►
From YouTube: Node.js User Feedback Initiative Meeting
Description
A
A
A
So
welcome
everybody
to
the
no
jess
user
feedback
initiative.
We
have
a
lively
discussion
that
we
are
carrying
over
from
github
and
it's
been
enjoyable
to
sort
of
dive
in
and-
and
you
know,
take
some
of
the
feedback
from
users
and
share
some
of
the
understanding,
so
we're
joined
today
by
Aaron,
Bartel
and
Aaron.
You've
teed
up
a
you
know
fairly
a
broad
set
of
topics
for
discussion,
and
you
know
why
don't
why?
Don't
you
give
us
a
bit
of
overview
about
yourself
and
then
you
know
help
us
choose.
B
So
my
name
is
Aaron
Bartel
I
work,
primarily
with
no
gsib,
my
platform,
formerly
known
as
iSeries
and
as/400.
We
just
got
up
to
version
8
of
node
on
IV,
my
but
then
also
as
of
April.
24Th
can
compile
it
ourselves
and
I.
Be
that's
very
good.
The
challenge
that
we're
having
and
what's
in
the
github
issue
is
that
and
IBM
is
traditionally
used
by
a
lot
of
insurance
companies.
Auto
manufacturers,
banks,
etcetera.
So
these
are
medium
to
large
companies
that
are
looking
at.
B
What's
my
modernization
path,
they're
looking
at
nodejs
they've,
traditionally
adopted
either
RPG
the
primary
language
in
the
platform
or
Java
they're,
seeing
some
of
the
same
things
with
Java
that
they're,
considering
it
a
little
long
in
the
tooth,
but
they're
not
it's,
also
very
solid,
so
like
well.
What
do
we
want
to
move
to
next
and
they're
reviewing
nodejs
but
are
struggling
through
all
the
things
that
they're
reading
online?
So
here's
the
the
big
issue,
some
of
them
were
true
at
a
point
in
time.
Some
are
no
longer
true.
Some
are
still
true.
A
Yeah,
let's,
let's
hit
threading,
so
you
know
that's
it's
a
discussion
that
you
know
in
general,
we've
had
you
know
round
node,
and
you
know
this
is
an
interesting
phenomenon.
You
know
happening
in
the
world
around
us
where,
rather
than
getting
bigger,
our
computing
in
the
environments
are
often
getting
smaller
right
now,
and
so
you
know
there's
this
competing
force
where
you
know
so
much
of
the
growth
that
we
experienced
when
you
know
Moore's
law
was
exponentiating.
Was
you
know
all
if
we're
just
gonna
be
able
to
grow
compute
as
much
as
possible?
A
But,
unfortunately
you
know
power
has
constrained
that
and
you
know
how
we
scale.
Things
has
shifted
from
monolithic
applications
to
orchestrated
applications,
so
in
those
orchestrated
applications
we
now
have,
you
know
frequently
very
you
know
thin
pods
of
execution
environment
that
are
being
scheduled
and
orchestrated
as
a
part
of
a
larger
reality.
So
you
know,
if
you
look
at
the
sort
of
macro
technology
changes
that
have
happened
around
node
many
of
the
operating
environments
have
actually,
you
know
been
optimized
down
to
you
know
a
single
single
threaded
event,
a
single
event
model.
A
As
being
you
know,
the
more
efficient
way
to
compute
in
your
today's
operating
environment,
so
you
know
my
best
answer,
for
you
know
why
node
in
you
know,
then,
in
this
new
operating
context,
is
actually
has
nothing
to
do
with
node.
It's
just
like
you
know.
The
world
is
moving
around
us
into
a
model
that
that
really
optimizes
for
nodejs-
and
you
know
it's
it's
a
better
fit
for
the
kinds
of
applications
that
were
writing
today.
A
B
Changing
your
authoring
and
deployment
strategies
is
probably
the
biggest
thing
of
you
know.
So
if
somebody's
coming
from
a
Java
servlet
container
background,
it's
changing
that
concept
and
moving
to
a
new
one.
That's
and
my
maybe
I'm
bastardizing
a
phrase
about
micro
services.
Are
you
leaning
more
towards
that
way,
or
are
you
talking
just
modularizing
your
business.
A
B
A
A
Gonna
spend
a
bunch
of
time
trying
to
figure
out
the
right
way
to
we're
gonna,
you
know
put
something
into
production
as
fast
as
we
can,
so
we
can
validate
it
in
front
of
our
end-users
right
and
that,
just
like
that
changes,
the
way
you
want
to
deploy
and
therefore
you're
you're
you're
forced
to
you
know,
consider
new
strategies,
and
you
know
you
know,
that's
that's
why
I
see
you
know
things
like
serverless
being
a
a
great
first
step,
for
you
know,
companies
that
are
looking
at.
