►
From YouTube: 2023-03-23 meeting
Description
Instrumentation: Messaging
B
B
A
B
A
B
B
C
B
Where
are
you
based,
Carlos
I
am
based
most
of
the
time
in
Czech
Republic
yeah
I
am
usually
traveling
to
Mexico
for
winter,
like
two
or
three
months,
but
other
than
that
I'm
in
Europe
yeah,
actually
yeah
yeah.
What
about
you?
Where
do
you
base
yourself
I'm.
B
Totally
for
me
than
most,
you
know:
I
never
got
used
with
cold
weather
myself.
I
mean
I,
know
that
I
have
some
friends,
not
only
in
Czech
Republic
Europe
in
general,
that
are
very
fond
of
winter
cold
winter.
You
know,
because
of
winter
sports
and
all
that
stuff,
but
myself
never
I,
never
clicked
with
winter.
B
I
saw
that
joao
is
not
joining,
as
you
know,
there's
still
this
mismatch
between
time
zones,
so
he
has
an
overlap.
He
said
he
left
some
comments
on
your
tab,
Johannes,
so
that's
great
and
doing
it.
So
your
your
message
following
up
on
destination
kinds,.
C
Yeah
yeah
I
saw
that
it's
been
removed
now,
which
is
again
I.
Think
it's
fine
I
just
propose
that
we
try
to
retain
the
information
that
it
had,
but
I
think
it
was
being
expressed
in
an
awkward
way.
So
I
was
we
had
two
weeks
ago.
I
was
away
last
week
on
vacation,
but
two
weeks
ago,
I
sort
of
took
an
action
to
propose
some
text
to
describe
the
alternate.
D
Yes,
so
as
we're
diving
I
can,
let's
have
a
look
at
the
Fortress
board,
which
should
hopefully
be
up
to
date
and
reflect
that.
So
here
we
see
that
the
was
working
on
and
we
were
talking
about
those
are
resolved
now.
So
there
is
a
the
message:
destination
was
removed,
the
pra
streamer
is
merged
and
there
is
I
think
it's
this
new
issue
that
Ludmila
submitted
that
covers
this.
E
D
D
We'll
be
assigned
it
to
you
and
we
move
it
up
the
queue
to
to
show
that
somebody
is
working
on
it.
So
we
see
that.
E
E
D
D
Maybe
let's
chat
about
it
later
and
then
there
is
the
re-allowing
spending
creation,
I,
think
one
of
those
Powers
got
reopened
and
this
one
working
together
with
fast
the
F
function,
is
a
Service
Group,
and
here
is
the
one
that
to
end
that
we
just
assigned
to
you
India
for
others
who
think
the
sne's
here
are
just
your
placeholders
I
think
nobody's
working
on
those.
D
So
if
there's
any
cyclists
left
it
can,
we
can
take
over
or
start
any
of
those
issues,
I'm,
not
sure
what
Melody
this
year
is
assigned
to
you.
How
to
trace
settle
operations.
A
D
D
I
just
want
to
say,
I
think
that
would
be
something.
Maybe
that
would
I
mean
I.
Think
from
my
point
of
view
from
stability.
I.
Think
one
of
the
most
high
priority
that
we
should
get
a
handle
on
is
is
the
the
metric
stuff
that
we
get
at
least
some
rough
idea
of
how
metrics
could
look
like
and
also
this
the
missing
set
operation
attributes.
That's
still
a
gap.
A
Yeah
I
can
actually
make
a
matrix
proposal
and
then
work
on
settled,
but
there
is
also
the
last
one
I
Christian
pointed
me
to
eat.
I
think
this
is
related
to
your
Otep,
directly
and
I.
Think
without
it.
We
oh
like
what
is
your
opinion
all
of
you
about
the
priority
of
this
one.
D
Just
in
terms
of
issues
I
think
there's
an
other
I
think
it
ties
also
into
this
one.
Doesn't
it
is
proposal
that
you
made
here.
A
A
So
the
question
is:
where
would
we
put
attributes-
and
it
sounds
like
one
of
the
solutions
that
we
might
need
to
change
this
pack
and
allow
links
to
point
to
nothing
so
that
we
can
record
attributes
on
them
and
I
think
it's
it's.
It
should
be
resolved
within
your
autop.
