►
From YouTube: Node js API WG Meeting 2015-10-23
Description
notes at https://github.com/nodejs/api/blob/master/wg-meetings/2015-10-23.md
A
A
A
couple
of
simple
things
will
start
with
some
introductions
for
review,
but
everybody
to
just
say
who
they
are
and
briefly
what
they
do.
I'll
go
ahead
and
start
on
patrick
miller.
I
work
at
med
source
I.
Do
a
lot
of
notes.
Stuff,
API
stuff
is
of
interest
to
me
and
so
I
thought
I
would
go
ahead
and
contribute
a
little
bit
by
organizing
this
call.
So
that's
me.
B
C
D
F
B
B
J
A
J
Interested
really
in
there
from
same
perspective
as
ours
is
there
in
Jack
score?
Oh
I'm,
writing
in
API
for
chrome,
packaged
apps,
it's
pretty
complete
and
and
had
a
long.
It
took
a
long
time
doing
it
because
there
was
no
testing
framework.
There
was
no
seventh
the
libraries
to
pull
in
the
same
way
as
we
do
with
leveldb
and
saw
this
crop
up
or
the
pic
chance,
contribute
my
findings
and
see.
What's
going
on.
A
Great
okay,
I
think
we
got
everybody,
so
I'm
I'm
really
new
to
this.
I
just
kind
of
watched
it
from
afar,
always
back
and
don't
don't
know
a
whole
lot,
so
I'm
not
going
to
be
like
driving
any
major
discussions
or
anything,
but
in
looking
at
just
what's
out
there
right
now
and
I've
since
I've
been
involved
in
a
couple
other
work
groups,
I
want
to
bring
up
a
couple
meta
issues.
I
have
those
up
on
the
on
my
screen
now
you
know
how
to
present
that
to
everybody.
This
is
just
the
google
talk.
A
This
pointed
to
in
the
chat
group
chat,
so
the
meta
issues
are
first
of
all,
I
I,
guess,
there's
some
kind
of
a
status
that
has
the
work
groups
have
in
terms
of
being
chartered
or
not.
This
is
something
I
just
found
out
about
yesterday
and
I.
Don't
really
know
what's
involved,
so
I
was
curious.
If
anybody
knows
if
this
working
group
has
already
been
chartered
or
the
idea
is,
this
is
a
and.
H
H
G
Need
to
be
chartered
chartering
basically
is
just
like
when
a
working
group
work
reaches
a
certain
point,
they
can
like
basically
have
authority
over
a
certain
point
from
the
tsc,
but
I
don't
think
we
need
that
yet.
Well,.
A
Okay,
that's
all
good
news
and
I
was
going
to
say
even
if
we
wanted,
you
know
become
chartered
certainly
seems
like
something
would
be
kind
of
a
low
priority
anyway,
and
I
could
start
to
look
into
it.
But
let's
just
move
away
from
that.
The
other
thing
I
was
wondering
the
the
name
on
the
repo
is
kind
of
weird,
because
its
ends
up
being
no
jsapi.
So
you
makes
you
kind
of
think
oh
well.
A
This
must
be
the
nodejs
api's
like
from
user
land,
right,
FS
and
HTTP
and
whatnot,
but
it's
clearly
something
a
bit
different
from
that
and
or
more
involved
so
I'm
wondering
if
we
want
to
actually
change
that
we've
I
mean
to
kind
of
reflect
that
it's
more
of
a
working
group
kind
of
thing,
then
something
that
that
is
just
API
and
with
that
in
my
out
so
yeah
so
I
mean.
Does
that
okay
or
I
mean
wouldn't
look
pop-up
I.
A
A
H
You
know
your
API
needs
to
return
this
thing
and
unfortunately
we
can't
it's
impossible
to
support
every
javascript
api
out
there,
and
so
the
purpose
of
the
javascript
api
is
to
remove
as
much
abstraction,
as
is
logical,
because
I
already
attempted
the
whole
libby
v
KS
thing
and
it
just
doesn't
work
nicely.
There
needs
to
be
a
small
amount
of
abstraction
for
it
to
be
usable,
but
to
remove
as
much
as
possible
and
use
the
fastest
methods
that
be.
