►
From YouTube: Lang Team Meta WG 2019.06.06
Description
Meeting from 2019.06.06, wherein we dug more into what exploration and stages might look like, and took a closer look at the errors proposal. Towards the end we had the first glimmers of what a "exploration RFC template" could look like.
A
A
A
A
Maybe
it's
fine,
but
I
think
we
do
want
something
of
one,
but
in
particular
where
you're
coming
up
with
use
cases
and
examples
in
kind
of
exploring
the
range
of
possible
solutions
and
and
maybe
starting
to
look,
we
were
saying
that
using
tests
might
be
a
really
you're
not
doing
as
much
test
writing
now
that
we
realized.
We
had
so
sure.
It's
essentially
examples
right,
but
writing
helping
people
understand
it
by
example,
and
then
actually
with
examples
that
we
can
actually
take
to
me
consuming.
A
That
seems
like
the
best
of
the
options
or
a
couple,
and
you
try
to
drill
in
more
detail
and
I
feel
like
this
right
around
the
somewhere
in
the
midst
of
this
phase
and
sort
of
where
our
existing
R&D
process
kicks
in
you've
got
enough
to
write
a
document
and
you
kind
of
have
a
fairly
specific
proposal,
and
so
an
idea
to
be
sort
of
to
change
the
alignment
a
little
and
then
the
evaluation
stage
would
be
like
actually
implemented
and
you're.
A
Looking
for
people
to
you
have
you
would
like
the
feedback
on
it
and
ideally,
at
this
point
you
would
want
to
have
enough
explainers
and
documentation
to
actually
get
feedback.
And
of
course
stabilization
is
the
final
one
mm-hmm,
and
so
we
were
talking
about
like.
When
do
we
move?
How
do
we
move,
but
is
it
part
of
the
stage
start,
and
when
do
we
move
between
stages,
one
and
the
kind
of
entrance
criteria?
A
A
So
the
idea
is
that
you
would
bring
that
right
up
with
some
set
of
people
like
who
and
the
Lange
team
is
for
the
sponsor
or
something
some
people
that
will
do
the
implementation,
which
needs
to
be
enough
people
to
do
the
implementation
and
to
get
it
reviewed
and
there's
other
bits
of
work.
I,
don't
quite
know
how
we
can
come
for
them,
but
the
idea
is,
you
should
have
sort
of
us
enough
people
that
you
can.
You
could
convince
in
case
that
the
work,
because,
right
now,
we
often
have
this
problem.
D
Yeah
I
I
watched
back
the
last
times,
video
and
something
I
found
useful
for
myself
is
to
not
necessarily
phrase
it
in
like
what
is
required
to
enter
this
stage,
but
to
flip
it
around
and
be
like
what
are
the
deliverables
from
the
current
stage
so
for
exploration
it
would
have
like
this
is
what
you
should
have
after
the
explore
stage
and
what
should
be
done.
A
sort
of
hide
to
that
something
that
I
was
missing
was
deliverables
in
terms
of
communication
like
we
know
that
stabilization,
or
at
least
currently
for
async/await.
A
Remembering
that
I
spent
a
while
thinking
about
this
when
I
wrote
the
staged
artsy
proposals
some
time
ago
and
I
broke
it
up
like
this
on
purpose,
because
exactly
around
this
now
there
was
this
that
tc39
does
not
describe
it.
This
way
and
I
was
confusing,
that's
sort
of
what
will
you
be
doing
in
this
stage
and
then
what
do
you
do
to
exit
this
stage,
but
you
bring
and
and
so
on,
I,
like
the
idea
of
I,
think
that's
a
good
point
that
we
should
separate
those
things
and
I
think
I'm.
A
A
A
That
was
the
other
thing
I
was
trying
to
do.
Was
I
had
the
Bing
name
for
like,
while
you're
doing
it
and
now
informed
non
gerund,
I
guess
they're,
both
nouns
when
for
the
thing
that
you
present
but
I
guess,
ideally,
we
would
have
some
sort
of
like
template
for
these
things.
