►
From YouTube: 2021-02-17 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).
A
A
A
I
have
written
it.
I
wrote
a
lot
more.net
earlier
in
my
career
than
I
wrote
java,
but
close
enough.
A
I
was
actually
writing
some
python.
This
weekend
we
were,
it
was
a
friend's
birthday
and
we
were
using
one
of
those
deep,
fake
apps
to
make
a
video
of
barack
obama
wishing
him
a
happy
birthday
and
the
code
for
it
is
all
python
and
it
was
the
worst.
A
Yeah,
actually
it
was
it
wasn't
the
full
defect
thing.
It
was
just
the
you
can
take
any
video
and
then
make
the
lips.
You
give
it
an
audio
file
and
a
source
video,
and
it
makes
the
lips
speak
to
the
audio
file.
B
I
really
wish
I
had
the
code.
I
wrote
for
my
phd
what'd,
you
do
it
was
in
astronomy,
so
it
was.
I
was
writing
code
to
to
model
my
like
write,
a
build,
a
model
of
what
might
produce
my
data,
so
it
was
some
ray
tracing
through
some
interstellar
clouds
of
varying
density
density
profiles.
B
A
A
That
was
my
implementation
of
opengl
and
it
could
render
like
a
spinning
cube
with
anisotropic
filtering
and
anti-aliasing
and
textures
and
shadows,
but
not
self
shadows
at
like
one
frame
every
20
seconds
on
my
laptop
it
was.
It
was
not
fast.
B
Well,
I
did
my
you
know:
university,
3d,
graphics.
I
did
take
a
3d
graphics
class,
but
it
was
like
it
was
measured
in
many
minutes,
because
this
was
in
1991..
Oh
yeah,
yeah.
A
B
B
A
B
A
A
Anyway,
we
should
get
started.
I
don't
usually
join
these
so
I'll.
Just
everybody
here
knows
me,
but
but
I'm
just
joining
cause
cause
john
and
bogdan
and
and
steve
at
splunk
kid
raids
to
me
saying:
hey
morgan,
you
should
join
the
the
java
sig
this
week
just
to
talk
about
timelines
and
and
requirements,
because
I
think
some
people
thought
there
was
like
a
lot
of
pressure
on
the
sigs
to
ship
their
gas
immediately.
A
So
I
can
start
with
that
by
saying:
there's,
there's
not
a
lot
of
pressure
on
the
cigs,
like
the
big
release.
That
happened
that
was
announced
yesterday
was
or
sorry
today
was
the
the
spec
going
1.0
I'm
expecting
the
ga
releases
of
the
sdks
to
come
out
through
the
coming
weeks,
maybe
the
coming
month
or
two
there's
there's
no
one,
at
least
in
the
project
who's
like
sort
of
cracking
the
whip
on
that
at
least
not
currently.
Obviously,
some
of
the
vendors
may
be
splunk.
A
We
don't
have
any
immediate
deadlines
at
splunk,
but
but
I
suspect
from
chatting
with
a
lolita
and
wreck
that
at
amazon
there
are
some
perhaps
stronger
expectations,
but
but
I
can
definitely
say
like
at
least
with
regards
to
the
project
itself.
You
know.
Obviously
we
want
these
out
relatively
soon,
but
there's
there's
no
like
sort
of
immediate
deadline
that
everyone
needs
to
jump
into
and.
G
A
A
That's
really
up
up
to
you
guys
as
the
maintainers
of
the
java
sdk,
but
for
the
project
you
know
I
I
would
optimize
if
this
is
something
on
the
order
of
like
just
just
a
couple
weeks
or
even
a
month
or
two.
I
would
optimize
for
quality
before
I
optimize
for
like
an
immediate
1.0
release.
B
B
It's
probably
worth
talking
about,
I
think
the
big,
thorny
one,
thorny
one,
that
I
see
that
there's
no
really
great
answer
to
is
the
resource
sbi,
and
what
to
do
with
that.
