►
From YouTube: CNCF Serverless WG Meeting - 2018-04-05
Description
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
E
B
B
B
A
B
A
B
It's
interesting
watching
these
people's
things
bounce
around
the
participant
list,
trying
to
keep
me
or.
E
B
K
B
B
A
Yeah
Shawn:
are
you
there
Shawn
I'm
here
excellent?
Thank
you.
Let's
see,
did
it
Clemens
and
Andrews
joins
uh-huh
Clemens.
Thank.
B
B
Okay
did
I
miss
anybody.
Did
it
do?
B
B
M
B
Case
cool,
let's
get
started,
I
assume
you
guys
can
see
my
screen.
Thank
you
awesome
all
right.
Let's
get
this
out
the
way.
It's
blocking
me
all
right.
So,
first
up
on
the
agenda
reminder
for
coop
con
as
of
right.
Now
we
are
still
planning
into
being
an
official
meeting
or
the
face-to-face
portion
of
our
event.
There
will
be
an
official
meeting
soon,
assuming
we
can
get
a
dial
on,
which
would
probably
be
the
Xoom
call
so
going
Wireless.
B
We
should
be
good
to
go
just
a
quick
question
and
in
a
previous
phone
call,
people
asked
if
they
could
get
to
the
face-to-face
meeting
without
actually
buying
tickets
to
the
conference
itself,
we
were
wondering:
is
there
actually
anybody
who's
planning
on
going
to
the
face-to-face
meeting?
Who
is
not
going
to
be
attending
the
conference.
B
G
B
B
Guess
it's
nothing
really
much
more
say
that
we
will
talk
about
later
on
the
call
you
know
what
we
want
to
do
with
that
I
just
want
to.
Let
you
guys
know
that
we
do
have
the
boss
session
set
up,
which
is
all
good.
So
before
we
get
into
some
of
the
other
meteor
discussions,
I
did
want
to
address
Kathy's
PR
that
was
supposed
to
have
been
talked
about
last
week,
but
unfortunately
I
messed
up
and
dropped
it
from
the
agenda
by
mistake.
B
B
B
It's
a
bill
referred
to
it
as
an
MVP,
but
whichever
way
you
want
to
think
of
it,
we
decided
on
this
week's
called
to
discuss
that
as
of
right
now,
I,
don't
believe,
I've
seen
anybody
comment
on
any
issue
or
PR
saying
they
wanted
included
aside
from
the
one
that
I
tagged,
which
is
to
get
an
HTTP
mapping
hosel
out
there
I'm
sorry
with
this.
One
want
to
comment.
I
agree,
okay,
so
so,
let's,
let's
start
with
the
one
that
was
tagged
as
wanting
to
be
quitted.
B
G
Yeah,
so
I
actually
made
progress
and
I
have
pretty
much
an
IETF
RFC
style
documented
about
this
now
and
I'm
gonna
make
available
later
today.
I
just
didn't
get
travel,
but
it's
in
the
spirit.
It
looks
kind
of
similar
to
the
strong
men
here
where
it's
not
exactly
the
same.
Some
variations,
because
I
thought
about
a
little
bit
a
little
bit
more,
but
this
is
effectively
taking
our
properties
that
we
have
and
creates
a
HTTP
mapping
for
them.
G
Where
I
have
two
models,
one
is
the
binary
model,
the
other
one
I
now
called
the
structured
model
where
you
can
take
cloud
event
that
has
data,
and
so
the
first
one
the
binary
one
is
if
the
data
is
something
from
let's
an
IOT
device
that
is
in
some
proprietary
format
that
you
simply
want
to
go
and
describe
with
metadata
but
then
route
on.
As
is
that's
what
the
binary
format
is
for.
So
all
the
for
the
HTTP
projection.
G
All
the
called
events
properties
go
into
HTTP
headers
I've
been
toying
around
with
there's
two
options
here.
One
is
to
go
and
make
a
single
header
and
then
put
Jason
into
the
adjacent
value
inside
that
header.
