
doing
that.
One
thought
crossed
my
mind
based
on
what
you
had.
This
is
Dave
altameyer,
based
on
what
you
had
up
on
the
on
the
screen.
I
think
it
would
be
useful,
maybe
to
descend
coast,
wid
and
really
software
descriptor
at
the
same
time,
because
the
software
descriptor
references
coast,
wid
and-
and
you
know,
since
they're
related
and
concept,
it
might
make
sense
to
tie
those
milestones
together.
G
G
Yes
or
a
draft,
or
is
this
something
we're
going
to
have?
We
have
these
work
items,
and
so
what
I
want
to
know
is
what
specific
documents
I
mean.
The
milestones
can
evolve
and
change
over
time,
but
it
seems
to
me
that
we
ought
to
have
an
idea
right
now
of
what
these
drafts
are
going
to
be,
and
some
of
our
drafts
already
map
into
this,
but
or
do
portions
of
this,
but.
M
So
this
is
Hank
hi,
first
of
all,
without
a
stable
architecture.
This
is
impossible
to
do.
Second
of
all,
I
think
that
the
yang
push.
What
we
do
is
in
the
orchestration
of
filter
expressions
for
young
push
and
subscribe.
Notifications
is
a
wait
who
orchestrated
structions
so
yeah.
That
will
be
a
draft,
become
a
draft
but
I
don't
know
exactly
when
it
depends
on
I.
M
H
So
I
was
I
was
just
gonna,
respond
to
the
first
first
point
about
providing
different
types
of
guidance.
So
we
we
have
Roli
as
a
general
purpose
method
of
allowing
any
kind
of
guidance
to
be
retrieved
for
use,
and
we've
been
working
on
information
types
with
enrollee
to
define
specific
types
of
security,
automation,
guidance
that
can
be,
you
know,
sent
to
a
collector.
So
an
example.
Draft
of
that
would
be
the
configuration
descriptor
draft,
which
is
a
current
personal
draft.
H
H
B
The
observation
seems
to
be
that
there's
a
lot
of
work
going
on
with
kind
of
yang
malls
coast,
wood,
etc.
That
to
me
feels
like
mechanisms
to
enable
collection
which
I
would
kind
of
think
of
you
know
if
I
look
at
some
kind
of
vertical
stack
of
operation.
Those
are
methods
that
are
at
the
very
bottom
and
then
I,
look
at
Raleigh
and
I
think
boy,
that's
kind
of
at
the
top.
B
But
it's
not
clear
to
me
that,
where
we're
coming
from
on
top
and
where
we're
working
at
from
bottom
actually
know
that
they're
going
to
a
single
spot,
I
think
we're
in
I
think
we've
gotten
to
the
point
via
hackathons
that
were
that
we're
going
to
the
same
neighborhood
at
least
that
that's
positive
I'd
like
to
end
up
in
the
same
house,
and
you
know
feel
free
to
disagree
with
me
on
that.
But
that's
kind
of
my
impression
at
this
point
now:
okay,
Adam
go.
F
Yeah
I
kind
of
feel
like
the
collection
we've
we've
got
so
to
me,
like
the
second
line
here,
define
extension
of
IETF
Nia
to
collect
to
deliver
information.
You
know
from
from
that
thing,
that's
actual
collection,
how
we
express
the
guidance
for
that
I,
don't
think
we
have
anything
really
set,
even
if
we
have
Roley
to
deliver
it,
we
don't
have
anything
set
for
that.
Yet
right
so
so
we
are
going
to
need
a
way
to
express
that
I.
F
H
Dave
altameyer,
so
this
text
says
to
find
a
way
to
provide
three
types
of
guidance,
and
so
that's
why
I
was
speaking
to
two
Rolly
I.
Think
I.
Think
it's
your
point
that
we
we
do
need
to
get
to,
and
this
is
maybe,
where
the
squishy
middle,
that
Chris
was
talking
about,
comes
into
play.
We
do
need
to
get
to
defining
what
that
guidance.
B
I'm-
okay
with
that,
so
so
now,
let's
pivot,
that
back
into
milestones.
So
to
me
it
sounds
like
you
know:
it's
gonna
be
really
hard
to
write
guidance
until
we
know
what
the
middle
looks
like,
we
can
make
some
version
of
progress
on
collection
and
understand.