H
It
gives
us
to
achieve
that,
so
that
anybody
who
wants
to
attach
their
own
api
or
interface
into
nodes
and
kernels
from
the
JavaScript
side
can
do
that.
The
other
one
other
purpose
of
the
this
is
to
make
it
easier
to
test.
Thirdly,
we
have
companies
like
Oracle
and
others
who
want
to
like
make
their
own
engines
underneath
right.
They
don't
even
use
v8,
and
so
they
want
a
clean
cut
between
the
native
and
the
JavaScript
side
and
then
a
clean
test
suite
that
they
can
run
to
validate
that
their
implementation
is
correct.
H
C
D
May
I
have
the
side
of
native
modules
that
no
Jas
exposes
an
epi
to
Native
module
auto.
That
also
exposes
the
v8
and
the
pvap
is,
and
so
is
subject
to
the
changes
that
in
the
in
the
v8
API
and
even
with
the
abstraction
layer
like
nan
there,
the
ABI
is
constantly
changing.
So
it
requires
rebuilding
modules,
a
deployment
time,
which
is
arguably
a
very
bad
deployment
story.
So
that's
another
problem
that
I
think
we
should
try
to
solve.
C
Now,
I,
in
my
impression
what
we
were
going
to
tackle
two
of
those
three,
so
you
wouldn't
necessarily
cover
the
external
api's
and
changes
to
those,
but
it
was
more
focused
on
the
two
other
layers.
You
know
the
one
that,
if
you're,
implementing
an
API
or
a
service
and
note
you
know
the
internal
set
of
api's
that
you
could
use
for
that
and
then
at
the
lower
layer.
The
sort
of
glue
between
the
v8
engine-
that's
seen
both
by
node
and
by
the
Martha.
H
H
Yeah
problem
is
some
hunters
that
use
in
v8
are
not
available,
for
example,
hooking
into
the
garbage
collectors
pro
prologue
and
resurrecting
the
object
just
before
it's
garbage
collected
right,
I,
pretty
sure
that,
like
spider
monkey
can't
do
that
I
think
I
II
might
have
just
barely
implement
to
that
because
of
their
IOT
push
and
getting
no
to
work
on
those.
But
anyway
it's
not
available
everywhere.
C
I
H
C
E
C
Yeah,
I
guess
I
was
thinking
to
two
different
ones,
one
which
is
the
one
you
just
said,
and
then
one
which
is
you
know
both
no
talking
to
you
know,
maybe
it's
because
I
think
some
of
these
are
the
same,
but
like
no
no
uses
some
be
a
teacher's,
a
layer
there
I
might
as
well.
As
you
know,
he
features
and
functions
are
exposed
to
launch
of
today's
to
wrap
both
of
those
you
end
up
having
a
separation
from
the
actual
JavaScript
yeah.
H
So
in
previous
discussions
on
threads
on
or
we
could
probably
abstract
the
current
node
API
for
module
authors,
because
that
doesn't
depend
on
many
v8
internals
right,
but
as
far
as
the
layer
that
we
use
between,
like
the
node
internal
libs
and
the
via
API
it.
It
just
won't
work
across
all
platforms,
right,
yeah,.
D
So
Trevor
those
features
that
don't
work
across
all
platforms,
I
mean:
is
it
possible
to
have
an
alternative
implementation
or
to
isolate
the
features
that
depend
on
it
so
that
we
can
say?
Okay,
if
maybe
this
feature
is
not
available
on
this
platform,
but
at
least
you
can
get
the
bulk
of
node
running
well,.
H
C
There's
also
two
aspects
to
that
layer:
one
is
that
you
know,
even
if
you
continue
to
use
the
eight,
if
you
instruct
that
layer,
you
can
get
to
the
point
where
you
can
help
when
you
go
to
the
next
level
v,
you
don't
have
a
chain.
You
know
you
don't
have
to
change
your
modules
or
necessarily
know
to
keep
up
with
that.
So
there's
still
a
benefit
looking
at
it
from
that
perspective,
and
then
once
you
have,
the
API
defined
those
specific
issues
that
you
know
that
you're
calling
out
I
think
will
stand
out
more.