D
C
D
A
B
A
Might
be
different
beyond
that
is
the
matter
is
sort
of
bias
like
an
RFC
today
is
biased
towards
a
primary
solution
and
alternatives
and
I
think
it
might
be
helpful
with
we
can
fun,
we
can
make
things
more
neutral
for
is
more
of
it
right
of
kind
of
possible
directions,
and
there
is
probably
bound
to
be
some
well
I
guess.
Hopefully,
there
arises
like
a
certain
preference,
but
I
suspect
it
will
not
be
uncommon.
Meters,
I'm,.
B
The
language
is
the
language
or
to
rephrase
it
sort
of
sort
of
I
feel
like
the
language
of
the
bias.
So
for
us
for
the
first
time
in
captions
of
solutions
like
maybe,
if
you
take
something
from
a
school,
it
has
more
probable,
it's
more
probable
that
maybe
fifth,
a
trust.
This
we
have
already
have
a
bunch
of
stuff
from
like
straits,
taking
something
from
that
nights.
A
Maybe
maybe
I
think
it's
I
think
it's
helpful,
but
I
think
it
can
be
done
poorly,
like
I,
think
I
found
if
I
try
to
write
neutral
text.
Sometimes
it
has
no
guts.
It's
just
kinda
colorless
doesn't
really
say
what
the
pros
and
cons
are
because
it's
hard
to
present
them
in
unusual
life,
but
I've
been
trying
to
like
overcome
that
by
just
making
a
strong
case
in
both
directions.
Right,
but
I
think
that's
I.
Think.
D
It
might
be
useful
to
focus
less
on
a
proposal
of
solutions
at
this
stage
and
more
on
the
like
hundred
percent
be
like
this
is
the
problem
space
we're
looking
to
solve
and
I
as
a
addition
to
that,
I
was
like
here's,
so
impossible
things,
we're
thinking
of
exploring
or
have
explored
or
like
whatever
and
like
in
the
stage
after
that,
like
in
the
design
implementation,
you
can
strongly
say
like
hey,
we
chose
this
solution.
Here's
the
alternatives.
We
not
chosen
these
alternatives.
For
these
reason,
I
feel
like
the
eye.
A
B
B
B
A
A
The
moment
one
thing
I
was
gonna
say
is:
maybe
the
answer
is
that
yes,
that
the
sort
of
the
exact
but
there's
some
wiggle
room,
basically
like
you,
have
to
be
describing
problems
and
you're,
maybe
just
cracking
solutions
and
like
yeah,
yeah
yeah,
totally,
no,
no
I'm,
thinking
of
something
like
async
IO
or
something
where
the
problem
space
is
actually
really
fast.
In
some
sense
right,
I,
wonder,
I.
Think.
B
D
D
D
A
D
There's
I'd
say:
there's
the
thing
where,
if
you
want
to
have
like
a
generalized
arrow
type,
I
was
like
kind
of
painful
right
like
a
box
of
din
stood
air
plus
end
plus
sink
plus
static.
Right
was
kind
of
painful
to
write
out
every
single
time.
D
A
D
Then
a
like
a
third
one
which
may
or
may
not
be
controversial,
is
that,
like
annotating,
like
we
have
a
sink
annotation
for
like
a
dysfunctions
like
a
sink
given
how
common
error
is
maybe
there's
improvements
we
could
do
that
would
make
it
easier
to
spot
like
a.
This
is
the
actual
type
of
returning
here's,
the
arrow
type
right.
Sorry
I
was
like
the
thing
of
like
a
here's,
the
improvements
we
could
probably
look
at
you.
A
A
D
A
We
go
to
solutions.
I
want
to
talk
about
one
thing,
so
central
brought
up
the
everybody
agrees
and
I
think
that's
important
thing:
yeah
I,
don't
know
how
it
fits
in
I.
Think
it's
important
to
list
the
problems
and
I
think
it's
also
important
to
talk
about
like
why
they
named
it.
