►
From YouTube: 2022-02-16 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
B
B
A
Yeah
well,
yeah,
there's
been
no
pretty
much
zero
activity
since
ted
made
those
comments,
hey
it's
it's
proving
hard
to
get
kind
of
agreement
or
you
know
even
much
activity
regarding
the
semantic
invention.
B
A
It
seems
like
it's
hard
to
get
all
the
different
kind
of
companies,
and
you
know
people
from
different
companies
to
agree
on
the
semantic
conventions.
A
Specifically,
I
wonder
what
yeah
I
wonder,
what
is
a
way.
A
B
Not
really,
maybe
you
know
you
know
so
currently
I
have
this
question
from
bogdan
or
a
suggestion
by
from
bogdan,
and
there
is
no
reply
for
two
weeks.
Maybe
we
can,
we
can,
you
know,
set
some
time
out
so
if
there
is
no
activity
or
reply,
we
can
just
close
the
conversation,
and
you
know
think
about
this
as
a
result,
but
yeah,
that's
something
that
probably
we
need
to
discuss
with
a
broader
audience.
A
C
I
mean
I'll
take
anything
bill.
It's
declaring
instrumentation
library
stable.
I
personally
don't
care
that
much
about
telemetry
being
stable,
because
I
can
somehow
accommodate
that
and
dashboard
and
stuff.
If
I
need
to
it's
not
ideal,
but
for
me
it's
not
a
blocker
for
rolling
out
instrumentation,
but
for
hotel,
like
they've,
already
declared
that
you're
not
supposed
to
label
it
stable
until
the
conventions
are
so
that's
the
main
motivation
for
me,
I
would
say
personally.
A
So
it
just
it
to
me.
They've
they've
done
that
they've
said:
okay!
Well,
the
you
can't
declare
instrumentation
stable
without
declaring
so
many
conventions
stable,
but
then
it
seems
like
there's
far
more
work
just
being
done
on
instrumentation
than
like
semantic
convention.
I
would
think
the
semantic
convention
is
like
kind
of
the
base
for
instrumentation
to
use.
So
I
don't
know
what
it
I'm
not
sure
why
there
isn't
more
focus
from
kind
of
the
community
on
getting
this
to
stable,
because
you
think
it's
in
everyone's
interests
right
but
yeah.
I
I.
C
Anyways
did
not
even
just
cement
the
conventions,
just
anything
in
the
spec
tends
to
flounder
quite
easily.
So
there
is
a
general
issue,
I
think
kind
of
tell
where
any
changes
the
spec
just
are
very
hard
to
get
through
yeah.
Yesterday
with
tigran,
he
said
he
doesn't
need
to
work
on
the
pr
workflow
or
the
spec
repo.
I
don't
know
if
that
will
help
or
not,
but
for
now
it's
sort
of
there's
no
process,
so
you
send
a
pr
someone
might
review
it.
Someone
might
not.
A
Yeah,
true,
that's
true.
I
did
listen
to
a
spec
meeting
the
other
day
where
they
argued
about
blogger,
name,
instrumentation
name
or
something
like
that
for
like
half
an
hour,
it's
just
kind
of
amazing
but
and
it
got
it
got
a
bit
passionate.
But
you
know,
I
guess
that's
good,
but
it's
also
kind
of
frustrating,
because
it's
like
well
it'd
be
nice
for
this
to
just
a
decision
to
be
made.
You
know.
A
A
B
Yeah,
so
we
do
have
these
two
pr's
open
and
we
have
two
more
items
to
be
kind
of
you
know
investigated
and
basically
the
second
one
after
the
retrace.
A
The
client
client
error
like
4xx,
should
it
be
marked
as
an
error
or
should
be
marked
as
like
onset,
that's
the
second
one,
and
and
to
be
honest,
I
don't
actually
feel
strongly
about
any
of
them.
The
most
important
thing
for
me
as
someone
who
would
be
using
it
is
just
that
it
becomes
stable
sooner
rather
than
later.
I
can
live
with
it
being
an
error
or
being
on
set.
A
What
I,
what
I
want
is
stable,
instrumentation,
so
I'll
put
the
let's
see
it
was
the
pr,
but
it's
just
hard
to
get
it's
hard
to
get
people
to
agree.
