►
From YouTube: 2020-09-23 meeting
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
It
is
a
good
time,
except
for
the
fact
that
I
have
a
meeting
11
hours
ago
in
the
middle
of
the
night,
to
support
asia
pacific
as
well.
So.
B
Well,
there's
four
meetings
total.
I
think
this
is
the
main
java
api
sdk
meeting
less
instrumentation
perfect.
That's.
C
Perfect
because
I've
been,
I
sent
you
an
email,
but
I
guess
I
sent
it
to
the
wrong
one,
because
there
is
this
doc
right
that
we
discussed
about
sleuthing,
open,
telemetry
potential
integration
and
open
telemetry
and
brave
bridge,
and
since
I
have
some
time
right
now,
I
started
to
do
a
spike
and
there
is
a
series
of
let's
say,
differences
between
brave
and
hotel
and
also
like
conceptually.
B
I
think
this
is
a
good
meeting
for
that,
I
think
also
well,
so
you
probably
sent
to
the
wrong
email
because
I
changed
jobs
this
week.
So
that's.
B
So,
that's
not
your
fault
for
sure.
My
personal
email,
which
I
think
is
attached
to
my
github
profile,
is
also
just
fine,
but
this
is
also
a
great
venue
and
it's
good.
Sometimes
it's
really
good
to
have
these
like
if
there's
a
general
issue,
creating
an
issue
and
we
can
discuss
it
there
as
well
or
on
getter.
All
those
things
are
great
places.
C
Okay,
of
course,
I
don't
want
to
hijack
this
meeting,
but
I
have
a
list
of
questions
I
could
ask
about
api.
B
Cool
well,
the
agenda
at
the
moment
is
very
short.
Let
me
share
my
screen
and
I
can
at
least
share
the
agenda.
B
B
So
if
someone
else
could
share
the
agenda,
that
would
also
be
great
morning,
carlos
or
afternoon
evening,
whatever
it
is
for
you.
E
B
E
Okay,
I
haven't
shared
my
screen
long
time,
but
let
me
try
my
best.
B
All
right
cool-
well,
sometimes
things
actually
work
cool
yeah.
So
oh
and
someone
left
a
blank
space
for
me
at
the
top.
Thank
you
yeah.
So
the
only
thing
I
had
on
the
agenda
today
is
just
to
mention
again
that
there
is
a
gi,
ga
burn
down
github
project
we
currently
have
in
the
to
do
column.
One
p1
issue,
20
p2
issues,
there's
three
in
progress.
One
of
them
is
in
progress
assigned
to
you.
Carlos
might
be
good
to
at
least
take
a
look
at
what
that
is
and
figure
out.
B
E
Yes,
not
correct
yeah,
so
just
for
your
information,
as
you
may
know,
starting
from
next
week,
I
will
hopefully
have
more
cycles
because
in
theory
the
remaining
very
important
items
in
this
specification
will
be
done
so
excellent
yeah.
So.
B
A
Hey
john,
I
noticed
the
zoom
meeting
number
you
had
it
in
the
getter,
but
in
the
calendar
entry
I
think
it's
still
the
old
one.
B
So
it
got
updated
yesterday,
if
you
copied
it
to
your
own
calendar,
that
doesn't
the
copy
doesn't
keep
up
this
keep
in
sync.
So,
okay,.
B
B
But
yeah,
so
we
did
have
to
change
our
zoom
meeting
because
we're
gonna
collide.
Did
you
show
up
in
the
node
the
node
meeting
this
morning.
A
No,
I
almost
did,
though,
because
I
went
to
the
meeting
calendar
first
and
then
I
saw
your.
I
saw
that
it
was
also.
You
said
that
it
was
also
in
the
agenda
notes,
so
I
went
to
the
agenda
and
clicked
from
there.
A
A
B
B
I
don't
know
if,
before
we
dive
into
that,
are
there
any
other
issues
that
people
want
to
bring
up
for
today?
I
think
one
thing
to
note.
Last
week
we
had
discussed
doing
a
quick
release,
a
9-0
release
for
to
support
the
conversion
to
java
8..
I
don't
think,
there's
a
lot
of
value
in
that,
because
we're
going
to
release
next
week
anyway.
B
C
Okay,
so
for
those
who
don't
know
me
or
who
haven't
heard
what
I
was
talking
about
at
the
beginning,
I
am
marching.
