►
From YouTube: CNCF Serverless Workflow 2020-03-09
Description
CNCF Serverless Workflow 2020-03-09
A
Ho
get
bit
of
a
funky
weekend
here
that
describes
it
like
on
the
weekend
hundred
meters,
200
meters
away
from
my
house.
Another
house
exploded
me:
how
did
that
help?
I,
don't
know
it's
not
yet
known.
I
guess
everybody
suspects
a
gas
explosion,
but
it
was
a
pretty
crazy
scene
so
bit
over
an
intense
weekend.
I
mean
nothing
happened
to
us,
but
we
are
definitely
sorry
for
something
like
that,
but
no
a
neighborhood.
B
C
B
D
A
B
B
What
we
have
is
good
enough
to
move
to
a
sandbox.
This
means
a
lot
of
things.
Hopefully
now
I
am
kind
of
new
to
this
too.
So
Doug
is
just
you
know.
It's
gonna
needs
to
help
us
a
lot
on
this,
but
from
what
I've
seen
it
means
our
own
github
repository,
which
by
itself
is
a
big
win
and
your
mailing
list
hopefully-
and
hopefully
our
dedicated
the
host
IP
for
a
website.
B
D
Yeah,
that
would
be
great
any
ideas.
I
am
very
new
to
the
CN
CF
myself.
Any
idea
how
what's
the
time
range
here?
Is
it
weeks
or
I
suspect
this
depends
on
how
often
the
TOC
meets
and
discusses
this
right,
and
then
we
need
to
move
through
six
stages
to
gather
support
for
becoming
a
sandbox
project.
Yeah.
B
B
D
E
D
Okay,
so
let
me
start
and
talking
about
the
TOC
EPO
is
actually
a
good
thing,
because,
along
the
way
to
become
sandbox
project,
where
we
need
to
show,
of
course,
this
community
be
also
a
little
bit
of
background
or
additional
material
than
just
the
spec
itself.
We
had
a
motivation
from
duck,
I
think.
D
Last
year,
when
there
was
the
first
vote
on
the
subgroup
preceding
as
an
individual
group
that
we
should
come
up
with
a
primer,
very
much
like
cloud
events
has
done
it
and
the
primer
document
is
as
I
understand
it.
It's
not
the
specifications
or
a
specification
would
have
a
normative
reference.
That
would
be
maybe
our
JSON
schema,
maybe
something
in
a
different
format.
D
You
know
we
do
have
natural
language
specification
as
well,
so
these
documents,
but
these
they
both
describe
the
state
of
the
current
workflow
specification,
but
the
primer
would
explain
the
would
motivate
first,
how
all
of
this
is
being
done
and
I'm.
There
are
few
mentions
of
why
we
do
this
in
form
of
goals
or
and
when
background
information
across
the
documents
we've
written.
D
But
it's
good
to
have
this
in
as
the
starting
point
of
the
primer
to
motivate
why
this
specification
came
about
and
then
also
to
motivate
the
design
of
the
specification,
so
why
we
have
chosen
to
describe
a
workflow
of
States,
and
this
maybe
requires
us
to
name
a
little
bit
the
roots
of
this
specification.
So
what
concept
it
originates
from
to
explain
why
we
are
adopting
this
weird
terminology
when
one
could
argue
that,
well
we
just
sort
of
copy
the
Amazon
States
language,
but
maybe
there
is
a
little
bit
of
background
to
that.
D
So
people
can
build
their
this.
This
understanding
in
their
mind
when
they
start
reading
the
primer
before
actually
diving
into
the
specification,
and
then
we
can
provide
a
little
bit
of
explanation
of
the
concepts
that
are
recurring
throughout
the
specification,
because
I
think
there
are
a
few
concepts
applied
here
and
then
we
can
also
point
out
the
affiliated
projects
and
how
we
or
how
the
IDI
specification
relates
to
these
other
projects.
D
I
think
this
is
very
important
also
to
show
that
it
integrates
into
the
CNC
F
and
it's
not
just
project
that
stands
on
its
own.
For
example.
Currently
it
is
affiliated
to
the
cloud
events
as
cloud
events
can
be
used
to
trigger
workflows
and
also
workflows
can
produce
cloud
events.
