►
From YouTube: CNCF Serverless WG Meeting - 2018-10-4
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
B
A
Thank
you
all
right,
I
want
to
go
and
get
started.
We
have
a
fairly
full
agenda
today.
Let's
see
actually
I,
don't
think,
is
anything
too
exciting
there
too
negative
in
yellow
other
than
Austin.
Were
you
gonna
set
up
another
SDK
caller
we're
gonna
play
that
by
ear,
I
could
not
remember
what
you
decided.
We.
C
Are
going
to
set
up
another
one
because
we
have
to
work
through
some
versioning
issues.
The
only
thing
I'm
kind
of
waiting
on
is
I
think
our
team
is
actually
gonna
submit
a
first
pass
at
the
JavaScript
SDK
shortly
here
and
I
just
want
to
see
what
learnings
we
get
as
a
result
of
that
before
jumping
on
the
next
call,
I
think
we'll
be
in
a
more
informed
position.
A
You
could
submit
be
ours
yeah.
Actually,
since
you
kind
of
jumped
into
the
essay
discussion,
let's
do
that
right
now.
I
did
create
the
github
repositories,
as
you
said,
except
for
the
goaline
one.
That's
actually
gonna
get
transferred
over
from
VMware
and
they're
still
working
on
the
legal
process
or
what
you
want
to
call
it
to
get
a
transfer
door.
However,
I
have
yet
been
given
the
github
IDs
of
people
who
want
to
be
made
maintainer
zorad
vendors,
or
you
want
to
call
it
on
those
repos.
A
So
I
still
need
people
to
send
me
a
note
or
ping
me
through
slack
telling
me
who
they
want
to
add
it
as
admins,
because
right
now,
no
one's
to
do
anything
with
PRS,
except
for
me
or
the
admin
or
the
other
admins
of
the
group.
So
I
just
need
some
github
IDs.
If
you
guys
want
to
send
those
to
me,
okay
I've
got.
A
A
A
D
A
Okay,
thank
you
much
all
right.
Next,
the
interrupt
demo
so
I
did
some
that
notes
to
people
talking
about
the
results
of
our
meeting
that
we
had
last
week
talked
about
ideas
around
that
the
next
Interop
event.
There
was
a
document
out
there
linked
in
the
agenda
right
here
right
now,
you
seem
to
be
heading
towards
some
two
different
possible
applications.
One
involves
human
language
translation
so,
for
example,
an
English
sentence
might
get
passed
through
a
bunch
of
different
nodes,
and
each
one's
translates
to
a
different
language
than
by
the
atomic.
A
Full
circle
is
transit
back
to
English
and
we
get
to
see
how
our
systems
butchered
it
as
they
went
for
one
language
to
another.
The
other
option
is
to
do
sort
of
a
mad,
libs
kind
of
a
thing
where
there's
a
sentence
with
gaps
in
it
where,
where
we
say
okay
verb
goes
here
and
now
goes
here,
an
adjective
goes
here
and
then
each
node
that's
participating.
The
demo
fills
in
part
of
the
sentence
and
we
get
to
see
you
know
what
kind
of
funny
symptoms
it
produces
in
order
to
do
either.
A
One
of
those
we
kind
of
need
to
know
a
couple
of
things,
I
should
also
mention.
We
talked
about
potentially
also
leveraging
different
transports
between
the
various
nodes,
just
to
sort
of
show
some
mobility
around
that
as
well,
but
in
order
to
decide
which
way
to
go
on
these
things,
I
think
the
biggest
decision
point
here
is
actually
things
like:
what
are
the
transports?
A
Can't
people
actually
support
in
time
for,
say:
coop
can't
see
at
all
which
companies
can
support
things
like
in
an
English
translation
or
just
language
translation,
those
kind
of
things,
because,
obviously,
if
you
only
have
one
company
that
can
do
one
particular
transport,
you're,
not
gonna,
get
much
of
a
neuro
ability
statement
there
right.
So
what
I
really
need
from
you
guys
is
to
fill
out
in
this
dock.
These
two
questions:
what
transports
can
you
support
in
time
and
do
you
support
language
translation?
Those
will
help
us
decide
whether
demo
is
going
to
look
like.
A
So
please
when
you
get
a
chance.
Let
us
know
what
your
company
can't
actually
support
and
now
help
us
make
a
better
decision,
because
you
want
to
make
it
as
inclusive
as
possible.
So,
for
example,
if
no
one,
except
for
three
people
support
language
translations,
we
may
not
go
that
path
right.
They
have
to
do
a
mad
libs
kind
of
a
thing
which
I
think
is
something
that
everybody
probably
could
support.
Relatives
would
easily
it's
just
randomly
picking
a
word
from
a
list,
but
anyway,
think
about
that.
A
C
Quick
question
Doug
mm-hmm
can
we
submit,
we
could
probably
still
go
submit
our
own
individual
talks
to
coop
con
right
and
if
because
our
company
has
service
for
most
version,
two
coming
out
of
which
cloud
events
is
kind
of
a
premier
concept
and
we've
got
some
really
cool
interoperability.
Stuff
just
show
off,
but
I
think
it's
best
if
it
probably
just
go,
submit
our
own.