You
know
how
they
evolve
their
technology
story
right.
A
A
You
know
collaborating
with
PayPal
back
in
2013
and
you
know
they
went
through
a
C++
to
Java
journey
and
then
you
know
kind
of
the
tipping
point
for
PayPal
was
at
this
moment
where
you
know
they
were
trying
to
make
a
single
line
of
text
change
in
their
deployment
and
it
took
eight
months
for
that
to
flow
through
to
production.
So
you
know
that
you
know
the
front-end
team
was
like
yo.
A
C
C
D
C
It's
magic
yeah,
so
I
that
big
transition
happened
before
I
started
at
PayPal.
I
started
about
a
year
and
a
half
ago,
but
so
what
I'm
kind
of
coming
in
and
looking
at
is?
Where
are
things
now
and
where?
Where
do
I
think
they
really
should
be
like
based
on
what
the
industry
has
been
up
to
the
last
several
years
you
know
before,
since
they
made
the
big
change,
and
definitely
that
has
been
very
positive
for
PayPal,
as
Dan
mentioned,
you
know,
release
cycles
were
insanely.
Slow
people
were
very
frustrated,
something
needed
to
be
done.
C
C
It
enables
teams
to
move
more
quickly.
You
know
you
don't
need
to
have
bulletproof
plans.
You
know
you
have
an
idea.
What
you
think
is
gonna
work.
You
can
get
a
pretty
good,
you
know
version
out
the
door
in
front
of
people
and
then
iterate
on
that,
and
you
just
can't
do
it
with
you
know
multi-month
release
cycles.
So
it
was
a
lot
of
education
that
came
with
it.
I've
been
interviewing
folks
around
the
company
in
the
last
few
weeks
because
they
gave
a
talk
recently
and
I
wanted
more
context,
and
that's
really
what
happened.
C
They
gave
like
workshops
and
classes
to
train
these
teams,
how
to
be
how
to
adjust,
how
they,
how
they
work
and
then
what
came
with
that
was
no
js'
which,
in
a
you
know,
was
it
an
enabler
of
being
a
little
move
more
quickly,
so
I
think
it's
been
successful
as
far
as
that
goes
at
PayPal,
and
now
you
know
we're
facing
different
kinds
of
problems,
we're
like
how?
How
do
we
want
to
take
it
to
the
next
level?
I,
don't
know
if
we
need
to
get
into
that,
but
yeah
it
was.
B
C
C
So
it
took
somebody
at
the
top
and
I
don't
know
who
who
ultimately
made
that
choice
again.
This
was
you
know
before
I
started,
but
I
know
it
was
like
VP
or
senior
VP,
something
up
there,
and
then
they
brought
in
I
think
bill
Scott
from
Netflix
and
a
few
others
too.
They
decided
as
an
orc
up
top
we're
gonna
make
this
change.
You
know
they.
They
saw
the
benefits,
they
had
considered
it
and
then
they
they
just
went
for
it.
Basically
I
had
a
pig
at
a
company
like
that.
You
know
I.
C
Think
big
change
really
needs
to
come
from
the
top,
though
it's
it's
I
mean
you
can
do
it
naturally,
but
in
a
hierarchical
sort
of
situation
where
people
have
to
kind
of,
like
you
know,
follow
along
the
mission,
it's
harder
I
think
for
that
stuff
to
bubble
up.
It's
not
impossible,
because
it
is
actually
happening
now
in
other
ways,
but
eventually
it
needs
to
get
up
there
so
that
it
can
roll
back
down
everywhere
else.
You
know
what
I
mean
and
that's
when
I
think
you're
now
you're.
A
A
E
Hello,
yes,
of
course,
so
we're
working
on
digital
transformation
for
large
ecommerce
companies
such
as
Macy's,
American,
Engels,
Toys,
R,
Us
and
aza,
and
also
we
decompose
our
application
to
smaller
apps.
But
within
each
request
we
do
work
for
CPU,
empty
collection,
logging,
SSR,
server-side,
rendering
garbage
collection.
All
those
execution
is
in
the
request
and
box
event
loop
and
block
servers
our
web
server,
so
the
solutions
that
we
use
is
to
process
or
move
the
execution
to
other
child
processes.
So
we
put
logging
like
a
pina.
We
use
acid
out
stream.
E
We
used,
we
used
stream
server-side
rendering
to
do
not
block
event,
local
garbage
collection.
So
we
a
team,
helped
us.