C
C
C
F
H
B
H
Basically,
node
it
as
it
stands
now,
the
lib
folder
would
just
be
rewritten
to
sit
on
top
of
this
new
level,
JavaScript
API
and
then
extracting.
If
we
were
to
abstract
v8
itself
from
even
the
low-level
JavaScript
API,
then
we
essentially
wouldn't
be
note
any
more.
We
could
fork
this
off
call
another
project
and
tight,
like
literally.
This
would
not
be
node
anymore
right
just
because
it
then
all
like
then
note
essentially,
is
the
lib
folder
in
the
current
repository.
H
We
have
we're
completely
yeah
we're
completely
changing
a
module
interface,
we're
completely
obstructing
way,
v8
altogether
we're
completely
changing
the
JavaScript
interface,
so
that
node
as
it
stands
today,
the
JavaScript
interface
can
sit
on
top
of
that,
like
we're
talking
about
so
much
abstraction
that
it
is
literally
no
longer
node
right
is
its
own
thing
completely.
We
could
do
all
this
work
outside
and
get
it
running
and
node
wouldn't
even
have
to
be
involved.
H
Okay,
abstraction
of
abstraction
of
the
module
API.
It
is
going
to
be
a
fair
amount
of
work
right,
especially
because
there
are
so
many
opinions
of
how
it
should
be
done.
Honestly,
I
think
the
bike
shedding
that
will
be
involved
in
doing
that.
We,
the
majority
of
work,
the
abstraction
of
v8
and
the
JavaScript
API
itself,
is
going
to
be
tremendous.
You
might
as
well
start
from
scratch,
define
the
lower
level
API
and
build
it
directly
on
top
of
the
extracted
API.
H
What
I
mean
that's,
essentially
what
we're
doing
right
that
is
like
and
then
what
we
do
after
we
create
the
abstractions.
Is
we
write
code
to
hook
v8
into
it,
and
we
rewrite
the
lib
folder
to
sit
on
top
of
the
JavaScript
API
and
we're
done
that's
essentially
what
we're?
What
we're
talking
about
doing.
D
F
Yeah
actually,
recently,
right
now
we
have
me
a
328
VA,
314
chakra,
actually,
its
uses
328
and
Mozilla
Firefox.
Sorry,
a
spider
monkey
sharing
the
same
interface
same
sort
code
base
by
using
different
proxies,
the
only
tricky
part
we
are
having.
It's
really
tricky
the
types,
the
native
types
they
they
are
really
tricky
to
handle.
F
Sometimes,
actually
we
could
have
some
sort
of
shimming
approach,
but
because
of
the
types
and
half
routing
is
different
among
the
engines,
it's
really
it's
really
tricky
part.
That's
why
we
depend
on
macros
a
lot
I,
don't
know
if
it's
a
good
starting
point
or
not,
probably
it's
open
source
you
can
see,
and
we
can
discuss
all
that,
but
I
am
also
open.
I'm
also
totally
agree
with
the
other
approach.
Today
we
can
do
it
from
scratch.
That's
also
fine.
D
F
Well,
actually,
after
proxying
three
different
engines,
my
own
opinion
is
I.
Don't
think
it's
possible
to
have
something
for
very
long
term
kind
of
a
bi-stable
I,
don't
know
if
it's
possible,
but
we
can
have
api,
stable,
I.
Think
it's
a
a
better
goal
to
have,
because
there
are
lots
of
native
modules
around
and
on
each
knee
of
node.js
release.
Some
people
are
expecting
to
know
module
developers
to
update
to
latest
not
J's,
API
or
v8
API.
Sometimes
it's
not
possible.
F
Sometimes
there
are
some
stability
issues
because
of
the
changes
and
that's
that's
the
crazy
part.
For
example,
if
you
don't
write
it
or
updated
by
our
own
using
spider
monkey
engine,
almost
none
of
the
native
motors
are
right
now
working
on
Josh.
Only
if
the
user,
you
pics,
we
ate
three
point:
14
release.
Okay,
they
are
working
with
check
our
other
than
that
they
are
totally
auto
auction.