Our
own
talk
for
that
is
this
well
I.
Guess
that's
just
standard
process
right.
We
can
go,
submit
separate
talks
for
this.
Oh
yeah.
A
They're
doing
having
reserved
slot
so
we
will
have
an
intro
and
a
deep
dive
session.
Yes,
so
if,
for
example,
well
so,
let's
take
example,
the
buffer
was
it
the
Europe
of
wanna
come
what
city
we
were
in
where
you
did
that
demo
Austin.
If
someone
wants
to
do
that
interrupt
demo
as
part
of
a
talk
that
they
had
accepted,
I
think
that's
great.
If
that
does
not
happen,
then
we
can
leverage
our
intro
and
or
deep
dive
sessions
to
show
the
interrupts
that
well,
so
we
do
have
it.
A
A
All
right
next
is
okay,
Shanghai,
so
trying
to
member
now
I,
don't
think
made
a
phone
call.
Yet
we
do
have
one
I
believe
right
after
this
phone
call
today
to
talk
about.
In
essence,
this
list
of
topics
that
people
put
down
for
potential
discussion
points
at
the
intro
and
face
to
face
for
our
workgroup
or
yeah
I
guess
work
group
in
Shanghai.
A
If
you're
interested
in
participating
in
this
discussion,
it
will
be
I
think
on
this
soon
call
right
after
this
one
you
feel
free
to
join,
even
if
you're
not
going
to
be
there,
but
we're
gonna
be
talking
about
what
topics
to
discuss
who's
gonna
volunteer
for
talking
about
each
topic.
You,
the
presentation.
We
only
have
35
minutes
per
in
per
session,
meaning
35
for
the
intro
35
for
the
deep
dive.
So
we
don't
have
a
whole
lot
of
time,
but
we're
gonna
be
talking
about
that
on
the
call
right
after
this
one.
A
A
Okay,
this
one
make
you
guys
aware
of:
what's
going
on
there
in
case
you
want
to
participate
next,
let's
see
if
we
can
get
through
some
easy
PR.
First
Before
we
jump
into
some
really
deep
ones.
So
last
week's
call
Kathy
had
a
PR
with
a
relatively
minor
change,
but
based
upon
a
either
miscommunication
or
I
just
misread,
or
some
like
that.
A
I
thought
that
there
may
have
been
a
possibility
that
other
people
may
have
misinterpreted
Kathy's
intent,
because
the
subject
of
her
PR
actually
didn't
say
a
whole
lot
of
them
and
she
was
just
trying
to
make
a
change
and
so
I
think
other
people
may
have
done
what
I
did
is
misinterpreted
where
she
was
going
with
it.
So
she
and
I
did
a
little
bit
talking.
A
We
came
back
with
some
slight
changes
and
I
want
to
make
sure
you
guys
are
okay
with
it,
since
this
is
a
change,
the
governance
talk
and
that's
kind
of
a
touchy
stuff
check.
Sometimes
I
don't
want
to
to
approve
anything
without
you
guys
getting
another
chance
to
see
it.
So
there
were
two
changes.
The
Katherine
I
talked
about.
A
First
was
to
add
the
phrase
in
aggregate
to
make
it
try
to
make
it
clear
that
when
we
take
attendance,
it's
either
the
primary
or
alternate
person
can
be
there
for
three
out
of
four
meetings
it
so
it's
in.
If
neither
one
of
them
is
there,
for
example.
Today,
then
that
company
gets
credited
for
being
there,
it
that's
the
first
one.
A
The
second
one
was
calculi
to
make
sure
that
she
can
basically
change
who
the
primary
and
alternate
people
are,
but
we
want
to
make
sure
that
we
don't
necessarily
encourage
game
playing
by
having,
for
example,
some
company
have
a
different
person
every
single
week
just
to
get
their
voting
rights.
That's
feels
not
quite
a
fickle,
so
we
decided
okay,
fine,
you
can
change
it.
A
Obviously
it's
people
move
in
and
out
or
maybe
they're
on
vacation
for
a
while,
but
we
want
to
limit
it
to
no
more
than
once
per
month,
making
those
kind
of
changes
and,
to
be
honest,
our
history
has
been
that
this
things
rarely
changed
at
all.
So
once
a
month
seemed
to
be,
you
know
something:
there
was
a
nice
middle
ground
position.
There
I
have.
H
A
Sometimes
weird
is
going
on
here.
This
isn't
this:
isn't
the
intent
of
what
we're
trying
to
do
here
of
having
you
guys
change
out
a
different
person
every
single
week,
because
it
doesn't
make
it
look
like
you're,
interested
and
I
needed
some
way
to
sort
of
push
back
on
that
kind
of
behave.
That's
all
I
was
before
okay
and,
like
I
said
this
hasn't
been
an
issue
because
being
the
list
has
really
changes
and
most
you
guys
are
are
I,
think
barely
Activant
and
doing
the
right
thing.
A
D
I'm
fine
I
think
the
the
whole
point
of
you
know
sometimes
need
to
change
it
because
the
primary
or
alternate
you
know
business
trip
together.
So
yeah.
J
J
A
A
A
A
All
right,
thank
you
guys
very
much.
Okay.
