►
From YouTube: GraphQL Over HTTP - December 17, 2019
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
B
B
B
B
C
B
D
F
A
B
H
F
H
G
F
B
E
Current
current
state,
like
I
Pinza
and
last
year,
waiting
for
apply
a
last
thing.
I
heard
for
them
is
that
they
need
to
change
founding
documents
for
crafty,
expect
foundation
to
a
wolf
or
additional
specs,
and
they
also
asked
me
to
provide
a
mission
statement
and
yeah
just
remember
about
that
and
actually
think
quick.
E
E
Tough
figuring
out
how
to
change
documents
and
the
kosslyn
foundation
join
this
okay,
maybe
yeah
yeah.
We
definitely
need
to.
We
actually
need
to
add
mission
statement
into
to
spark
any
to
the
despair
and
we
can
discuss
it
inside
its
action
items.
Action
item
for
me.
So,
if
condition
have
GDK
organization
for
expect
development.
They'd
like
me
to.
E
Out
how
to
do
it
properly,
so
it's
a
motive
stuck
one
another
thing:
I
wanted
to
discuss,
actually
discussed
threesome
and
because
I
initially
actually
wanted
to
create
a
spark
on
the
graph
kill
organization.
But
at
that
time
it
was
work
and
they
Facebook
and
not
really
ownership.
Question
was
not
really
transparent,
so
I
created
under
a
piss
guru,
so
his
guru
is
basically
mean
my
friend
is
just
a
brand
name.
Is
not
a
company
website,
so
a
question
here:
si
si
will
probably
stuck
with
non
official
repo
for
next
couple
months,
at
least
through.
E
We
move
it
to
make
new
organizations
I
think
like
graph
key,
lower
GDP
or
everybody.
Okay,
with
it's
stand
on
the
article
rank
you
way
we
can
move
it
to
craft
a
foundation.
Like
so
question.
Do
we
need
the
intermediate
organization
who
is
couple
months
or
we
can
just
wait
for
official
adoption,
I'd.
A
D
Not
two
more
I
think
it's
definite
that
you
know
we
intend
to
move
it,
but
Sooners,
probably
better
I,
don't
know
I've
me
and
all
my
friends
have
followed
the
nodejs
modules
working
group
and
shared
a
lot
of
links
between
ourselves
and
I.
Don't
know
if
that,
if
that
had
have
moved
around
I
think
a
lot
of
people
would
have
lost
track.
C
E
Yeah
so
the
question:
is
it
worth
additional
confusion
to
create
like
new
organization
just
for
a
couple
of
months
or
wait
tables
up
this
book
because
I
graph
kill,
like
everybody,
agrees
it?
It
goes
into
official
graphical
work,
just
a
matter
of
time,
so
for
us
to
cooperate
and
this
month
should
we
move
a
lot,
which
is
a
question.
D
E
Yeah,
if,
like
nobody
awake
like
any
concern
of
contributing
to
his
guru,
reprise
I
said
it's
not
the
company's.
There
is
no
legal
status,
just
Brendon
so
and
I
can
put
whatever
way:
cool
and
magnific
work
yeah.
We
can
put
a
new
license
or
code
of
conduct
or
anything
that
your
company
require.
If
you
need
something
for
contributing
before
before
we
give
it
away
to
craft
care
foundation,
we
can
make
put
like
temperature
documents.
So
if
anybody
have
any
concerns,
just
open
an
issue
may.
D
E
E
Any
other
questions
or
stuff
that
I
can
raise
with
Linux
Foundation
guys
about
this
process
yeah
when
I
will
get
updates
from
them.
I
will
probably
open
an
issue
just
to
update
like
political
status,
so
everybody
in
sync
and
understand
what's
happening
but
like
maybe
you
want
to
actually
sum
when
I
will
ping
them
yet
another
time,
maybe
I
need
to
raise
something
else,
for
anybody
have
an
expression.
B
B
Well,
so
I
have
the
updated
version
of
the
roadmap.
Nice
bring
me
write
down
preview
over
here,
so
what
I
did
was
in
this
PR,
just
kind
of
summarized
the
different
things
that
we
talked
about
and
took
an
approach
that
it's
kind
of
a
bit
opinionated
I
know
we
had
different
perspectives
about
whether
version
one
should
have
like
what
arrives
in
version
one.
So
I
kind
of
wanted
to
like
stay
out
clearly
a
possible
set
of
things
to
include
in
version
one
and
discuss
on
it.
B
I
think,
before
we
kind
of
dig
into
the
details,
I
think
it
would
be
really
important
for
us
to
you
have
a
shared
understanding
and
a
shared
goal
for
whatsit
version
one
and
how
the
different
things
that
we're
all
interested
fit
into
the
roadmap
long
term
so
that
we
can
feel
like
we
have
a
shared
understanding.
So
maybe
we
don't
go
with
this.