This
is
specifically
stated
in
the
current
specification.
B
If
I
came
just
couldn't
just
jump
in
for
just
a
second
please,
yes,
two
things
for
I've
been
thinking
about
this
primary
instance.
Since
the
scope
of
service
work,
you
know
we
have
to
see
is
much
bigger
than
cloud
events
itself.
It's
a
format
Jason
of,
however,
they
also
have
like
10
times
the
team
that
we
currently
have
and
have
contributed.
So
I
would
really,
if
possible,
put
this,
and
if
you
guys
agree
in
two
steps.
One
is
a
short-term
one
is
a
long-term.
B
The
primer
for
this
specification
can
be
miles
long
pages,
long
and
very
in-depth.
If
we
wanted
to
go
in
there,
where
everybody
can
contribute
it,
we
got
some
people
in
place
like
yourself
and
thoughtful
and
others
who
have
very
deep
knowledge
into
this.
To
contribute.
However,
for
again
going
back
to
this
TOC
and
short
term,
we
need
some
sort
of
primer
for
short
term
and
that's
what
I
put
in
a
document.
We
need
something
to
show
some
background.
B
I
think
they
will
look
at
that
and
I
think
for
that,
maybe
comparing
it
even
stating,
like
you
said.
Yes,
our.
There
is
a
lot
of
interest
in
this
there's
a
lot
of
different
workflows
for
service
integration.
We
don't
really
have
to
say
everything
why
and
what
was
the
reason
and
we're,
but
comparing
it
to
or
three
things,
even
if
they're
completely
similar
or
you,
which
might
people
say
they
are
some
not
but
giving
examples
currently
in
the
short
term.
I
think
is
the
best
idea
and
then
we
can
move
and
evolve
this
document.
B
This
does
not
have
to
be
a
document,
it's
that
it's
done
right
now,
but
if
we
can
put
milestones
behind
it,
like
let's
say
my
stone,
one
is
eggs,
milestone.
2.
Is
why
and
and
and
stuff
like
that
within
then
we
can
all
involve
it,
but
we
need
something
by
the
end
of
the
month,
probably
which
I
assume
we
will
get
the
TOC
call
to
actually
show,
and
at
that
point
we
can
just
say.
Yes,
we
have
a
primary
it's
in
its
early
stages,
but
at
least
we
have
it.
D
Don't
think
this
needs
to
be
comprehensive.
It's
I
agree
that
we
don't
need
to
be
exhaustive
here
and
right
pages
long
documents,
I,
think
books
have
been
waiting
on
on
this
and
there
is
a
huge
amount
of
research
and
but
really
what
we
need.
What
I
believe
we
need
to
give
is
some
motivation
to
why
we
need
the
workflow
specification.
D
It
may
be
why,
for
example,
it
needs
to
be
specific
to
service
I.
Had
this
one
comment
from
Scott
that
yeah,
why?
How
is
he
your
invocations,
that
you
orchestrate?
How
are
they
service?
Is
this
and
I
think
he
was
getting
to
whether
this
is
just
by
assumption
or
whether
there
is
actually
something
to
this
back
and
yes,
we
can
say
there
is
something
to
respect,
for
example,
our
embedding
in
our
sorry
interoperability
with
cloud
events,
something
that
makes
this
okay,
then
Scott
might
argue
it's
cloud.
D
A
I
also
don't
have
any
ambition
to
write
a
book
about
the
reasoning
for
a
way
we
have
the
separate
language
I
think
we
should
focus
on
one
concise
document.
However,
that
looks
like
that
describes
to
the
TOC
why
this
matters
otherwise
I
mean
you
know,
writing
a
book
about
it.
It's
nothing.
That's
it's
necessary.
Maybe
later
on,
one
can
have
some
more
explanation,
respect,
but
for
now
I
guess
we
just
put
our
seals.
B
B
That
was
the
whole
point
of
this
is
to
kind
of
keep
moving
us
forward.
We
have
to
move
forward
in
order
not
to
stagnate
number
one
and
number
two
is
the
main
reason
that
I
think
and
you
I
guess
this
is
a
meeting
for
that.