I
B
And
I
don't
know
that
it
I
mean
we
know
we
have
a
few
changes,
whether
they're,
with
the
specific
details
of
them,
probably
are
not
super
important
about
compared
to
what
we
need
to
whether
we
need
to
we
whether
we
are
willing
to
do
another
rc,
and
I
think,
from
my
perspective,
the
big
thing
is
that
I
think
we
need
to
draw
a
line
in
the
sand.
At
some
point,
I
think
we
can't
continue
to
do
rc
after
rc
after
rc
after
rc
for
the
next
three
months.
B
So
yeah,
I
think
that's,
I
think
that's
my
only
concern
is
how
do
we?
How
do
we
know
when
we're
done
so
it
sounds.
A
B
I
What
I'm
hearing
trying
to
to
make
sure
I
understand
correctly,
we
may
be
willing
to
do
this
as
long
as
we
ensure
that
the
all
all
the
people
in
this
room
are
are
reasonably
commit.
Not
reasonable
are
committed
that
this
is
the
last
one
and
if
we
screw
up
with
something
we'll
deal
after
our
after
ga
and
that's
it
is
that
what
I'm
hearing
from
everyone
is
is
a
reasonable
conclusion
or
not.
I
I
I
haven't
looked
into
too
many
issues
since
last
friday
or
thursday.
Are
there
any
other
issues
filed
in
the
repo
that
may
worth
changing
things
or
or
require
changing
things.
B
I
don't
think
there's
been
anything
new,
there's
trasks
issue
around
put
on
the
trace
state
builder.
J
B
J
B
Yeah,
I
I
don't
think
that
nothing
that
I
know
has
been
added.
I
don't
think
so.
F
There
was
one
the
obfuscation
issue,
but
that
wasn't
really
an
api.
A
breaking
api
change,
yeah.
B
I
I
think
I
think,
for
that
we
should
make
one
release
to
work.
One
release
do
not
work
back
and
forth,
or
a
random
initialization
at
runtime.
You
know
like
if
it's
flips
one
it
doesn't
do
that
just
to
confuse
people
so
to
ensure
that
they
don't
do
this.
B
Yeah,
I
already
put
the
pr
in
for
that
one.
It's
super
simple.
F
G
B
I
I
Okay,
but
no
no
jokes
aside,
I
think
we
should
probably
have
somewhere
written,
maybe
maybe
as
an
issue
that
we
are
willing
to
to
have
this
or
we
are
going
to
do
this,
and
this
is
the
last
chance
for
for
for
things
and
and
have
a
deadline
as
well,
for
when
this
rc
will
happen
just
to
ensure
that
people
are.
If,
if
anyone
comments
after
this,
it's
clear
that
is.
K
B
Yeah
there's
an
open
discussion
right
now
about
the
1.0
release.
It
would
be
a
great
place
to
put
that.
I
think,
okay,
that
I
created
a
while
ago,
yeah
we're
by
the
way
we're
we're
a
day
late
on
our
1.0
milestone,
which
it's
fine
putting
a
date.
Putting
a
date
was
good.
B
It's
fine,
but
yeah.
We
have.
I
have
this
here
and
please
add
comments
here,
hasn't
been
much
discussion
on
it.
Aside
from
me,
25
41,
25
41
are
the
ids
the
same
across
discussions,
issues
and
pr's.
I
think
yeah.
E
E
B
I
It
is
number
five
it's
it's
here.
Okay,.
E
I
Right,
yeah
one
or
the
other,
I
don't
care
which
one
it's
just
like
it's
confusing
when
there
are
both,
and
I
try
to
explain
my
confusion
there
in
the
examples
I
don't
have
a
strong
opinion.
I
think
I
think
I
was
proposing
to
builder
to
be
removed
just
to
have
a
more
consistency
with
spam.
So
it's
it's
way
more
consistent
with
the
spam
builder
and
and
yeah.