That's
you
know
it's
open
as
a
PR,
it's
not
merged
into
the
repo
yet
but
kind
of
open
that
up
for
conversation.
B
So
we
should
probably
just
replace
this
section
with
the
mission
statement
that
I'm
in
you're
gonna
you
come
up
with
included
some
kind
of
guiding
principles
so
over
in
the
graphic
UL,
the
main
graphic
you
are
working
group
there's
a
lot
of
emphasis
on
kind
of
building
development
of
the
spec
based
on
use,
cases
and
I
think
that's
helpful
for
us
to
articulate
changes
and
new
things
that
we
want
to
add.
Why
would
be
adding
how
they
help
different
people
who
are
using
the
spec
and
also
wanted
to
outline
that
we?
B
B
E
So
like
from
previous
experiences
graph
here
Spock,
we
sometimes
actually
do
breaking
changes,
but
we'll
try
to
minimize
it
so
I
there
is
the
juror
breaking
changes
and
there
is
de
facto
breaking
changes
is
a
UI
so
like,
for
example,
we
discussed
an
alien,
a
new
directive
code
described
by.
We
should
break
in
change
because
people
can
already
have
with
the
routine.
So
in
a
sense
but
probabilities
that
somebody
have
described
by
directive.
So
I
think
we
we
should
try
for
backward
compatibility
and.
E
B
B
So
this
is,
this
is
basically
just
kind
of
the
the
main
parts
of
the
document
that
are
outlined.
I
just
wanted
to
kind
of
show
like
okay.
This
is
the
details
that
we
would
go
into,
which
are.
These
are
mostly
headings
directly
from
the
current
version
of
the
spec
and
I
think
these
are
the
things
that
we
will
be
looking
for
to
happen,
and
so
there's
some
meat
in
here
some
ideas
and
thoughts
about
how
we
might
go
about
that.
B
Maybe
let
it
just
kind
of
gloss
over
that
for
the
high
level
and
then
so
version
two
and
beyond
and
I
just
tried
to
capture
the
different
things
that
we've
discussed
in
the
previous
working
group
at
least
kind
of
throw
them
out
here.
These
will
kind
of
look
at
how
we
manage
these
in
the
future
in
version
two
and
beyond,
and
if
we
can
do
that
in
such
a
way
that's
backwards
compatible
like,
for
example,
something
like
versioning.
B
If
we
say
like
they
must
communicate
this
way
for
versioning,
then
we
still
want
to
kind
of
grandfather
in
most
of
the
popular
graph
QL
servers,
as
even,
if
they're
not
yet
doing
that
they
at
least
are
compliant
with
the
initial
version
of
the
spec
and
can
additionally
start
supporting
these
other
features
as
they
get
standardized.
Do.
D
H
D
D
B
Probably
want
to
have
to
think
deeply
about
those
two
things
and
how
they
tie
together.
That's
partly,
why
I?
Why
I
think
that
we
should
probably
do
that
is
version
two
like
we
should
absolutely
be
thinking.
We
need
to
do
that,
but
if
we
do
it
as
a
version,
one
allows
us
to
get
like
the
mechanic
or
the
administrative
things
all
out
of
the
way
and
structure
the
repo,
so
that
we
can
then
put
a
lot
of
focus
and
emphasis
on
delivering
merging
and
modularity
in
a
way
that
really
it's
a
really
good
way.
B
We've
already
started
to
deliver
value
from
people
who
can
see
the
spec
it's
already
published.
They
can
see
that
we're
proposing
new
additions
to
it.
If
you
look
here
at
some
of
the
actions,
I
think
something
will
be
helpful
is
to
have
some
test
suite.
We
can
already
automate
basic
compliance
with
this
back
and
then
we
could
add
in
versioning
and
I.
Think
that's
going
to
take
us
some
time
to
really
get
something
that
solid.
E
E
And
no,
but
we
discussing
it
so
for,
but
we
have
respect
text
and
reference
implementation
and
we
only
try
all-italian
cuz,
the
guy
who
wrote
sangria
for
years.
We
try
to
create
one
graph,
kill
cuts,
but
for
my
inspectors
work
uncomplicated,
because
it's
the
bundle
resolvers
it's
hard
to
test.
So
we
have
some
points
to
test
like
some
parts
of
spec
and
make
it
bigger.
So
it's
something
that
you
want
to
do
for
my
spec,
but-
and
we
discussed
like
five
times
on
working
groups,
but
it's
really
hard
to
do
for
with
graphical
diversity.
Tb.
E
It's
like
very
simple,
because
the
so
I
curse
DDP.
You
know
what
you
testing,
you,
don't
care
about,
like
resolvers,
MERS,
over
object
and
promises
and
like
I,
think
I
think
stuff.
So
it's
very
simpler
to
do
for
graphical
over
TTP
and
we
can
do
have
already
reference
implementation
in
form
of
expression
of
care,
because
all
the
server's
initially
started
from
Express
graphical
is
the
baseline
and
they
all
support,
always
feature
set.