Why
people
think
their
home
the
why
they
don't
I,
don't
know
if
that
would
be
just
like
subheading,
like
certainly
in
this
base
in
particular,
there's
a
lot
of
disagreement
or
there's.
A
A
B
A
One
thing
I
tried,
I
sort
of
liked
it
I
did
a
deal
and
I
found
it
somewhat.
Successful
is
taking
what
I
call
the
narratives
right,
they're
kind
of
lay
out
the
sort
of
problems,
and
maybe
this
religions
neighbors
talked
about
those
in
a
second
and
then
try
to
say
like
here's
one
camp,
they
kind
of
feel
like
these
are
the
biggest
problems.
These
are
less
big
and
therefore,
then
and
I
find
out
pull
to
me
to
try
to
write
out
what
are
the
narratives.
Why
was
the
story
to
it
to
give.
A
The
nuances,
maybe
yeah
I
mean
I,
does
seem
risky
to
group
like
camps
proceed.
One
of
mindset,
maybe
but
I,
also
think
it's
like
useful
to
describe
what's
real,
but
there's
there's,
like
so
collected
different
ways
to
look
at
the
problem
and
anyway,
let's
talk
about
the
solutions,
so
we
actually
were
saying
something
like.
D
A
B
B
A
D
B
Solutions,
yes,
notably
we're
the
solutions
of
different
problems,
so
the
procedural
macro
is
something
that
you
could
do
in
the
library.
Presumably,
if
we
just
be
a
great
and
the
shorthand
for
for
trying
and
like
equals
to
mine,
would
be
nice
sugar,
perhaps
nice
sugar
for
much
and
one
one-line
functions
and
it's
not
specific
to
around
them,
but
it
helps
around
I.
D
A
Think
yeah
I
was
just
thinking
that,
first
of
all
that
we
have,
we
haven't
addressed
all
the
problems
and,
secondly,
that
it
feels
like
this.
It's
very
fine
grains
illusions,
but
you
sort
of
also
want
to
be
having
that
story.
I,
don't
know.
Maybe
it's
just
a
way.
The
way
you
presenting
a
common
set
of
things
you're
trying
to
show
I
think
this
is
how
it
would
look
and
work
in
these
different
ways.
D
So
I'm
kind
of
feeling
like
after
a
problem,
if
initial
you
don't
necessarily
come
up
with
like
one
RFC,
you
can
come
up
with
a
family
of
RFC's
like
multiple
separate
ones,
are
unrelated,
but
work
towards
the
same
solution
or
like
a
chain
or
you're
like
hey,
we
sort
of
see
a
future
path
here.
My
assumption
was
that
the
group.
A
A
Right,
but
they
should
somehow
all
there
is
a
like
a
coherence.
You
should
be
able
to
tell
a
coherent
story
of
like
here's,
where
we're
going
to
end
up
here.
Are
the
pieces
along
the
way,
I
think
a
good
example
of
this
is
infiltrate
where
I
think
we
have
not
done
an
amazing
job.
What
we
are
like,
we
have
a
lot
of
moving
parts
around
infiltrate.
Where
there's
a
lot
of
places
you
can
use
it
and
they're
each
sort
of
their
own
are
I've.
Seen
that's
good.
A
A
Exactly
this
is
very
close
to
what
I
wanted
to
see.
I
was
imagining
that,
just
to
sketch
out
what
that
could
look
like,
but
you
could
imagine
having
a
program.
That's
like
like
a
website,
that's
showing
all
the
implicates
places
you
could
put
it
with
like
links
for
each
one
kind
of
telling
you
like
in
this
position.
This
is
what
it
means:
here's
the
RFC,
the
current
status,
just
some
way
to
get
to
convey
that,
in
terms
of
speak,
only
infiltrate
kids
right,
it's
nervous
what
already
butchered
yet
yeah.