We
the
main
reason
to
even
do
the
specification
regardless
we
right
why
this
is
done
like
that
or
or
the
other
in
much
detail
is
because
every
month
there
is
a
new
workflow
language
coming
out
and
it
called
themselves
server
lists,
and
there
is
no
standardization.
B
Otherwise
we're
just
one
of
many
many
other
work
for
languages.
However,
we
tend
to
standardize
things
across
the
board,
and
this
is
why
we're
standardization
we're
not
an
implementation
right
here,
so
that,
if
we
can
put
in
the
primer
as
the
top
kind
of
paragraph
and
then
evolve
it
into
the
future,
maybe
with
some
examples
between
what
we
have
compared
to
two.
Maybe
that's
enough
for
now
other
implement
workflow
models.
I
will
be
very
happy
with
that
right
now
and
then
later
on,
we
can
now
we
have.
D
B
I
mean
I'll,
be
if
it
is
happy
to
have
it
and
then
then
I
think
TOC
will
be
happy
just
to
see
that
we're
all
working
on
this
together
and
then
keep
moving
forward
and
every
every
release
we
got
to
focus
and
also
releasing
version.
0.1
and
I.
Think
having
a
primer
is
a
big
deal
for
that.
So
if
we
can
have
some
sort
of
description
of
other
languages,
how
many
that
more
incoming
and
why
a
standard
is
needed
for
portability
you
vendor-neutral
model,
blah
blah
blah
will
be
very
good.
B
D
Blah
blah
blah,
if
it
is,
if
no
vendor
implements
it
and
I
mean
I,
have
it
so
there
is,
it
needs
endorsement,
and
this
also
I
think
why
we
should
really
motivate
and
give
enough
background
and
sure
the
the
specification
that
we
have
right
now.
We
can
move
this
to
version
0.1
and
also
the
timeframe
I
think
it's
okay,
a
little
bit
of
pressures
or
whiskey,
put
a
deadline
out
and
stuff
will
be.
It
will
be
done
so.
A
D
Presentation,
yeah
I,
think
if
you,
if
you
look
at
the
TOC,
the
governance
and
the
project,
how
you
can
go
from
like
become
an
incubator
and
project
and
graduate
you
find
you
enough
information.
I,
don't
want
to
discuss
the
TOC
PR
too
much
in
detail
here.
I,
rather
want
to
focus
on
what
we
should
put
in
the
primer.
A
D
A
A
A
D
D
It
might
be
good
if
we
use
the
suggestion
mode
and
then
discuss
it
or
maybe
have
a
comment
on
this
and
if
people
sign
off,
then
we
were
okay
with
it.
If
people
have,
because
if
stuff
is
changing
too
fast,
we'll
lose
people
on
the
way
so
I
myself,
I
only
maybe
I-
can
maybe
spend
one
or
two
days
a
week
on
this
at
most
and.
D
So
yes,
I
have
started
with
affiliated
specifications,
because
I
thought
that
these
are
very
good
to
explain
how
group
the
workflow
specification
embeds
into
the
CN
CF
landscape
and
how
it
relates
to
to
other
things.
Maybe
we
want
to
what
you
don't
they
asking
do
you
think
it's
good
to
adopt
the
the
structure
of
the
cloud
events?
Primer,
I
can
take
an
action
item
and
you
apply
the
structure
of
that
primer
here,
yep
yeah.
E
A
D
Comes
about
with
a
history,
and
so
the
history
of
this
document
we
already
have.
The
original
is
a
design
document
that
Casey
I
think
she
opened
it
up.
At
least
I
have
read
access
to
that
service.
One
function,
workflow
design
document.
He
please
try
it
it.
It's
like
a
primer,
it's
like
I
think
it.
It
also
served
as
the
workflow
specification
proposal
and
it
uses
an
example
and
already
states
what
it
is
that
this
vacation
should
cover.
So.
D
B
Second,
that
would
focus
on
and
we
can
go
through
that
if
you
need
to
stateless
versus
stateful,
I
think
that's
a
big
deal,
because
a
lot
of
languages,
even
AWS,
which
is
the
leader,
has
a
separate
implementation
for
one
nd
or
the
other.