So
there
is
a
thread
pool
in
v8
and
it
helped
a
lot
and
for
APM
for
matrix
collection.
We
also
use
external
process,
so
we
just
send
them
and
around
and
collect
those
in
matrix
have
passed
to
serious
time.
Zeros
database,
so
I'm
interesting
in
an
inability,
may
be
of
any
work.
Your
API.
This
is
just
an
interesting
just
to
create
our
application
of
pasta,
because.
E
D
D
Yet
it's
like
a
real
concrete
use
case
that
says
like
here's
one
where
you
know
we
do
this
today,
but
we
think
it
would
be
more
efficient
if
we
had,
like
you
know
in
process
worker
threats,
and
so,
if
you
have
dodd
and
can
help
sort
of
capture
that
that's
one
of
the
ways
that
I
think
it
would
be
useful,
you
know
this
group
could
could
capture
one
or
more.
Those
would
be
like.
D
A
D
So
it's
sort
of
capturing
a
bunch
of
those
where
you
could
then
experiment
with
well.
We
could
do
that
today,
like
with
the
cluster
modules
and
with
you
know,
you
can
say:
okay,
but
then
here's
some
bottlenecks
and
then
there
are
some.
There
are
already
all
are
some
like
NPM
packages
that
will,
let
you
you
know,
use
worker,
threads
and
stuff.
We
could
compare
those
just
sort
of
help
in
terms
of
the
process
of
making
the
case
for
why
you
need
them
and
why
they're
important
and
great.
A
So
I'd
love
to
sort
of
to
try
to
connect
the
dots
a
bit
between.
You
know,
Aaron's
comment
on
threading
and
Nikolai's.
You
know
it's
a
deep
dive
into.
You
know
a
technical
solution,
so
you
know
one
thing
that's
happening
today:
in
front-end,
JavaScript
is
we've
introduced
into
the
the
front-end
platform,
a
serviceworker,
and
you
know
that
is
an
off
main
thread
thread
and
you
know
that
begins
to
introduce
a
whole
nother
possibility
of
compute
model
for
the
platform,
and
you
know,
is
it's
not
quite
multi-threading.
A
D
A
D
D
B
One
of
the
things
that
I
do
will
I
to
speak
on
the
topic
of
integrating
in
Quebec
and
legatus
legacy
systems
and
on
I,
be
my
dad's
communicating
with
RPG
business
logics.
There's
millions
of
lines
of
this
code
and
I
just
shared
a
slide
deck
slide,
where
it
shows
an
example
of
communicating
with
that
RPG
business
logic,
so
it's
same
machine,
a
different
runtime
environment,
those
same
operating
system.
So
it's
two
different
processes,
so
are
spanning
processes
in
this
scenario.
B
But
that
could
be
a
case
for
maybe
a
use
case,
and
this
is
a
frequent
UK
use
case.
That's
gonna
happen
on
IBM
I,
so
I
don't
know.
If
that
would
work.
It
sounds
like
there's
two
different
things.
Are
we
talking
about
web
workers
as
separate
threads
within
the
same
process,
but
then
also
web
workers
also
can
be
a
separate
process.
Yeah.
D
I
mean
you
can
implement
them
as
threads
or
processes.
So
there's
like
advantage
pluses
and
minuses
that
way.
There's
you
know
the
kind
of
API
is
that
you
need.
So
today
you
could
use
the
cluster
module
or
child
processes
to
actually
already
kick
off
things
that
would
work
on
their
own,
and
you
know
you
could
you
know
you
could
basically
say?
Okay,
you
know,
as
Dan
mentioned
it,
you
know
you.
You
also
then
get
into
like
the
multi-threading
stuff.
That's
looked
at
is
more
message-passing
oriented
like
not
just
sharing
the
stock
I.
D
D
D
That
well
is
the
ergonomics
of
that
bad
enough
that
it's
worth
having
a
new
API
that
lets
you
do
it
much
more
easily,
because
it's
a
common
and
important
use
case,
in
which
case
like
the
workers,
one
may
let
you
do
it
a
lot
more
easily
without
having
to
have
as
much
knowledge
and
so
forth.
But
it's
it's
difficult.
It's
it's
like
one
of
the
areas
where
you
know
I,
think
people
figure
it
out,
but
there's
not
a
good.
D
You
know
how
to
guide-
or
you
know
here-
are
the
cases
that
we
can't
do
and
so
like,
hey
I
would
have
I
would
have
moved
over
to
node
except
I.
Couldn't
do
this
and
you
know
that's
that
having
those
captured
somewhere,
so
we
can
have
as
input
I
think
is,
is
where
I
could
see.