This
one
was
raised
by
me.
Basically,
I
was
going
through
our
roadmap
to
trying
to
see
where
we
were
and
I
realized,
that
there
was
nothing
on
our
actual
roadmap
doc
that
gave
somebody
new
to
the
group
indication
of
how
far
along
we
are
in
roadmap,
so
pretty
much
I
just
wanted
to
add
to
them.
That
said,
help
by
the
way
these
two
are
completed
and
here's
the
dates
that
I
think
we
completed
them.
This
one
was
easy
because
we
actually
have
a
formal
release
for
it.
A
This
day
was
a
little
bit
fuzzier
for
me,
so
I
kind
of
guessed
based
upon,
we
had
a
license,
talk
and
I
kind
of
completed
them.
I,
don't
think
the
dates
on
this
setup
matter
too
much,
but
it's
fine
to
get
in
the
pattern
of
actually
saying
we
actually
completed
these
things
went
for
so
someone
looking
at
our
group
can
see
where
we
are
on
our
timeline.
A
Hopefully
this
isn't
controversial
any
questions
or
comments
on
this.
Any
objection
to
adopting
alright.
Thank
you
guys.
Next
one
this
one.
It
was
in
response
to
issue
268
where
I
came
her
the
gentleman
who
raised
it,
but
he
basically
wanted
a
slightly
different
layout
for
the
list
of
specs
or
listened
documents
we
have
but
I
think
in
the
before
this
PR.
This
was
just
a
flat
list
and
a
table
for
him.
A
There
was
a
flat
list
and
we
he
didn't
really
think
it
was
clear
which
specs
were
sort
of
required
or
core
versus
optional,
and
so
that's
what
I
tried
to
do
here.
I,
try
to
make
it
clear
that
there's
a
core
spec,
which
is
the
kind
of
in
spec,
and
then
he
had
a
list
of
optional
specs
which
are
the
transports,
and
then
we
have
additional
documentation.
A
Alright,
any
objection
to
approving
then
looks
good.
Alright.
Thank
you
guys
very
much.
Those
are
relatively
easy.
All
right
mm-hmm,
so
we
I,
was
actually
trying
to
go
through
our
backlog
of
issues
trying
to
see
where
we
were
trying
to
easily
just
clean
them
up
and
identified
five
different
issues.
A
That
I
was
hoping
we
could
go
through
relatively
quickly
to
close
out
if
we
end
up
starting
to
deep
dive
on
one
of
these
I'd
like
to
defer
that
discussion
for
a
later
time-
and
we
can
just
skip
that
issue,
but
I'm,
hoping
that
some
of
these
we
can
just
close
out
really
really
quickly.
Hopefully
most
you
guys
have
looked
at
this
because
I
did
send
out
a
note
about
it
and
I
haven't
seen
any
comments
on
these
issues,
the
saying
that
they
would
think
that
she
would
kept
open.
A
So
the
first
one
is,
you
know
what
open,
if
anything
is
not
to
alert?
What
is
you
know
what
care
event
is
not
now
we
now
have
a
non
goals
section
in
our
primer
and
I
thought.
That
was
a
good
enough
step.
Obviously
people
can
add
budget
OPR
later
to
enhance
that
I
thought.
That
was
good
enough
to
feel
like
we've
actually
addressed
this
issue
as
a
starting
point.
A
A
Any
objection
to
close
them
for
no
action,
then,
on
these
two
you
guys
are
going
home.
Just
being
too
friendly
today,
this
is
great.
Thank
you
all
right,
log
level
attribute,
so
this
one
actually
really
really
easy.
There
was
issue
63,
which
actually
was
the
PR
for
this
issue
and
that
one
it
was
basically
trying
to
add
a
log
level
attribute,
and
we
decided
to
did
it
do
to
close
this
in
favor
of
the
tracing
extension,
which
I
believe
Thomas
added
for
us.
A
I
A
A
C
A
Right,
any
objection,
then,
to
closing
this
one
all
right
and
as
a
side
note,
I
actually
did
get
the
stickers,
so
I
will
be
rigging.
Those
two
coupons
Shanghai
can
Seattle
and
actually
for
those
of
you
who
might
actually
be
in
Switzerland
next
week
for
the
CS
summit.
I
will
try
to
remember
to
bring
some
there
if
you,
if
you
find
me
I'll
I'll
hand
out
some
of
those
to
you
guys.
A
K
So
I
just
put
I
just
pasted
a
blame
link
for
the
message
or
specification
for
a
particular
line,
which
literally
lifts
text
from
our
spec,
which
kind
of
suggests
to
me
that
the
open
messaging
spec
is
kind
of
feeding
of
some
of
the
work
that
we're
doing
and
this
this
is
before.
If
you
recognize
this
is
before
my
any
change
that
I
just
did
that
we
accepted
I
think
last
week,
but
this
is
the
literally
same
text
and
which
already
says
that
there's
a
bunch
of
overlap
between
what
we
do
and
what
open
messaging
tries.
K
She've
and
if
messaging,
wants
to
be
a
similar
type
of
abstraction
for
messaging.
As
what
we
do
for
eventing,
then
that's
fine.