Like
always,
this.
A
C
Need
your
help,
but
like
bringing
this
to
the
spec
meeting
is
a
serious
thing
like
these
are
not
moving
forward.
Should
we
assign
actual
owners
that
have
more
authority
on
each
category
so
that,
when
it's
ready
to
merge,
emerge
and
base,
if
people
have
want
to
speak
up,
they
need
to
speak
up,
and
if
they
don't,
then
after
it's
merged,
they
can
still
try
to
address
their
post
merge.
B
Sorry,
I
just
wanted
to
say
that
it
will
be
good
to
have
some
some
process
like
that,
so
at
least
it
can
guarantee
us.
You
know
that
we
can
move
forward.
A
Yes,
I
don't
know
where
you're
based
dennis,
but
for
me,
and
I
think
it's
similar
for
anarch
the
spec
sig
is
on
at
3.
Am
so
mostly.
I
just
can't
attend
it,
but
it
seems
like
then
yet,
but
tigran
seems
to
like
what
you
said
seems
to
know
that
there
is
a
problem
getting
changes
and
there
is
a
kind
of
slowness
because
there
isn't
such
as
the
nature
of
kind
of
open
source
project.
There
isn't
like
one
person
who
can
just
decide
right.
A
You
need
to
have
some
kind
of
consensus,
but
I
don't
know
what
the
process
is
there.
If
there
is
a
disagreement
and
a
decision
has
to
be
made,
what
is
the
process
yeah?
Does
it
go
to
a
vote
or
something
I
I
don't?
I
don't
know,
but
this
is.
These
are
things
that
need
to
be
resolved
and
just
because
not
everyone
agrees
doesn't
mean
a
decision
shouldn't
be
made.
B
Exactly
yeah,
for
example,
for
for
for
this
pr,
for
it
tries
to
redirect.
I
already
like
I
have
four
or
even
more
approvals
and
just
one
open
discussion,
which
I
cannot
just
resolve
right,
because
there
is
no
reply
from
background.
Yet
so,
basically,
I'm
stuck,
and
I
have
no
idea
how
to
move
forward.
A
B
A
It's
it's
sometimes,
I
feel,
like
the
people
from
the
same
companies
don't
agree
with
each
other
as
well
like,
so
it
is
really
hard
to
to
get
agreement
even
within,
like
I've
seen
multiple
times,
people
from
like
splunk
just
they
don't,
they
can't
even
agree.
So
how
can?
How
can
we
all
agree?
It's
hard!
Hey,
it's
quite
hard.
So
yeah,
I
don't
know
how
they
do
it
in
other
open
source
projects.
What
what
do
they
do?
If
you
know
two
people
in
the
linux
kernel
disagree.
B
Like
they,
they
have
a
pretty
much
straightforward
process
there
so
like
some
of
them,
for
example,
linus
just
come
and
you
know
make
decisions
whatever
yeah.
C
Yeah
sorry,
it's
like
we
can
have
one
person
deciding
everything
hotel,
but
I
think
for
semantic
convention.
Maybe
there
is
still
the
possibility
to
have
some
defined
owners,
because
again
the
people
that
are
doing
the
work
should
have
more
sway
than
the
people
that
don't
do
any
work.
I
think
that's
one
of
the
tenants
open
source
and
yeah.
C
A
C
A
A
Well,
I
think
the
only
action
we
have
now
is
just
to
make
more
noise
and
try
and
get
agreement
be
a
bit
more
vocal
but
yeah
anything
else
to
anything
else
on
the
agenda
for
today.
B
Yeah,
maybe
we
can
spend
like
a
10
minutes
or
something
for
topics
that
we
do
have
in
in
this
roadmap,
for
we
won
for
first
table
version,
so
I
can.
I
can
share
my
screen
and
maybe
we
can
spend
some
some
time
on
it,
so
this
screen
so
yeah
this.
This
is
the
old
tap.
That's
well
like
I
got
merged
and
for
we
want,
we
do
have
this
stuff.
So
basically
it's
like
a
dpr,
that's
james.