You
know
you
guys
getting
in
people
getting
involved,
and
maybe
we
can
put
something
together
on
that
front.
Sure.
D
B
D
D
Guess
the
the
other
expect
that
we
haven't
touched
on
in
the
discussion.
There's
the
technical
aspect,
there's
also
the
educational
aspect
of
you
know:
if
people
come
to
know
it
and
still
have
a
hard
time
getting
over
the
well,
what
do
I
do
because
a
multi-threaded
there's,
maybe
still
still
some
communication
aside
like
if
it
turns
out
to
be
a
heart,
a
real
issue.
Yes,
if
it's
like,
oh
you
need
to
think
a
little
differently.
A
Right
yeah,
the
one
thing
I've
found
that
that
I
and
I've-
never
really.
You
know,
actually
gone
and
and
sort
of
forced
myself
to
write
this
down.
So
I
I
did
the
the
node
training
there
at
PayPal.
You
know
taking
that
that
first,
you
know
group
of
cohort
of
folks
through
the
journey
from
Java
to
to
know
Jess,
and
so
much
of
it
ends
up
being
unlearning
right.
You're.
Trying
to
you
know
you
know
impart
the
best
practices
in
node
and
you
you
convey
that
information
of
you
know.
A
How
do
you
do
this
right
and
that's
that's.
You
know
good,
but
you
know
in
helping
folks,
you
know
really
embrace
that
you
know
those
Breck's
best
practices
oftentimes.
There
are
practices
that
they're
used
to
in
other
platforms.
That
just
aren't
aren't
necessary
and
you
know
if
you
kind
of
strip
away
some
of
those
expectations
and
get
down
to
this
very
raw
computing
platform
that
we
have
and
no
js'.
Then
you
you
come
back
to
you
know
much
more
simple
state
where
you
can
be.
B
Yeah,
that's
very
true
I'm,
just
as
you
were
saying
that
I
was
remembering
when
I
went
from
raw
CGI
programming
to
Ruby
on
Rails,
and
it
was
painful
because
I
was
like
well.
How
does
it
accomplish
that?
There's
so
much
magic
going
on,
but
I
did
have
another
question
I'm
wondering
if
Aaron
you'd
be
willing
to
answer
the
other
big
thing
I
get.
Is
you
shouldn't
store
any
business
logic
in
JavaScript
and
I'm
wondering
if
you
have
a
comment
on
that,
based
upon
your
experience
with
PayPal.
C
C
Primarily
it
was
C++
and
then
Java
node.
Up
to
this
point,
I
mean
honestly,
you
know
coming
in
the
door.
I
I
was
surprised
at
how
how
low
a
percentage
of
that
logic
was
a
node.
But
then
you
know
once
I
was
there
a
little
bit
and
looked
around
and
realized?
Well,
you
know
this
is
years
and
years
and
years
of
stuff.
Okay,
no
everything
is
gonna,
move
over
right
away
and
the
other.
The
other
thing
is
notice,
primarily
used
as
a
UI.
C
C
Maybe
there
aren't
a
lot
of
folks
doing
it,
but
I
do
I'm
getting
more
and
more
feedback
that
folks
are
interested
in
going
further
that
direction
and
that's
something
we're
looking
at
right
now:
I'm
trying
to
get
a
sense
of
how
serious
these
teams
are
and
whether
or
not
we
need
to
invest
more
there,
because
mostly
it's
like
orchestration
layer
stuff,
where
you
call
you
know
a
few
services
or
a
dozen
services
or
whatever,
based
on
what
you're
doing
and
then
you
have
to
you
know
manage
whatever
that
isn't
take
some
specific
actions,
but
it's
you
know
most
of
that,
like
payment
logic,
happens
further
down
the
stack
but
yeah
it's
an
it's
an
ongoing
evolution
but
curious
to
see
how
much
you
know
demand
is
at
there
and
now
they
were
starting
to
hear
more
about
it.
C
You
know
because
if
you
look
at
the
surveys
and
things
like
that,
what
do
people
use
node
for
well
I?
Think
majority
is
like
API
layer
stuff
and
you
don't
see
that
same
percentage
within
the
cut
within
PayPal
right
now.
It's
mostly
like
just
a
gateway
or
like
thin
layer
in
front
of
the
other
fbi's
sure.
B
C
Some
of
our
busiest
teams
are
currently
considering
moving
from
Java
to
node
for
the
API
service
layer,
which
is
a
pretty
big.
It
would
be
a
big
commitment,
I
hope
they.