So
I
think
it's
it's!
An
wine
is
a
gross
of
Waynesburg
I.
A
Actual
graph,
QL
and
implementation
in
order
to
serve
as
our
reference
implementation,
because
this
spec
specifically
is
serving
graph
QL,
which
we
can
just
call
it
graph
QL,
but
it's
really
the
black
box
that
is
graph
QL
over
HTTP.
So
this
is
everything
outside
of
graph
QL,
so
we
could
just
have
something
in
there
that
says
this
starts
graph
QL
and
that's
no
longer
a
purview.
A
D
So
I
I
think
there
would
be
a
huge
value
to
having
like
a
just
a
fake
schema,
just
just
to
use
to
demonstrate
different
client
tech
and,
like
I'm
thinking.
Maybe
this
test
suite
could
be
that,
because
you
know
I've
just
used
graph,
QL
Pokemon
all
the
time,
but
you
know
that
doesn't
always
suit
all
the
use
cases
for
demonstrating
like
mutations
or
something
and
so
like
when
I
was
writing.
D
My
spec
I
just
made
up
fake
sort
of
scheme
of
stuff
that
it
isn't
even
a
real
schemer
I,
just
called
it
like
single
upload
or
multiple
upload,
but
like
if
there
was
a
functioning
actual
reference
graph.
Ql
schema
like
that
could
be
used
consistently
in
all
the
code.
Examples
of
the
spec
that
probably
be
value
even
just
outside
of
our
documentation
purposes,.
D
B
E
E
We
need
some
kind
of
schema
to
test
some
things
like
error,
propagation
and
subscribe
for
most
of
the
tests
done
in
the
schema,
but
for
some
tests
you
need
but
a
problem
with
with
kima.
If
we
use
too
much
features,
we
exclude
some
of
limitations
like
not
every
implementation
even
support
this
year,
some
of
whom
limitations
cost
based.
So
until
recently
I
think
white
Python
implementation
was
caused
by
standard
office
terabytes.
E
D
C
E
It's
an
idea
for
future
discussion.
Maybe
we
can,
since
we
have
now
list
of
implementation,
as
maybe
we
can
have
a
wait
for
proposals
of
alternative
permutation
me.
Maybe
we
can
have
like
feature
matrix
and
I
could
have
marking
that
with
wide
widths
implementation
for
his
budget
and
other
other
projects,
so
just
for
for
future
reference.
So
I
right
now
the
idea
to
include
non-controversial
stuff
so
everybody's
back
component
I
like
Benjy
last
time
you
called
the
threat,
reactive
spec
or
such
and
Seymour
I,
forget
and
I.
C
So
there's
actually
an
issue
in
this
repo
that
talks
about
them.
They,
they
aren't
very
similar,
but
they
work.
They
have
different
things
like
the
ordering
of
mutations
versus
not
the
ability
to
change
things
like
exporting
variables
from
one
and
feeding
them
into
the
other
versus
not
doing
that,
and
a
few
other
subtleties
like
that.
But
yes,.
I
Exactly
and
I
mean
we
net
implementation
or
chocolate,
we
we
are
implementing
two
of
them
and
we
have
the
export
stop.
We
are
very
similar
in
the
way
to
sangria,
but
like
Apollo
user.
This
uses
batching
just
to
reduce
requests
where
we
actually
can
change
operations
and
then
export
the
results
from
one
operation
into
the
next
as
variables.
So
there
are
different
variations
that
you
can
do
there
and
I
think
for
a
first
version
of
the
spec.
We
should
keep
it
simple
and
just
that's
the
most
basic
basic
use
cases
in
there.
B
B
E
A
B
H
One
of
the
bits
of
that
discussion,
we
were
just
having
restated
something
I
think
we
talked
about
last
meeting
as
well,
which
was
the
feature
matrix
I,
think
that
would
be
I'd
like
to
see
that
I
think
is
an
action
item
that
we
kind
of
work
together
to
figure
out
which
implementations
we
want
to
assess
and
which
features
we
want
to
look
at
and
if
that
feature
matrix
does
have
three
columns
for
batching.
That's
fine
if
we
could
figure
out
which
invitations
are
using,
which
of
the
three
ones.
B
H
Can
certainly
help
fill
in
the
details
for
all
of
the
Apollo
details
on
that
feature.
Matrix
I,
don't
know
if
I
am
able
to
speak
for
every
implementation.
I
think
I'd
be
great
if,
if
people
who
know
like
a
particular
implementation,
best
kind
of
like
help
fill
in
the
gaps
there,
if
we
just
add
a
lot
of
extra
columns
and
their
features
that
maybe
are
only
unique
to
a
particular
implementation.
That's
fine,
but
maybe
that's
a
better
cooperative
way
of
getting
this
matrix
filled
out.
Oh
one,.