B
I
So
that
that's
why
I
said
no
remove
the
two
builder,
because
that
was
something
that
we
don't
have:
a
creed
child
on
the
on
span
or
equivalent.
I
Yes,
but
there
is
already
a
builder
and
there
is
a
possibility
on
the
builder
to
set
apparent.
So
it's
instead
of
say
and
usually
people
are
working
with
context,
so
to
use
this
one
they
would
have
had
to
do
anyway,
baggage
from
context
and
then
to
builder,
because
you
have
context
in
place.
So
this
way,
you
just
say
baggage.builder
setup
and
this
context
and
it's
essentially
the
same
thing.
E
E
This
does
feel
like
it's
okay
to
be
like.
I
find
a
static
method,
implicitly
getting
something
from
the
context
to
be
more
magical
than
I'd
expect
like
if
you're
getting
a
builder,
it's
supposed
to
just
be
a
empty
builder
right,
like
that's.
Every
other
dot
builder
in
our
code
base
does
that,
but
this
one
will
automatically
appear.
F
F
B
E
I
F
How
do
you
set
no
parent?
I
Is
there
is
a
concept
because
you,
you
have
a
baggage
in
the
context,
and
you
always
want
to
add
more
things
to
that.
So
that's,
essentially
the
parent
or
whatever
you
call
it.
Maybe
the
terminology
of
parent
is
not
good,
but
if
you
have
couple
of
baggages
coming
from
from
from
the
wire
or
from
something
you
most
likely
want
to
append
to
that,
not
to
to
start
from
new.
B
Yeah,
I
think,
actually
that's
a
really
good
point
that
if
it's
not
apparent
like
this
fan,
like
the
whole
parent
idea
of
span,
is
makes
a
lot
of
sense
because
parents
fans
have
parents
baggage
doesn't
like
intrinsically
have
something.
That's
apparent.
It's
just
so
it
was
just
a
convenient
way
to
get
all
of
the
attributes
from
the
existing
one,
so
I'm
leaning
towards
option
c
actually
to
try
to
not
make
it
look
like
span
at
all,
since
it
isn't
really
a
span.
It's
just
a
some
data
that
you
can
stick
in
the
context.
F
B
I
The
only
part
that
will
be
weird
a
bit
to
be
honest
is
the
build,
returns,
the
context
and
not
just
the
baggage
from
here,
because
because
because
build,
let
me
explain
because
build
now.
You
you
don't
have
a
context
instance
to
to
put
that
into
the
the
context.
Make
sense
like
build
half
has
to
return
a
baggage
here.
Correct,
okay,.
I
E
I
I
B
So
this
one
so
often
this
one-
I
commented
quite
a
bit
on
this.
The
issue
after
bogdan
tank,
hopefully
did
some
research.
So
I
think
if
we
change
the
trace
date
builder
to
be
put,
then
everything
is
put
except
the
span
and
the
span
names
are
all
set
attribute,
they're,
not
just
a
bear
set.
So
I
think
that
actually
is
a
difference
in
an
api
that
is,
it
deserves
to
be
a
little
bit
differently,
a
little
bit
different
because
it's
not
it's
not.
The
span
is
not
just
a
map
right.
B
I
No
just
put
attribute
because
behind
the
scene
kind
of
we
signal
the
attribute
is
part
of
a
map,
but
I
would
say,
add,
link
because
link
is
part
of
a
list
and
maybe
add
event.
I
I'm
not
seeing
I'm
just
I'm
just
trying
to
be
the
the
it
was
devil's
advocate.
I.
L
I
Does
it
seem
that's
the
same
behavior
for
any
setter.
I
Signals
that
behind
the
scene,
there
is
a
map
concept,
adds
more
or
less
signals.
There
is
a
list,
I
think,
and
instead
it's
just
a
property,
the
property
that
all
right
anyway,
these.
E
B
E
I
I
guess,
but
it
seems
we
mentioned,
that
attributes
act
like
a
map
correct
because
by
by
the
key
like
it
replaces
the
value
that,
for
that
corresponding
key,
it
does
not
replace
the
the
value,
the
the
one
value.