D
D
And
this
case
we,
we
would
basically
run
into
this
case
when
we
receive
messages
that
don't
have
a
context
attached
to
them.
Yes,
and
then
the
idea
would
be
that
we
create
a
link
to
some
kind
of
a
empty
or
ephemeral
context
to
still
capture
that
there
was
a
message
that's
handled,
and
we
also
put
those
attributes
that
are
relevant
on
this
link.
D
D
I
I
definitely
think
it's
a
valid
console
I'm
just
now
thinking
whether
it
makes
sense
to
pack
that
into
that
old,
tip
tool
or
or
not,
I
get
your
point
where
it
should
be
in
the
old
tab,
because
here
we
are
talking
about
order
or
the
link
related
stuff
there.
So
it
would
make
sense
to
also
answer
that
question.
D
A
D
B
A
B
A
I
think
this
was
our
working
proposal
so
far.
B
E
D
A
A
D
I
see
so
that
is
basically
the
kind
of
the
I
I
know.
I
see
nobody
say,
that's
the
other
way
around
here
we
basically
here
we
here
we
create
and
invalid
I
mean
in
quotes
and
not
an
English,
but
to
spend
less
context.
I.
D
D
Yeah,
just
look
out
for
like
both
both
Solutions
are
not
really
like.
They
are
similar
in
terms
of
use
cases,
but
they
don't
like
they
they're,
not
really
Alternatives,
that
we
say
okay,
all
or
the
other
source.
The
problems
that
we
have
I
think,
let
me
go
is
the
one
we
still
have
problems
that
are
that
are
solved
that
are
not
solved,
but
that
are
solved
by
the
other
Solutions
so
that
it's
just
or
to
put
it
differently.
They
kind
of
dissolve
different
problems.
Yes,.
A
A
I
think
it
will
be
answered
with
your
autop
right
so,
like
the
the
Euro
tab,
implements
the
0.1
in
this
proposal
right.
So
sorry,
if
you
just
scroll
up
so
the
first
picture,
so
for
each
message
we
are
sending,
we
are
suggesting,
as
one
of
the
solutions
to
link
to
it
and
include
all
per
message,
attributes
on
link
attributes.
A
D
D
This
here
and
then
the
the
Alternatives
have
the
Alternatives
here
that
we
have
is
either
yavi
I
mean
here
proposed
that
we
have
this
invalid
span
context
on
links
where
we
bought
the
attributes
on.
The
first
alternative
is
that
we
just
ignore
messages
with
no
context
when
it
comes
to
tracing.
That's
probably
the
worst
one,
and
the
other
solution
is
that
we
use
events
to
put
like
the
message.
Specific
information
on
yeah.
D
In
order
to
be
here,
that's
that's
a
problem
exclusively
for
batch
cases,
because
when
we
have
a
single
message,
we
anyway
can
put
everything
on
the
span.
D
I
mean
what
is
the?
What
is
the
benefit
that
we
get
from
using
or
putting
the
attributes
on
links
instead
of
putting
the
attributes
on
events?
I
guess
it's
consistency
with
other
approaches
where
we
have
where
we
have
links,
or
we
have
the
context
that
we
can
create
links.
A
There
was
not
instrumented
yesterday,
but
now
it's
instrumented
and
yesterday
you
got
Advance
now
I
got
links
without
getting
into
the
discussion
between
those
two
like
it's
just.
This
would
be
difficult
for
backgrounds
to
maintain
and
it
will
be
difficult
for
users
to
understand.
A
And
if
we
decide
that
links
go
to
like
the
the
problem
that
today,
if
we
do
add
links,
if
we
will
edit,
then
they
will
go
to
the
same
places,
traces
and
events.
You
know
nobody
knows
where
they
go.
D
D
To
to
to
make
that
clear
in
the
old
tab,
I
mean
that's
basically
also
like
one
other
requirement
that
then
we
would
have
in
the
old
tab
that
we
would
need
to
have
those
invalid
span
context
just
to
preserve
link,
attributes.
D
I
can
bring
up
I'm
into
spec.
Pr
I
guess
would
be
would
be
rather
small
because
it
would
kind
of
then
probably
say
here
that
implementations
must
not
ignore
links
with
invalid
span
context.
