►
From YouTube: 2021-03-23 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).
A
Hey
good
morning
or
good
afternoon
for
some
of
us
we
are
four
minutes
past
the
start
time,
but
let's
wait
one
more
minute
and
please
add
yourselves
to
the
attendee
list.
Meanwhile,
please,
if
you
didn't
do
that.
A
A
B
B
B
But
we
want
to
formulate
that
just
give
you
the
idea
here
we
put
a
note
here
saying:
please
take
care
and
don't
try
to
implement
that
at
this
moment.
So
probably
the
idea
is
to
remove
this
line
entirely
and
introduce
a
separate
field,
saying
experimental,
but
not
recommended
something
like
this
just
want
to
get
people's
idea,
see
which
approach
we
prefer.
A
I
will
provide
some
background
on
that
folder,
so
it
was
originally
meant
to
prototype
things
you
know
or
to
try
to
specify
those
prototypes,
and
one
of
the
examples
was
see
pages
which
at
the
moment
this
was
still
in
the
works
and-
and
this
is
like
a
post,
for
example,
to
logs,
like
we
know
or
metrics,
that
we
know
that
there
will
be
part
for
sure
of
the
specification
and
we
just
we
know
that
they
are
experimental,
because
the
api
itself
is
not
finished,
but
we
know
that
we
are
going
with
it.
A
B
So
one
observation,
we're
saying
is
the
experimental
folder
seems
to
be
a
place
where
people
put
stuff
that
cannot
be
easily
categorized
and
since
the
experimental
folder
doesn't
seem
to
make
progress.
So
people
start
to
wonder
if
this
should
be
part
of
old
tab,
just
move
back
to
old
type.
So
for
things
that
we
know
we
will
move
forward
and
there
are
like
a
lot
of
people
driving
this
effort.
We
probably
can
use
the
experimental
status
tag
for
the
document
and
for
things
that
are
not
yet
ready,
not
like
making
good
progress.
A
C
C
B
C
B
Yeah,
so
I
I
think,
like
what
I'm
saying
is
for
the
semantic
convention.
We
probably
need
we
are
facing
the
same
problem
as
the
matrix
one
like
a
lot
of
people
come
and
see
this
the
spike
semantic
convention,
all
the
semantic
conventional
markets
experimental,
but
I
believe
some
are
relatively
mature,
like
http
and
some
are
brand
new,
so
people
shouldn't
start
so
how
about
we
start
from
that?
B
C
I
think
you
are
addressing
two
different
issues
here
and
you're
talking
about
two
different
things,
so
indeed
we
probably
need
something
to
clarify
the
maturity
of
the
semantic
conventions,
which
is
not
the
same
for
every
file
and
second,
is:
do
we
keep
this
folder
or
not
yeah.
B
You're
right,
I
I
I
try
to
bring
this
together
because
I
want
to
avoid
the
situation
where
I
start
to
like
tell
people.
The
spike
has
experimental
something
still
early
draft
don't
implement
and
something
is
ready
experimental.
We
want
people
to
implement
that
across
multiple
languages
and
then
there's
question
coming
and
say:
hey,
there's,
experimental
folder.
Why
don't
you
just
put
stuff
in
the
folder?
So
I
I
think
I
want
to.
D
B
C
B
C
Look
at
look
at
the
logging,
libraries
or
tab
it
was
merged,
but
it
clearly
says
that
this
is
an
attempt
for
an
experiment
because
it
clarifies
whatever
we
want.
We
believe
it's
useful,
but
we
still
have
to
prove
unless
we
want
to
make
the
process
of
hotep
six
months
and
keep
that
open
for
six
months
until
people
are
implementing
the
the
the
thing
that
we
can
do
is
create
a
tab
saying
that
hey
this
is
the
initial
experiment
we
will
implement
couple
of
prototypes.
C
A
That
that
sounds
good
to
me.
If
we
could
just
have
a
note
on
that
clarification
of
yours,
really
it
would
be
nice.
F
A
D
Hey
hello,
thank
you.
Okay,.
A
D
A
D
Of
time
way
too
much
like
like
900
seconds
so,
which
I
believe
is
way
too
much,
so
I
was
interested
in
like
what
does
the
specification.
C
G
I
guess
we
didn't
specify
it
out
right,
like
you're,
still
able
to
have
a
back
off
maximum
time,
or
is
this
saying
that
the.
C
Think
I
think
what
he
says
is
if
people
are
choosing
randomly
probably
would
not
be
the
right
choice,
which
I
don't
know.
We
can
argue
about
people's
quality
of
choosing
the
defaults
for
this
kind
of
things.
But.
G
Well,
it's
it's
also
you're
choosing
how
reliable
versus
efficient
you
want
your
things
right
like.
I
think
this
might
be
a
thing
that
we
want
to.
Let
people
have
some
configuration
on,
so
you
can
provide
a
bounds
or
like
a
default,
but
I
don't
think
you.
I
don't
think
you
lock
people
down.
I
I
agree
there
should
be
a
good
default.
A
really
good
default
specified.
F
This
honestly
feels
like
it
should
be
the
domain
of
the
specific
back-end
implementation
like,
for
example,
I
would
expect
honeycombs
versus
you
know
somebody
else's
implementation
to
to
require
different
back
off
parameters
and
then
add
on
to
that.