That
has
the
problem
that
that
header
value
may
run
very
long
and
then
may
run
into
very
practical
limitations
of
HTTP
servers.
Also,
people
may
want
to
use
do
some
filtering
or
anything
up
in
terms
of
logic,
on
HTTP
server,
based
on
the
values
that
the
header
values,
without
forcing
them
to
parse
Jason.
G
That's
why
I
exploded
them
into
headers.
The
other
one
is
a
more
compact
one
or
your
friendly
win
for
routing,
where
you
have
a
Jason
protection
in
this
case
for
the
event
properties
and
the
Jason.
What
I
chose
to
do
and
what
you'll
see
today
when
I
go
and
put
that
PR
in
is
actually
submit
two
documents:
one
is
the
HTTP
mapping,
the
other
one
is
adjacent
event
format.
The
Jason
event
format
basically
codifies
what
I
have
here,
and
obviously
this
is
based
on
an
earlier
graph
that
we
have,
which
effect
you
says.
G
This
is
how
you
take
these
events
that
we
have
so
far
defined
and
how
you
put
them
into
how
you,
first
of
all,
define
them
individually
or
Express
them
individually
as
Jason
values,
and
then
how
you
put
an
envelope
around
them
and
also
how
you
specifically
treat
the
data
property.
If
the
data
property
contains
Jason,
then
it's
going
to
be
projected
as
it
is
here.
If
the
data
property
contains
something
that's
not
Jason,
it
will
be
basics
before
encoded
and
then
the
content
type
will
describe
what's
in
there.
G
That's
also
addressing
a
question,
an
issue
question
that
dog
you
raised
this
week
on
how
data
should
be
represented,
so
I
have
the
so
that's
the
Jason
event.
Mapping
I
have
that
almost
done,
and
then
the
HCP
mapping
bassy
picks
that
up
and
then
uses
that
mapping
to
do
what
I
call
the
structured
mode
so
you'll
be
able
to
go
and
read
this
sometime
at
the
end
of
the
day.
G
The
reason
why
I
split
this
up
in
this
way
is
first,
we
certainly
want
down
the
road
we
want
to
go
and
define
transport
mappings
for
A&T
P&M
TTT.
It
might
actually
go
and
immediately
following
through
a
straw
man
for
one
of
the
two
just
make
sure
that
the
HP
mapping
doesn't
have
any
weird
HTTP
isms
in
it
or
sorry
that
that
the
the
that
we
know
that
these
can
be
mapped
to
two
transports
and
then
for
the
for
the
Jason
event.
G
G
F
All
right
to
chime
in
on
Doug's
question,
Clemens
I
think
you
did
a
great
job
here.
These
simple
examples
are
worth
you
know,
20
pages
of
documentation,
I
think
we
should
absolutely
make
shipping
these
examples
with
the
first
version
of
the
specification
mandatory
because
they
explain
everything
in
a
very
simple
manner.
B
G
If
you,
if
you
a
dog,
can
you
give
me
the
screen
for
a
second
yeah.
G
Just
just
for
you
to
get
a
feel
of:
what's
there,
okay
see?
Yes,
okay,
so
so
this
is
the
HTTP
transport
binding
and
I,
basically
have
so
I,
just
let
you
and
just
scroll
through.
So
you
have
a
sense
of
what's
there
I
defined
the
relationship
to
http.
What
I'm
not
doing
is
I'm
not
prescribing
any
use
of
any
methods
or
how
the
URS
should
be
shaped.
I'm,
basically
just
making
this
a
mapping
to
http
messages.
This
can
be
met
to
the
request.
G
This
can
be
mapped
to
response
and
I'm,
not
making
any
hard
constraints
about
whether
you
should
go
and
send
an
event
with
a
foot
or
with
a
post.
Whether
you
can
you
can
return
an
event
from
from
a
gets.
It's
just
really
how
the
the
event
is
exploded
on
to
the
HTTP
message
and
then,
if
we
want
to
go
and
create
a
further
respect
that
defines
how
a
web
app
functions.