D
And
must
enter
and
they
always
must
then
consider
these
links
and
send
or
format
those
links
to
the
exporter,
because
I
think
now
they're
dropped
before
I.
Think
a
question
will
then
come
up
here.
Our
back-ends
treat
those
tweet
those
links
with
link
on
its
main
contexts.
A
Maybe
we
can
also
do
the
following.
Maybe
I
will
think
and
try
to
prepare
some
some
understanding
how
it
can
work
as
events
paint
and
Maybe.
A
D
Oh
yeah,
that
is,
that
is
an
other
good
question
that
would
we
need
to
figure
out,
I
mean
currently
I,
don't
have
any
back-end
that
supports
attributes
and
links.
D
So
it's
it's
a
good
question
here
whether.
D
D
E
D
Reason
why
here
I
I
put
this
command
here
is
also
that
yeah
when
I
think
about
back-ends
supporting
this
and
yeah
I'm
working
on
one
of
those.
Since
yeah,
you
need
to
figure
out
what
to
do
with
those
invalid
spawned
context
when
it
comes
to
like
yeah
having
our
spans
and
phrases
go
through
index,
Source
or
visualization
tools,
or
you
build
a
topology.
The
question
is,
you
need
to
have
then
some
special
handling
there
for
those
invalid
contexts.
A
A
D
Would
it
be
a
breaking
change?
Yeah
I
mean
I,
I
think
the
language
what
we
propose
in
the
old
tip.
Now,
let
me
check,
but
I
think
we
we
we
more
or
less,
allow
anyway
instrumentations
to
either
create
links
or
have
events,
if
you
remember
correctly,
but
let
me
check.
D
D
We
always
require
a
link.
Yes,
but
I
think
we
don't
make
this
requirement
of
creating
a
link
explicit
when
there
is
no
context
on
the
message.
I
think
we
are
deliberately
kind
of.
A
D
So
that
that's
actually
that's
actually
a
a
case,
that's
not
handled
by
this
PR.
A
But
essentially,
if
we
allow
on
the
on
the
receiving
site
to
create
links
or
events
right,
then,
however
bad
the
experiences
with
it
we
won't
be.
We
will
be
able
to
improve
it
later
by
saying:
okay,
now
it's
on
the
links.
D
D
I
think
that
is
something
that
that
that
we
should
strongly
recommend
I,
wouldn't
say
required.
I
strongly
recommend
that
if
you
get
the
context
there,
you
should
create
a
link
and
open
question
is
here,
but
what
what
do
you
do?
If
you
get
a
message
and
there
is
no
or
a
text
request
question?
Is
you
get
a
badge
and
there
is
no
context
information
on
the
messages
in
the
batch?
What
do
you
then?
What
do
you
do
then?.
A
I'm
Sudan
The
Proposal.
Let's
say
that
you
create
the
the
instruments.
Instrumentation
should
create
an
event
and
then
back
ends
would
expect
to
get
a
link
or
event
wait,
and
then
we
might
need
to
think
about
the
special
semantics
of
this
event,
or
we
might
think
that
Google
probably
should
describe
like
the
name
of
the
event
and
the
attributes
go
there,
and
then
we
can
see
as
backends
if
they
on
board
onto
this,
if
they
give
any
feedback
or
they
don't.
A
If
it
becomes
problematic,
then
based
on
this,
we
can
consider
this
problem
more
serious,
because
it's
my
assumption
that
it
will
be
difficult
to
embrace.
A
D
A
B
D
Awesome
thanks
that
maybe
we
can
yeah
or
something,
let's
move
that
in
a
bit
up
the
queue.
D
Okay,
then,
let's
clarify
this
here
and
that
also
will
Daniel
go
like
any
results.
You
build
and
put
into
this
PR
here
and
then
the
other
other
things
I
mentioned
this
year
regarding,
like
just
a
russet
here,
oh
I,
think
there's
an
other
like
two
things
that
I
think
we
should
look
into
this
product
either.
One
is
the
getting
at
the
about
metrics
and
the
other
is
the
like.
The
ECS
compatibility.
D
I
mean
I,
don't
know
for
the
ECS
compatibility
I,
guess
it's
just
you're
going
through
the
elastic
common
schema
going
through
our
attribute,
see
if
there's
any
overlap
and
seeing
whether
there's
anything
that
we
both
need
to
change
to
a
line
or
to
avoid
confusion.