We
have
a
bunch
of
different
mechanisms
that
look
like
they're
moving
forward
on
collection.
We
have
some
mechanisms
that
look
like
they're
moving
forward
kind
of
on
sharing.
B
We
can
potentially,
in
my
mind,
set
some
reasonable
targets
around
that
and
then
there's
the
hard
work
that
needs
to
be
done
to
be
able
to
define
how
we're
going
to
talk
about
after
we
do
those
collections
in
different
formats,
how
we
create
guidance
and
repository
to
be
able
to
work
on
that
middle
part.
Does
that
sound
reasonable
to
people?
F
B
So
so
we
have
so
Hank
you're
doing.
Your
comment
was
that
you
know:
if
we're
gonna
do
the
yank
election,
we
need
to
really
have
yang
push
get
to
last
call.
So
if
that
comes
up
for
last
call
this
meeting,
then
you
know.
Presumably
we
can
have
something
a
little
firm
for
whatever
that
means
by
next
meeting
on
on
yank
election.
Does
that.
B
So
firm
means
that
if
you
have
working
group
last
call
on
Yang
push,
then
a
lot
of
the
ambiguity
comes
out
of
what
you
can
write
for
your
yang
push
collection,
and
that
means
that
say
bye
at
least
by
London.
We
can
have
something
that
we
can
really
argue
about
and
that
by
Montreal
we
could
potentially
anticipate
last
call.
Does
that
seem?
Does
that
timeline
seem
crazy,
I.
M
R
B
Think
so,
I'm
jumping
in
line
Hank
I
would
think
you
yeah
attestation
is
in
scope,
I
think
there's
a
nuance
between
when
you
collect
data.
That's
attestation
versus
just
data
collection.
Attestation
comes
with
some
kind
of
you
know,
kind
of
security
bound.
You
know
ability
to
assert
the
the
level
of
trust
in
that
data,
and
it's
not
clear
to
me
that
the
goal
isn't
to
be
able
to
collect
both
and
put
them
side-by-side.
M
Hi
this
is
Hank,
and
this
is
my
IOT
director
and
TCG.
There
is
a
head
on.
It
happens
to
be
that
I'm
kind
of
familiar
with
attestation
and
I
think
today
the
rats
email
list
will
start
it's
a
non-work
email
list
of
about
remote
attestation,
which,
for
example,
is
I,
think
they're
talking
about
here
and
not
at
the
station
and
I.
Think.
The
important
things
for
second
are
the
conveyance
of
evidence
and
not
attestation.
M
We
are
solely
in
charge
of
producing
a
thing
that
is
producer
processing
it
because
evidence
appraisal
by
a
verifier
in
a
remote
attestation
procedure
is
super
complex
and
it
would
be
I,
think
half
of
a
complexity
we
already
created
in
second
already
and
just
saying
yeah,
we're
going
to
also
do
that
is
a
huge
scope
creep
so,
but
be
very
careful
how
we,
how
much
verification,
that's
or
appraisal
we
are
very
thick
verify.
We
include
here.
H
I
think.
Another
thing
that
we're
trying
to
do
here
is
define
a
method
of
evaluation.
So
when
you've
collected
data,
you
want
to
be
able
to
figure
out
if
that
matches
some
expectation,
some
expectation,
and
we
we
have
some
ideas
on
how
to
do
that.
That
evaluation,
how
to
make
a
statement
of
what's
expected
and
had
to
compare
that
against
what
has
been
collected
and
then
I
think
we're
also
talking
about
you
know
providing
data-driven
guidance
we
like
to
use
terms
like
imperative
and
declarative
and
and
and
and
and
all
of
that.
H
But
what
I
think
we're
trying
to
do
is
where
we're
trying
to
both
describe
and
provide
the
the
data
that
is
necessary
to
have
a
data-driven
approach
to
to
collecting
the
posture
to
performing
the
evaluation
and-
and
so
those
are
the
pieces
that
I
think
we're
talking
about
here.
In
very
simple
terms,.
M
Hi,
this
is
Hank
I
think
here
we
could
maybe
used
to
actually
use
it
from
baseline.
So
maybe
second
provides
a
set
of
baseline
evaluation
capabilities,
and
then
you
can
build
on
that.
Okay,
I
heard
a
lot
of
ie.