B
Thing
I
was
thinking
that
we
could
maybe
do
is
because
what
I
worry
about
is
maintaining
this
table
and
I
was
wondering
if
we
could
automate
this.
The
creation
of
the
table
I
like
if
it's
like
the
basic
things
here
like
so
it's
growing
up.
If
we
can,
we
add,
like
the
features
that
we're
looking
at
here,
baby
sick
compliance
with
this
bag.
B
We
could
automate
most
of
those
things
most
of
compliance
with
that,
and
therefore
you
just
run
the
test
against
the
implementation
running
the
basic
schema
and
you're
you're
able
to
determine
whether
or
not
it
has
the
feature.
So
then,
if
someone
was
adding
batching
as
a
new
feature,
maybe
they
add
tests,
but.
H
D
B
There's
kind
of
like
a
lot
of
different
options,
but
like
we,
it
would
be
good
if,
like
when
we
get
to
that
point,
we
have
someone
kind
of
just
like
propose
like
propose
some
different
options
for
how
we
could
proceed.
Maybe
even
a
couple
people
tag-team
on
it,
because
I
could
see
like
at
least
a
couple
of
different
ways.
They
could
work,
but
I
haven't
thought
deeply
enough
on
it
to
feel
like
I
have
like
a
yeah.
That's
a
great
way
to
do
it.
B
D
F
Yeah
you
much
work
for
us
to
to
put
all
the
tests
and
centralized
repository
or
to
run
all
the
tests
and
centralized
repository.
So
what
we
could
have
is
like
a
very
basic
set
up
that
makes
it
easy
to
for
the
implementers
to
run
the
tests,
so
I
don't
have
any
specific
dependencies.
But
just
that's
actually
I
said.
I
B
Yeah
I
think
separate
the
like
the
examples
of
the
different.
You
know
possible
implementations
from
the
test
suite,
so
you
could
fire
that
test,
but
you
give
it
here's
the
URL
and
maybe
it
in
the
future
when
we
say
and
also
test
these
extensions,
these
you
know
separate
things
and
then
it
just
runs
to
that
URL
and
it
does
the
test.
It
gives
us
the
results
rather
than
us
having
to
like
tie
those
two
things
together.
So
then
I
could
you
could
just
fire
it
off
it?
G
I
G
E
Would
suggest
like
very
simple
criteria,
a
person
who
volunteers
to
to
do
this
job
choose
like
a
first
tech
stack
and
when
we
iterate
based
on
web
because,
like
I,
think
way
to
bootstrap
something
it's
important
to
start
starting
and
when
we
can
figure
out.
If
it's
working,
not
so
I
think
it
will.
It
would
be
a
responsibility
of
a
person
who
who
here
to
do
that.
You
choose
tech,
stack
and
how
it's
done
and
when
we
penetrate
so
it's
not
fixed
version.
We
didn't
and
west
a
lot
of
time.
E
Suggest
work,
I
posted
very
basic
scheme
on
this.
One
field
suggest
like
if
we
create
example,
servers
like
a
poor
expressed
rocky
post
graph,
Iowa
pop
or
ones
like,
especially
if,
after
so
and
here's
I
can
do
one
for
expressed
graph.
Kayla
know
how
to
control
write
one
and
we
create
work
with
this.
E
F
E
H
A
F
D
D
E
D
D
That
will
just
hit
that
test
request
with,
and
you
know
the
devops
around
having
a
big
list
of
all
the
URLs
that
you
know
are
served
by
all
the
different
server
implementations
and
like
maybe
we
don't
even
need
to
get
as
far
as
having
the
first
server
that
is
being
tested
yet,
but
just
the
DevOps
of
by
p.m.
to
automate
that.
E
Yes,
I
like
to
fully
test
specifications
the
rest
like
he
thinks,
for
example,
what
if
what
if
we
have
personnel?
So
we
need
to
test
that
we
don't
have
data,
we
have
errors.
We
cannot
test
exact
snapshot
because
a
different
implementation,
little
different
person
colors,
but
we
need
to
test
it
easier
and
we
need
to
desert
status
code.
This
not
200.
It's
like
I
forget
which
one
but
not
200,
I,
think
inspectors,
say
400
so
like
we
need
to
just
such
things.
So
we
cannot
do
exam
like
snapshot
based
testing
code
or
just
hitting
URLs.
E
If
we
really
want
so
like
a
go
of
that
shoot,
at
least
like
when
I
suggested.
This
idea
goal
was
to
make
sure
that
current
spec
test
is
compatible
with
all
servers
and
we
cannot
achieve
it
and
you
we
write
a
test,
shoot
that
way,
send
some
stuff
and
check
responses.
So
it's
both
ways.
But
let's
understand
the
basic
question
is
like
we
have
two
ways
to
write:
an
attitude.
We
can
just
to
a
great
proof
of
concept,
with
write:
three
five,
ten
basic
tests
and
around
that
against
five
servers.