So
so
that's
that's
why
anyway
java
decided
to
do
this,
but,
but
I
think
I
think
the
put
on
the
trace
state
for
me
is
the
simplest
one.
I
I
think
that's
no
question,
that's
the
very
inconsistent
with
all
the
other
builders
and
simple
maps
that
we
have,
and
I
think
that's
that's
yes
for
myself
for
the
span.
If
I
were
to
change
them,
as
I
said,
I
would
have
put
attribute,
add,
link
and
add
event
to
to
kind
of
signal
the
the
the
the
some
more
than
just
the
fact
that
I
mean
the
semantic
behind
the
scene
that
it's
it's
at
least
it's
a
map
and
so.
I
I
B
B
B
Event,
well,
I
think
an
instrumentation
said
attribute
it's
going
to
be
much
more
common
than
said
event
than
that
event.
It's
actually
added.
It
is
already
had
event.
It
is
at
event,
yeah.
It
is
elementary
and
for
link
do
we
have
settling
or
ad
link.
I
think
it's
like
that's
only
on
the
builder
also,
and
but
I
think
it
is
ad
link
on
the
builder.
Yes,.
E
I
I
B
B
B
I
Now,
I'm
also
very
happy
that
only
attributes
we
didn't
follow
the
thing
on
the
span.
I
just
double
check
now
for
everything
else.
We
have
add
for
the
least
semantic
and
yeah,
so
only
only
attribute
is
it
doesn't
follow
the
java
thing.
I
B
Yep
yeah!
No,
it's
true.
I
think
it's
never
looked
weird
to
me
when
I've
been
writing
code
like
it's
never
felt
strange,
it's
never
felt
unnatural.
B
F
My
the
set
attribute
seems
fine
to
me
the
attribute
if
you're
working
in
attribute
like
creating
attributes
and
doing
attribute
dot,
put
to
create
them,
and
then
that
lack
of
similarity
there,
I
think,
is
a
little
not
ideal.
But
I'm
not
sure
that
that's
worth.
F
Yeah
because
you're
you're
you're
putting
attributes
into
the
builder
when
you're,
if
you're
working
with
the
attributes,
object
but
then
you're
setting
them,
but
I
also
don't
know.
Like
I
mean
in
in
java
world,
I
mean
the
most
the
the
biggest
example
I
can
think
of.
Is
system
set
property
right,
which
is
it's
not
put
property.
B
I
It
is
overwriting
the
attributes,
it
is
the
same
thing
like
if
you
are,
if
you
are
writing
the
key,
the
thinking
you
will
override
that.
E
B
I
don't
know
anyone
who
argues
that
java
apis
are
consistent
or
make
any
kind
of
sense
clearly
hasn't
been
working
in
java
very
long.
The
sdk,
the
built-in
jdk
apis
are
just
madness.
They're
all
over
the.
B
K
I
I
spend
tonight
thinking
about
four
and
I
can
bring
you
a
couple
of
more
arguments
across
funds
for
this
or
or
we
need
to
make
a
decision
on
this
one
right
now.
I
wanna
have
some
time
to
think
about.
If,
if
not,
we
can
make
a
decision,
but
I'm
asking,
if
possible,
to
wait
until
tomorrow
morning
to
provide
more
thoughts
or
or
do
you
think
it's
we
want
to
make
the
decision
right
now
on
this.
F
Sport,
if
we're
gonna,
my
recommendation
would
be
if
we're
going
to
push
it
beyond
this
meeting.
We
should
schedule
a
meeting
for
some
time
tomorrow
to
force
the
issue,
not
because
otherwise
it
can
drag
on
and
get
home.
I
B
I
Okay,
so
then
I
let's,
let's
keep
it
as
a
tentative
now
and
if
we
find
good
business
and
if
I
find
good
reasons
I
will
share
with
everyone
and
but
let's
keep
it
as
a
tentative,
no
make
sense.