G
We
can
do
that,
but
this
thing
here
kind
of
composes
with
anything
you
want
to
do
with
HTTP
with
events,
whether
you
want
to
send
them
where
we
want
to
solicit
them,
and
this
year
describes
what
I
just
explained
with
the
structured
in
the
binary
content
modes
kind
of
the
overview
talks
about
the
event
formats.
Then
references,
the
JSON
format.
The
JSON
format
is
started
here.
G
G
Minute,
okay
go
ahead
and
then
I
define
how
we're
using
the
properties
I'm
not
calling
them
all
out,
because
I
want
to
keep
that
that
thing
flexible
for
all
the
extensions
and
changes
we
want
to
make
sure
I
just
call
out
what
I
need
to
call
out
and
then
I
define
the
message,
mapping
and
then
at
down.
The
bottom
I
have
affecting
the
examples
that
you
just
saw
in
the
in
the
issue.
G
B
B
G
So,
based
on
what
I
just
said,
I
think
we
need
to
define
the
type
system
and
I
can
basically
what
I'm
thinking
is
the
the
text
that
I
wrote
for
the
Jason
mapping.
I
want
to
have
that
sitting
in
the
in
the
base
bag,
so
I'm
going
to
break
that
out
into
the
PR
and
the
the
type
system
really
is
like
we're
using
string
and
your
I
and
timestamp
and
map,
and
then
we
have.
We
have
beta
as
loosely
defined
as
arbitrary
data
and
I
think
that
should
be
object
and
so
I.
G
B
I
G
Of
this
is
how
you
take
it
cloud
events,
events
and
send
it
over
HTTP.
That
is
the
scope
of
that
that
one
specification
and
I'm
writing,
and
then
there
can
be
other
specifications
that
say:
here's
how
you
take
a
cloud
events
event
and
about
that
over
mqtt,
and
this
is
how
this
is,
how
you
do
it
over
AMQP.
A
I
B
E
G
B
So
let
me
get
my
circle
back
around
then.
Do
we
conditionally
approve
Clemens
idea
for
defining
our
type
system
and
obviously,
as
I
said,
I
hear
in
the
notes
we
can
revisit
that
decision
if
it
ends
up
being
something
that's
beyond
what
we
want
to
do
for
0.1
but
conditionally
approve
it
is.
Is
there
any
objection
to
that.
K
G
K
What
I
posted
in
the
chat
like
I,
noticed
in
this
examples
that
it's
referencing
some
ideas
that
you
wanted
to
promote
like,
instead
of
this
source?
Being
your
URL,
where
I
like
having
topics
and
subject
to
stuff
like
they're,
all
good
to
just
include
examples
to
what
we
already
agreed
and
not
like
zero
two
in
there
and
then
we
release
it
or
one.
The
the.
G
B
Right
so,
let's
put
it
this
way.
Is
there
any
disagreement,
then,
with
the
idea
of
just
making
sure
that
whatever
form
of
serialization
stuff,
we
talk
about
whether
it's
Clemens
PR,
the
existing
document
we
have
in
there?
We
need
to
make
sure
those
are
all
in
sync
by
0.1.
Yes,
any
disagree
with
that.
F
Quick
comment
on
this
Doug
we
have
in
our
roadmap.
We
have
almost
a
kind
of
a
list
of
criteria
that
we
like
to
ship
with
0.1
Clemens,
just
added
a
few
things
for
including
specifications
for
mapping
cloud
events
to
HTTP
and
for
mapping
cloud
events
to
JSON
and
defining
a
type
system.
Can
we
add
these
into
our
0.1
roadmap
I.
F
F
G
And
it's
yeah,
it's
really!
It's
really
just
that,
because
we
are
using
a
type
system
already.
I
just
want
to
go
and
have
a
declaration
section
of
fries
that
says
this
is
these:
are
the
values
you
can
use?
So
if
we
we
currently
have
like
event
type,
an
event
type
is
a
string,
but
nothing
says
what
the
string
is.