E
H
D
A
E
Document
when
we
say
like,
for
example,
with
first
thing,
server
support
yet
and
like
second
sub
test,
it
should
support
to
get
with
operation.
Name
should
support,
like
just
say,
get
quarry
without
operation
name.
We
this
type
of
test
right.
We
discussed
in
this
one,
so
writing
force
no
comment
and
another
person
just
attacked
in
a
way
how
to
how
to
stop.
Okay,
yeah
I
think
it's
make
sense
yeah.
If
somebody
want
yours
to
actually
why
do
is
so
extent
step
by
step
very
data,
one
basically
to
turn
spec
text
into
test
cases
here.
Would.
B
B
We
kind
of
have
kind
of
being
off
road.
A
little
bit,
I
think
we're
kind
of
discussing
next
steps,
but
I'd
like
us
to
kind
of
get
kind
of
a
clear
grasp.
It's
like
yeah
like
could
we
just
kind
of
like
finish
up
like
the
implementations
we
can
talk
through
the
API
and
then
we
can
like
get
a
clearer
document
that
has
here's
the
next
steps.
We're
gonna
do
towards
these
actions
in
the
next
month.
With
that,
can
we
kind
of
focus
around
that
a
little
bit
yeah.
A
Can
I
make
one
comment
about
the
test
suite
before
we
do
move
on,
though
sure
we
spoke
in
a
previous
meeting
about
this
at
least
the
first
version,
not
being
a
restrictive
spec
in
that
there
are
so
many
implementations
and
right
now
we
don't
want
to
start
right
off
the
bat
and
say
this
is
what
must
be
true,
because
there
may
be
something
out
there
that
doesn't
currently
fit
that
and
we
want
that
to
be
okay.
We
wanted
this
to
be
a
suggestive,
spec
or
a
spec.
A
E
B
So
I
just
wanted
to
share
with
you
guys,
because
you
hadn't
seen
it.
This
was
some
of
the
work
that
we
did
to
look
over
like
the
graph
QL
over
HTTP
implementations,
so
this
kind
of
begins.
If
we
were
to
have
some
matrix
that
showed
support,
we
could
add
it
to
this,
but
what
I?
What
I
did
was
I
just
found
all
the
things
that
I
could
think
of
that
I
that
were
linked
either
from
gravity
wells
main
page
of
code,
because
it's
a
page
there
as
well.
D
Question
is
like,
like
part
of
the
goal
of
having
these
server
and
client
implementation
lists
like
should
we
also
have
anything
like
any
package
or
module
that
just
touches
on
the
HTTP
layer,
for
example
like
the
graphical
upload
middleware,
it's
not
a
whole
server,
but
it's
it
like
does
deal
strictly
with
HTTP
stuff.
So
it
is
an
implementation.
D
B
That's
that's
what
introduces
complexity
because,
like
let's
say,
for
example,
another
PR
I
opened
was
with
the
API
the
list
of
like
popular
api's,
but
it's
really
like
you
can
look
at
any
of
these
servers
and
the
server
on
its
own
without
a
schema
without,
like
whatever
you
know,
gateway
that
it's
behind
and
all
these
other
things
it's
like,
they
actually
could
transform
it
completely.
You
know
you
take
Express
graph,
QL,
plus
your
custom,
middleware,
plus
your
own
schema,
that's
doing
random
stuff
like
it's.
B
E
E
B
D
Know
and
it
jumps
to
mind
the
graphical
upload
middleware,
because
it's
it's
used
in
a
lot
of
the
servers
and
like
that.
That's
where
users
will
complain
if
we,
if
we
have
a
spec
that
diverges
from
that
implementation,
because
it
directly
relates
to
the
structure
of
the
post,
multi-part
request.
I
would.
B
Wonder
it
because
if
we
put
it
in
this
list,
it's
gonna
get
lost
in
in
the
mix
like
that.
I
would
guess
that
sometimes
like
the
servers
that
may
use
that
I'm
more
populated,
perhaps
than
that
particular
library,
and
so
therefore
it
could
get
further
down
the
list.
So
I
wonder
if
it
would
make
sense
to
have
like
a
separate
section
of
like
popular,
like
supporting
libraries,
that.
E
D
Works
with
any
nerd
server,
it's
it
works
with
all
of
them.
I've
used
it
with
the
Express
and
with
color
and
with
the
polar
stuff.
So.
E
Like
I
think
we
don't
want
to
introduce
new
dimensions
so
right
now.
It's
always
like
is
a
language
based
awake
at
least
so.
I
keep
a
system
by
so
a
ports
on
a
system
in
a
sense,
so
actually
like
a
grieve.
If
it
will
be
like
support
inquiries
or
like
some
additional
category,
where
it's
like
augmented,
it's
not
done
for
Minecraft
killed
over
HTTP.
It's
out
mounted
it.
E
D
Depends
on
what
the
intention
of
the
utility
of
this
list
is
because,
in
my
view,
it's
just
a
good
place
to
gather
together
common
patterns
in
the
industry
and
make
sure
that
we
cater
to
what
people
are
doing.