D
F
F
F
D
F
Meanest,
no
no
I
know
I,
don't
use
the
API
from
api.
What
I
say,
for
example,
in
some
cases
they
say
we
depend
on
like
nine
or
JX
go
using
macros,
let's
say
or
some
inline
functions
but
considers
we
may
change
a
little
bit
or
heavily
change
them.
What
or
what
they
are
doing
because
of
the
engine
updates.
D
E
A
C
A
C
A
C
H
Yeah
and
that's
what
man
tries
to
achieve
right
like
API,
to
not
a
B
I'm,
just
API
compatibility,
and
they
did
pretty
well
right
up
until
they
decided
to
introduce
the
maybe
locals,
which
just
destroyed
everything,
but
not
like
that.
They
like
eat
eight
to
twelve
months,
they're
able
to
maintain
a
plaque
compatibility
which,
from
our
track
record,
is
really
good.
But.
D
If
we
could
piggyback
on
that,
you
try
cracker
then
say
achieve
the
same.
Api
eligibility
would
be
a
face
that
was
not
ready
to
one
another
approach
what's
worth
considering
is
that
of
not
exposing
so
much
of
the
entry
in
eternal
as
much
as
we
do
now,
but
use
a
foreign
function
interface,
which
believe
for
the
great
majority
of
native
modules?
There
are
not
there
to
extend
notes
functionality
in
the
sense
of
exposing
additional,
no
internal
functionality.
It
could
be
sufficient.
D
A
A
We
also
had
ffi
and
turns
out
nobody
ever
used
the
internal
yeah,
because
the
FFI,
which
is
so
much
easier
to
use
right,
you
didn't
have
to
write
any
C
code
at
all,
so
that
I
think
that's
actually
a
great
idea,
but
that's
not
kind
of
the
way
na
manja
lobsters
kind
of
think,
and
we
don't
have
an
epiphyte
and
core
yet
but
I
I
honestly
do
think
that
is
a
pretty
good
idea.
So.
H
H
Like
the
type
coercion,
an
object
munching
you
have
to
do
under
the
hood
is
super
easy,
expensive,
comparatively
and
feel
like
the
it's,
not
even
the
cost
of
the
FO
fire
in
the
JavaScript.
It's
a
cost
like
maybe
another
engine
could
do
it
faster,
but
v8
at
least
being
able
to
do
all
those
conversions
into
the
interface
on
the
natives.
Iphone
javascript
is
just
too
slow.
I
mean.
D
A
Epiphyte
typically
has
like
fairly
f
of
ice,
typically
have
kind
of
fairly
restructuring
capabilities
that
the
current
ffi
does.
They
actually
have
a
separate.
They
called
resources
or
something
let
you
essentially
build,
strikes
and
deal
with
pointers
and
players
instruction,
etc.
So
it
can
get
rich
to
trevors,
know,
I'm
sure
them
I'm
sure
the
perf
is
bad.
I,
totally
trust
you
in
that
space.
A
However,
I
did
some
timing
a
while
back
on
it
and
yeah.
It
is
slow.
However,
you
know
if
you
look
at
where,
like
in
the
past,
where
I've
used
f
of
5
was
for
things
like
database
in
vacations
and
stuff
like
that,
where
the
time,
because
we're
not
talking
about
it's
taking,
like
you,
know,
hundreds
or
even
tens
of
milliseconds
to
make
these
calls
it's
a
couple
milliseconds.
A
So
if
you,
if
you
look
at
that
in
the
context
of
these,
like
Amongus
lee
long
running
things
thats
that
particular
time
doesn't
make
a
difference,
so
I
understand
that
it
certainly
does
like
you
wouldn't
want
to
build
a
FS
employment
at
this
moment,
but
it
might
be
appropriate
for
for
some
folks
to
do
it.
I'm
not
a
scientist.
Somebody
already
handling
ffi
this
you
know,
should
we
do
fi
a
note,
or
is
that
another
thing
that
we
should
kind
of
there's.
H
Been
a
discussion
/?
How
about
there's
been
a
discussion
on
that
in
more
though
in
a
PR
actually,
and
we.