I
just
know
that
ECS
also
has
like
terms
I,
think
for
source
and
destination.
D
I.
Think
here
will
be
worth
looking
better.
Any
changes
from
our
site
would
be
needed
and
the
second
part
I
think
that
would
be
important
to
get
an
idea
before
stability
is
to
just
get
a
recipe
about
how
metrics
should
look
like.
A
D
D
And
for
four
four
metrics
I'm
just
curious
to
is
there
already
something
that
can
that
can
serve
us
as
a
prototype
from
Azure
SDK
site.
A
Yes,
we
do
have
national
drink
metrics
in
Java
and
I
can.
Maybe,
if
we
have
time
next
time,
I
can
share
the
findings
there.
What
you've
done
there
or
hear
your
feedback
and
then
come
up
with
the
proposal
based
on
this
feedback.
D
C
I
sorry
I
mean
with
our
messaging
product.
I
mean
I,
definitely
have
a
lot
of
experience
with
metrics,
but
in
terms
of
open,
Telemetry,
metrics
I,
don't
have
any
experience
with
but
like
with,
with
implementing
any
anything
related
to
open
Telemetry
specifically,
but
I
mean
certainly
our
our
product
that
I'm
very
familiar
with
has
all
sorts
of
metrics
there's
not
exposed
through
the
open
Telemetry
standard
of
metrics.
D
Oh,
yes,
that's
that's
cute,
but
I
think
that
it
will
just
be
then
immensely
helper
when
you
work
on
that
to
see
to
get
like
a
your
experience
in
there
kind
of
starting
to
come
up,
I'll
sing,
it
open
Telemetry.
That
kind
of
also
captures
your
needs.
Then.
D
But
awesome,
then:
let's
then,
let's
kick
that
off
next
time.
I
will
put
that
on
the
on
the
agenda
and
let's
have
a
Time
boxed.
Maybe
then,
let's
reserve
the
last
I,
don't
know
30
or
20
minutes
to
start
brainstorming
about
that
and
let's
let
love
Miller
kicked
it
off.
D
Awesome,
okay,
so
great,
we
have
a
quite
some
stuff
in
Flight
here.
So
let's
see
that
looks
great
thanks.
Everybody,
then
what
I
yeah
one
question
I
put
here
is
that
I
saw
other
projects
have
like
a
for
example.
Http
stabilization
has
a
project
project
tracking
issue
and
that's
mostly
I
guess
a
question
also
for
you
Carlos.
Do
you
think
we
need
something
similar
to
formalize
our
working
group
here
or.
B
Oh
yeah,
yeah
I
saw
your
question
by
the
way.
Last
week,
I
forgot
to
answer:
I
think
that
we
can,
if
we
want
now,
usually
these
ones
are
created
when
we
want
to
also
Define
deadlines,
and
everything
and
I
know
that
we
are
kind
of
blocked
so
to
speak
on
the
HTTP
semantic
conventions,
work
group,
but
for
probably
for
clarity
reasons,
even
if
we
don't
have
the
Appliance
on
timeline
and
anything
like
that,
we
can
do
that.
B
So
in
that
case,
if
that
sounds
right,
I
can
go,
and
you
know
at
least
you
know,
do
a
kind
of
intention
of
the
you
know,
Declaration
of
instance,
of
intention.
Sorry,
just
mentioning
what's
like
this.
Is
your
table?
We're
working
on
this
is
the
the
dashboard.
We
don't
have
a
you
know
timeline,
but
we
are
working
on
that.
D
B
Though,
alternatively,
even
though
it
could
be
kind
of
weird,
but
probably
it's
a
valid
thing
that
we
said
timeline
and
it's
like
to
redefine
like
waiting
for
the
for
the
HTTP,
cementing
semantic
conventions
group
to
to
finish
their
stuff,
that
also
could
be
a
thing.
B
However,
one
of
the
things
by
the
way,
which
is
also
well-
and
actually
you
can,
let
me
know
what
you're
feeling
about
that
these
are
in
theory,
each
working
group
will
have
approvers,
you
know
or
people
that
can
actually
you
know,
help
with
these.
B
So
that
means
that
if
we
get
ever
get
blocked
because
we
don't
have
enough
people
approving
the
PRS,
we
can
probably
have
people
there
and
I.