C
My
last
name
is
impossible
to
pronounce
and
I'm
the
lead
of
spring
cloud
sleuth
and
I'm
doing
a
spike
because,
like
there
was
a
there,
was
a
discussion
at
some
point
where
we
took
into
consideration.
Checking
out
what's
happening
there
in
the
autos.
All
the
world
and
we've
been
discussing
that
sleuth
underneath
uses
brave
brave,
is
massively
used
because
massively
used.
C
So
we
want
to
even
see,
if
that's
possible,
for
the
users
to
let's
say
if
we
migrate
and
sleuth
to
hotel
in
the
core
of
how
would
these
users
be
impacted,
etc?
Right
and
we
had
a
couple
of
meetings,
there's
a
talk
where
we
discussed
how
what
are
the
differences
between,
let's
say,
current
sort
of
implementations
in
some
places
and
what's
missing
there
in
hotel
and
vice
versa,
and
somebody
was
supposed
to
try
to
do
the
brave
bridge.
C
So
let
me
dig
into
my
notes,
because
I
have
a
an
adobe
file
with
questions
so
to
begin
with,
because
I
read
the
let's
say
the
documentation
on
on
the
differences
between
api
and
sdk
jars,
etc,
etc.
C
So,
let's
start
with
the
hotel
api.
An
extremely
simple
question
is:
how
can
I
continue
a
span?
What
does
it
mean?
Let's
say
I
have
a
context
span
context
or
I
have
a
span
object.
They
are
not
put
into
the
context
yet
right.
I
have
just
an
object
span
or
an
object
of
context
type.
How
can
I
continue
it
right?
So
let's
say
having
a
span
context
object.
I
would
like
to
make
a
span
out
of
it
without
creating
a
child
span
right.
C
So
that's
my
first
question
because
I
I
haven't
seen
such
in
such
a
such
a
way
to
do
it.
What
I
could
do
is,
let's
say,
put
the
span
in
scope,
then
get
current,
and
then
I
have
let's
say,
access
to
the
span,
but
I
can't
really
modify
it
to
let's
say
in
the
way
that
I
can
do
with
the
span
builder.
So
the
question
is:
how
can
I
continue
to
spawn
in
such
a
way
that
I
can
modify
it?
For
example,
add
a
kind
to
it.
C
Yeah
and
now
you
need
to,
for
example,
with
messaging.
Let's
say
you
have
the
propagation
via
some
message:
headers
whatever
and
you
let's
say,
got
the
the
message
you
received
it.
You
want
to
continue
the
span,
but
you
want
to
add
it
a
kind
consumer
or
kind,
for
example,
produce
let's
say
kind
consumer.
I
can
do
it
because
what
I
retrieve
is
spam
context
or
spam
and
the
only
way
to
add
a
kind
is
to
your
spam
builder.
B
E
Yeah
yeah,
it
does
not,
and
I
think
that
if
I
remember
correctly
from
sipkin
in
sifting,
you
can
basically
start
spam,
then
propagate
the
spam
context
and
then
the
other.
The
next
process
will
get
the
spam
context
and
will
finish
the
spam,
but
it's
so
the
fun
context
represents
the
same
span.
Right,
marching
correct,
exactly
yeah.
We
don't
support
that.
E
E
Okay,
yeah,
okay,
let's
yeah,
let's
fill
an
issue
for
that.
There's
no
problem,
probably.
C
I
will
come
back
to
sdk
because
I
don't
understand
that
yet
how
why
certain
things
are
in
sdk,
for
example,
sampling.
I
have
no
idea
why
it's
in
sdk-
it's
not
in
api,
but
I'm
going
to,
let's
say
finish
the
api
part
and
then
let's
go
to
that.
I'm
going
to
take
a
note
just.
A
I
have
a
question
about
the
messaging
example,
so
you
mean
it's
broken
because
it
doesn't
support
the
particular
zipkin
model
of
sharing
a
span
between
producer
and
consumer
or
it's
broken
period,
because
we
do
support
we
do
like
in
the
auto
instrumentation
we
have
instrumentation
of
jms
kafka,
rabbit,
mq
and
that
all
works.
We
we
propagate
the
trace
context
across
you
know
in
headers,
and
we
create
a
new
span
in
the
consumer
whose
parents.