H
The
the
implementation
in
the
PR
was
kind
of
scattered,
and
we
said
if
ffi
is
implemented,
it
has
to
be
localized
just
the
ffff
I
module
because
of
the
amount
of
exploits
that
are
available
through
that
I
mean
using
that
I
could
actually
write
a
string
to
memory
and
then
execute
it
from
JavaScript.
Yes,
it's
dangerous,
absolutely,
yes,
yeah!
So
yeah.
We
basically
decided
for
any
enterprise
customers
to
make
it
simpler.
It's
like
okay,
you
just
eat
the
FFI
module
and
startups
and
nobody
has
access
to
it.
H
G
No,
the
other
thing
about
ffi.
Is
it
can't
be
our
only
option
because,
as
I
mentioned,
there's
people
who
use
this
stuff
on
machines
are
much
lower
power
that,
even
if
your
machine
can
do
it
relatively
faster,
what
you're
doing
it
for
some
of
these
robotics
types
applications
simply
can't.
It
can't
do
that
on
the
platforms,
a
run
on
yeah.
D
E
J
A
You
know
essentially-
and
that
might
be
another
thing
that
we
want
to
look
at
and
that
might
end
up
being
sort
of
a
bridge
for
some
of
these
kind
of
diagnostic
api's
that
we
have
in
the
user
land
today.
That
probably
don't
belong
there,
because
they're
more
like
in
perspective,
unplugging
diagnostic
oncology,
not
application,
building
things
so
I
think
if
we
actually
had
something
like
that,
that
was
useful
would
probably
make
sense
to
make
some
of
those
existing
api's
out.
Well,.
H
H
We
are,
we
very
well
might
be
bringing
in
v8
inspector.
There
will
be
pulling
it
out
of
the
chromium
project
and
making
it
a
self-maintaining
module
so
that
we
can
actually
bring
that
in
and
then
that
way
we
can
just
send
out
all
the
debugging
information
to
using
standard
API.
Now,
if
that's
brought
in
there's
already
a
ton
of
tooling
around
that,
because
it
lives
in
chrome
but
also
all
of
their
javascript
files
are
also
stand
alone.
So
you
can
embed
it
into
anything
else.
H
I
Think
we
need
like
two
separate
buckets
that
one
is
like
this
is
our
promise
to
API
that,
like
we
can
just
stick
this
folder
of
stuff
on
top
of
your
lower
level
thing
and
it'll
work,
and
then
another
stuffed
it's
just
like
implementation-specific
stuffs
like
we
already
have
that
like
v8
file,
that,
like
you
said,
has
a
bunch
of
stuff
in
it
that
is
v8
specific
and,
like
all
this
extra
debugging
stuff,
there's
always
going
to
be
platform.
Specific
things.
B
A
A
F
H
H
They're
right
there
you
bring
up
a
really
good
point:
I
mean
there's,
there's
plenty
of
language,
specific
features
that
we
can't
control
you
that
you'd
have
to
like
go
back
in
time
to
a
specific
version
of
the
engine,
so
that
all
of
them
support
the
same
feature
set
right
like
the
support
maps
and
set
some
of
that
spider
monkey
supports
freaking
everything
they
are.
It's
a
developer's
playground.
I
remember
working
Mozilla
there
were
features
implemented
that
weren't
even
a
speck.
Yet
for.
H
D
Hopefully,
if
the
belt
between
the
engines
is
not
too
huge,
some
compile
time
switching
could
be
visible
like
to
enable
some
newer
features
under
if
death.
D
H
The
ok
so
in
VA-
and
this
has
been
proven
many
times,
the
only
anything
behind
the
flag
will
break.
We've
had
large
number
reports,
people
trying
harmony
whatever
and
then
the
process
of
courts
or
rashes
or
whatever
right
like
none
of
nobody
in
the
tsc
would
ever
advocate
for
enabling
those,
because
they
will
crash.
In
some
scenario
the
owner
makes,
I
think.
C
That
may
be
slightly
different
in
canadian
misunderstanding,
but
I
think
it's
like
a
runtime
flag
that
says,
if
you're
on
this
engine,
I'm
going
to
do
something.