Think
that's
Pro
I,
don't
remember
like
Who
besides
lewd
meal
and
me
are
actual
approvers
in
the
spec.
At
this
moment,.
A
D
D
So
it's
up
to
you
yeah!
So
let's,
let's
wait
with
that!
Yeah!
Okay!
Perfect!
That's
right!
Here!
Wait
until
we
have
a
timeline!
Okay,
great!
Then
there
is
this
PR
here
I
wanted
to
talk
a
bit
about
Miller
commanded,
as
I
said
yesterday
that
Miller
commented-
or
there
are
some
other
comments-
comments
too.
A
Yeah,
so
it
sounds
like
the
client.
Id
is
not
messaging
specific
problem.
D
D
That's
what
the
Christian
says,
but
I
think
your
point
is
still
valid,
that
it's.
It
should
not
be
messaging
specific,
but
it
should
be
something
that
is
generic
and
can
be
used
for
multiple
different
areas.
E
D
D
Or
just
to
give
context
to
others
here
what
this
PR
does
it's
it's
there's
this
issue
here
that
submitted
by
Christian,
and
here
we
meet
the
attributes
here,
a
bit
out
of
date.
We
have
a
DOT
here
now
and
here's
in
instead
of
the
underscore,
but
it's
pretty
much
the
same
thing
with
like
consumer
ID.
We
have
client
ID.
D
We
actually
also
have
Kafka
client,
ID
and
Christian
asks
to
have
a
messaging,
a
generic
messaging
client
ID
instead
and
generic
messaging
client
group
I
didn't
take
into
the
client
group
because
I
think
consumer
group.
This
is
this
is
something
different.
You
wouldn't
expect
the
consumer
group
to
bring
client
group,
but
I
think
the
client
ID
is
valid
to
pull
it
out
of
messaging
specific
stuff
and
also
replacing
this
consumer
ID
with
client
ID.
D
It's
not
it's
the
only
thing
with
our
current
definition,
where
consumer
ID
is
different
from
kind
that
is
Kafka
because
for
Kafka,
according
to
the
definition,
we
pack
also
the
consumer
group
into
this
consumer
ID,
which
right
and
later
here
propose
not
to
do
I
trust
in
your
makeup
proposal
to
kind
of
ditch
this
consumer
ID
have
a
client
ID.
Instead,
a
generic
one
and
yeah.
D
Simplify
by
ditching
consumer
ID
and
making
those
kind,
that
is
generic,
so
that
in
that
way
we
will
place
three
attributes
with
just
one
and
then
the
question
here
is
whether
this
kind
ID
attribute
that
Luke
Miller
brought
it
up,
whether
this
should
even
be
messaging
specific
or
whether
we
can
find
like
kind
of
a
generic
definition
for
this.
This
can
share
can
be
shared
with
others
like
this
Cosmos,
for
example,
who
also
have
a
we
have
a
cosmos,
DB,
client,
ID
and
yeah.
Maybe
we
can
find
a
generic
place
to
have
that.
A
Of
I
was
looking
mostly
into
resource
attributes,
but
it's
a
road
question
that
it's
probably
not
the
resource
attributes
and
then
we
should
probably
Define
a
new
or
namespace.
D
D
D
Yeah,
it's
a
good
question:
I
mean
I.
I.
Think
your
your
your
point
here
is
definitely
valid
because
I
mean
when
we
we
already
have
two.
We
have
it
in
a
cosmos
DB,
and
we
have
it
here
and
that's
where
the
indicator
and
I
can
see
this
coming
up
in
many
other
many
other
instances
too
so
I
mean
the
other
approach.
D
Would
would
that
we
say:
okay,
we
keep
it
as
messaging
kind
ID
for
now
and
in
future,
if
something
crystallizes
to
make
this
generic,
we
can
still
handle
this
via
schema
conversions,
that
it's
moved
out
into
a
more
generic
place,
but
I
think
that's
something
I
think
we
we
want
to
avoid
kind
of
putting
something
in
there
when
we
think
okay
in
future,
maybe
it
will
change
I
guess
it
might
make
sense
too,
at
least
think
about
possibilities
of
a
of
a
more
stable
solution
right
away.
A
Yeah,
well,
it's
a
counter
argument.