You're.
F
Gonna
have
a
hard
yeah,
yeah
you're
gonna
have
a
hard
time
over
turning,
but
that
sounds
reasonable.
I
B
I
I
F
C
B
And
for
non-english
speakers,
ambivalent
means,
I
feel
strongly
about
both
sides,
not
that
I
don't
know
the
answers
like.
I
have
arguments
for
very
very
strongly
on
both
sides
of
of
what
to
do
here.
B
Yeah,
I
I
think
that.
I
N
I
And
if,
if
the
get
default
from
the
resource
returns,
only
the
properties
that
we
know,
we
want
to
have
all
the
time
like
the
service,
I
think
that
was
the
only
decided
one
that
we
want
to
have.
I
think.
I
I
B
I
think
they're
certainly
possible.
My
issue
is,
I
think,
that
resources,
that
is,
the
concept
that
people
have
the
most
trouble
understanding
at
the
moment.
I
think
of
anything
I've
seen
in
open
telemetry
people
have
been
most
confused
about
what
a
resource
means
what
it
represents,
and
I
fear
that
if
we
make
it
difficult
for
people
to
configure
them,
they
won't
use
them
and
a
lot
of
the
value
of
our
kind
of
auto
detection
of
resources
will
kind
of
get
lost.
B
I
We'll
be
in
auto
configuration,
so
if
people
use
auto
configuring
will
become,
there
will
be,
there
will
be
in
the
java.
Agent
will
not
be
only
if
people
manually
configure
the
sdk
to
clarify
like
the
proposal
is
to
not
completely
remove
the
all.
The
spi
is
just
moving
to
the
auto
configuration
thing
which,
which
the
java
agent
most
likely
will
offer
via
auto
configuration
or
other
packages,
the
same
functionality.
I
I
That's
that's
a
that's
a
win
for
us,
because
then
we
are
spi
free.
The
other
thing
is
all
these
aspis,
it's
a
bit
harder
for
me,
especially
with
merge
to
guarantee
the
order
of
of
merging.
B
E
That's
a
good
point,
I
think.
Currently
our
nvar
might
be
wrong,
because,
if
it's
just
using
spi,
the
user
can't
override
any
of
the
other
resource
attributes
so
that
one
way
to
fix
it
is
having
an
order
for
providers.
But
another
way
is
to
just
move
it
all
into
the
auto
configuration
or
it
can
be
controlled
more
directly.
B
I
We
we
can
make
it
a
module
in
auto
configuration
that
people,
even
if
they
depend
so
so
we
can
even
auto
configuration
split
it
into
auto
configuration
for
resource
out
of
configuration
for
trace
that
depends
on
resource
and
so
on.
So
people
that
really
want
they
can
just
have
the
auto
configuration
for
resource
package
and
said
that,
and
that
does
the
magic.
But
that
does
not
it's
again.
I
B
Yeah,
I
mean
it's
kind
of
structured
that
way
already,
although
I
think
there's
only
is
there
only
one
public
method?
Maybe
there's
I
don't
remember
if
the
rest
of
the
one
or
no
yeah,
but
there
certainly
could
be.
I
think
we
will
want
to
probably
expose
auto-configured
metrics
separate
from
tracing
eventually,
but
we
don't
have
that
at
the
moment.
B
So
I
would
be
okay
if
we
move
it
out
as
long
as
we
do
expose
it
as
something
that
you
can
get
directly
at
auto
configuration,
because
I
can
definitely
see
a
use
case
where
someone
wants
to
auto
configure
the
resource
but
configure
everything
else
by
hand,
especially
in
like
distributions
I
know,
vendor
distributions
is
something
that
might
be
very
useful
to
have
like.
I
want
to
configure
my
custom
exporter.
I
That
sounds
very
reasonable
and
also
another
advantage
by
removing
it,
we
remove
a
bunch
of
environment
variables
from
the
sdk
that
people
don't
need
to
know
java
properties
and
stuff.