So
the
point
is
I
just
want
to
have
a
preamble
that
says:
here's
a
set
of
here's,
a
set
of
types
that
the
following
properties
are
using.
B
B
N
G
N
More
so
talking
about
the
I'm,
not
talking
about
the
specific
statement
of
like
a
malicious
middleware,
that's
reaching
in
and
updating
fields
on
the
payload
more
just
that
we
can
verify
specifically
that
no
bits
have
been
mangled
along
the
way
or
if
we
want
to
try
and
guard
against
something
like
malicious
middleware,
I
think
they're,
separate
use
cases,
but
just
curious
if
they've
been
talked
about
no.
G
N
B
Okay,
I
think
deciding
whether
it's
in
or
out
of
scope
is
probably
a
good
topic
for
later
on,
when
I
think,
when
I
think
people
have
a
chance
to
think
about
it.
But
let
me
I
said:
let
me
ask
more
focus
question.
Is
there
anybody
on
the
call
who
believes
that
dealing
with
integrity
should
be
a
mandatory
40.1.
B
N
B
E
F
F
G
G
I
have
a
I
call.
These
I
call
the
protocol
I'll
call
the
transport
bindings
and
format
so
I
have
currently
I
have
HTTP
transport,
binding
and
then
I
think
we're
gonna
have
an
MPD
transport
binding
and
an
MPP
transport
binding
and
then
I
have
a
JSON
format
and
I
think
we're
gonna
have
a
message
back
for
a
while:
okay.
B
All
right,
in
that
case,
I,
believe
we're
done
relatively
short
list
of
my
opinion,
so
hopefully,
Clemens
PR
will
be
out
there
today.
Please
review
it
as
always:
it'd
be
nice.
If
we
can
get
that
in
there
relatively
soon,
because
the
next
thing
I
want
to
talk
about
is
what
are
we
going
to
do
for
coop
Khan
and
this
our
scene
CF
con
in
particular
I'm
interested
in
knowing?
B
B
F
Yeah
one
quick
comment:
I'm
not
sure,
is
anyone
else.
Speaking
about
this
at
cloud
native
gaunt,.
F
F
How
we
do
that
I'm
open
to
all
suggestions.
You
know
some
some
goals,
I
was
kind
of
thinking
about
is,
would
love
to
you
know,
show
off
something,
that's
pretty
exciting
and
would
love
to
include
a
lot
of
the
people
and
their
organizations
and
their
technologies
in
this
Interop
demo,
although
that
could
get
very
complicated
but
I'm
open
to
that
I
think
it
would
be
very
cool
if
we
almost
had
you
know
a
nice
group
showing
of
all
of
all
of
all
of
us
working
together
doing
some
type
of
interoperability,
demo
and
I.
F
Don't
know
what
that
looks
like,
but
I'd
love
for
that
to
be
kind
of
the
basis
of
the
talk
and
open
to
even
having
other
people
come
up
on
stage
and
talk
a
bit
about
it,
but
that
you
know
we
only
have
limited
time
so
I'm
not
sure
how
how
well
that'll
scale,
but
anyway,
I
just
want
to
put
that
out
there.
So
everyone
knows
and
I'm
all
in
favor
of
discussing
this
further
on
separate
calls.
Yeah.
B
B
I
B
Okay,
so
I'm,
assuming
during
that
call
the
little
more
concreteness
will
appear
around
what
we're
doing
around
an
interrupt
type
of
event
and
related
to
Austin's
talk.
Are
there
other
things
we
should
be
looking
at
doing
so,
for
example,
we
have
a
bee.
Oh
we
have
a
bird
of
a
feather
session.
Is
there
anything
we'd
like
to
do
in
preparation
for
that?
So,
for
example,
I
was
thinking.
We
maybe
should
have
a
potential
list
of
topics
or
questions
that
we
could
ask
ourselves
that
people
in
the
audience.
B
B
Usually
it's
more
people
come
and
ask
questions,
but
if
they
don't
have,
if
the
audience
members
don't
have
specific
questions
of
the
people
running,
the
show
then
usually
they'll
have
a
list
of
questions
to
help
prompt
people
to
ask
questions
or
to
just
provide
some
overview
of.