They
want
to
go
that
direction
because
I
think
that's
a
good
fit
for
what
they're
trying
to
do
and
it's
exciting
to
me.
It's
just
interesting
use
for
it,
but
we're
not
there
yet
sure.
B
And
one
of
the
things
I've
been
doing
it
kind
of
further
justify.
This
line
of
thought
is
put
together
a
list
of
existing
open-source
full
ERP,
or
what
have
you
applications
that
are
larger,
that
have
done
the
business
logic
thing
I
just
posted
a
link
into
the
chat,
but
if
anybody
else
had
additional
things,
open
source
repositories
to
add
to
that
that
are
full-fledged
business
apps,
please,
let
me
know
so.
I
can
forward
those
to
these
different
IB.
B
My
companies,
because
I'm
constantly
facing
a
littered
legitimization
commentary
of
was
it
is
just
out
of
you
know,
fisher-price
mode.
You
know,
there's
no
jest
thing.
Can
it
really
be
used
for
these
types
of
things,
because
a
lot
of
them
say
well
I've
only
ever
seen
it
implemented
for
this
web
chat
thing
on
socket,
IO
only
use.
F
B
They're
decently
high
level
mm-hmm
and
when
I'm
going
into
a
situation-
and
these
are
week-long
engagements-
where
we're
talking
in
depth
about
all
these
different
technologies,
it
ends
up
being
well.
How
we
get
down
to
the
thread
talk
like
well.
How
does
it
do
this
in
threads,
so
it
gets
that
deep
throughout,
but
those
case
studies
don't
go
quite
that
deep.
So
it's
great
for
the
CEO,
because
he
wants
to
feel
comfortable
that
it's
you
know.
B
Pay
Pal
went
before
him
and
JP
Morgan
went
before
him,
but
once
it
gets
down
to
the
geeks,
the
Java
coders
that
love
their
environment
today
they
want
to
see
that
justification,
so
a
testimonial
of
an
existing
Java
coder
who
gives
examples
of
here's?
How
I
succeeded
in
my
career
and
helped
my
business
moving
to
no
js'?
There's
things
like
that
that
could
be
put
on
a
legitimate
site.
Ie
nodejs
org
would
mean
a
lot
right.
D
The
thing
that
came
to
mind,
while
you
were
talking
is
like,
if
you're
getting
the
same
questions
over
and
over
again.
If
you
could
capture
that
list,
and
then
you
know
is
it
you
know,
as
a
group,
we
could
come
up
with
answers
and
and
as
to
the
next
point,
I'm
sure,
once
we
do
that
we
can
get
those
on
to
the
the
know,
Jess
web
sites
and
stuff
like
that,
because
you
know
the
foundation
is
very
interested
in
helping
to
push
those
kinds
of
things
as
well,
but
yeah.
D
F
F
A
You
know
dotnet
engineer,
so
people
don't
want
to
go
and
do
Java
anymore
in
certain
your
regions,
and
you
know
that's
where
pearl
is
several
pearl
shops
that
you
know
have.
You
know,
abandoned
ship
and
it
moved
over
to
node,
mostly
because
they
couldn't
hire.
You
know
the
type
of
pearl
programmer
anymore
that
you
know
had
been.
You
know
that
their
driving
force-
you
know
with
their
you
know,
build
out.
C
Mirror
that
that's
exactly
the
reason
that
team
I
mentioned
earlier
is
heavily
considering
switching
akram
job
over
to
node.
It's
because
they're
having
trouble
hiring
engineers
that
are
interested
in
working
on
that
anymore.
They
feel,
like
their
perception,
was
communicated
to
me
yesterday
that
these
do
good
engineers
have
moved
on
to
like
machine
learning
or
you
know,
big
data
analytics
risk
all
that
kind
of
stuff,
and
they
just
not
work.
C
You
know
otherwise
they're
not
really
interested
in
working
on
Java
or
api's,
and
things
like
that
services,
so
they're
considering
node
node,
isn't
the
only
thing
they're
considering
they're
trying
to
follow
like
again
really
outlined.
You
know
where
the
these
great
server
engineers,
where
do
they
go?
Maybe
we
should
do
that
and
we,
as
a
company
have
already
invested
in
node.
So
note
is
very
high
consideration
there,
but
there.
If
we
hadn't
been
I
would
imagine
you
know.
Golang
would
be
on
the
list
as
well
or
whatever.
C
E
A
That's
right:
yeah
I,
I
just
hosted
an
event
up
in
Toronto
and
we
had
ed
Cudahy,
the
CTO
of
Conde
Nast.
You
know
come
and
share.
You
know
some
of
the
ways
they're
that
they're
adopting
you
know
emerging
technologies.