You
know,
I,
don't
know
if
we
intend
to
necessarily
convert
this
exact
list
over
to
the
the
to
the
matrix
that
we,
you
know
or.
F
F
Decks
are
in
relation
to
each
other
so
because
the
first
version
of
the
specification
is
going
to
be
a
compromise
between
something
that's
at
least
minimally
useful,
because
if
you
don't
specify
anything,
of
course,
we
are
compatible
to
everything
and
I
think
there
has
to
be
some
breakage
at
some
point
like
breakage,
in
the
way
that
we
specify
something:
that's
not
implemented,
absolutely
100%
each
and
every
implementation.
F
A
B
A
I
think
I
think
it's
about
the
direction
of
what
you
guys
are
thinking
of
this.
If
you're
thinking
of
this
as
a
restrictive
spec,
then
there's
no
way
we
can
possibly
write
one
that
makes
it
so
that
everybody
is
compliant,
but
if
the
spec
is
written
such
that
this
is
not
what
we
expect,
but
this
is
what
is
a
allowed
that
does
that
says
nothing
about
what
is
not
allowed
only
what
is
allowed.
E
I
can
speak
a
little
bit
and
I
try
to
be
very
short,
so
I
initial
motivation
for
I,
don't
know
like
common
motivation
for
his
group,
but
my
initial
motivation
when
I
wrote
it
was,
it
was
another
company
in
our
co-working
I,
come
to
them
and
help
them
with
graph
K,
but
they
use
postpone
force
any
inquiries
as
and
while
used
in
post
one,
and
they
knew
both
graphical
and
I.
Try
to
use
playground
and
it
wasn't
working
and
I
spend
a
half
hour
debugging.
E
It
just
found
out
that
they
capitalized
property
argument
names,
so
they
have
a
brush
name
from
capital
letter
and
quite
a
variable
from
a
wager.
So
in
a
sense,
I
deal
like
my
year
when
I
wrote
it
initial
one
and
can
it
can
be
made,
but
idea
unusual
was
to
specify
restrictively
common
set
of
Popo
our
servers
so
X
like
what
general
community
agrees.
E
So
in
a
sense,
if
we
discuss
with
use
case,
we
break
in
server
that
these
guys
in
my
co-working
world,
because
it's
non-standard,
a
used
on
some
non-standard
Ottoman
things
without
we
can
write
anything
like
we
need
to
break
some.
Some
like
weird
servers,
and
actually
it's
a
good
thing
because,
like
another
thing,
like
background
dependence
and
behavior
and
graphical
the
panels
and
behavior
and
other
coins
dependence
under
here
so
I
make
a
system
always
like
98,
97
percent,
agreed
and
stuff.
We
just
need
to
capture
with
97
percent.
F
E
I
C
I'm,
just
a
button
here
representing
graphical
graphical,
doesn't
care
about
HTTP.
Currently
it
will
in
future,
but
at
the
moment
it's
all
done
with
a
fetcher
function
which
could
read
from
the
file
system
or
send
things
over
network
sockets
in
other
protocols.
It
doesn't
care
about
HTTP
at
all.
I
know.
I
E
E
B
D
B
A
A
E
A
B
B
C
Then
I
didn't
add
this
to
the
agenda
earlier,
but
for
anyone
who
was
reviewing
the
pull
request
that
I
raised
earlier
bring
sorry,
yes,
so
I
raised
a
pull
request
regarding
making
clearer
a
bar
coding.
The
URLs
and
I
wanted
to
know
whether
what
our
stance
should
be
on
putting
code
examples
into
the
spec
I.
C
C
C
But
there
is
debate
over
whether
or
not
we
should
actually
include
something
like
sample
implementations
like
this
in
the
spec
or
where
those
things
should
go.
I'm
personally
happy
either
way,
I
don't
mind
whether
it's
in
there
or
not.
I,
just
think
that
it
may
be
something
we
should
discuss
because
having
code
samples
is
often
a
lot
clearer
than
actual
spec
text,
particularly
for
people
who
don't
speak
English
as
their
primary
language.
Absolute
spec
takes
the
primary
language
I.
D
A
E
I
feel
like
I
need
to
present
opposite
side,
because
it's
actually
objected
code
sample
in
this
PR
and
in
a
sense
it's
like
example.
So
it's
not
prescribing
you
to
use
white
whisper
typical
sample
just
example:
how
quick
and
can
implement
creating
URL
URL
encoding
of
parameters
so,
but
my
concern
here
is
like
just
we,
after
that,
we
have,
like
example,
requests
and
what
I
should
think
fill
it
sufficiently
illustrate.
So,
what's
going
URL
encoded,
unencoded
an
encoded
parts,
so
we
can
see
how
it's
encoded
and.