What's
going
on
with
that
piece
of
work,
at
least
that's
been
my
experience,
maybe
other
do
all
have
had
different
experiences,
so
maybe
else
want
to
speak
up
yeah.
J
That's
been
spend
about
my
experience
as
well
as
sometimes
the
presenter
will
yeah
just
just
give
the
shortest
of
of
intros
or
maybe
have
or
maybe
walk
people
through
a
couple
of
slides,
but
in
what
was
the
intention
of
stopping
and
having
a
relatively
lengthy
discussion
at
any
point
in
the
slides
and
really
really
trying
to
facilitate
discussion
more
the
slides
being
there
to
facilitate
the
discussion.
Whether
in
the
house,
yeah.
B
B
We
can
maybe
talk
to
that
for
people
who
are
completely
clueless
as
to
the
exciting
work
that
we're
doing
here,
but
then
beyond
that,
maybe
just
have
a
list
of
potential
list
of
topics
to
bring
up
for
the
audience
editors
to
get
more
information
about
it,
but
Austin
I
would
think.
If
you
actually
have
a
formal
talk,
we
may
be
able
to
steal
some
of
your
slides
right
for
that
kind
of
thing
to
provide
an
overview.
F
E
F
B
If,
if
the
audience
members
aren't
as
engaged
as
we'd
like
them
to
be
just
to
keep
the
conversation
going
and
then
beyond
that,
we
can
look
at
what
you
come
up
with
for
your
talk
and
see
whether
we
can
steal
that
as
a
quick
intro
and
if
not
then
yeah,
we
can
create
a
our
own
version
of
a
short
intro
just
for
the
birds
of
a
feather,
if
necessary,
but
I
think
we
have
time
to
decide
that
later.
That's
okay
of
people.
I
F
F
B
J
B
F
F
However,
at
the
same
time,
I've
been
talking
to
a
lot
of
people
in
this
working
group
and
a
lot
of
them
as
long
as
we
can
get
this
release
out
and
they're
comfortable
with
it
feel
comfortable
with
their
companies
kind
of
being
showcased
as
being
associated
with
this
and
I.
Don't
know
if
there's
any
other
way
to
do
it,
then,
for
me
to
just
go,
I
guess
check
in
with
them
and
say:
hey.
Are
you
okay?
F
F
B
B
B
Okay,
so
the
last
topic
I
had
in
this
section
was
the
face
to
face
meeting
itself
it.
What
were
people's
thoughts
in
this
space,
where
people
assuming
it
was
just
a
regular
face
to
face
a
regular
working
group
meeting
and
we're
just
gonna
go
through
the
list
of
topics
like
we
do
normally
are
they're
very
specific
things.
People
would
like
to
do
there
because
we're
at
a
conference.
B
F
I
think
the
the
goal
I
have
in
mind
for
the
talk
I'm
gonna,
give
is
to
show
this
to
the
world,
show
this
to
new
people,
and
you
know
excite
them
as
to
its
potential
by
showing
them
some
some
examples
of
interoperability.
I.
Imagine
that
the
face-to-face
meeting
was
just
for
us
to
talk
about.
You
know
just
to
meet
up
with
each
other,
get
to
know
each
other
and
perhaps
probably
discussed
what
0.2
or
two
point
looks
like
yeah.
G
I,
though,
face-to-face
meetings
very
productive
for
just
figuring
out
what
the
future
will
hold-
okay,
I'm
Rob-
that,
rather
than
going
through
the
process
of
you,
know
checking
off
list
item
items
from
the
list,
but
rather
just
go
and
do
the
it's
really
hard
in
the
format
that
we
have
to
go
and
speculate
about
things
or
throw
ideas
around,
but
I
think
in
the
face-to-face.
It's
going
to
be
easier,
so
I
would
I
would
probably
do
scoping
discussions
and
then
see
in
the
scope
that
we
kind
of
figure
out
together.