B
You
had
created
and
hope
we
can
finalize
it
soon,
that's
another
one
for
it
right
and
redirects,
but
we
do
have
two
more
items
left
here.
So
this
is
the
first
one
required
attribute,
stats
and
context.
Propagation
is
something
also
probably
we
want
to
discuss
and
make
it
more
kind
of
clear
how
to
how
to
put
this
to
specification.
B
So
we
have
this
one
and
this
probably
something
that
maybe
we
can
take
and
also
like,
create
a
pr
for
this
one.
So
the
overall
idea
here
was
that
we
can
actually
kind
of
you
know,
decouple
those
things
or
just
separate
this
attributes.
B
So,
basically,
in
the
current
version
of
specification,
it
says
that
we
do
have
different
fields
here,
like
url,
and
all
of
them
are
not
required
target
hosts,
schema
and
whatnot,
and
we
also
have
this
paragraph
here
for
http
client,
saying
that
at
least
one
of
the
following
sets
of
attributes
is
required.
B
So
the
idea
was
that
at
least
what
we
can
do
is
like
to
say
that
for
http
client
sites,
basically
this
one
is
required,
and
that's
it
just
to
basically
just
to
clarify
this
this
part
here,
so
this
one
is
required
and
that's
it,
but
on
http
server
we
can
say
that
okay,
so,
for
example,
kind
of
there
is
a
dependency
between
different
attributes,
and
we
can
have
it
like
this.
This
set
is
required,
but
in
case
like
some
value
here
isn't
available.
B
For
example
hosts,
then
you
can
use
this
one
or,
and
if
some
of
this
is
not
available
like,
for
example,
pure
name,
you
should
use
peer
id.
So
this
one
is
kind
of
controversial
right
because
it
says
one
of
the
following
and
there
is
no
priority
so
at
least
for
http
server
side.
We
we
can
build
this
priority,
saying
that
it's
like
this
set
is
has
priority
one.
This
had
this
set
has
priority
two.
This
is
priority
three,
so
it
will
be
clear
and
for
client.
A
Yeah,
I
think
the
second
one
there
is
the
ideal
one
right,
that's
what
you
you'd
want.
A
If
I,
if
I
want
a
server
that
is,
of
course
that's
that's
the
one
I
would
personally
want
but
yeah
and
then
obviously
you
can't
always
have.
I
think
http
target
is
sometimes
not
always
available
like.
A
Targeted
except
the
first
one
though
or
maybe
maybe
the
host
wouldn't
be
available,
but
yeah
for
client.
I'm
not
what
I
wonder
why
they
specify
those
for
things,
because
you'd
think
the
client
wouldn't
necessarily.
A
A
B
Yeah,
that's
a
separate
topic
right,
so
that's
a
kind
of
you
know,
sensitive
information
and
all
this
synthesization
that
we
can
probably
apply
to
this,
but
at
least
like
at
this
this
part,
so
the
the
the
problem
that
I
probably
wanted
to
to
be
so
to
be
resolved
before
we
can
get
this
to
stable
version
is
like
at
this
to
make
this
clear,
more
cl,
to
bring
to
bring
more
clarity
here
to
this,
this
particular
product
paragraph.
B
But
then,
in
this
section
we
have
another
information
with
just
you
know,
tries
to
clarify
something
but
it,
but
from
my
point
of
view,
like
a
brings
more
controversy
even
right,
so
what
we
can
do
is
just
like.
I
just
remove
this
and
make
it
less
controversial
or
just
super
clear
what
exactly
should
be
done,
and
that
will
be
it.
So,
basically
that
that's
exactly
what
this
this
item
is
about.
A
So
for
the
server
spec,
what
does
it
say
to
say
the
exact
same
thing
in
terms
of.
C
B
So,
like
a,
we
have
some
information
here
like
http
url.
It's
usually
not
really
well
available
on
the
server,
which
is
totally
true
right
so
because
it's
like
only
available
on
the
client
side.
So
on
the
client
side,
you
have
to
string
url
string
and
that's
it,
but
on
the
server
side
you
do
have
this
kind
of
additional
stuff,
and
you
see
this
is
the
exactly
the
same
thing
here
for
for
for
a
client
for
a
server.
So.