But
then
that
kind
of
excludes
precludes
that
we
have
some
integration
and
if
things
are
at
the
same
level
of
abstraction
and
there's
that
they
are
and
then
a
mapping
makes
very
little
sense.
K
I
think
I
think
from
a
technical
perspective
and
first
of
all,
the
the
the
the
mapping
that
we
have
is
I'm,
not
sure
what
it
says
really
because
it
seems
to
be
mapping
to
this
to
this
abstract
model.
And
so
it's
a
it's
a
mapping
between
an
abstract
model
and
an
abstract
model,
and
that
seems
very
strange
to
me
and
there's
very
little.
You
can
really
do
with
that,
spec
and
and
again,
if
this
wants
to
do
messaging,
that's
fine
but
I,
don't
think,
there's
a
there's
a
there's,
a
way
to
make
those
things
match.
A
J
A
Yes,
okay,
Clemens
came
to
me
a
favor
since
you
articulated
your
concerns
very
well,
and
people
seem
to
like
the
way
you
phrase
it.
Could
you
add
a
comment
to
the
PR?
Why
would
you
do
that
right
away?
Okay,
thank
you
very
much.
So
let
me
just
make
it
a
little
more
official.
Is
there
any
objection,
then
to
closing
as
pull
requests
with
no
action
and
Clemens
will
have
the
comment
explaining
why.
L
K
K
It
well
sorry,
but
that's
just
that's
just
the
example,
though
it
was
the
closest
and
so
it's
very
difficult
to
define,
define
a
specification
that
anybody
could
go
and
implement
for
wire
compatibility
and
but
you
can
go
and
use
MVP
all
right
or
you
can
use
even
the
M
capiz
0.1
0.9
drafters.
She
wanted
to
because
that's
wire
specification,
but
anything,
that's
that's
proprietary
I
mean
if
you
go
to
JMS,
then
even
and
you
can
potentially
map
to
the
Java
API.
K
But
the
state
of
GMS
is
not
such
that
it's
actually
reliably
compatible
with
all
products.
So
I
find
I
would
yeah.
I
think
I
think
we
should.
We
should
really
be
principled
about
wire
compatibility
and
not
trying
to
go
and-
and
you
know,
create
that
compatibility
with
some
with
some
api,
where
we
need
to
have
be
lucky
that
the
drivers
actually
do,
what
what
you're
expecting
them
to
do
and
basically
constrain
it
to
a
particular
language,
etc.
L
Okay,
so
I
get
that
and
I
understand
your
concerns.
I
I
guess
is
there:
do
we
have
room
in
this
group
to
support
people
that
want
to
do,
for
instance,
the
JMS
binding?
Would
that
would
that
would
just
be
something
that's
out
there
in
reference,
but
wouldn't
be
governed
or
or
sort
of
formally
supported,
I
guess.
K
M
K
I
think
inside
this
group
is
to
create
interoperability
across
languages
and
runtimes
and
end
the
wire
and
I'm,
not
sure
how
much
you
know,
JMS
mapping
given
given
where
we
are
today
right,
if
you,
if
you
had
races
ten
years
ago
or
five
years
ago,
I,
would
have
been
leaning
more
towards
agreeing
with
you
or
sorry.
But
to
say
you
know,
JMS
is
something
that
we
should
support.
Well,
we're
now
so
far
down
the
road
with
the
MGP
that
comes
point
yeah.
L
N
One
clear
way
to
say
that
is
to
say
that
it's
not
really
a
transport
binding
that
you're
trying
to
make
it's
a
mapping
to
another
event
or
another
messaging
format,
so
that
that's
its
name
transport
binding.
What
you're
trying
to
create
is
not
a
transport
binding
and
as
such,
it
doesn't
meet
the
bar
yeah.
L
A
A
I
A
K
Yeah
this
this
section
made
some
made
some
changes
down
downstairs
to
the
properties
so,
instead
of
getting
into
into
case
sensitivity
or
case
insensitivity,
which
is
complicated
because
ultimately
we're
mapping
to
and
we're
using
multiple
protocols
where
we
can't
change
their
stance
on
case
sensitivity,
sensitivity
or
insulin,
sensitivity
and
we've
had
now
the
the
heroic
attendance
in
the
PR
that
I'm
referencing
for
at
the
number
of
trying
to
rescue
effectively
the
casing
and
through
HTTP,
by
introducing
dashes
I'm
going
effect
in
this
PR
I'm
going
the
way
I'm
saying
you
can
only
use
lowercase
and
which
means,
if
I'm,
restricting
the
character
set
effectively,
which
means
no
matter
where
the
K
says,
K,
sensitive
or
case
insensitive
is
not
allowed.
K
You
can
only
use
lowercase
and
therefore
it
doesn't
matter,
and
even
if
someone
do
does
K
and
it
seems
like
there
is
there's
case
folding
being
done
by
some
frameworks
which
then
go
and
try
to
be
smart
about
it
and
make
it.
You
know
mixed
case
or
make
it
or
a
full
upper
case
down.
I'm.
Just
saying
the
only
way
to
deal
with
those
properties
is
for
them
to
be
lower
case.
If
you
see
them
in
any
other
casing
use
some
framework
dealt
with
them.
K
K
So
I
also
made
a
rule.