One.
E
I
B
I
E
I
think
so.
There's
some
tasks
here.
One
is
to
move
the
spi
interface
to
auto
configure
and
the
second
is
I'm
going
to
split
each
of
those
files
into
one
that
returns
the
resource
statically
so
that
users
use
it
in
sdk
and
one
that's
the
provider,
implementation
and
then
the
third
one.
I
think
we
can
get
rid
of
that
get
telemetry
sdk
resource,
then,
because
that
will
be
equivalent
to
the
default.
If
I'm
not
mistaken,.
B
E
B
L
Let's
it
could
be
trust
you
can
come
and
see
that
that
controls
last
thing.
Last
thing
I
wanted
it
was
correct.
Yeah,
that's.
I
Correct
last
for
the
os
resource
and
other
things
is
that
package
stable
or
alpha
it's
people.
If
that
is
stable,
then
we
need
to
expose
direct
access.
Give
me
the
entire
os
resource
not
get
that
attribute
yeah.
I
don't
know
if
you
want
to
have
an
extra
point
here
or
travel
by
the
first
one.
I
don't.
G
B
Making
sure
it's
not
publicly
public
access,
it's
just
stuck
into
the
into
the
default,
and
if
someone
has
a
need
for
it,
we
can
add
it
later.
I
Yeah
I
brought
this
even
though
it
was
closed
just
for
for
for
us,
because
I
I
think
we
we
can
say
no,
it's
fine,
but
I
was
like
okay.
This
was
we
said
no,
because
we
didn't
want
another
rc.
Do
we
want
to
consider
this
now
that
we
we
do
another
rc.
B
I
B
Yeah,
okay
number:
eight:
I
have
a
draft
jar.
B
B
So
I
guess
I
was
not
helping
with
the
leg.
Sorry!
No,
so
I
actually,
I
think
the
word
count
there
is
dumb,
it
doesn't
add
any
value,
so
I
explicitly
wanted
to
get
rid
of
the
word
count,
because
I
mean
it's
a
number
I
mean
I
don't
know.
I
didn't
didn't
seem
like
there's
a
lot
of
reason
to
have
the
word
count
sitting
in
there.
H
D
Spawn
attribute
limit
can
be
confused
either
the
number
of
attributes
or
either
the
length
of
the
attribute.
B
B
B
D
E
I
E
B
B
E
And
this
it
does,
I
guess,
attributes
prevent
limit
the
attribute
count.
E
F
E
I
Also,
by
the
way,
how
confusing
do
you
think
will
be
if
we
keep
the
max
thing
and
we
use.
L
E
B
E
And
then
one
other
point
just
a
random
one,
like
at
least
for
me.
I
consider
english
readability
to
be
more
important
for
code
than
variables
like.
I
think
there
is
this
expectation,
the
variables
contract,
remove,
verbs
or
remove
adjectives
and
stuff
and
methods.
Often
don't
so.
I
would
expect
some
difference.
Anyways.
B
F
You
know
my
I'd
say
my
concern
about
like
tying
them
together
is
just
addressed
by
just
dumping
the
you
know,
environment
variable
or
the
spec
name
in
the
javadoc,
just
just
so
that
people
have
a
way
to
link
them
together.
It
doesn't
have
to
have
the
same
name.
G
G
I
Okay,
the
last
one
related
to
this,
the
value
thinking
do
we
keep
it.
B
B
Nikita
you
were
the
one
who
pushed
for
that
originally
for
being
able
to
limit
the
the
length
of
the
attributes.
D
B
I
Yeah,
but
is
it
not
more
expensive
because
in
language
like
java,
you
already
have
the
string
format
constructed
if
you
just
keep
a
reference,
and
you
do
all
these
things
on
the
exporter
side?
Is
that
not
better
than
doing
it
in
the
sdk.