C
And
you
spend-
and
I
don't
want
to
create
a
new
span-
that's
the
thing,
but
so
answering
your
question
I
would
like,
because
for
for
spring
for
spring
users,
compatibility
is,
like
is
extremely
important
thing.
I
don't
want
the
users
to
of
sprinkled
sleuth.
You
suddenly
see
loads
of
new
spans
right.
So
what
I'm
doing
right
now,
as
I
said,
it's
just
a
spike.
C
I
really
doubt
that
anything
works
right
now
with
what
I
did,
because
when
I
try
to
make
a
bridge,
it's
almost
impossible
for
it
to
work
right
now,
but
coming
back
to
the
messaging
part.
Currently
it
works
like
this
that
we
send
a
message.
We
mark
this
as
a
producer
kind
and
then,
when
the
message
is
received,
the
same
span
is
marked
with
kind,
but
it's
on
the
you
know
the
other
side
right.
So
it's
the
same
span
in
two
pieces
and
then
zipkin
merges
it
and
and
understands
those
two
things.
B
I
wonder
if
we
this
could
be
salt,
so
I
mean
the
open
source
model
is
obviously
a
little
bit
different
and
I
think
the
model
in
this
case
is
more
to
create
a
span
link,
create
links
between
those
spans,
and
I
think
then
the
assumption
I
could
be
wrong
here.
The
assumption,
then,
would
be
the
back
end
or
whatever
is
or
an
exporter
would
then
kind
of
do
that
reconciliation
and
link
like
take
the
link
and
make
it
into
a
single
span
and
zipkin
world.
C
Maybe,
but
so
I
would
create,
I
would
create
a
new
span
on
the
consumer
side,
with
a
link
to
what
really.
B
I'm
saying
you
could,
if
you
needed
a
different
if
you
needed
a
different
model,
another
if
you
didn't
want
to
do
parenting
another
way
you
could
join
them
together
as
be
a
linking.
A
So
I
don't
think
merchant's
issue
is
with
the
parenting
it's
with,
and
I
think
I've
seen
like.
So
we
create
one
span
for
the
producer
on
the
producer
side
and
we
create
a
second
span
on
the
consumer
side
that
parents
the
producer
and
I'm
not
very-
I'm
not
familiar
with
the
zipkin
model,
but
it
sounds.
I
feel
like
I've
seen
that
before
where
it
was
different
and
even
the
b3
headers
are
different
right
that
you
you
propagate.
A
So
you
don't
in
zipkin,
doesn't
differentiate.
You
can't
tell
the
difference
between
how
long
the
client
took
and
how
long
the
server
side
took,
because.
C
A
E
Interesting
by
the
way,
I
remember
I
do
remember
some
discussion
on
the
collector-
I'm
not
familiar
with
the
corrector,
but
I
remember
a
discussion
there
about
trying
to
merge
spans
based
on
some
spam
processors,
but
I'm
guessing
that
you
guys
are
not
expecting
that
users
will
have
to
use
a
collector
correct
they
just
using
zip
code.
E
C
A
E
B
G
It's
not
broken
because
zipkin
added
two
years
ago,
a
support
to
to
have
the
spans
different
or
to
be
mergeable.
There
is
a
flag
that
you,
you
have
to
set
or
a
property
on
the
span
telling
zipking
is
it
expecting
to
to
be
mergeable
spans
or
not?
Mergeable
experience.
G
G
C
Yes,
I
was
saying
because
maybe
maybe
you
join
in
the
meantime-
I'm
doing
this
spike
because
I'm
still
waiting
for
you
to
talk
to
adrian,
but
I
don't
wanna
wait.
I
want
to
play
around
with
stuff.
A
C
I'm
trying
to
to
rip
out
brave
from
core
put
it
into
a
separate
module
and
I'm
going
module
after
module,
and
I
started
to
see
these
problems.
In
I
mean
creating
bridge
is
hectic
right
now.
I
have
a
lot
of
return.
C
Nil,
return,
zeros
and
stuff
like
that,
because
I
can't
get
data,
but
conceptually
often
in
in
this
brave
world
you
can
continue
a
spam,
so
you
get
an
object
of
spam
type
or
context
type,
and
you
can
just
continue
it
and
you
can
modify
some
stuff
on
it
and
in
auto
I
can
do
some
stuff
on
this
one
builder
side.