D
A
What
like
it's
more
the
beginning
of
the
design
and
the
end
of
exploration
like
it
should
well
the
reason
I'm
separating
that
is
I
think
there's
some
room
for
like
doing
exploration
where
we
don't
expect
to
start
writing
the
code
right
away.
My
Chronicle
example,
for
this
is
I,
want
us
to
think
about
what
it
means
to
anybody
with
residency,
but
I
don't
want
to
spend
time
writing
lessons
in
C
code,
because
there
is
no
residency
proposal,
but
I
wanna
have
a
story,
but
when
there
is
one
there
we
can
give
useful
feedback
and
I.
A
B
I
think
of
the
same
feature,
but
the
staging
implementation
staging
of
that
feature.
I,
don't
consider
infiltrate
embarrass
positions
to
be
different
features.
You
could
see
them
that
way,
but
I,
don't,
whereas
something
like
the
ergonomics
initiative
would
be
more
like
a
collection
of
features
working
towards
like
the
problems,
other
languages.
A
D
A
Is
like
the
best
thing
that
ever
exists
in
terms
of
showing
a
road
map
right,
it's
it's
bigger
than
a
lot
of
what
we
do,
but
not
all
of
it
like
you
could
imagine
it's
well.
Some
of
these
boxes
might
be
things
like
I,
don't
know
for
something
like
async
Rio
or
something
I.
Think
if
this
scale
is
not
but
trying
to
tackle
here
are
some
you
skied
in
series
and
pieces.
They
come.
B
B
A
Like
that's
the
point
of
it
essentially
right
knowing
that
there's
there
are
many
different
domains
of
things
you
might
do
with
lesson
and
here's
what
they
how
they
relate
to
one
another,
and
here
are
features
that
benefit
more
than
one
and
and
that
sort
of
biases
you're
trying
to
push
things
up
the
tree
to
the
most
to
the
right
spot.
So
I
think
the
goal
there
is
probably
to
have
skill
trees
around
an
area,
but
that
it
should
integrate
it
to
a
larger
skill
tree.
That
is
like
the
language
and
alright.
A
A
D
A
B
Interesting,
if
you
think
about
this
sort
of
Shane
and
trees
or
solutions
here,
is
that
the
if
you
just
have
the
short
time
function,
syntax,
where
you
say
e
equals
try,
but
you
don't
have
tried.
Then
you
can't
write
equals
try.
You
can
just
write
like
much,
but
if
you
want
to
try
than
just
to
interact
and
if
you
sort
of
would
have
point
and
phosphates
much,
then
we've
sort
of
gets
try
match
for
free,
more.
A
Less
that's
a
good
example
of
where
I
imagine
this
sort
of
iterative
nature
of
like
we
come
in
with
error
handling.
We
come
up
with
this
idea.
We
can.
There
are
problems
that
are
not
really
a
terre
handling
that
that
are
potentially
addressed
by
this
idea
and
I
feel
like
it's
legit
sort
of
in
some
way
to
talk
about
them.
Well,
one
thing
I
noticed
we're
missing
from
this
something
I've
heard
about
239
in
both
good
and
bad
ways.
A
Is
that
the
staging
process
since
were
bias
towards
action
because
it,
especially
as
you
move
towards
the
later
stages,
it
becomes
harder
to
business
like
well
the
time
for
that
was
pager,
which
is
a
good
thing,
but
also
maybe
a
bad
thing,
and
that
one
thing
we
don't
we're
not
talking
about
here
is
like
positives,
not
just
problems,
but
like
strengths.
We
want
to
preserve
or
dangers
or
something
I
think.
That's
all.
That's
kind
of
related
to
this
conservative.
C
C
A
Mean
for
everything,
what
might
it
be
in
it
I
think
one
might
be,
but
there's
a
I
think
there
is
a
transparency
towards
the
current
system
right
of
like,
in
other
words,
you
see
that
how
it
works
and
you
kind
of
can't
help
it
see
always
nothing
else.