D
That
I
don't
have
a
strong
opinion
about
where
exactly
that
should
happen,
regard
time
or
export
time.
I
don't
have
opinion
on
that.
I
I
No,
we
can
build
a
small
span,
exporter
implementation
of
that
interface,
that
that
accepts
the
next
call,
and
that
does
only
the
copy
and.
I
B
I
Is
a
discussion
and
most
likely,
I
will
ask
exactly
this
question
that
I
just
told
you,
because
I
was
reading
through
the
pr
a
bit
while
nikita
told
me
that
there
is
a
pr
postponed
and
I
I'm
like-
I
would
do
this
later
in
at
least
in
languages
where,
where
strings
are
already
constructed,
I
don't
have
to
copy
that
data,
because
if
I
would
have
done
an
implementation
that
would
have
copied
things
in
another
buffer,
then
yeah.
I
would
have
not
copied
the
entire
thing.
E
D
I
N
B
A
rename
of
get
default
not
again
empty
to
empty.
This
is
changing,
get
default
to
be
empty.
I
B
I
But,
as
I
said,
they're
worth
checking
for
the
same
thing
in
a
couple
of
other
places,
I
think
that's
an
action
item
for
whoever
picks
this.
I
put.
I
put
some
comments
that
there
may
be
other
classes
where
we
have
yet
default.
That
actually
is
the
empty
thing
that
we
return.
E
I
By
the
way,
can
we
actually,
I
would
be
our
proposed
other
thing:
can
we
remove
tracer,
get
default,
it's
and
keep
the
provider
get
default,
and
if
you
want
to
have
a
default
racer,
you
just
do
three
trace
provider
get
default.get
like
you
go
through
the
entire
api,
how
you
get
a
trade
from
the
provider.
E
I
G
G
I
I
think
it's
an
inheritance
from
before
yeah
we
should
just
be
moving.
I
think
it
started
from
the
fact
that
initially
we
didn't
have
the
provider
and
I
think
it
it
backs
an
year
and
a
half
ago,
or
something
like
that.
That's
legacy.
M
B
I
For
the
moment
currently
unsampled
yeah,
but
if
we
add
another
beep,
has
default
one.
For
example,
the
get
default
will
return
one
for
that.
E
I
Sure
the
other,
the
last
one
we
and
I
proposed
the
not
versus
get
default
for
the
for
the
tracer
provider,
meter
provider
and
all
these
not
things
that
will
the
top
level.
No.
No
things.
I
K
L
I
B
L
E
B
I
Yes,
fine
one
thing
last
thing
that
I
would
like
as
number
10
is
not
nothing
changes,
but
I
think
everyone
agrees
on
that.
Can
we
double
check
that
we
don't
have
any
public
static
instances
or
or
public
static,
especially
instance,
so
we
have
a
bunch
of
like
singletons
with
get
instance
and
I'm
worried
by
mistake.
We
forgot
the
public
static
instance
like
the
the
static
variable
and
expose
that
as
the
way
to
to
access
the
instance
versus
get
instance,
method.
I
Okay
and
I
think
we
need
to
check
all
these
singletons
and
the
I
think
the
only
public
static
number
that
we
want
to
expose
are
constants
correct
in
semantic
conventions.
Everything
else
should
be
exposed
by
methods
I'll
file,
an
issue,
but
I
I
would
like
to
double
check
on
this.
B
So
the
only
place
we
have
a
public
instance
is
inside
this
open
tracing
shim.
At
the
moment,.
I
I
I
Very
happy
we
said
number
four:
we
have
24
hours
if
we
find
very
good
results
only
for
number
four,
but
but
so
far.
The
answer
is
no
for
that.
But
if
we
can
have
a
story
there,
we
may
see.
L
Final
rc
to
release
tomorrow
tomorrow,
night
yeah
tomorrowland.
I.
I
J
E
E
Let's
see
so
it's
end
of
day
26th
in
the
us.
I
guess
I
was
thinking.