If
I'm
a
different
engine,
I'll
do
something
else
is
supposed
to
being
on
or
off.
And
then
you
know
if
we
are
trying
to
support
multiple
engines
with
our
CI
to
run
the
test
we
on
all
of
it,
but.
C
C
Is
okay,
you've
got
the
same
problem
in
the
modules
and,
like
you
said,
I
mean
I.
Think
at
that
level
it's
kind
of
a
documentation
anywhere.
You
know
if
you
want
your
module
to
let
across
you're
either
gonna
have
to
do
the
same
kind
of
thing
where
you're
aware
and
can
see
which
engine
you're
on
and
choose
d
the
future
or
not
or
you
I
guess,
come
to
the
most
common
denominator
in
delft
use
certain
peak.
H
E
D
F
F
Yes,
stronger,
let's
implemented
on
spider
monkey
I,
don't
know,
for
example,
I
mean
what
we
are
doing
all
what
we
are
going
to
do
is
kind
of
hidden
related.
What
needs
to
be
introduced
into
node.js
coffee
right
because,
let's
say
today:
strong,
not
yeah
the
other
day
extra
mod
and
hey.
We
decided
to
introduce
extra
mode,
okay,
api
phase,
our
API
ting
everything
face
because
there
is
no
extra
more
days
of
shooting
exams.
It's
just
something
good
for
v8
or
something
they
want
to
do
for
some
reason.
C
D
B
C
E
H
My
what
I
was
thinking
well,
one
of
the
other
reasons
why
I
wanted
hit
the
JavaScript
API
first
is
because
I
felt
like
if
all
of
the
right
now
node
has
some
API
is
that
go
directly
in
the
native.
But
if
everything
can
be
a
JavaScript
call
first,
then
anybody
can
write
something
underneath
to
accept
those
calls
and
basically
asked
the
test
and
I
figured
the
easiest
barrier
to
entry
would
at
least
say:
okay,
we
can
run
all
JavaScript
modules
on
whatever.
I
D
D
H
Yeah,
because
in
in
in
the
worst-case
scenario
right,
if,
if
you
didn't
want
to
say
you're
working
out
of
fraction
and
you're
finding,
it
has
performance
issues
whatever
or
also
I
mean
there
are
many
modules
out
there
today
that
either
are
or
used
other
native
modules
and
the
likelihood
of
them.
Switching
to
a
numerical
api
is
low.
It
would
more
depend
on
new
modules
being
read
to
do
so.
D
D
D
C
A
So
I
think
we
probably
have
some
work
to
do
in
terms
of
figuring
out
what
the
chunky
bits
here
are.
There's
kind
of
you
know
what
Trevor
said
he
you
know
was
was
wanting
to
work
on
and
that's
some
very
interesting
work
and
then
there's
other
things
that
have
been
discussed
that
maybe
other
other
people
can
kind
of
spend
a
little
bit
of
time
on
think
about
and
write
some
notes
up
on
so
well.
I
think
we
should
try
to
we've
got
a
few
minutes
here,
try
to
settle
on
on
maybe
a
couple
of
those.
A
A
Yeah,
once
a
month
is
kind
of
my
default
take
on
how
often
to
do
things
like
this
looking
out,
we
probably
we
have
the
u.s.
holidays.
The
Thanksgiving
is
coming
up
a
month
and
that,
of
course,
December's
kind
of
it
that
a
lot
of
people
off
so
we
probably
sorry
have
time
to
maybe
fit
in
a
meeting
somewhere
there,
either
before
thanksgiving
or
very
early
december.
So
all
I
will
start
to
put
a
doodle
together
for
that
next
meeting,
I
just
for
some
dates
that
won't
affect
too
many
people
and
then
about
the
action
items.
D
A
F
Interesting
thing,
actually,
which
is
say
if
you
are
going
to
do
this-
we
are
going
to
do
something
totally
separate
and
it
is
not
going
to
be
not
or
actual
note
itself.
I,
don't
know
if
it's
very
early
to
say,
but
do
you
think
it's
possible
to
make
ever
or
call
it
like
not
five
point
or
not
six
point
or
something
so
this
work
group
has
a
mission
to
accomplish
right.