So
you
know
to
modernize
the
business.
You
know,
Conde
Nast
across
the
board,
moved
the
platform
over
the
node
to
just
you
know,
to
embrace
the
agility
of
the
platform
ends
aligned
with
you
know
what
developers
were
looking
to
do,
but
obviously
they
are
now
incorporating
blockchain
and
AI
and
other
compute-intensive.
A
You
know
project
into
their
their
ecosystem,
and
you
know,
rather
than
force
all
of
the
the
you
know,
machine
learning,
folks
to
adopt
node.
The
approach
that
they
you
know
use
around
AI
is
we're
gonna
hire
talented
machine
learning
developers
and
you
know
let
them
work
in
the
native
language
that
they
are
proficient
in
and
then
everything
that
they
deliver
to
you
know.
The
applications
here
is
all
exposed
via
AP
is,
and
most
of
that's
written
with
node
right,
and
so
that's
how
you
know
the
specialization
is
integrated
into
a
you
know.
A
D
Sorry
Wyatt
talk
to
me,
but
anyway,
if
you
can
come
across
the
across
the
board
with
the
questions,
I
think
that's
something
that
then
we
could
look
at
together
and
say.
Okay,
do
we
have
good
answers
for
that,
and
you
know
figure
out
a
you
know.
That
would
be
a
concrete
thing
to
to
try
and
document
and
move
forward
on
yeah.
D
The
other
one
too
is
like
if,
if
you
know
in
terms
of
the
threading,
if
they're
you
know,
oh
there's
a
good
there's
a
good
answer
in
many
cases,
but
if
there's
use
cases
that
that
people
here
have
come
across
where
they
do
think
that
things
like
the
the
service
workers
or
you
know,
some
other
sort
of
api's
would
help
to
push
forward.
Capturing
and
documenting
those
would
be
good
input
input
to
feed
into
the
effort.
A
You
know
feels
like
one
of
those
things
that
you
know:
we've
we've
really
beaten
soundly
to
death
as
a
topic
for
someone
who's
been
around
for
a
while,
but
yeah
I'm
happy
to
get
it
I'd
like
to
remind
folks
that
you
know
it's
been
growing
a
hundred
percent
year-over-year.
So
that
means
you
know.
Every
year
we
have.
You
know
a
new
group
of
individuals
that
are
new
to
the
platform
as
big
as
we
had.
You
know
you
know
a
community
last
year,
so
you
know
I'm
happy
to
dive
in
that
or
we
can
get
insights.
A
F
B
F
I
I
think
one
of
the
biggest
pieces
of
context
here
is:
it
gets
really
hard
to
debug,
there's
in
node,
there's
no
source
Maps
and
not
fully
implemented
source
Maps.
There
there's
some
implementation
that
we've
done
with
some
other
features,
but
we
can't
actually
fully
map
an
error
from
the
compiled
sort
or
the
transplant
source
to
the
original
code.
F
So
effectively
you
lose
a
lot
of
those
debugging
steps
and
it
really
does
throw
up
a
barrier
to
understanding
your
code
and
then
being
able
to
fix
it
appropriately,
and
so
that's
one
of
the
biggest
issues
right
now
with
something
like
toast
about
typescript
or
something
like
battle
or
something
like
you
know,
caught
even
CoffeeScript.
Going
back
to
that,
you
know.
F
That's
a
I
I
think
it's
a
barrier
to
adoption
and
if
we
were
able
to
land
that
you
know
that
would
I
think
make
that
much
more
common
and
for
what
it's
worth
I
am
seeing
more
users
of
typescript
with
node
like
it.
That's
definitely
increasing
as
I
think
it
goes
hand-in-hand
with
more
Java
developers
and
more
dotnet
developers,
and
you
know
more
external
developers
starting
to
use
node.
F
D
F
But
you
know
we
just
need
to
kind
of
finish
that
implementation
and
that
would
help
us
complete,
complete
that
step
and
really
get
that
debug
ability
back
for
the
for
those
transplant,
languages
and
honestly
I
think
that
would
help
a
lot
with
just
general
adoption.
You
know
it's
it's
a
relatively
simple
feature
well
in
terms
of
like
saying
in
the
implementation:
it's
not
from
what
I
understand
like
you
know.
It
is
something
that
will
help
expand
the
user
base
and
the
people
willing
to
do
things
with
note.
D
Right
because
I
I
mean
have
to
have
the
talks
I've
been
to
on
typescript,
you
sort
of
see
the
audience
falling
into
two
categories.
You
know
people
who
love
their
types
and
the
people
are
like
well.