No,
no,
no
here
next
to
me,
neutral
parties
and
just
the
thought.
Just
a
comment
also
I,
don't
know,
really
know
what
data-driven
means
so,
meaning
that
we.
H
H
H
M
Can
totally
understand
content-driven
if
you
define
information,
types
and
content
types,
enumerate
them
and
then
somehow
tell
you
how
they're
related
and
should
be
and
compared,
for
example,
I'm.
Absolutely
fine
with
that.
This
gives
structure
to
the
process,
but
but
just
saying
yeah,
you
should
be
free
to
do.
Compare
everything
with
everything
is
like
doing
nothing,
yeah,
exactly,
there's
a
problem
we
are
trying
to
solve.
So
if
you
have
a
list
of
information
types
and
content
types,
critics
economy-
maybe
with
that
and
then
say
this
type
of
information
typically
requires
this.
M
H
H
G
C
H
So
what
I
was
saying
before
I
was
proposing
that
we
that
we
can
use
Rolly
as
a
method
of
providing
that
that
content
and
that
we
work
on
drafts
to
describe
information
types
which
are,
you
know,
classes
of
content
that
that
can
be
provided
over
Roley,
and
that
may
be
at
some
point
in
the
future,
we'll
get
to
defining
the
form
of
of
those
specific
kinds
of
content.
But
that's
not
a
that,
doesn't
necessarily
have
to
be
a
priority
right
now.
B
B
M
I,
think
that
is
the
the
information
type
content,
type
of
correlation
we
just
were
talking
about.
This
seems
to
be
the
metadata.
You
just
mean
meant
yeah,
but
there
has
to
be
things
to
connect
in
order
to
draw
a
line
and
I
would
I
am
here
and
that
the
mic
I
wanted
to
highlight-
and
maybe
I
didn't
push
it
to
the
list,
because
I
was
tired.
I
pushed
the
first
yang
module
force
which
submitted
it
before
the
submission
event.
M
R
R
R
L
F
Okay,
good
I,
don't
know
how
to
say
it.
I
want
to
say
like
we
need
some
kind
of
like
wouldn't
want
it.
We
talked
about
the
different
transfer
mechanisms
right
like
what
I
think
the
ECP
I
think:
okay,
there's
you
know
we're
collecting
some
stuff.
It
gets
sent
up
to
a
server
and
then
ECP
I
think
actually
talks
about
storing
into
a
repository
right.
F
J
H
F
H
This
is
Dave
again
one
of
the
things
that
I
think
we
need
to
come
to
terms
with
in
in
this
draft
is
I,
think,
there's
a
role
for
guidance
to
be
exchanged
between
organizations
as
well
as
within
an
organization
and
the
the
transports
that
we
may
want
to
use
may
may
vary
based
on
on
those
those
needs,
as
well
as
the
discovery
mechanisms.
The
the
ways
of
you
know
initially
determining
where
to
go
to
get
those
versus
tracking
changes
that
are
occurring
within
repositories.
H
So
one
of
the
things
that
I've
thought
about
over
the
longer
term
is
you
know,
roley's,
really
great,
for
discovering
sort
of
an
initial
set
of
content
that
an
organization
may
provide,
but
it's
not
great
for
identifying
changes
that
are
occurring
as
they
occur.
You
know
within
within
that
repository,
and
so
maybe
there's
a
way
also
to
sort
of
connect
together,
Adam
pub
for
syndication
and
XMPP
for
notification.
As
part
of
that
I
think
Peter,
st.
H
J
F
B
So
chair
hat
on
I'd
like
to
continue
that
the
hackathons
provide
some,
you
know
drivable
moments
for
us
to
actually
resolve
issues
and
make
tangible
progress
that
informs
kind
of
you
know
draft
writing.
You
know
and
what
we
can
build
so
I
kind
of
agree
with
Adams,
keep
it
simple,
stupid
and
then
we'll
add
on
the
complexity
later.
F
G
H
Think
we
can
get
to
that.
This
is
Dave
I,
think
we
can
get
to
that
soon
and
I
think
we
actually
do
have
some
that
we
that
we
can
continue
to
work
on
like,
like
the
configuration
descriptor
work,
like
the
you
know,
like
some
of
the
ongoing
you
know,
Roley
work
that
work
that
we're
already
doing
I-I-I
think
we
should
also
maybe
add
some
calendar
milestones
for
doing
things.