A
E
G
E
To
read
it
so
I'm,
not
a
native
English,
speaker
and
I'm,
not
native,
though
my
best
bigger
brother.
I
speaker,
because
I
like
most
of
my
career
forecast
and
budget
engineer
so
for
me,
like
I
can
recognize
this
function,
but
it's
not
something
that
I
use
daily,
so
I
think
like
in
a
sense
I'm.
Okay,
with
saying
in
an
example:
Kevin
Kevin
a
sentence
any
weight
in
JavaScript.
You
can
use
this
function.
For
example,
you
can
use
functions
from
Thunder
fireplace
for
your
own.
Couldn't
can
in
javascript.
E
Is
this
function
but
I'm
guessing
you
didn't
quite
entire
code
sample
because
it's
it's
it's
like
quote,
inspect
and
ii.
Think
it's
like
it's
big
and
third,
it's
like
in
JavaScript,
and
I
think
we
should
not.
He
was
like.
I
already
have
a
program
that
people
saying
graph
here
should
be
a
gradual
process
should
be
implemented,
JavaScript
or
crafty
little
should
be
implemented.
Javascript
and
I.
I
B
B
Know
I
I
agree
that
in
general
to
I
think
that
the
examples
I
want
to
see
are
can
I
see
what
the
HTTP
request
was
and
what
the
HTV
response
should
be
like
that.
What
I
would
like
to
see
in
examples
more
around
thin,
the
I
know
it's
going
to
be
probably
obtuse
in
some
ways
when
it
has
to
be
URL
encoded
according
to
RFC
3986,
but
that
would
I
think
be
more
informative
for
a
future
person
implementing
this
back
to
seem
like
what
is
the
request?
What
should
the
response
be?
Json.
E
C
E
D
E
B
D
E
C
C
So
I
think
it's
just
going
to
be
like
a
should
escape
in
many
cases
like
the
code
that
you'd
already
written
even
in
the
in
the
readme,
would
probably
have
worked
in
most
like
web
browsers,
for
example,
but
may
not
have
worked
in
other
libraries,
I
guess,
I'm,
not
sure
and
also
wouldn't
have
worked
if
we'd
had
an
equal
sign,
for
example,
in
the
text.
So
yeah.
C
E
Think
it's
brilliant
question
why
it's
so
important
to
have
this
case,
because
if
we
had
the
test
case,
we
can
actually
tested
against
at
least
like
couple
example.
Servers
in
this
brink,
like
Sammy,
wanted
to
raise
a
question
about
next
steps
and
I.
Think
it's
important
because
we
like
hour
and
a
half
in,
and
we
need
to
wake
finish
and
before
worth,
to
discuss
what
we
need
to
do
next
month.
B
Yep
what
I
was
gonna
do
is.
Let
me
share
my
screen,
but
I'm
just
going
to
share
the
document
that
we're
you.
Anyone
can
edit.
So
you
should
all
just
see
the
same
document
and
I
paste
it
in
here.
These
are
the
actions
that
we
had
from
you
know
just
for
context,
not
gonna,
say
what
we're
gonna
do
in
the
next
month
towards
our
next
meeting
and
I
know.
B
E
F
E
Will
take
me
awake
less
than
an
hour?
Definitely
so,
and
we
sure-
and
please
don't
wait
for
this
first
version-
it's
not
important
to
create
a
docker
file
and
if
you
just
create
a
folder
that,
if
you
have
everything
necessary
installed
on
your
system,
you
can
run
a
so
it
was
a
gradual
Java.
So
if
you
need
install
JDK
and
other
stuff,
but
like
the
entire
code
of
of
a
library
of,
for
example,
is
in
this
folder
and
I,
don't
think
we,
the
support,
ripple,
that
we
can
do
it
in
the
same
repo
and
I.
E
E
D
E
E
Like
same
I,
think
like
the
JavaScript
to
have
liked
issued
and
like
JSON
schema,
have
a
test
shoot
and
they
just
publish
it,
but
to
write
with
desk
issue
to
me
to
complementation,
to
run
it
against
and
I
can
think.
We
need
some
so
awake.
It's
obvious
to
choose
away
pop
or
one
or
or
implementations
that
we
have
a
volunteer
to
maintain
so
I'm
totally.
Okay,
with
maintain,
express,
crafting
and
think,
like
other
people,
can
help
me
in
well,
because
javascript
is
popular
and
express
that
I
feel
is
pretty
simple
and
stable
library.
Same-Same.
D
E
E
B
So
I
would
not
seem
like
particular
like.
Maybe
these
are
not
specific
enough
for
an
action
point
that
someone
can
own
I
mean
maybe
like
something
like
what
the
feature
matrix
is
like
yeah
I
can't
do
that,
but
I
would
really
like
us.
I
think
we
kind
of
talked
about
a
couple.
People
who
said
I
could
do
something
or
I
could
do
this,
but
I'd
like
us
to
capture
that
in
some
way,
so
that
we
kind
of
can
know
like.