They're
all
there
is,
but
you
see
that
errors
are
just
an
error,
is
a
type
and
if
you
can
jump
into
it,
I
don't
know
if
it's
something
we
necessarily
want
to
preserve.
There
I
think
it's.
B
A
I
think
yeah
it
feels
like
it
would
be
important
to
me
to
lay
out
that
argument.
It's
you
know,
but
you
know
the
good
way
like
separately
from
just
why,
okay
or
having
like
medicine
to
be
problems
like
it's
more
argument,
if
this
is
a
strength,
is
something
Josh
values
that
some
of
these
solutions
might
injure.
B
I
think,
in
terms
of
going
back
to
the
original
coins
about
about
the
Train
and
and
having
you
want
time
to
stop
it
once
you've
reached
a
certain
point.
One
thing
which
Mike
out
is
that
you
are
clear
about
sort
of
what
things
are
experimental
and
what
things
are
more
like
you're
sure
about
this
yeah.
A
A
We
think
this
is
a
good
solution,
because
X
Y
Z,
because
we
hypothesize
the
following,
like
that
this
will
be
very
common
or
did
this,
and
you
would
want
to
be
able
to
talk
about
that,
but
also
like
one
other
thing
that
I
remember
people
bringing
up
is
that
just
it's
very
useful
early
in
the
design
process.
Someone
might
raise
an
objection
like
well.
This
could
be
unsound
if
you
like,
failed
to
take
into
account
whatever
something.
B
C
C
C
D
Something
that's
very
interesting
from
the
tc39
state
proposal
or
a
staged
thing.
Is
that
there's
one
phase
where
it
requires,
if
possible?
Sometimes
it
isn't,
but
generally
it
is
to
have
a
babel
transform
where
babel
is
more
or
less
than
the
javascript
compiler
yeah
I
like
actually
I.
Should
air
quote
that
because
you
know
reasons,
but
it's
like
a.
A
D
I
think
there
is
a
stage
before
that:
yeah
like
sort
of
right,
because
I
feel
like
sometimes
like
nightly,
is
more
blessed
than
we
would
like
it
to
be
where,
if
something's
in
nightly
getting
it
out
of
nightly,
is
like,
oh,
you
know,
people's
built
break.
You
know
something
through
that
that
I
can,
but
if.
D
A
That's
actually
two
points,
that's
that's
a
really
good
point,
so
approximate
gross
and
the
other
thing
would
be
I
made
it's
more
of
a
general
thing,
but
looking
at
in
general,
factoring
out
I
think
around
the
solution.
Space
like
we
might
bias
towards
solutions
that
enable
experimentation
via
proc
micros,
to
feed
into
leaders.
B
B
A
B
A
B
Well,
so
in
terms
of
exploration,
then
macro
is
probably
too
soon
as
well.
That
feels
more
like
checking
whether
the
implementation
makes
sense
talk.
Macro
ship
is
probably
not
a
good
idea
to
try
out
whether
the
solution
makes
sense,
because
it
won't
feel
but
actually
different
if
it's
implemented
in
the
language,
whereas
I
have
implemented
in
in
in
I,
have.
D
But
so
something
we
were
thinking
of
adding
is
like
hey
what
what
if
we
could
like
add
arguments
to
Maine
right,
so
you
can
give
it
like
implemented
traits,
and
then
you
can
pass
arguments
remain.
It's
kinda
nice
for
command-line
programs
right.
That
was
very
well
to
implement
like
using
protocol,
and
we
were
like
oh
cool.
Let's
built
this,
try
it
out
gather
feedback
like
let's
give
it
six
months
or
something.
D
A
A
D
Yeah,
like
people
seem
to
like
it
so
far,
but
the
second
one
we're
doing
is
like
using
doc
comments
to
add
context,
attributes
to
all
arrow
types.
So
if
we
go
back
to
like
the
Airy
stuff
with
what
was
it
not
the
back
traces
but
the
other
one,
the
annotations
right,
so
you
like
use
doc
comments
for
that
purpose.