A
A
You'd
want
yeah,
I
guess
for
me
when
I
read
that
I
kind
of
view
the
first
one
as
the
best
one
to
have
so
I
kind
of
viewed
it
as
a
priority
already,
but
maybe
that
should
be
specified
well.
B
Yeah
definitely
so
it's
it
feels
like
this
is
kind
of
you
know
some
kind
of
priority
list
or
something,
but
definitely
will
be
much
better.
If
we
can
make
this
super
clear.
A
B
Actually,
I
don't
know
for
to
be
honest,
I
was
not
participating
at
that
time,
so
I
have
no
idea.
C
In
java,
we
mostly
ignore
this
stuff.
We
do
make
its
own
clients
it's
on.
We
only
accept
http,
url
and
server.
We
only
accept
the
scheme
host
target
triplet
yeah
and
we
treat
hcp
and
net
separately,
so
it
might
also
have
the
net
stuff,
but
we
don't
tie
the
http
together,
because
that
just
makes
things
too
confusing.
A
C
B
Than
providing
too
many
options
yeah,
I
heard
the
same
exact
feedback
from
from
azure
sdk
team,
and
you
know
it
was
also
discussed
several
times.
Probably
people
also
like
confirmed
that
that's
exactly
the
experience,
so
it
looks
like
we
just
have.
You
know
some
some
additional
information
here,
which
we
can
just
just
remove
just
to
make
it
clear-
and
you
know
straight
forward.
B
There
is
no
issue
yet
it
wasn't
just
the
item.
Sorry
this
this
is
item
in
this
old
tap.
So
basically
that's
it
and
or
maybe
there
is
a
related
issue.
Sorry,
I
missed
that.
So.
B
Yeah,
it
was
created
by
ludmila
and
there
is
a
discussion
here.
Oh
it's
really
long
discussion,
but
yeah
ludmila's
actually
wanted
to
send
dpr
to
the
start.
Okay,
I
remember
the
one
so
I
can
actually
follow
up
with
her
about
this
and
try
to
get
more
information.
A
Ludmil
is
a
microsoft
as
well.
Isn't
she
right
yeah
she's
actually
from
azure.
B
Sdk
team
she
works
there
right
now.
Are
you
in
the
azure
sdk
yep,
oh
cool?
She
works
on
java
java
version
of
azure
sdk.
A
B
Yeah
and
basically
the
the
item
that
we
also
have-
and
probably
that
was
not
discussed
yet-
is
how
to
propagate
contacts
between
tries
so,
for
example,
it's
really
related
to
this
one,
but
it
can
be
kind
of
you
know.
Some
decisions
that
we
can
make
here
can
be
useful
in
many
different
scenarios,
including
retries
redirects.
B
So,
for
example,
if
I
have
a
library,
an
instrumentation
library
and
we
create,
we
are
creating
some
some
stands
out
of
it,
for
example,
for
it
rice
how
to
make
sure
that
we
propagate
information
that
we
require
between
these
spans,
for
example,
if
we
are
reusing
the
same
kind
of
instance,
how
we
make
sure
that
it's
clean
up,
if,
if
we
are
not
reusing
it,
which
storage
we
can
use
to?
Actually
you
know
propagate
is
between
between
tries
or
between
between
different
things
that
we
are
doing
in
in
session.
B
For
example,
it
also
can
be
useful
when
it
comes
to
http
2..
Maybe
it
can
be
separately
discussed,
but
we
also
have
this
here.
For
example,
if
you
have
this
kind
of
session
stuff
and
we
are
creating
several
spams
within
the
same
kind
of
connection-
how
we
can
make
sure
that
we
propagate
this
session
id
or
something
like
that
like
within
this
within
the
this
kind
of
window
or
something
so
this
this.
B
This
item
is
still
here
and
there
is
no
much
clarity,
so
maybe
we
can
also
have
just
you
know,
brainstorm
discussion
about
that
next
week,.
B
All
right
so
I'll
stop
sharing
yeah,
actually
guys.
I
need
to
drop
right
now.
I
have
another
meeting
and
that
is
like
a
heart
stop.
So
I
will
put
these
these
questions
or
actually
some
plans
for
the
next
one
to
the
document
and
we
can
take
it
from
there.