If
we
try
to
say
what
it
is
in,
like
the
general,
there
are
no
one
either
in
messaging
or
in
for
for
all
Technologies.
A
A
D
Yes,
I
get
your
point
when
you
take
it
out
of
context
and
just
look
at
this
attribute
without
context
to
messaging
or
or
database
like
Cosmos
DB
here
about
the.
What
what
does
this
attribute
tell
you.
D
A
C
I
mean
I
think
where
this
could
be
useful
is
like.
If
someone
is,
you
know,
querying
the
back
end
like
they
they're
they're
debugging,
a
problem
or
something
unusual
happened,
they're
searching
it,
they
find
a
trace
and
they
kind
of
want
to
know
like
you
know
who
did
this
right
like
this
is
something
that
they
want
to
be
able
to
trace
it
back
to
to
an
instance.
C
A
D
I
mean
I
mean
regarding
mcquester
I,
definitely
think
yeah,
this
kind
ID
is,
it's
is
useful,
I
think
it.
It
definitely
makes
sense
to
capture
the
information.
I.
Think
it's
just
the
the
question.
In
what
form
do
we
capture
the
information
like
in
the
message
specific
namespace,
or
do
we
go
the
route
like
that
seems
more
economic
and
try
to
come
up
with
a
generic
attribute
that
captures
it.
D
But
then
other
thought
here
that
comes
to
me
is
when
we
find
the,
if,
if
that,
if
there
would
be
a
generic
attribute
and
it's
not
specific
to
messaging-
and
that
is
kind
of
shared
with
others
as
a
wonderful
actually,
if
that
is
really
helpful
to
people
who
use
this
attribute,
because
I
wonder
if
there
is
a,
for
example,
a
scenario
where
you're
going
to
use
this
attribute
the
in
in
use
cases
where
you
don't
have
messaging
or
database
specific
use
cases
administer
any
use
case
where
you
would
want
to
where
you
would
want
to
search
for
this
client
ID
in
an
in
an
overarching
day,
covering
both,
let's
say
a
database
and
messaging,
and
make
sense
of
that
I'm.
D
Not
sure
of
that
so
in
that
regard,
maybe
it
makes
more
sense
to
have
it
under
specific
namespaces,
but
yeah.
A
D
D
Yes,
I'm
in
the
middle
ground
with
the
I
think
Ken's
initially
initial
proposal
for
this
consumer.id
was
that
he
said
okay.
He
wants
to
have
like
this
messaging
system,
independent
attribute
where
he
can
identify
consumers
in
a
more
or
less
unique
way.
I
mean
I
think
it
worked
for
his
use
cases,
but
didn't
work
for
generic
use
cases
and
I
mean
the
point
I
was
making
here
that
I
was
like
okay.
D
This
client
ID
does
the
job,
basically
as
well
as
this
consumer
ID,
and
there
is
then
there's
there
is
a
whilst
they're
able
to
remove
this
ambiguous
consumer
ID,
but
yeah
I
also
get
a.
D
Your
point
look
Miller
that
you
say:
okay,
then
this.
This
moves,
this
kindly
concept
to
a
middle
ground,
where
it's
kind
of
in
the
overall
view
on
the
semantic
convention
is
not
really
clear.
A
Yeah
and
like
if
we
can
define
a
some
generic
attribute,
however
value
it
is
the
specific
systems
like
Kafka
can
reference
this
attribute
and
say:
okay,
this
means
exactly
this
thing
in
case
of
Kafka,
so
this
this
would
be
fine
to
have
a
very
vague
Theory
defined
I
mean.
A
Maybe
we
can
think
about
the
the
namespace
for
it,
and
if
we
have
a
good
suggestion,
then
it
will
resolve
the
the
concerns.
Maybe
we
can
do
it
offline
and
see
if
we
can
come
up
with
something.
D
D
Let's
think
about
it.
I
will
capture
that
the
capture
data
as
a
comment
in
this
PR
we
talked
about
and
then
let's
see,
I
will
think
about
the
possible
top
level.
Namespace
I
mean
I
I
like
that
idea.
From
a
conceptual
point
of
view,
it's
from
a
pragmatic
point
of
view
yeah.