B
Okay,
so
it
sounds
like
what
I'm
hearing
so
far
is.
We
would
like
to
keep
the
face-to-face
meeting
to
be
more
of
sort
of
an
internal
discussion.
Obviously,
other
people
are
free
to
join,
but
it's
mainly
for
us
in
terms
of
planning
purposes
or
design
discussions
or
anything
else
related
to
our
work
itself.
It's
not
there
necessarily
to
to
showcase
our
work
to
other
people.
It's
for
us
to
actually
get
some
work
done.
Is
that
fair?
Is
there
any
disagreement
with
head
in
that
direction?.
B
F
F
D
B
Pr
reviews
a
couple
of
hopefully
easy
ones:
someone
just
opened
up
today.
It
should
be
used
instead
of
use
in
this
sentence.
Here,
everybody
just
take
a
quick
look
at
that.
I
was
gonna,
prove
it,
but
I
thought
it
would
wait
for
at
least
one
more
LG
TM
that
I
never
got
so
I
decided
to
force
it
on
you
guys
right
now.
B
All
right,
thank
you,
guys
very
much,
oops.
Okay,
this
next
one
was
syncing
up
our
Jason
with
some
of
the
PRS
that
we
resolved
last
week.
Unfortunately,
when
I
first
added
the
Jason
serialization,
there
were
just
some
fields:
I,
just
flat-out,
missed
like
event
time
kind
of
type
I
have
no
idea
how
I
miss
that,
but
it
was
just
additions.
The
biggest
change
was
changing
source
to
be
a
URI
and
getting
rid
of
namespace.
B
B
Some
of
it
is
just
this
doesn't
apply
anymore
and
given
all
the
other
stuff
that
we
have
related
to
this
between
our
milestone
documents,
our
issue
list
and
stuff
like
that,
I
just
don't
think
we
need
the
section
anymore,
so
I'm
proposing
that
we
remove
it
any
questions
on
that
and
particulars
ends.
It
also
talked
about
the
scope.
We
already
have
a
scoping
section
in
the
spec,
so
we
don't
need
to
duplicate
it.
B
So
I
was
going
to
give
everybody
told
next
week's
call
to
to
see
there's
anything
in
the
list
they'd
like
to
include,
and
if
so,
open
up
an
issue.
Otherwise
I'd
like
to
propose
next
week
they
were
remove
this
section
I'm
not,
and
you
clear,
I'm,
again
I'm,
not
asking
people
to
vote
on
this
today.
But
there
are
there
any
questions
on
this.
Is
there
any
objections
were
heading
down
that
path.
B
F
Okay,
oh
I,
think
these
are
important
ones.
I
wish
I
wish
they
were
on
the
call
today,
but
I
think
everyone
should
go
and
check
this
out.
So
I
think
these
are
some
of
the
more
controversial
pieces
of
0.1,
so
everyone
can
weigh
out
weigh
in
on
these.
That
would
be
super
helpful.
Probably
the
most
helpful
thing
you
could
do
to
move
this
forward,
I
think
and
so.
G
I
think
the
the
the
first
one,
adding
names
basic
to
the
event
type
is
will
be
very
helpful.
The
other
one
is
kind
of
sparking
a
discussion
about
what
are
we
going
to
do
with
custom
mix
with
custom
headers
and
that's
something
that
might
be
a
good
topic
to
discuss
face
to
face
right
now.
Extensions
are
very
good
for
that,
because
I
can
I
can
see
very
many
different
contexts
in
which
you
want
to
have
extensibility,
and
so
therefore,
I
don't
want
to
have
ten
buckets
of
of
custom.
B
So
the
message
just
a
quick
question
and
I
don't
like
because
Thomas
isn't
on
the
call
I,
don't
think
it's
fairness.
They
have
a
deep
discussion,
but
I
am
kind
of
curious
just
to
get
a
sense
of
from
the
group
on
the
namespace
or
adding
namespace
to
the
event
type.