Oh
okay,
someone
is
going
to
do
that,
and
this
is
who
we
do,
that.
C
Yeah,
this
has
worked
pretty
pretty
well
with
the
input
Union
spec
on
the
graph,
Co
spec
working
group
at
the
moment.
So
someone
started
with
just
a
basic
document
with
a
few
examples
and
then
others
died,
contributing
more
I.
Think
that
should
work
quite
well
rather
than
taking
it
out
of
github
I
think
we
will
typically
follow
github
for
this
spec
anyway.
That's
where
we're
centered
on
so
doing
it
on
a
central
document.
I
think
makes
sense.
E
E
E
Yeah
I
think
like
it's
like
two
separate
documents.
One
document
says
to
write
test
cases
and
another
one
is
to
just
capture
whatever
implementation
supports
in
case
of
a
poet's
great,
because
I
have
documentation.
So
we
can
reference
like.
If
you
have
automatic
persistent
queries,
you
have
a
documentation
page
about
automatic,
persistent
queries,
so
I
think
is
keys
to
to
documents.
C
E
So
I
think,
like
a
first
version
until
next
month,
we
need
to
focus
on
the
stuff
that
written
inside
aspectx
so
like
for
feature
matrix.
We
cannot
doing
all
the
stuff
we
know
but
for
right
hand,
particular
tests
I
think
we
need
to
scope
ourselves
to
would
currently
written
in
black
text.
So,
for
example,
we
much
like
URL
encoding
and
we
need
to
add,
like
euro
encoding
description
for
Euro,
including
test
cases.
C
E
I
was
thinking
like
expect.
Number
of
test
would
be
like
hundred
less
than
hundred
or
hundred,
so
we
can
I
think
we
can
pretend
in
one
document
and
if
it's
become
to
be
cooking
splitted,
but
one
thing:
I:
wake
from
input,
unions,
process
on
minecraft
case
practice
that
we
have
yeah
we
cooperate,
but
there
is
somebody's
making
decision
instead
of
we
discussing
everything
and
so
I
I
feel
like
if
somebody
wants
to
want
just
to
be
champion
for
like
every
every
point
and
we
can
just
cooperate
on
Wed.
E
B
E
F
E
E
E
E
A
I
just
want
to
throw
this
out
there.
I.
Don't
think
that
anybody
is
here
asking
for
a
commitment
for
delivery,
and
it
is
just
a
commitment
to
pay
attention
or
work
on
this
as
far
as
test
cases
I,
don't
even
think
we
need
any
actual
test
cases,
maybe
just
a
list
of
tests
that
we
feel
are
appropriate
and
we
can
iterate
on
that
itself
before
even
writing
any
test.
Yeah.
B
I
think
that'd
be
a
list
of
tests
that
we
should
implement
like
if
we
had
someone
work
on
building
a
technical
framework
which
I'm
kind
of
interested
in
this
one.
Next
meeting
we
might
say:
okay.
Well,
let's
take
a
list
of
the
test
cases
someone's
been
doing
some
thinking
about
what
we
should
be
testing.
We're
gonna
actually
implement
those
tests.
You
know,
and
they
could
be
this
little
streams
could
say
so.
D
Regarding
the
sample
schema
it
deep,
so
there's
two
different
flavors
paths
we
could
go
down.
We
could
either
have
it
totally
abstract
from
any
sort
of
realistic.
You
know
thing
like
comparing
Star
Wars
a
Pokemon,
how
they're
sort
of
a
fictional
use
case
versus
a
totally
abstract.
Like
caller
mutation,
you
know,
create
thing
or,
like
you
know,
upload
one
file
or
instead
of
like
add
avatar
to
Star
Wars
news
or
you
know
what
I
mean
like
do.
E
Think
we
need
to
just
go
to
an
example.
Schema
and
test
schema.
So
examples,
Kumar
should
be
a
said.
It
should
be
something
quick.
You
know,
for
example,
examples
in
the
spec.
It
would
be
great
if
you
can
actually
run
them
against
some
examples
or
worse,
but
for
test
server
if
tester
or
actually
makes
sense
it's
great.
But
if
we
need
to
add
some
weird
strange
things
that
we
cannot
like
find
a
good
name
so
funny
example
or
something,
but
if
you
need
to
put
it
to
test
it,
you
know
so.
E
I
was,
as
I
said
before
it's
like
foot
for
Tess
kima.
We
need
to
add
stuff
only
if
we
need
to
test
it,
because
we
will
put
additional
pressure
on
people
commenting
like
example,
servers
and
we
put
additional
pressure
on
the
wipers.
So
if
you
put
subscription
somebody
dumped
them
for
my
subscription
or
I
gave
report
like
yeah,
so
I
would
keep
just
asking
my
minimal,
but
we
can,
as
a
separate
issue,
can
think
about
example,
schema
which
illustrate
thousand
feature
which
is,
for
example,
tasks.