A
C
G
Is
this?
Okay?
That's
fine,
but
but
I
think
we
we
can
talk
to
adrian
about
this,
but
I
think
zipkin
works
in
both
modes.
You
can
continue.
C
If
it
does,
then
that's
not
a
problem
right,
because
I'm
gonna
do
this
in
the
let's
say
auto
way,
I'm
gonna
say
supports,
join
false,
yes,
whatever,
and
that's
it
right.
Yeah.
G
Yeah,
I
think
I
think
it
should
work
based
on
my
understanding.
It
should
work,
maybe
I'm
wrong,
but
let's
let
me
I'll.
G
Let
me
also,
maybe
maybe
schedule
a
short
meeting
with
adrian
as
well
to
talk
about
this
because
I
think
will
help.
C
Of
course
I
mean,
as
I
said,
I'm
doing
just
a
spike
I'm
playing
around
with
brave,
but
adrian
is
crucial.
Here
I
mean
because
I
don't
want
to
have
it
on
the
side.
It
should
be
somewhere.
You
know
either
on
the
auto
side
or
the
brave
side.
It
shouldn't
be
a
side
project,
side,
side,
module
in
sloof,
right,
yeah,.
G
Yeah,
no,
no,
no,
no!
No!
We
and
I
also
by
the
way
I
started
to
play
a
bit
with
the
brave
stuff
to
to
make
hotel
api
being
implemented
by
brave.
So
far.
I
did
not
find
any
problems
except
except
two
things.
So
there
are,
there
are
in
brave.
There
are
some
some
special
things
like
annotate
endpoint
or
something
like
that,
which
I
don't
know
if
you
are
using
in
smooth.
C
A
C
Take
a
look
when,
when
I
did
the
spike
I
mean
I'm
still
am
doing
so.
I
had
the
biggest
problems
with
so
you
have
a
spun
exporter
and
a
spawn
processor
and
and
zipkin
has
as
brave
as
spam
handler,
and
these
are
a
little
bit
conceptually
different
things.
C
So
I
have
a
lot
of
to
do's
here,
but,
okay,
that's
actually
not
that
terrible.
But
that's
the
terrible
thing
is
yeah.
Go
ahead.
G
C
G
C
See
because
because
that
was
easier
for
me
and
speaking
of
which
that
was
something
I
wanted
to
to
talk
to
you
in
a
in
a
in
a
second
actually
there's
quite
a
few
things
that
are
there
in
the
sdk
jar
which
conceptually
I
don't
get
that,
for
example,
I
have
a
span
which
is
an
interface,
but
I
can't
really
read
from
it
because
there's
another
thing
readable
span,
that
is
in
the
sdk
and
it's
not
in
the
api
right.
C
So
if
you're,
creating
a
library
or
a
bridge
between
libraries
from
from
the
docs
that
I've
read,
you
were
supposed
to
use
the
api,
but
you
can't
read
from
a
spawn,
so
you
can't
do
a
translation
from
myspan
to
your
span
and
vice
versa.
But.
G
I'm
not
doing
that,
I'm
not
doing
that,
but
I
am
what
I
was
thinking
is:
I'm
gonna
have
a
brave
wrapper
span
that
that
actually
implements
the
api
and
redirects
everything
to
the
brave,
so
I
don't
need
to
to
read
so
if,
if
I
configure
brave
everything
will
get
too
brave,
so
all
the
data
will
be
only
stored
in
brave.
C
C
B
It
is
purely
for
instrumentation
authors
and
it
should
be
and
will
be
right
only
and
no
and
no
read,
and
then
the
sdk
is
an
implementation
of
that
api,
which
may
may
or
may
not
have
ways
to
read
that
data
internally
but
the,
but
the
sdk
implementation
exposes
certain
extension
points
like
like
the
spam
processor
and
samplers,
and
exporters
that
have
access
to
the
span
data
in
a
more
robust
way
in
an
sdk
specific
way.
B
So
this
is
a.
This
is
a
pretty
strong
philosophical
underpinning
to
the
way
that
the
the
apis
are
designed
that
they
should
be
right
only
with
no
reading.
C
Sure,
most
likely
I
mean
if
bogdan,
took
a
look
at
the
brave
stuff
and
it's
feasible.