If
you
go
to
scroll
up
again,
I
also
made
a
rule
that
says
yeah,
you
know
you
should
be,
you
should
be
greedy
and
you
should
not
exceed
20
characters.
Length
reason
for
that
is
simply
that
if
we're
sending
Jason,
you
know
canonical
Jason
across
the
wire
we're
sending
hundreds
or
millions
of
them,
and
that
gets
very
long,
very,
very
quickly,
so
being
a.
N
K
A
K
A
Our
our
property
types
that
are
maps
going
to
be
encoded
with
dashes
or
something
else
or
are
they
going
to
be
included
as
a
JSON
object
as
a
value
or
you
didn't.
You
didn't
address
that
part
of
the
issue
at
all,
I'm
wondering
whether
that's
something
that
we
should
resolve
separately
from
this
or
should
I
be
incorporated
into
this.
K
K
The
core
spec
for
the
purpose
of
the
initial
discussion,
so
basically
there's
a
there's,
a
1
there's
a
there's
a
march
once
through
the
the
rest
of
the
specs
to
go
in
and
adjust
that
I
didn't
want
to
do
this,
all
in
one
giant,
PR
what
we
don't
know
what
we're
talking
about
but
but
but
to
address
your
question.
I.
K
Think
these
the
model
with
the
dashes
for
ACP
will
will
still
be
the
case
that
doesn't
go
away
because
you,
you
just
mapped
them
into
the
transport
and
then
the
s
you
lift
things
again
off
the
transport
into
an
abstract
program
model.
What
you
need
is
you
need
to
have
a
notion
of
like
so
that,
for
instance,
the
Cee
prefix
in
for
HTTP
headers
will
stay,
because
the
reason
for
that
existing
is
that
you
can
go
and
and
have
an
abstract
programming
model
which
only
deals
with
our
properties.
K
You
can
go
and
project
it
into
HTTP
and
none
of
those
will
clash
and
on
the
other
side,
if
you're
collecting
up
your
properties,
you
know
that
everything
is
prefixed
with
C
E,
we'll
have
to
go
back
into
that,
so
that
property,
it's
in
a
property
bag.
Your
representation
of
the
cloud
event
that
you're
then
presenting
off
to
the
program
model.
K
A
M
This
is
neck
I.
Personally
think
this
comment
is
irrelevant.
The
the
two
issues
that
are
brought
up
down
here,
the
the
second
ones
kind
of
the
easiest
one
once
they're
upgraded
right,
they're
still
all
lower
case
on
the
wire.
It
doesn't
matter
that
we've
gone
from
an
extension
into
the
kind
of
the
main
spec
because
on
the
wire
they're,
all
still
just
lower
case
in
the
SDK.
If
we're
you
know,
we
know
about
that
new
property
in
the
new
version.
M
You
know
it's
the
SDKs
job
of
getting
that
lower
case
name
and
presenting
it,
in
whatever
appropriate
case
for
the
language,
is
correct
and
then
for
the
extensions
again
I.
Don't
think
it
necessarily
matters,
because
the
in
all
the
SDK
models
you
are
just
going
to
be,
comparing
everything
against
the
lowercase
version
of
the
you
know.
If,
if
you're,
whatever
map
representation,
you
allow
people
to
put
in
the
camel
cased
or
snake
case
whatever
names,
it's
an
your
job
to
convert
that
back
to
the
simple
lower
case
and
we're
done
it
I.
K
All
of
our
attribute
names
would
be
turning
into
numbers,
and
then
we
would
surface
those
numbers
using
those
names
in
the
SDK,
but
they
would
never
turn
into
those
names
on
the
wire,
so
they
would
turn
into
those
numbers
and
that's
the
same.
That's
the
same
concept
here
right
we're
making
we're
making
expressions
identifiers
that
work
on
the
wire
and
that
you,
if
you
are
working
close
to
the
wire,
you
can
go
and
interact
with
them
easily.
K
You
can
identify
them
easy
easy
easily,
but
how
they
surface
up
in
the
SDK
will
not
be
a
concern,
and
if
you
look
at
very
many
different
API
s
and
how
they
represent
wire
concepts,
they
use
their
most.
You
know
the
idiomatic
representation
in
their
respective
languages
and
runtimes
and
not
necessarily
how
the
wire,
what
the
wire
representation
looks
like
so
I
agree.
K
A
Thank
you
any
other
questions
or
comments
for
Clemens
on
this
PR
keeping
in
mind.
This
is
just
the
initial
past.
The
PR
he'd
mentioned
that
the
digital
changes
wouldn't
be
made,
but
the
biggest
point
here
is
to
get
a
sense
from
the
group
in
terms
of
whether
they
like
this
direction
of
just
blower
casing
everything
and
to
refresh
our
various
memory.
This
is
gonna,
be
compared
against
Christoph's
PR,
which
we
talked
about
last
week,
which
is
D,
I,
believe
the
Matsuda
alphanumeric
and
a
before
all
capital
letters.
L
Sorry
I
forgot
my
hand
up
protocol
I,
said
I
I,
don't
necessarily
disagree
with
this,
but
I
and
I
could
buy
into
it.