It
means
that
we
introduce
and
other
top
level
namespace
that
maybe
for
messaging
we
depend
on
and
that
we
will
also
need
to
declare
stable
if
we
declare
messaging
stable
so
from
this
pragmatic
Viewpoint,
it
makes
our
life
a
bit
harder
but
yeah.
A
D
C
My
thought
on
this
one
is
that,
although
like
different
messaging
systems
might
have
their
own
way
of
sort
of
identifying
clients,
I
think
the
concept
of
a
messaging
system
having
unique
identifier
for
a
client
isn't
really
terribly.
You
know,
like
I,
think
it's
very
common
and
so
I
I
think
putting
it
rather
than
relegating
the
same
thing
off
to
every
system.
If
we
can
just
agree,
this
is
a
common
concept
and
put
it
in
the
messaging.
C
Namespace
is
a
good
thing
if
it
should
be
outside
of
messaging,
then
I
think
it
would
be
nice
to
decide
that
before
we
declare
this
table,
although
to
johannes's
point
I
think
you
know
that
that
may
make
our
lives
more
difficult
to
try
to
go
in
that
direction,
but
it
would
be
nice
to
make
a
decision
before
you
go
stable,
which
way
we
want
to
go
foreign.
B
B
If
we
can
do
that
like
if
me
or
Dwayne
or
let
me
or
somebody
would
want
to
go
and
document
these
things
and
write
a
prompts
and
cons,
thing
least
story.
Then
we
could
think
about
that.
If
we
don't
have
such
time,
let's
say
because
we
want
to
go
a
stable
and
we
have
a
deadline.
Then
we
should
just
go
with
the
conservative
way
and
just
keep
them
separated.
C
I
would
also
support,
like
sticking
with
a
messaging
client
ID.
D
A
B
By
the
way,
can
we
ask
Christian
for
help?
Maybe
basically
somebody
would
need
to
go
and
check
and
tell
them.
I
mean
I
know
that
he
created
the
issue.
Maybe
he
will
be
interested
I
I
mean
I
honestly,
don't
know
how
busy
he
is,
but
probably
he's
he's
motivated
enough.
D
I
will
ask
him:
I
mean
he
he
commented
on
the
issue.
I
mean
he
gave
the
initial
I
mean
he
initially
proposed.
Messaging.Client
ID,
but
I
will
I,
will
summarize
our
discussion
here
and
then
ask
him
and
yeah,
maybe
when
we
maybe
even
then
in
the
in
the
summer,
in
the
summer
of
the
discussion
when
we
propose
this
kind
of
global
client
attribute,
maybe
other
people
will
chime
into
and
say
that
this
is
a
good
idea
or
bad
idea.
So
maybe
we
get
some
vital
feedback.
B
The
egg
would
also
suggest
that
we
don't
make
a
quick
decision.
I
mean
I,
know
that
there's
the
pr
itself,
but
since
we
don't
have
General
agreement,
I
suggest
we
wait.
You
know
enough
time,
so
we
you
know,
we
take
our
time.
Thinking
about
the
options
here.
D
D
B
D
We
discussed
I
didn't
get
to
make
the
changes
there,
but
yeah
I
will
I.
Will
this
work
this
week
work
through
the
commands
and
resolve
the
smaller
one
and
then
yeah
the?
If
there's
any
controversial
issues,
let's
go
over
those
next
week,
I'll
be
prepared.
B
B
D
D
I
I,
don't
have
a
strong
opinion
here.
So
for
me
to
trust
yeah
from
the
from
the
normative
point
of
view
should
be
clear
that,
like
the
creating
link,
is
a
it's
a
strong
like
short
requirement
and
that
you
can
have
parent
child
relationships.
In
addition
to
that,
it's
like
an
option
but
yeah
I'm
I'm
not
opposed
to
adding
any
kind
of
language
that
gives
explanation
for
their
dot.
It
mentions
cases
where
we
are
adding
parent
track
relationships
is
beneficial.
D
D
Okay,
as
I
will
said,
I
will
go
through
the
open
comments
here
and
kind
of
consider
or
like
the
the
smaller
knits
and,
let's
see,
if
there's
anything
controversial,
and
then
we
can
go
over
that
next
week.
D
Awesome
so
thanks,
everybody
yeah
we're
making
steady
progress
great
to
see
with
lots
of
stuff
assigned
and
in
flight.
So
thanks,
everybody.