So
that's
a
big
deal,
I
think
and
the
second
of
course,
the
cloud
events
integration
and
not
only
integration
but
kind
of
the
tie
in
that
you
pretty
much
have
to.
If
you
use
events,
you
have
to
use
cloud
events
so
link
the
cloud
event.
B
Specification
would
be
a
good
thing
and
then
focus
in
the
beginning
chapter
so
I
would
say
this
is
just
my
opinion.
You
guys
can
tell
me
I'm
crazy
on
describing
or
giving
an
example
like
you
already
have
a
BPM
and
to
be
AWS,
maybe
conductor-
maybe
you
know
the
ones
that
you
have
found
some
good
examples
and
the
another
focus
that
you
probably
have
to
put
a
little
bit.
Can
you
say
well,
can
you
use
other
things
to
create
several
work
who's?
B
The
answer
is
yes,
so
we're
not
saying
that
there
is
no
other
options
for
designing
service
and
specifically,
service
work
was,
however,
given
the
state
of
the
landscape.
Currently,
this
is
what
people
are
using
and,
however,
this
is
why
there
is
no
standard,
and
this
is
why
a
standard
is
needed
and
kind
of
leave
it
at
that
for
one
and
keep
moving
for
her.
Then,
with
with
description
of
details,
that's
just
my
opinion,
but
but
you
you
guys
do
figure
it
out.
D
D
Might
want
to
because
there
was
a
lot
of
discussion
on
the
terminology
that
this
has
adopted
right.
The
way
we
called
state
states
and
using
transitions
is
also
nice,
because
it's
enlightened
with
the
state
diagram
terminology,
but
yeah.
We
call
the
entire
thing
workflow.
So
do
you
think
it
make
sense
to
at
least
reference
these
concepts
to
describe.
E
D
A
You
know
I,
think
that
makes
sense
especially
I
mean
the
terminology
is
one
part,
but
maybe
we
should
really
clearly
state
under
which
assumptions
we
are
designing
this,
and
also
what
is
the
niche
we
are
targeting
because
Yuki,
you
guys
keep
saying
that
we
are
like
a
smaller
targeted
language
for
a
particular
niche,
and
that
should
be
well
identified.
A
thing
right.
B
Yeah,
basically,
which
you
can
write
there
is
something
don't
quote
me,
but
you
know.
Currently
you
have
to
look
at
how
people
develop
service
applications
and
you
can
maybe
define
even
I'll.
Give
you
I
think
I
did
I
shared
their
slides
from
my
presentations
where
they
I
think
we're
not
presented
it.
B
You
had
an
example,
and
you
can
show
that,
similarly
to
typical
applications,
there
are
non
serverless
developers
have
to
write
both
business
logic
and
orchestration
logic
in
in
same
codebase,
maintain
it
how
the
problems
around
that
or
many
the
same
reason
why
workers
are
used
in
traditional
development
environments
as
well,
but
in
typically
in
surpluses.
You
have
this
code
that
you
deploy
their
code
functions
and
again
serverless
work.
They
are
specific
for
those
environments
to
to
distinguish
orchestration
logic
from
business
logic,
so
again
we're
allowing
developers
to
focus
on
their
business
and
the
core
code.
B
They
are
wanting
to
develop
and
we're
offloading
all
this
other
stuff,
orchestration
data
management,
control
flow
logic.
You
can
say
even
all
kinds
of
stuff
in
there
from
that
base
development
effort.
The
developers
need
to
focus
on
so
that's
kind
of
like
the
main
topic
on
what
we're
trying
to
do
and
again
say.
Yes,
that's.
C
This
is
Gemma,
I
was
going
to
be
listing
a
lawyer
and
I
think
you
guys
are
making
valid
points.
I
think
the
word.
When
people
see
the
term
workflow
orchestration,
they
immediately
will
leap
to
you
know
large
industrial
scale,
business
process,
automation,
I,
think
what
you're
referring
to
is
really
thing
is
what
the
previous
speaker
was
talking
to
you.
You
read
about
function,
orchestration,
yeah
or
micro
process,
orchestration
or
something
and
sort
of.
C
D
B
D
D
D
I
think
I
missed.
Oh,
maybe
I
may
have
missed
your
point.