The
whole
reason
you
use
JavaScript
is
so
that
you're
not
typed,
so
it's
never
come
through
to
me
as
a
like
a
really
clear
message
that
says
the
future.
Is
you
know
types
or
or
not,.
F
D
I'm,
just
I'm
just
wondering
if
there's
like,
is
there
more?
Is
it
that
that's
shifting
over
time
that
we're
seeing
more
and
more,
you
know
sort
of
demand
or
pull
for
typing?
And
if
so,
how
do
we
sort
of
capture
and
quantify
that
so
again,
that
you
know
is
data
for
the
project
to
say:
oh,
wait,
a
sec,
you
know
if
you,
if
you
knew
that,
like
80%
of
people
wanted
to
use
types
or
that.
F
Think
one
of
the
for
me
at
least
one
of
the
useful
things
here,
would
actually
be
topping
talking
to
the
typescript
team
about
this
and
getting
their
feedback
on
adoption
for
types
and
like
the
reasons
people
want
types-
and
you
know
even
their
reasoning,
freezing
node
and
you
know
that's
definitely
doable
girl
would
probably
be
happy
to
help
us
set
that
up.
But
you
know
III
I
do
that
that
that
it
types
are
necessarily
the
only
value
out
of
certain
things
like
out
of
these
different
kinds
of
toolings
their.
F
You
know
one
added
reason,
so
I
do
think
that
if
we
are
when
we
are
talking
about
something
like
typescript
or
or
you
know
anything
else
that
can
help
you
introduce
types,
we're
talking
higher
or
at
a
different
level,
but
they
do
have
context
on
the
reasoning
for
types
and
III.
Do
think
that
you
know
that
that's
useful
context
for
that
lower
level.
A
C
I
was
gonna,
mention
personally,
not
really
caring
too
much
about
typescript
until
recently
here.
Well,
the
flow
was
another
option
that
said
even
given
my
context
of
working
on
mongoose
for
several
years,
a
while
back
and
really
one
of
the
big
drivers
for
having
that
tool
was
that
it
basically
gave
you
types
for
for
working
with
data
interfacing
with
the
database.
Where
types
really
mattered.
C
That
said
so,
you
know
coming
out
of
that
now,
there's
typescript
and
flow.
You
know
I
kind
of
didn't
feel
too
strongly
about
it
other
than
those
interface
layers
like
where
I'm
receiving
data
from
some
other
process.
You
know
everything
is
a
request
coming
in
or
whatever
I
need
to
validate
that
or
I'm
pushing
data
down
somewhere
else,
I
want
to
make
sure
I
store
it
right
or
whatever.
Those
are
the
layers
where
I
think
it
mattered,
and
personally
you
know
I've
always
just
kind
of
winged
it.
C
It's
sort
of
say
you
know:
I
didn't
I've,
never
actually
used
typescript
or
flow.
However,
I
see
a
lot
of
teams
around
PayPal,
starting
to
use
it
more
and
more
particularly
I
mean
there's
no
clear
Victoire
yet
but
I
would
say,
typescript
is
gaining
momentum
and
it's
because
people
are
hitting
bugs
related
to
oh.
If
I
had
types
I
wouldn't
have
had
this
I.
F
C
So
now
now
that
you
see
like
a
dollar
sign
behind
you,
like
a
saved
me
time,
I
could
be
much
more
productive
and
move
faster
again.
That
was
a
big
driver
for
node.
I
want
to
be
able
to
move
quick,
but
I
don't
want
to
do
it.
You
know
it's
not
a
move
fast
and
break
things.
That's
a
lie.
You
know
it's
move
fast
and
not
break
things.
So
so
types
are
becoming
a
more
interesting
factor
for
teams-
and
you
know,
there's
a
few
folks
who
are
like.
C
Maybe
I
should
use
something
else
instead
of
node.
You
know,
because
of
that
you
didn't
want
to
use
Java
either
so
they're
looking
at.
Maybe
they
want
to
use,
go
again
or
whatever
like
so
there's.
There
is
a
need.
Not
everybody
shares
it.
It
depends
on
like
how
much
they're
doing
on
the
back
end,
I,
think
and
how
much
data
they're
working
with
I.
D
D
Couple
people
in
the
room
here
who
said
hey
it's
it's
we
get
lots
of
questions
we
think
would
be
better
if
we
actually,
you
know,
as
a
team
here,
crafted
a
set
of
questions
which
we
think
would
pull
out
answers.
We
could
then
have
a
survey
go
out
and
if
we
can,
you
know
if
we
get
a
large,
a
reasonable
number
of
responses.