Instead
of
okay,
we
are
going
to
do
something
in
uncivil
future.
F
Instead
of
that,
can
we
see
or
can
we
define
the
thing?
Okay?
This
can
be
not
five
point,
because
this
is
a
breaking
change,
whatever
working
for
not
4.2,
long
term
release,
it's
not
going
to
work
with
this
time,
and
this
will
be
the
actually
real
long-term
release
because
after
this
point,
any
module
any
applications.
Switching
to
this
version
of
note
will
be
compatible
to
future
versions
of
not
right.
A
It
might
be
a
good
idea.
That's
come
up
with
a
completely
different
name.
Oh
you
know,
that's
not
that
great
of
a
name
as
Trevor
mentioned
this.
Isn't
this
doesn't
know.
This
is
really
you
know
it's.
It's
kind
of
a
node
compatible
thing
or
dancing
attention,
but
it's
not
really
the
note
that
we've
been
using
all
these
years,
but.
C
A
The
other
thing
I
was
going
to
add,
is
you
know
it's
way
too
early
to
be
suggesting
anything
like
that
right,
but
I
mean
just
as
a
kind
of
like
thinking
far
ahead.
What
can
possibly
happen
and
it'd
be
more
like
a
Python
2
versus
python3
switch
or
something
I
mean
a
difference
right.
The
language
probably
have
all
the
language
will
change
by
then.
But
you
know
it's
that
sort
of
a
sea
change
in
terms
of
the
differences
in.
G
G
C
H
H
Yeah
I
mean
yeah
I
mean
as
a
starting
point.
Would
it
be
possible
for
anybody
coming
from
other
backgrounds
to
look
at
man
and
be
like
we
could
like
on
this?
It
like
Jim
our
stuff
into
name
right
right.
If
nan
has
to
make
some
adjustments,
shirt
Australian
like
make
those
adjustments,
I
think
that'd
be
a
feasible
starting
point,
saying
that
we're
going
to
replace
all
the
internals
I'm,
not
even
saying
I'm,
opposed
to
that
I'm
saying
with
that
at
right.
H
D
H
A
A
Maybe
reorganize
this
into
like
a
docs
directory
and
if
anything
becomes,
you
know
a
thing
in
and
of
itself,
and
it
works
out
I
kind
of
hate
to
start
just
creating
like
new
repos
all
over
the
place,
without
really
having
a
history
behind
it.
That
we
need
to
do
that
because
oftentimes,
they
kind
of
peter
out.
H
Yeah
this,
the
the
oh,
my
god,
son,
shocker
anymore.
What's
it
called
now.
H
Know
it's
it's
for
edge
right.
What
was
to
be
ancient
for
edge
chakra?
Sorry,
it's
still
chakra
I
thought
they'd
replace.
Okay
anyway,
the
team
for
the
IOT
devices
actually
wrote
a
crazy,
impressive
compatibility
sioner
for
v8
to
get
it
working
which
are
I,
was
honestly
really
impressed
by
and
I
would
love
to
hear
feedback
from
them
and
get
it
like
get
some
information
back
from
them
about
how
it
was
how
it
went.
H
I
mean
one
thing
is:
I
was
actually
I
was
in
the
thread
of
the
what
of
the
the
new
was
spec,
that's
being
written,
and
one
of
the
I
was
talking
about
how
not
all
browsers
have
garbage
collection,
detection
or
object
detection
where
you
can
notify
and
one
of
the
Microsoft
dudes
jumped
in
and
said.
Actually
we
support
that
now
because
we
had
to
write
the
famous
burger
for
node,
so
I
I
know
that
they
did
a
large
amount
of
work
to
get
that
working
and
I'd
love
their
feedback.
No.
D
A
A
A
All
right
so
will
get
the
meeting
minutes
out,
get
the
YouTube
movie
published
and
start
thinking
about
next
meeting
and
looking
at
billions
for
information
about
what's
going
on
here
over
the
next
month,
so
locked
everybody
next
month.
Probably
you
guys
everyone.