They
didn't
on
goal,
which
we
somehow
need
to
make.
Maybe
a
section
for
this
regarding
pictures.
Examples
are
nice,
but
I
think
examples
are
very.
They
cover
a
lot
of
space.
So
if
we
can
put
this
in
a
short
paragraph
instead
of
an
exhaustive
example,
then
it
would
be
easier
to
digest
the
primer.
So,
as
you
said
to
me,
I
don't
want
this
to
be
100
pages
long
document.
B
B
D
B
D
D
It
seems
very
related
to
and
I
know,
cube
flow
is
gaining
some
importance,
at
least
in
the
ml
community,
but
to
using
or
is
there
any
project?
So
we
named
conductor
and
I
still
have
this
on
my
to-do
list.
You
look
at
the,
which
is
the
open,
whisk
sequence
or
open
whisk
function.
Service
function,
orchestration
right
there
conductor
is
there
any
other
project
that
you
would
deem
important?
D
B
B
Just
Netflix
conductor
has
one
example
in
their
documentation,
which
is
also
very
good.
That
is
kind
of
called
the
kitchen
sink
right,
where
they
use
everything
every
one
of
their
constructs
so
pick
legs
to
just
to
that,
you
can
find
documentation
for
and
put
it
on
there,
and
then,
let's
see
how
we
can
all
work
together
to
come
up
how
we
can
and
please
guys
stop
me
at
any
point
I'm
talking
too
much.
But
our
goal
for
this
is
not
an
implementation.
B
Based,
go
we're
not
worried
about
implementation
and,
quite
frankly,
a
recently
updated
our
specification
to
say
we
don't
care
this
whole
argument
between
workflow
versus
state
machines.
We
don't
even
care
about
that
honestly.
At
this
point,
probably
what
we
care
about
is
that
our
specification
modeling,
which
can't
describe
what
other
workflows
are
doing
right
because
dents
that's
kind
of
like
the
goal
right
now.
B
If,
if
some
languages
out,
there
have
much
better
much
more
descriptive
manner
of
doing
many
more
things
they
might
not,
we
might
not
be
targeting
them,
but
we
can
target
Netflix
and
we
can
target
it
obvious
hundred
percent.
Currently,
so
maybe,
let's
focus
on
those
now
right
and
anything
about
implementation
like
context
the
instances
of
workers,
let's
keep
that
out
of
here
for
now,
because
quite
frankly,
I
don't
think
we're
there,
yet
they
just
focus
just
on
the
model.
How
can
we
describe
this?
How
can
we
describe
that.
B
D
Is
invocation
of
functions
alone,
so
invocation
of
the
service
functions?
We
currently
have
the
your
eyes,
so
not
even
I
think
I
had
one
yeah.
It's
right
here.
I
had
mentioned
services,
but
this
is
not
to
orchestrate
or
is
it
is
it
to
orchestrate
between
functions
and
services
and
do
we
do
we
even
make
it
distinction
between
something
running
as
a
service
or
not?
D
B
Whether
it
is
a
service
function
or
micro
service
or
anything,
we
do
not
care,
we
should
not
even
care
and
put
down
specific
function.
Well,
one
of
the
things
that
our
actions
can
invoke
are
functions
or
micro
services
or
whatever
the
only
thing
that
we
are
100%
kind
of
tied
to
our
cloud
events,
format
and
that's
it.
Everything
else
is
whatever
implementations
want
to
focus
on
in
this
case,
because,
quite
frankly,
in
our
implementation,
for
example,
that
we're
working
on,
we
can
invoked
an
HTTP
request
or
we
can
call
a
web
hook.
B
D
A
Is
pretty
interesting,
so
you
know
if
we
say
and
I
agree
it.
What
we're
designing
here
doesn't
necessarily
need
to
focus
exclusively
on
functions,
but
then
we
are
entering
in
this.
The
general
space
of
service,
orchestration
and
I
think
then
it
will
be
difficult
to
differentiate
between
other
orchestration
languages.
In
this.
A
B
But
we
the
way
we
scope.
This
is
by
our
definition
of
a
function
which
has
a
resource,
location,
right
and
parameters.
So
the
this
is
kind
of
like
with
AWS.