As
far
as
more
more
hackathon
work
I'd
like
to
see
us
I'd
like
to
see
us
get
better,
though
about.
H
About
getting
knowledge
out
of
the
hackathon
work
and
I
don't
mean
to
take
us
from
The
Shining
Tower
to
the
pit
of
despair
here,
but
you
know
we
we're
doing
a
lot
of
interesting
things
in
the
hackathon,
but
I
don't
think
we're
doing
a
good
job
at
communicating
all
of
this
integration
work
that
we're
doing
across
all
of
these
different
groups.
You
know
to
in
a
way
that
that
folks,
that
are
not
intimately
involved
in
the
hackathon
can
actually
understand.
H
We
didn't
have
any
slides
at
this
meeting
to
talk
about
the
work
that
we
did
at
the
hackathon
and
that
would
have
been
a
really
great
way
to
actually
brief
people
about.
You
know
this
is
what
we
did,
and
these
were
the
successes
that
we
had,
and
these
were
the
problems
that
we
found,
and
these
are
the
things
that
we
think
should
be
worked
on
and
I
think
that's
a
missed
opportunity.
I
think
we
should.
We
should
find
ways
to
do
more
of
that.
So.
G
Maybe
we
can
get
it
resolved
shortly,
but
I
to
be
honest,
I
thought
this
would
be
like
a
five
minute
exercise
and
we
would
go
down
through
those
work
items
we're
like
okay,
this
maps,
to
this
this
maps
to
this
this
map.
So
this
may
be
done
and
now
we're
twenty
or
thirty
minutes
into
this,
and
we're
not
done
that's
kind
of
it
shouldn't
be
this
hard.
K
G
J
G
K
I
mean
if
we
can
get
the
milestones
to
me
by
Thursday
night
right
match
back
to
the
Charter,
and
it's
it's
real
clear
that
you
and
Chris
work
I
think
that's
going
to
help
the
rest
of
the
IETF
understand
what
sacrum
does
write
the
comments
you're
seeing
from
the
IAS
G
is
it's
a
little
nebulous?
They
don't
quite
get
it
all
right
and
it's
because
it's
a
space,
none
of
them
have
worked
in
so
I.
Think
it's
our
job
to
make
it
clear
to
them
the
importance
of
this
work
and
showing
the
concrete
deliverables.
K
G
Mean
I
would
like
to
see
us
I
would
like
it
to
remain
on
the
agenda
for
Friday
I'm,
not
willing
to
give
up
on
that
quite
yet
great
and
I'm
not
saying
that
the
milestones
that
we
have
to
define
have
to
fully
flesh
out
every
one
of
those
work
items.
But
there
needs
to
be
a
clear
mapping
between
these
abstract
work
items
that
we've
defined
and
concrete
milestones
right
may
I'm,
not
all
of
them,
but
at
least.
K
G
K
G
M
This
is
Hank
yeah,
just
connecting
to
the
requested
from
Benoit
I
think
that
it's
kind
of
easy,
because
this
is
the
work
I,
think
we
have
the
most
concrete
progress
made
in
and
Eric
I
think
you
would
agree
with
that.
So
this
is
very
visible
and
then
addressing
Dave's
comment.
First
of
all,
I
we
have
hackathon
slides.
M
We
could
have
just
pulled
them
up.
Second
of
all,
I
was
at
the
LCF
meeting
on
Friday
ended
a
PR
interrelationship
presentation
of
en
coche
in
our
work.
In
second
year
we
have
slides
net
conf
that
highlight
the
hackathon
output.
We
have
slides
and
core
that
elaborate
on
how
this
will
be
done
in
kamae.
We
have
slides
in
this
interesting
research.
B
G
G
H
H
G
H
G
All
right
so
so
I
think
we've
had
a
fair
conversation
on
this
I
think
we
need
to
continue
it
on
the
list.
Chris
is
busy
over
here
typing
up
a
list
of
potential
milestones
and
I'm
sure
he
will
email
it
to
all
of
us
shortly
and
we
can
discuss
on
the
list,
but
I
mean
they
don't
have
to
be
specific
draft
names,
but
they
need
to
be
a
you
know
like
four
or
five
or
six
words.
It
says
we're
going
to
do
something:
we're
going
to
do
a
document
that
addresses
this.