Maybe
I
I
did
something
wrong.
Maybe
I
tried
to
do
it
both
ways.
This
is
why
we'll
check
it
out
sampling
api
should
be.
Why
isn't
sampling
api
in
in
api?
Why
is
it
sdk.
B
B
G
Marking
when
we
we
talked
about
instrumentation
author,
we
talked
about
the
sloot
part
that
just
goes
into,
let's
say:
http
calls
and
instruments
to
create
a
span
to
to
annotate
the
span
and
so
on.
When
we
talked
about
library,
application,
owner
or
things
like
that,
we
talk
about
the
slot
part.
That
is
configuring,
how
the
how
the
implementation
should
work,
how
to
implement
how
the
how
the
things
should
work
together.
Does
it
make
sense.
C
A
A
G
C
However,
so
so
maybe
maybe
I'm
trying
to
make
it
too
abstract,
because
I
thought
that
we
have
sprinkled,
let's
see
core,
we
have
brave
and
autumn
right
right.
A
C
Because
there
will
be
like
two
implementations-
maybe
I'm
exaggerating
here-
maybe
I
wanted
to
make
it
too
abstract,
but
I
thought
that
you're
going
to
have.
Let's
say
I
mean
for
me:
it's
going
to
be
easier.
If
we
don't
do
it
like
this
right,
because
there
will
be
no,
let's
say
so.
If
you're
choosing
brave
you'll
have
to
use
the
brave
api
to
do
the
sampling
right,
correct.
G
You,
if
you
choose
brave,
you'll,
have
to
configure
brave
assembly.
If
you
choose
to
use
hotel,
you
will
have
to
do
hotel.
You
have
to
configure
hotel
sampling.
C
G
I
I
think
I
think
this
way
is
more
abstract,
because
the
core
does
not
have
a
knowledge
about
sampler
or
anything.
The
core
is
just
pure
instrumentation
is
pure
spring,
giving
the
data
to
the
to
the
library
and
now
the
way
how
the
library,
the
library
acts
on
this
data
doing
sampling
doing
all
the
things
it's
a
property
of
the
implementation.
G
So
so
the
core
is
just
the
the
the
the
small
subset
of
the
functionality
that
just
passes
the
data
from
screen
to
the
api
and
now
and
now,
based
on
the
starter
that
you
use.
If
you
use
hotel
starter,
you
can
configure
sampling
and
whatever
you
want
to
configure.
If
you
config,
if
you
use
brave,
you
can
configure
also
sampling.
That
brave
has.
B
Makes
sense
you
could
imagine,
I
mean,
like
you,
could
imagine,
having
like
a
a
core
sampling
config,
but
that
wouldn't
be
a.
That
would
be
something
at
a
different
level
like
is
like
spring
spring
sampling
config.
It
would
be
like
how
you
can
how
you
do
that,
but
and
it
could
be
bridged
into
both
brave
and
the
open
spelmatry
configuration
for
spring.
C
Yeah,
that's
what
that
was
what
I
think
what
I
was
thinking
about
right
that
you
have
this
obstruction,
because
there
are
certain
since
free
cloud,
sleuth,
let's
say
in
the
core:
for
us.
Let's
say
sampling
is
something
that
people
can
override
right
and
if
we
used
this
is
what
I
thought
that
would
make
sense.
If
auto
provided
an
interface
for
that
like
api
spec,
then
we
wouldn't
care
what
kind
of
a
library
is
used
underneath
right
if
it's
compatible
with
the
api,
I
mean
we
don't
care.
C
However,
we
can
not
care
even
more
by
just
saying
hey.
There
is
no
sampling
interface
that
you
should
implement.
C
We
are
redirecting
the
problem
to
let's
say
a
given
library
of
your
choosing
right,
which
yeah.
G
Yeah,
I
think
I
think
I
think
that's
that's
a
bit
more
generic
in
a
way
that
we
we
don't.
We
are
not
opinionated
about
the
fact
that
there
has
to
be
sampler
or
not
and
based
on
whatever
implementation
you
choose
that
implementation
has
or
has
not
support
for
assembly,
and
I
think
in
that
regard
it's
a
bit
more
generic
doing
this,
but
it's
not
that
opinionated,
which
I
don't
know,
but
but
the
starters
for
sure
the
starters
will
come
with
defaults.