Does
right,
I
mean
you
can
with
AWS
call
us
or
a
service
or
anything,
that's
behind
a
what
is
it
called
a
RN
right
and
R?
We
do
not
know
what
that
is,
so
we
cannot
say,
but
we
don't
describe
our
function,
execution
with
a
class
path,
okay
or
some
sort
of
other
way.
We
we
have
a
specific
resource
which
is
drink.
B
D
We
might
need
to
be
a
bit
more
specific.
Currently,
the
function
of
f
only
carries
a
uniform
resource
identifier,
so
similar
like
in
a
RM,
that's
not
even
location
bound,
so
it
could
be.
It
could
identify
pretty
much
everything
and
we're
not
so
the
invocation
semantics,
except
for
the
parameters
that
are
being
passed.
They
are
not
covered
by
the
specification,
and
so
you
said
cloud
events
is
I
know
cloud
events
can
trigger
workflows
and
workflows
can
produce
cloud
events,
but
the
function
of
vocation
is
not
using
conferences
or
is
it
you.
B
Know
what
I'm
saying
is
the
only
thing
that
our
workflow
definition
understands
is
when
it
comes
to
invent
consumption,
is
the
cloud
event
format
yep
for
event,
states
and
stuff,
like
that.
That's
the
only
bound
that
we're
putting
on
to
implementations,
because
and
only
because,
our
in
our
workflow
modeled
events
are
described
using
the
context
variables
defined
in
the
cloud
event
specification
yeah.
D
We
actually
tried
how
we
implemented
ASL
ourselves
and
then
what
we
wanted
to
show
is,
like
you,
could
quickly
convert
from
step
function,
definition
to
wanting
it
on
our
system,
but
a
problem
rarin.
So
they
are
Amazon
specific
and
this
you
cannot
just
orchestrate
any
set
of
services
and
functions
with
that,
because
the
a
and
eventually
you
have
to
translate
it
to
something
that
is
known
to
your
own
system.
D
B
Specifically,
if
you
look
at
the
schema
for
functions
and
I'm
sorry,
this
is
sidetracked.
We
can
take
this
offline.
If
you
want
look
at
the
type
parameter,
we
in
addition
to
an
Arn,
which
is
kind
of
like
the
only
thing
that
AWS
allows
you
to
do.
We
do
allow
implementations
to
put
in
a
specific
type.
B
D
I
think
because
this
is
at
the
boundary
of
the
work
flows
back
much
like
the
cloud
events
to
invoke
a
workflow
under
the
workflow
to
produce
I
think
this
is
this
needs
some
emphasis
in
this.
This
motivation
so
just
outlining
what
is
the
boundary
between
a
workflow
definition
and
a
and
its
execution
and
the
outside
world?
D
D
How
do
you
feel
about
the
concepts
that
we
do
so
we
mentioned
event,
consumption
and
producing
events?
There
is
control
flow,
definitely
and
the
use
of,
for
example,
that
the
choice
state
or
that
that's
the
transitions
are
something
that
we
have
with
every
state
and
I
think
it
has
its
own
explanation
in
this
back.
D
Should
we
describe
the
concept
of
states
or
pick
up
the
concept
of
states
and
transitions
between
states
like
in
the
finite
state
machine
to
motivate
the
design
of
the
workflow
specification
language
or
using?
This
is
unnecessaries.
You
mentioned
earlier
that
for
now
we
don't
care
for
now.
We
want
just
something
that
works.
B
Yeah,
wouldn't
put
this
whole
state
machine
in
there
at
all,
we're
not
machine
description
to
begin
with,
because
we
cannot
invoke
any
state
at
any
point
in
time
we
do
have
a
dedicated
starting
state
or
node
or
whatever
you
want
to
call
it.
So
we
are
not
fully
a
state
machine
in
my
opinion
and
that's
a
discussion
that,
like
everybody's
gonna,
have
a
different
opinion
on
I
will
keep
that
out.
Implementation
implemented
whatever.
B
D
Nice
thing
about
transitions
is
that
actually,
at
the
end
of
execution
or
evaluating
every
state,
we
can
have
a
transition
to
any
other
state
and
a
finite
state
machine
more
or
mealy
automata.