All
saying
that
you
know
typing
isn't
important
to
this
to
the
future.
That's
great!
C
So
I
had
the
other
little
bit
of
feedback.
I'd
say
about
that
as
ideally
so
I
like
I,
think
typescript
is
taking
a
really
smart
approach
and
I
told
him
that
a
few
years
ago-
and
they
were
like
announcing
this
thing
and
getting
feedback-
is
that
you
don't
have
to
like
use
it
everywhere
and
so
I
would
prefer.
You
know
if
we
had
this
built
into
the
language
honestly
like
it
just
comes
out
of
the
box
with
v8
or
whatever
and
I
know
that
that
would
take
a
while.
C
D
C
D
The
kind
of
thing
we
should
come
up
with
enough
questions
that
tease
out
that,
like
basically,
is
it,
you
know
people
want
types
but
they're,
not
really
gonna
use
it
seriously
until
it's
part
of
the
language
or
is
running
an
extra
tool
that
helps
you
use
them
actually
acceptable
or
not,
because
it
would
help
us
decide
where
we
should
advocate
right.
Like
you
know,
if
there's
a
project,
we
thought
type
is
important
for
the
success
of
node.
Maybe
we
should
be
pushing
a
tc39
so.
A
The
you
know
just
just
to
bring
in
the
context
of
proposing
design
it's
worth,
noting
that
it
was
proposed
and
did
fail
as
an
extension
of
what
was
it
used
strict
and
then
there
was
use.
You
know
some
other.
You
know
ratcheted
up
pragma
from
there
and
you
know
setting
that
as
a
you
know,
an
additional
pragma.
It
was
just
too
many.
A
E
Thank
you.
It's
really
the
best
for
what
we
do,
for
example,
in
Russia.
What
but
I
like
a
freedom
of
weak
type
or
streak
type
in
JavaScript
and
in
Russia.
We
have
lot
of
companies
that,
for
example,
use
a
closure
to
Jay
s
because
they
have
a
closure
developers
or
types
creep,
or
we
have
Reich.
They
work,
they
work
with
dart
and
they
are
to
chat,
for
example,.
E
A
A
And
I've
definitely
seen
you
know,
especially
amongst
the
financial
institutions
and
we've
had
some.
You
know
financial
institutions
that
so
Ernest
is
a
lending
platform
and
they
are,
you
know,
a
node
from
the
ground
up
shop.
So
you
know
their
legacy
is
node.
You
know
now,
a
few
years
in
and
you
know
they
they
definitely
hit
a
point
in
their
API
evolution.
A
A
So
I'm
very
open
to
the
you
know,
setting
up
a
survey
and
helping
manage
do
that.
I
would
love
to
to
farm
that
out
to
another
group
to
go.
Do
that
work
like
we
did
with
the
benchmarking
group,
and
you
know
this.
This
means
that
we
take
on
double
duty.
If
we,
you
know,
go
and
design
the
other
survey.
So
if
we
could
find
a
or
organize
a
party
of
individuals
to
you
know
take
that
initial
action.
You
know
that
helps
me
prioritize
that,
with
you
know,
our
small
team
of
mysticism.
D
I'm
I,
don't
know
other
than
this
group,
like
it's,
the
user
user
facing
people
like
throughout
the
project.
The
people
working
on
core
aren't
writing
code
and
not
interacting
necessarily
with
the
people
who
who
actually
are
the
ones
who
are
saying
you
know
running
into
developers
who
say
well,
I'm
not
going
to
use
it
because
it's
not
typesafe
or
so
I
I,
don't
I,
can't
think
of
another
group
in
the
project.
That's
actually
going
to
be
able
to
be
informed
about
like
what
questions
will
pull
out,
support
or
non
support
versus.
You
know,
I.
D
A
D
D
F
F
D
D
Because
I'm
thinking,
if
we,
if
we
did
that
plus
then
advertise
to
whoever
else,
would
like
to
come
from
the
tooling
team
and
from
electron
or
whatever
that
you
know
I'm
hoping
that,
like
you
know
in
that
one
hour,
you
could
work
through
and
say:
okay,
here's!
Here's,
a
good
cut
at
what
the
questions
would
look
like
great.
A
Thank
you
so
we're
just
about
a
time.
I
hope
everyone
had
a
fruitful
discussion.
I,
you
know
really
enjoyed
touching
on
this
and
you
know
getting
the
perspective
of
you
know
whole
new
generation
of
users
that
that
adopting
node,
yes
and
you
know
helping
them
deal
with
that.
The
issues
that
they
need
to
us
just
as
to
you
know,
dr.
platform,
I.