Like
other
things,
so,
like
honestly,
I
think
there's
one
define
like
a
backup
configuration
defined
in
the
grpc
spec
I
mean.
Could
we
just
borrow
that.
C
Yeah
but
but
there
is
a
difference
I
mean,
if
you
are
using,
the
collector
in
between
your
back-end
and
stuff
is
probably
different.
If
you
are
talking
over
the
internet
with
your
back
end
is
probably
a
different
default
and
yeah.
So
there
are
multiple
scenarios
if
you
are
in
the
same
network,
if
you
are
over
close
domains
and
stuff
and
so
on
so
forth.
So
I
don't
know,
I
will
just
say
it's:
it's
too.
C
A
Yeah
actually,
my
question
for
the
for
the
for
this
case
is
that,
honestly,
if
there
is
any
specific
urgency,
but
otherwise
I
could,
I
would
not
work
on
them.
D
There
is
no
urgency
so
so
there
is
an
issue
in
python's
sake
like
900
seconds
is
like
too
long
is
there
any
you
know
different?
Is
there
any
value
defined
by
the
specification
that
that
the
sick
could
follow
so
yeah?
There's.
A
A
Yeah
I
mean
if
you,
if
you
in
the
in
the
near
future
or
anytime
in
the
future,
you
or
anybody
else
happens
to
have
a
specific
use
case
scenario
that
it's
not
easy
to.
You
know
to
solve,
and
we
need
this.
Of
course,
just
please
ask
what
you
are
seeing
the
issue,
but
I
think
that,
as
spoken
said,
we
had
a
lot
of
stuff
to
do
so
yeah,
and
I
don't
think
we
have
time
for
this
and
it's
not
easy
at
all
because
of
the
backend
part
that
oren
also
mentioned.
B
B
Yeah,
so
should
we
make
a
comment
here
I
can
summarize
what
we
discussed
and
at
the
labels.
Yes,
we
don't
have
to
communicate
that
again,
okay,
I'll
I'll,
do
that
if
we
have
other
topics,
please
continue.
A
Yes,
we
have
one
more
by
tyler
tyler
you're
around
it's
about
baggage.
H
It's
really
quick:
it's
should
the
baggage
propagator
clear,
the
pre-existing
baggage.
I
open
an
issue
that
issue
just
identifies
the
fact
that,
in
the
baggage
specification,
the
extract
functionality
of
the
baggage
says
that
it
needs
to
clear
I'm
sorry.
It
needs
to
return
a
context
with
no
baggage
in
it
if
it
cannot
extract
the
baggage,
but
I'm
wondering,
if
needs
to
return
a
contact
with
no
baggage
in
it
implies.
It
needs
to
clear
any
context
baggage
that
pre-existed
that
was
handed
to
the
function.
C
So
I
think
this
this
came
from
the
fact
that
we
wanted
to
be
implemented
to
support
multiple
propagators
at
the
same
time.
So
the
idea
was
that
the
idea
was
not
to
clear
but
more
or
less
to
not
change
the
context
or
not
put
an
empty
baggage
if
you
failed.
A
H
H
That's
what
I
thought
as
well,
but
the
way
that
it's
specified
doesn't
state
that
it
states
the
opposite.
No,
I
think
I
think,
I'm
sorry
not
the
opposite
it
just
it
just
means
that
like
yeah,
so
if
you
do
have
multiple
baggages
right
now,
the
way
the
specification
is
written.
If
you,
you
know
correctly
deserialize
one
and
then
pass
it
to
another
it'll
clear
that
first
one
if
it
was
compliant
with
the
specification
currently.
C
C
C
It
shouldn't
be
so
so
if
you
fail,
if
you
fail,
you
shouldn't
touch
the
context.
That's
that's
the
phrase
that
probably
we
we
miss,
but
I
think
it's
very
clear.
The
in
the
trace
that
one,
I
think
we
clarify
on
the
tray
side
or
maybe
I'm
wrong,
but
oh,
we
did.
H
That
okay,
I'd
appreciate
it,
I'm
happy
to
correct
the
language.
I
was
just
like
very
kind
of
confused
on
that.
A
C
D
C
One
second:
to
clarify
this
title:
I
sent
you
the
link
on
the
the
generic
api
propagators
on
the
extract.
It
says:
let
me
present
quickly.
It
says
if
a
value
cannot
be
parsed
from
the
carrier
for
a
cross-country
concern.
The
implementation
must
not
throw
an
exception
and
must
not
store
a
new
value
in
the
context.
C
Okay,
you
know
preserve
any
previous
existing
valid
value.
So
it's
clarifying
that
you
want
to
preserve
any
previous
existing
thing.
So
so,
essentially
it's
exactly
what
I
was
thinking
that
if
you
fail,
you
should
not
touch
the
context.
H
C
C
H
We
do
that
for
sure
yeah
go,
we
do
it
as
well,
but
I
we
were
auditing
and
I
didn't
like
didn't
want
to
release
something
that
wasn't
compliant,
and
so
I
was
like
this
doesn't
sound
right.
So
I'm
just
double
checking.