They
would
define
a
starting
point
and
then
transition
to
States.
The
only
difference
here
is
I
think
there
can
be
an
end
state,
but
this
can
go
on
like
forever.
D
B
B
D
B
I'm
talking
I'm
talking
specifically
about,
for
example,
the
realized
state.
Yes,
that's
that
was
specifically
done
in
mind
with
thinking
about
testing,
so
you
can
inject
your
own
data
into
that.
Your
X
testing,
with
into
your
workflow
and
and
and
see
how
your
work
will
behaves
and
then
there's
different
conditions
there
being
different
data
inputs.
B
So
then,
so
that's
another
thing
you
might
want
to.
If
you
wanted
to
say
that
we're
receivable,
we
has
testing
in
mind
another
thing
that
would
bring
in
to
your
document.
If
you
just
want
to
describe
what
we're
targeting
is
the
extensible
aspect
of
it.
So,
in
addition
to
and
that's
something
that
a
lot
of
people
lack
and
don't
have,
is
that
you
get
a
model
definition,
you
have
to
use
what
they
give
you.
B
B
A
B
Are
currently
every
extension
has,
if
you
have
a
unique
ID,
we
also
have
a
PR
open
for
metadata,
which
extensions
currently
are
extension
elements
you
can.
Think
of
those
like.
You
know
the
extension
elements
like
PP
sim,
for
example,
but
there
was
a
PR
to
also
do
a
property
extension
which
is
on
hold,
but
if
you
guys
want
to
resurrect
that
I
think
it's
a
good
idea,
we
can
resurrect
that
as
well.
D
I'm
not
sure
I
always
disliked
vendor
specific
extensions.
It's
I
personally,
but
this
is
just
my
personal
opinion.
I
always
consider
the
this
standard
body
to
be
a
little
bit
of
a
failure.
If
you
have
an
extensive
use
of
these
extensions,
so
yeah
I
mean
unless
the
the
standard
can
be
just
taken
to
the
next
version
to
include
what
people
want.
I,
don't
see
a
point
in
overlaying
it
with
these
extensions.
D
D
This
is
not
the
same
as
an
extension,
so
if
a
function
binding
you
mean
the
type
of
function
invocation
there.
It
is
and
the
realization
of
performing
this
function
invocation.
Then
we
do
have
a
type
field
and
everything
used
in
there
is
already
vendor-specific,
but
the
property
extension
would
be
to
introduce
new
properties
of
states.
A
D
D
These
sort
of
things
I,
don't
know.
If,
because
eventually,
what
you
want
is
to
also
verify
a
description
and
then
having
it
extensible
makes
it
very
difficult
to
run
the
same
verification
on
it
over
and
over
so
yeah.
It's
my
personal
opinion,
really
I
I
don't
want
to
stand
in
the
way
of
first,
we
need
to
come
to
a
1.0,
and
then
you.
B
Point
if
I
can
just
bring
it
up,
sorry
that
you
can
put
down
there.
What
distinguishes
this
from,
let's
say
AWS
is
that
we
describe
our
model
via
JSON
schema.
They
don't.
Okay,
Netflix
doesn't
so
we
have
an
actual
schema
description
of
our
modeling
of
our
model,
which
is
very
useful
for
implementation.
A
D
D
D
B
Please
look
at
the
specification
if
you
understand
it,
if
you're
new
or
not
spellcheck
it.
If
you
find
any
issues
or
problems,
misspelled
words
things
you
don't
understand,
either
raise
an
issue
or
be
brave
and
actually
do
a
PR.
It's
mostly
for
the
new
people.
We
really
need
right
now
we're
in
the
status
where
every
small
thing
just
formatting
making
things
right
will
make
us
look
better
for
your
COC
review.
B
Please
I
think
nobody
here
objects
to
add
Falco,
who
has
contributed
to
PR,
recently
contributor
and
four
new
people
here.
If
you
just
do
the
small
as
PR,
there
is
any
contribution
matters
and
your
name
will
be
added
to
the
list
of
contributors.
So
you'll
have
your
name
somewhere
and
ciencia,
hopefully
sooner
sandbox
project.
So
if
that
have
little
motivating
you
people
to
look
at
it
and
contribute
things.