How
do
people
feel
in
general
about
heading
that
direction,
ignoring
some
of
the
the
nitpicky
details
about
some
us
versus
may
and
all
that
other
stuff,
just
in
general,
adding
a
namespace
to
the
event
type?
K
Think
nobody
happy
with
his
source
as
a
URI
and
if
nobody's
happy
and-
and
there
is
an
alternative-
why
not
I
was
that?
How
do
you
feel
that
nobody's
happy
with
sources?
You're
I
know
just
from
the
comments
I
saw
and
like
everybody
coming,
but
at
least
there
are
two
known
proposals:
how
to
change
from
using
sources
your
eyes
like?
Well,
we
just
accepted
that
last
week.
Didn't
we
yeah
I,
know
but
I
know
I'm,
that's
my
impression
that
nobody's
really
happy
and
we
are-
we
are
excited
about
it.
K
B
Love
it
okay,
anyway,
I
think
I've
got
my
answer.
People
seem
to
be
generally
okay
with
the
namespace
in
the
event
type.
We
just
need
to
work
on
the
actual
wording.
Wording,
which
is
fine
and
I,
know
the
other
one's
a
little
more
controversial.
So
I
won't
ask
about
that
one.
But
before
we
move
on
just
reiterate
what
Austin
was
saying,
please
review
those.
So
we
can
have
some
good
discussions,
preferably
offline,
if
possible,
and
then
maybe
even
resolve
them.
Next
week,
yeah.
F
One
other
note
on
the
source:
labels
topic,
I,
appreciate,
Thomas's
kind
of
thought,
leadership
here
and
the
problems
that
this
could
potentially
solve.
I'm
I'm
kind
of
okay,
with
passing
on
it
40.1
if
necessary.
Just
because
I
do
feel
like
this
is
something
that
we're
kind
of
adding
at
the
time,
and
we
could.
We
could
potentially
add
it
later
well,.
B
F
G
B
F
So
especially
because
Thomas
and
our
Google
colleagues
will
will
hopefully
be
on
that
call,
but
in
general,
just
as
we
approach
each
thing
right
now,
we've
got
20
like
26
days
or
something
to
finalize
the
specification
build
some
type
of
demo
based
off
of
it.
So
you
know
each
each
thing
we're
looking
at
right
now
we
really
need
to
I
believe
raise
the
question.
Do
we
actually
need
to
focus
on
this
right
now?
Yep,
okay,.
B
Yes,
yep
definitely,
okay!
Next
one,
this
one
Leo
Leo
put
up
a
pull
request
against
the
working
group,
not
the
cloud
event
spec,
but
the
main
working
group
thing
dad
a
non
goal.
As
identified
I'm.
Sorry,
the
non
goal
is
identified.
One
service
project
to
rule
them
all
just
said
I
think
may
have
been
originally
added
as
a
joke,
but
it
may
actually
be
a
good
thing
to
include
in
there
now
before
I
just
approved
it
I
want
to
get
people
to
take
on.
It
was.
B
J
A
B
Okay,
any
objection
done.
Okay.
Now,
unfortunately,
we
don't
have
very
much
time
left
side.
I,
don't
think
we're
gonna
be
able
to
get
into
some
of
these
other
ones
that
are
kind
of
meaty.
But
please
do
take
a
look
at
those
PRS
and
get
some
discussions
going
out
there.
We
trying
to
do
as
much
work
offline
as
we
can
before
I
circle
back
around
and
do
attendance.
Are
there
any
other
topics
that
are
very
quick?
It
would
like
to
bring
up.
B
B
B
Excellent
okay,
my
case,
please
remember
to
go,
fill
out
mark
Peaks,
doodle
poll
for
the
offline
meeting
to
discuss
in
a
event
we're
hoping
to
get
that
going
relatively
quickly,
so
I
think
he's
hoping
to
get
something
going
early
next
week
and
with
that
I
believe
we're
done.
Thank
you
guys
very
much
I
appreciate
it.
We
got
a
lot
done
today,
amazing
yeah,
very
good.
Thank
you.