So
so
the
starter.
C
I
mean
yeah,
I
mean
from
my
perspective.
I
can
imagine
that
the
users,
the
current
users,
if
they
migrate,
they
will
change
the
library
they
are
using
right
now
so
from
sprinkler
starters,
love
to
sprinkle,
stars
brave
whatever,
and
things
should
just
work
correct
and
it
should
work
nothing.
Nothing
should
be
done.
G
Yes,
that's
that's!
That's
how
I
see
one
thing
that
you
can
you
may
want
to
answer
me
is
in
the
core
right
now:
do
you
have
any
dependency
on
sampler?
Do
you
call
the
sellers.
C
Of
course
that's
the
thing,
but
it
does
it
so
we're
just
defining
a
sampler
right
and
you
can
provide
properties
for
the
sampler,
but
but
they
can
be
brave
specific.
Yes,
that's
not
a
problem
yeah!
It's
a
problem.
G
Perfect,
I
will
show,
as
I
said,
I
will
show
you
more
details
and
once
I
put
together
my
view
on
things,
we
can
we
can
sing
and
stuff,
but
I
think
we
are
doing
good
progress
overall
and
and
I'm
happy
that
that
we
are
having
the
discussion
I
will
take
as
an
action.
I
can
schedule
a
meeting
with
adrian.
I
will
send
an
email
right
now
to
find
a
good
time
for
for
for
him
and
you
I
don't
know.
C
I
mean
he's
in
malaysia.
I
know.
G
Two
and
a
half
hours,
okay,
is
he:
is
he
okay,
that
that
should
be
fine,
because
that
means
seven
a.m
here,
which
is
fine.
C
Okay,
I
have
one
more
question
that
I've
found
out
here
so
also
with
messaging,
but
maybe
again
I'm
I
was
trying
to
be
too
generic
here.
So
we
have
messaging.
Is
there
an
option
to
actually
retrieve
the
propagation
keys,
because
this
is
what
we're
doing
in
in
with
brave
that
if
you
say
that,
for
example,
your
trace
id
span
id
whatever
keys
are
x
y
z,
you
can
define
them
somewhere
and
then
at
runtime.
C
B
C
Yep
yep
and
that's
that's
pretty
much
it
from
my
side.
I
mean
I'm
trying
to
make
the
code
compile,
like
I
said,
maybe
maybe
I
was
trying
to
be
because
I
tried
to
do
the
bridge.
Both
ways
which
apparently
was
a
was
a
terrible
idea
and
yeah
I'll
try
to
to
make
things
like
that
are
very,
very
specific.
Make
them
auto
compatible,
but
maybe
that's
not
necessary.
C
B
Absolutely,
and
if
you
you
mentioned
at
the
beginning
or
earlier,
you
wanted
to
create
an
issue
around
the
span,
continuation
or
whatever
the
support
is
for
that.
I
recommend
creating
an
issue
in
the
specification
repo
to
discuss
that,
because
it
will
be
okay,
because
it's
kind
of
a
general
zipkin
concept
that
we
need
to
make
sure
that
is
done
consistently.
B
A
A
A
Yeah
yeah
yeah
do
you
know
asir.
A
G
I
will
maybe
trust,
take
an
action
item
to
to
introduce
them.
A
B
All
right,
there's
any
I
don't
know
I
haven't
looked
at
the
agenda
in
a
few
minutes-
doesn't
look
like
anyone's
added
anything
any
other
topics.
Anyone
wants
to
discuss,
or
can
we
wrap
up
a
bit
early.
B
Not
any
further,
we
talked
about
it
last
night,
with
with
anarag
on
the
late
night
meeting.
Okay
he's
moving
forward
he's
he
was
on
vacation
for
a
few
days,
so
he
hasn't
been
doing
anything
with
it.
So
what
he
cannot
go
on
vacation
tell
him
no
I'll
make
sure
we'll
make
sure
to
relay
that.
H
Yeah
but
he's
he's
picking
it
up
again
this
week,
so
he's
starting
thanks
for
moving
forward.
Thank
you.
So
much.
C
I'm
gonna
send
you
on
getter
the
invite
to
zoom,
so
we
can
chat.
Okay,
perfect,
okay,
thanks
great.