I
think
there
are
a
couple
of
other
issues
floating
around
around
this
whole
subject:
area
and
I
actually
open
one
earlier
in
the
week,
which
is
somewhat
related
to
this,
but
not
indirectly.
So,
given
the
amount
of
churn
there's
going
to
be
when
this
stuff
gets
implemented
in
a
an
all-encompassing
PR,
is
it
worth
looking
at
those
other
issues
and
then
rounding
them
all
into
one
PR?
L
K
L
A
K
Yeah
I
have
a
question
and
then
they're
also
placed
into
probably
into
the
next
one.
That's
that's
up
there.
I
have
been
thinking
about
separators
and
I.
Think
I'm,
even
discussing
that
here
like
can
we
allow
a
dot
or
a
dash
or
or
an
underscore
and
I
found
that
that's
problematic
in
all
kinds
of
different
light
and
languages
and
runtimes.
If
you,
if
you
convert
those
names
into
identifiers
and
one
particular
question
I
have
is
for
for
all
the
Scala
fluent
people,
because
I'm
not
what
impact
underscores
have
in
Scala.
P
Is
a
bloody
mirror,
I
can
confirm
in
Scala
does
not
matter
you
can
have
underscores
now.
You
also
have
a
one
quite
interesting
property,
and
that
is
if
you
have
something
that
looks
like
identifier,
surrounded
with
backticks
that
can
actually
contain
even
spaces
and
and
other
strange
things.
So
there
is
a
great
flexibility
of
mapping
lanes
from
various
protocols
and
standards
to
something
that
works
as
a
scholar,
identifier,
yeah.
K
Because
at
the
only
language
where
I
found-
and
that
was
not
deep
research,
where
I
was
a
little
worried
about
underscores
with
Scala,
because
Scala
apparently
has
this
pattern
matching
with
that
is
driven
by
underscores
and
I.
Wonder
I
just
wondered
how
much
for
practitioners
that
may
be
confusing
if
you
are
using
the
underscore
across
the
language
and
then
also
have
identifiers
that
are
using
yes,.
P
N
N
A
These
are
good
things
to
remember.
As
the
SDK
developers
start,
you
know
playing
with
their
stuff,
let's
see
how
painful
it
is
for
people.
This
is
all.
This
is
all
be
good
feedback
all
right.
What
I'd
like
to
do
now
is
Joshua
I,
don't
think
Josh
was
on
the
call,
but
Joshua
opened
up
an
issue.
I
think
it
was
last
night
suggesting
that
maybe
we
should
consider
a
snake
casing.
Instead,.
A
A
N
J
A
J
K
Q
I
mean
I'm
intrigued
by
his
argument
of
the
of
communicating
a
little
bit
more
about
the
field.
With
this
with
the
underlying
that
previously
Keane
was
the
case.
I,
don't
know
if
that's
enough
to
convince
me,
but
I'm
intrigued
by
it
and
I
haven't
thought
it
all
through
the
transport
binding
soil
deferred
to
Clemens
on
that
topic.
K
So
dashes
are
a
problem
for
language
bindings
and
dots
are
problem
for
language
bindings
and
the
underscore
is
really
the
only
thing
that
is
that
that
works,
and
then
you
have,
you
know,
underscore
being
in
non-permissible
characters
and
other
api
is
and
that's
why
that's,
how
I'll
and
in
there
I
don't
know
how
language
is
so
I
don't
have
a
complete
overview
of
all
the
languages
that
are
being
used
and
that's
why
I
asked
that
that
question
about
scala
so
so
underscore
seems
to
be
a
choice.
K
G
If
we
preferred
a
short
like
leslie,
your
recommendation
for
less
than
20
characters
and
avoiding
the
use
of
underscore
is
also
recommendation.
So
if
there
are
companies
that
want
to
go
beyond
20
and
years
under
schools,
they
can,
but
it
goes
against
the
guidelines.
So
it's
still
permissible
in
some
cases,
but
we
don't
recommend
it.
Yeah.
K
N
A
A
Think
people
probably
need
some
time
to
go
back
digest
this
think
about
it
with
their
respective
teams.
What
I'd
like
to
do
is
perhaps
on
next
week's
phone
call
see
if
we
can
come
to
a
resolution
if
possible.
If
not,
we
haven't
neither
had
four
discussions
we
well
obviously,
but
what
I'd
like
to
encourage
people
to
do
is
to
add
comments
to
respective
issues
or
a
pull
request.
A
So
we
get
some
conversations
going
there
just
out
of
curiosity,
it
was
okay
with
you
guys,
I'd
like
to
get
a
general
sense
of
where
you
guys
are
leaning
towards
right
now,
just
for
the
guys
on
the
call
right
now.
So
just
out
of
curiosity
are
there
people
on
the
call
right
now,
who
would
prefer
say
Joshua's,
snake
case
solution.
P
Yeah
this
is
Vladimir
I
agree.
If
he
does
not
break
transports.
I
would
prefer
the
snake
case
because
of
much
easier
readability
and
I'm
concerned
about,
for
example,
the
aspects
of
tracing
particularly
distributed
tracing,
where
we
are
tracking
events
that
are
going
through
a
number
of
systems.