Like
hey,
is
this
great?
We
don't
know,
let's
try
it
out
right,
so
we're
currently
also
monitoring
that
and
there's
kind
of
an
experiments.
B
C
B
D
It's
it's
very
much
not
like.
Would
a
macro
solution
suffice
but
like
what
would
a
possible
language
feature
feel
like
right,
the
closest
we
can
get
to
with
that?
Like
start
messing
around
with
a
compiler
just
like
with
the
async/await
syntax,
it
was
first
introduced
as
a
prop
macro,
which
is
an
ideal,
but
you
get
a
feel
for
how
this
would
be,
which
helps
inform
help
strengthen
like
the
direction
you're.
D
Taking
sorry
sure
it
just
helps,
inform
the
conversation
right,
because
people
actually
get
to
try
it
out,
not
saying
that
this
is
the
only
way
to
do
it
not
saying
this
is
applicable
to
every
language
proposal,
but
I
think
it
has
its
place
and
if
possible,
it
ought
to
be
encouraged.
Probably
very
sugary.
Things.
A
D
A
I
guess
to
make
you
can't
well,
you
could
see
even
something
like
canary
disoriented,
I
guess
plausible.
If
you
could
read
for
cedral
members,
there
are
they.
There
argue
shipping's,
there's
it's
very
complex,
but
if
you're
our
limits
now
far,
you
can
go
along
these
lines
regardless.
I
want
to
use
the
last
four
minutes
to
look
at
this
and
see
if
we
actually
got
on
something.
When
that
my
feeling
is
so
I
separated
out
problems
in
and
around
the
space
positives
and
strengths
to
consider
solutions
and
how
they
relate
to
one
another
unknowns
and
constraints.
A
Things
to
be
resolved
later
and
I
also
have
this
road
map.
Recommendation
narrative
I
have
a
lot
of
question
marks
here,
because
I
think
this
part
is
by
far
the
least
clear
and
I
suspect.
That's
okay,
like
I,
think
he
can
vary
a
lot
between
proposals,
but
I
I.
Definitely
think
the
idea
of
saying
you
you
might
come
in
with
I
have
a
lot
of
answers
or
a
lot
of
questions.
There
are
more
questions
than
answers.
There's
like
the
space
of
solutions.
We
ourselves,
don't
know
which
one
is
best.
A
We
want
to
run
these
experiments
and
so
on.
That's
probably
something
that
should
still
be
considered
exploration,
though
I
don't
know,
I'm
trying
to
figure
out
how
much
of
an
answer
you
need
here
before
you
move
out
of
I,
don't
know
it
also
could
be
something
that
me
just
try
it
a
few
times
and
shape
the
answer
based
on
how
it
seems
to
be
working.
A
Like
one,
if
we
agree,
this
is
sort
of
vaguely
good,
there
is
I.
I
would
think
that
some
homework
we
could
do
would
be
trying
to
write
actually
go
ahead
and
write
this
right
up
for
a
few
things
we
sort
of.
Did
it
bear
handling
a
little
bit
here,
but
I
would
be
interested
in
trying
it
out
on
some
other
problems,
I'm
also
giving
it
to
Josh
to
do
for
clarity,
FFI,
which
I
I
know
he
already
sort
of.
Has
a
road
map
up
there,
but
I
think
Josh
would
be
excited
about
it
too.
A
D
A
B
A
B
A
Be
happy
if
you
want
to
point
at
other
things
by
the
way
it
happens
to
be
a
well
like
it's
a
good
example
in
a
world
document,
one
but
I,
don't
think
so
like
in
particular.
They
had
this
website,
for
example.
This
is
like
something
I
obviously
strongly
wish.
We
had
communicating,
there's
a
bunch
of
proposals.
Here's
the
author's
champions
test
status,
spectra
dates,
different
state
deals
and
you
can
click
on
one
like
oh
I'd
like
to
learn
about
decorators,
and
then
you
go
to
its
associated
repo.