B
Yeah
unrock
just
changed
all
the
build
script,
so
we
hopefully
the
the
publish,
will
actually
work.
I
End
of
friday,
san
francisco
or
west
coast,
whatever
just
to
make
sure
it's
not
end
of
friday,
japan,
because
that's
that's
right
now,.
B
So
we
do
have,
we
do
have
some
documentation
tasks
that
are
left
before
1.0.
But
that's
that's
fine!
So
we
have
time
we
could
do
they're
non-breaking
things.
We
can
get
documentation
and
that's
not
for
tomorrow
right
right,
yeah,
yeah,
I'm
just
saying
stuff:
we
need
to
get.
We
want
to
get
done
before
that
friday,
some
documentation
and
such.
I
Okay,
super
happy
with
this.
Thank
you
guys.
Nikita.
Do
you
have
10
minutes
to
discuss
about
me
being
an
to
your
pr
yeah,
absolutely.
I
So
yeah,
I
blocked
that
pr,
because
you
keep
asking
me
if
it's
blocking
or
not
do
you
understand
my
decision?
Why
I
blocked
it?
Yeah
jokes,
like
you
are
asking
for
it,
but
do
you
unders
so-
and
this
is
probably
important
for
java,
for
everyone
like
nikita-
is
trying
to
to
have
the
what
it's
called
engine
or
application.
F
We
all
watched
the
spec.
I
Creeper
so
but
yeah,
so
I
blocked
this
this
evening
because,
based
on
all
the
examples
and
all
the
documentation
that
is
available
there,
it
seems
to
be
only
a
java,
ish.
D
It's
currently,
it's
currently
certainly
a
java
issue
because
it
was
born
in
java
world
and
it's
currently
implemented
in
in
distro
for
java
yep.
So
I
have
to
get
back
to
evo
on
this
and
ask
him
again:
how
hard
does
it
does?
He
want
to
have
them
across
all
languages,
and
if
he
does
want
them
across
languages,
then
he
has
to
reach
to
other
languages
and
get
proper
examples
exactly
if
not.
If
not,
then,
yes,
we
will
then
reformulate
that
for
java,
only.
I
Yeah,
but
but
it's
not
about
that,
I
mean
to
be
correct
and
to
be
nice
to
the
community.
I
think
he
needs
to
reach
to
a
couple
of
languages
where
he
believes
this
is
the
case
and
ask
anyway,
because
even
if
we
go
with
the
java
specific
thing
now-
and
we
end
up
with
a
java
specific
thing,
a
ruby,
specific
thing,
a
python
specific
thing
is
not
going
to
be
good,
so
so
more
or
less,
even
if
you
follow
that
is
a
java
specific.
You
have
to
prove
that
is
a
java.
I
Only
problem,
it's
not
like
you
don't
choose
that,
because
you
are
lazy
and
not
doing
it.
I'm
I'm
trying
to
to
give
feedback
to
you
so
so
make
sure
you
explain
to
to
evo
that
he
needs
to
prove
us.
It's
either
a
global
problem
or
a
java,
only
problem
and
that's
what
I
did
not
see
from
from
the
content
of
the
the
issue
of
the
pr
there.
I
No
okay,
it
doesn't
make
sense
for
me,
it
makes
sense,
sorry
and
sorry
for
being
late.
I
I
mean
I
know
I
know
after
I
end
end
with
specs
last
thursday
or
something
I
just
wanted
a
break
from
the
specs
until
I
was
like
not
reviewing
anything
in
the
space
for
five
six
days
now
now,
I'm
back
and
I'm
trying
to
to
catch
up
with
things.
So
sorry
for
that:
okay,
okay,
thank
you,
everyone,
and
thanks
anurag
and
everyone
for
for
all
these
documents
and
stuff.
I
I
F
I
G
F
D
E
What
are
we
talking
about?
I
don't
know
what
we're
talking
about
sorry.
I
have
here
on
the
spec
dad
how
to
instrument
aws
sdk.