It
will
be
much
easier
for
the
human
reader
to
to
correlate
and
see
what
is
going
on.
Okay,.
R
A
A
A
H
He's
limping,
can
I
push
her
this
like
yeah,
instead
of
just
saying,
I
prefer
this
one,
which
is
like
one
like.
We
could
settle
it
that
way,
but
it
would
like
I
would
prefer
if
we
like,
tried
to
summarize
the
implications
of
each
and
then
and
then
I
imagined
that
there
will
be
a
winner.
What
like
a
clear
answer,
will
emerge.
I'm
like
this
might
be.
H
This
might
be
optimism,
but
I
think
that
the
the
implications
a
snake
case
will
outweigh
the
pros,
but
I
might
be
wrong,
and
so
I
would
I
would
really
appreciate
if
we
could
like
try
to
summarize
why
people
like
if
people
on
each
I
could
try
to
summarize
wide
like
what
they
see.
As
the
overwhelming
case
like
in
very
concrete
terms,
I
would
find
that
easier
to
like
be
open-minded
to.
S
I
can
suggest
what
I
would
observe
a
kebab
case
like
the
dashes
are
going
to
mean
that
you
can't
use
them
as
bear
or
like
bear
access
in
languages
like
JavaScript,
where
you
can
map
and
either
access
it
with
indexes
or
with
the
period.
A
snake
case
has
a
nice
benefit
of
making
human-like
multi-word
things
were
readable,
but
the
downside
is
it
makes.
Multi-Word
thing
is
much
more
likely
and
more
expensive,
because
they're
bigger.
N
K
I
would
so
if
I
I'm,
gonna
I'm
gonna,
make
my
edits
I
would
say
and
I'm
gonna
I'm
gonna
do
a
PR
on
top
of
my
PR
BC
I'm
gonna
do
a
fork
once
I'm
done
and
I'm
gonna
add
the
camel
case.
I'm
gonna
add
effectively
the
underscore
to
the
permissible
characters
and
then
we
can
go
in
and
because
then
in
the
end,
then
is
the
the
snake
case
right.
So
I
can
go
in
and
effectively
expand
mine
to
include
the
underscore
and
then
we,
as
all
arguments
have
been
exchanged
for
against.
A
I
think
you're
rich
wants
to
get
it,
but
I
say
that
after
I
got
a
sense
of
who
was
in
favor
of
what
so
I
make
sure
that
there
were
some
people
max.
It
did
want
each
one
of
those.
So
yes,
please
put
comments
into
each
of
the
PRS
or
the
issue
in
Joshua's
case,
so
we
can
get
some
conversations
going
there
in
particular
list
the
things
that
you
think
are
an
advantageous
or
problematic
with
each
one.
A
So
people
can
see
your
point
of
view
all
right,
I'm
gonna
see
what
we
can
do
about
trying
to
resolve
that
next
week.
That's
sure,
okay,
I,
don't
think
we
have
time
to
dive
into
your
next
PR
Clemons.
So
let
me
just
quickly
go
back
and
oh
cool
people
and
their
names
there.
That
is
wonderful.
Alright
s,
Lomb
are
you
there.
A
L
A
Can
all
right
and
what
about
the
renato
is
viewed
excellent.
Thank
you.
Simon
either
I
think
I
heard
the
Simon
already
once
David.
A
E
H
A
G
D
A
E
A
A
A
A
A
Can
do
that
one?
A
few
minutes?
Okay,
skipping
this
one
for
a
sec,
so
status
of
cloud
events-
I'm
assuming
this
is
probably
gonna-
be
related
to
cloud
event
over
you,
so
it'll
probably
get
lumped
together
in
some
fashion.
Yes,
workflow
I
know
Kathy.
You
probably
want
that.
That's
it
that's
an
easy
one
for
you.
Anybody
else
have
any
objections,
including
that
as
one
of
our
futuristic
topics.
T
D
T
Much
time
do
you
think
you
need
a
melon,
so
how
many
total
time
45.
T
A
Well,
let's
see
I
know
plays
out
cuz
I,
think
I
think
we
may
have
to
twiddle
the
times
a
little,
but
let's
start
out
with
10
to
15
minutes
and
see
yeah
okay.
So
what
about
demo
I'm
thinking?
This
may
not
be
the
best
form
for
that
I'm,
not
sure
we
nasai
have
time
to
do
it,
but
I
want
to
get
your
guys
to
take
up
on
it.
R
D
K
R
A
A
K
A
A
O
A
D
I
D
K
A
K
A
Okay,
that
so
that's
interesting,
so
it
seems
to
me
in
an
intro
session.
If
we
have,
you
know
one
or
two
minutes
of
what
was
certain,
what
the
service
working
group
did
and
how
it
led
to
cloud
events,
and
we
do
some
amounts
of
overview
and
status
in
an
intro
that
might
be
a
nice
introductory
session
yeah
and
then
a
in
a
deep
dive
talk
about
some
of
the
deeper
level
things
like
some
more
deeper
dive
on
the
transports,
for
example,
get
into
the
workflow.
A
K
D
D
Also
in
the
intro
I
think
we
also
need
some
overview
of
the
workflow.
Otherwise
it's
a
little
bit
weird.
Why
am
I
indeed
I
would
mention
workflow,
but
it
intro
we
do
not
mention
I,
think
we
should
also
same
thing.
We
should
talk
a
little
bit
about
you
know,
workflow,
introduction
about
workflow,
so.
F
A
Austin
is
not
going
right.
Awesome
not
going
I
can
definitely
talk
to
that
in
terms
of
what
they've
been
doing
because
I've
been
attending
those
meetings.
I
don't
think
that's
a
big
deal,
but
I
do
like
the
idea
of
a
little
bit
of
a
one
to
two
minute:
tease
of
work
for
SDK,
to
bring
people
back
and
then
talk
about
them
in
the
deep
dive,
yeah
and
I.
A
I
D
A
I'll,
try
to
run
what
I
was
thinking
of
Amit
I.
Think
I
wrote
that
I
think
I
was
just
trying
to
give
I
think
I
just
want
to
show
people
what
cloud
events
look
like
and
I
think.
That's,
probably
that's
probably
automatically
there
when
you
start
talking
about
what
our
cloud
events
and
they
were
view,
this
back
I,
don't
think
we
have
to
call
that
out.
Okay,
I
think
I.
Think
we
hear
that.
D
A
A
A
But
yeah,
but
I
would
like
the
Q&A
session
at
the
end
of
each
to
be
almost
in
a
bird
of
a
feather.
So
it's
interactive
okay,
anything
else,
I
feel
like
the
deep
dives.
Still
a
little
light.
I
know
Clemens.
You
said
we
could
ramble
for
quite
a
while
there,
but
are
these
two
sufficient?
You
think
four
I.
A
K
If
we,
if
we
only
show
the
demo
the
demo
kind
of
from
the
surface
and
say
missus
the
scenario
and
explained
the
scenario
basically
make
this
this
session,
a
as
we
would
call
it
level
of
200,
the
first
intro,
which
is
I,
take
the
light
on
tech
and
probably
doesn't
show
much
code.
Then
the
deep
dive.
The
deep
dive
starts
with
explaining
that
code.
That's
already
taking
some
time
and
then
and
then
I
would
and
then
I
would
go
into.
K
A
A
D
A
D
A
A
D
E
A
D
K
No
well,
it
doesn't
help
anybody
who's
in
the
room
that
we
had
is
good
enough
and,
and
if
dog
has
has
it
I
would
I
would
literally
just
just
use
that
as
the
cliffhanger
like
it's
a
it's
a
1
minute,
2
minute.
Oh
look
at
what
we
have
here
and
you
know
talk
about
the
scenario
per
se
and
then
say:
if
you
want
to
see
it,
if
you
want
to
see
the
detail
that
comes
with
next
session,
so.
K
K
A
D
K
K
Doug
just
show
that
so
the
integration
scenario
that
we
had
is
this
is
the
this
is
the
non
Twitter
or
Twitter
version
of
this
right
yep.
So
that's
what
we're
gonna
show
we're
going
to
explain
how
this
is
cut.
We're
gonna
in
the
intro
session,
we're
gonna,
we're
gonna,
explain
that
there
are
various
publishers
that
are
raising
events
and
they're
being
pushed
and
they
composed
the
street,
and
so
we'll
talk
for
a
minute
about
that.
But
we're
not
going
to
talk
about
how
that's
being
done
for
that
people
need
to
come
to
the
next
session.
K
D
N
A
A
F
A
A
B
A
A
A
K
B
A
A
N
K
K
A
D
A
A
D
A
A
Okay,
anything
else
yeah,
okay,
so
tell
you
it!
Why
don't
we
do
this?
Why
don't
we
now
stop
here
and
then
we
will
go
off
and
think
about
this
in
our
particular
sessions,
and
maybe
we
could
talk
again
after
next
week's
phone
call
just
like
we
did
today
and
to
see
if
people
have
any
concerns
about
the
flow
or
questions
or
something
like
that.
I'll
talk
about
the
next
steps
that
sound
good
that
works.
Okay,.
D
Sdk
I
might
need
your
help.
You
know,
because
I
did
not
run
that
this
group
meeting
yep.
D
A
That's
fine
I,
don't
yeah!
That's
no
problem!
I
was
just
thinking.
I
think
Scott
actually
might
be
really
busy
with
other
things,
so
he
may
not
be
able
to
talk
Varun.
You
said
you
don't
feel
cool
about
that.
That's
fine,
Nik,
I
haven't
I,
don't
know
if
Nick
was
talking.
Let's
assume
not
because
I,
don't
I,
don't
even
think
Nikki's
actually
been
in
our
phone
calls,
so
he
may
not
be
a
good
person
to
do
it
anyway
and
I.
A
Don't
know
who
Roux
is
so
that's
guy
I
wanna
make
sure
that
we
didn't
exclude
somebody
from
the
potential
speaker
list
so
I
think
we're
good
okay,
okay,
yeah!
So
let's
do
that.
Let's
talk
again
after
next
week's
phone
call
assuming
that
works
for
you
guys
I'll
send
out
an
invite
and
then
we
can
see
if
we
have
any
additional
questions
or
suggestions
for
changes.
At
that
point,
that's
not
good
yeah.