►
From YouTube: 2023-04-03 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
B
A
So,
let's
get
started,
apologies
I,
don't
know
my
camera
on
today.
I
was
planning
on
going
for
a
run
after
this,
my
hair
is
the
best,
so
I
have
to
make
dude
without
the
camera,
but
let's
get
going
through.
The
list
looks
like
we
have
a
lot
of
topics
today.
So
starting
with
the
trace
special
or
starting
with
the
spec,
looks
like
the
semantic
convention.
A
Changes,
including
some
breaking
ones,
have
been
merged
into
the
spec,
and
we've
released
PR
coming
later
today,
spec
folks,
I
guess
I
usually
turn
to
Carlos
for
these,
but
really
anyone
from
the
spec.
How
breaking
are
these
breaking
changes?.
E
In
the
area
of
HTTP,
but
there
are
also
some
some
other
breaking
changes
where
some
elements
have
been
removed,
and
there
is
something
about
grpc
span
status,
for
example,
so
those
would
need
to
have
instrumentations
updated
afterwards.
E
Okay,
and
also
there
are
some
PR
Surs
around
specifying
what
semantic
conventions
actually
enforce
and
things
are
being
marked
stable,
like
service
namespace
and
Telemetry
SDK,
those
that
are
like
transitively
stable
already,
because
they
are
part
of
the
SDK
spec.
So
that's
what's
what's
going
on
at
the
moment,.
D
The
other
thing
is
that
yeah
you
go
ahead.
Yeah
basically
yeah
like
just
bear
in
mind,
for
if
people
are
approving
the
merchant
PRS,
please
take
a
second
to
ask,
probably
and
consider
whether
we
should
include
some
changes
or
not.
For
example,
there
is
a
change
to
updated
your
APC
codes
and
how
they
should
be
handled
by
client
and
server
and
I
was
asked.
D
Integrand
like
this
is
something
that
should
include
or
not,
if
you
have
any
opinion
on
on
any
of
the
open
PRS
or
you
can
still
review
that
when
I
create
a
PR
leader
today.
Please
review
that
consider
you
know,
especially
if
you're
a
maintainer,
whether
you
think
that's
there's.
The
expectation
is
clear
on
what
these
changes
will
be.
C
A
Right
logs,
we
have
pull
requests
here
on
log
API
naming
Jack
I'm
guessing
you
added
to
Steve
Jetta.
You
want
to
talk
about
this.
G
F
F
You
know,
there's
there's
a
comment:
a
few
weeks
back
in
the
specification
Sig
that
whatever
we
do
with
logs
today,
the
logs
Bridge
API
is
not
intended
to
be
user
facing
today,
but
we
should.
We
should
have
a
strategy
for
you
know
if,
sometime
in
the
future,
we
want
to
add
a
user
facing
logs
API
like
how
would
the
the
bridge
API
that
exists
today
interact
with
that
and
what
I
lay
out
in
this
PR
is
that
the
logs
Bridge
API
is
actually
like
a
natural
starting
point.
F
Exactly,
and
so
what
I
lay
out
here
is
that
it
when
you're
naming
your
artifacts,
your
components
that
get
published
is
a
part
of
this
logs,
Bridge
JPI,
don't
include
the
words
Bridge
or
Pender
in
the
artifact
name
or
class
names,
and
essentially
what
this
would
allow
you
to
do
is
is
evolve.
The
logs
Bridge
API
into
a
user-facing
API
someday
in
the
future.
F
If
that,
if,
if
that
comes
to
pass
without
having
to
have
you
know,
confusion
caused
by
Bridge
or
a
Pender
being
in
the
name,
and
so
what
I
kind
of
advocate
for
is
to
make
it
clear
to
users
through
documentation
and
examples
that
this
for
now
is
not
intended
to
be
user-facing
and
rather
than
relying
on
naming
of
artifacts.
A
F
H
Three
weeks
in
a
row
there
Morgan
it's.
A
C
D
Sorry,
sorry
before
we
move
on
sorry
yeah
yeah
Jack,
are
you
confident
on
whether
we
should
merge
this
PR?
If
we
get
to
know
approvals,
or
should
we
post
it
and
include
that
in
the
mail
release.
F
F
A
E
E
C
A
Hp
we're
getting
closer
to
GA
wow,
that's
very
exciting!
You
want
to
review
of
the
repo.
Do
you
want
this
done
by?
Are
you
looking
for
like
maintainers
review
or
the
more
formal
TC
review.
A
Anyone
on
this
call
wanna
chat
with
Bob
sometime
over
the
next
few
weeks
and
do
a
brief
review
of
the
repo
and
Bob.
Do
you
want
to
talk
to
like?
Maybe,
if
and
for
first
pass
the
things
like
if
they're
specific
areas,
you're
looking
to
get
reviewed
because
I'm
guessing
most
of
the
potential
reviewers
are
not
going
to
be
PHP
experts.
C
Yeah,
you
know
just
following
patterns
making
sure
that
nothing's
really
outstanding
obviously
spent
a
lot
of
time
with
this
repos
now
I
think
we're.
The
PHP
group
is
confident
in
this
repo
to
go
ga,
but
you
know
we
definitely
need
to
get
a
second
pair
of
eyes
from
another
maintainer,
but
that's
it.
D
You
yeah
yeah,
totally
yeah.
Okay
I
will
mention
that
a
bit
to
see
I
would
probably
do
that
already
myself,
because
I
would
talk
to
the
TC
first
in
case
somebody
else
has
Cycles.
H
One
thing
that
we've
done
in
the
past
to
help
the
TC
is
to
just
open
an
issue
in
the
community
repo
and
give
a
high
level
overview
of
the
repo
itself.
Just
you
know
kind
of
what
Morgan
was
saying
so
that
the
reviewer
knows
where
to
look
and
that
kind
of
thing.
D
No,
it's
totally
fine
I
mean
usually
that's
where
the
DC
phones,
the
review
of
the
you
know,
but
you
can
we're
not
as
well
go
both
ways
like
you
open,
an
issue
and
then
they
they.
You
know
the
DC
comments
in
the
issue
later.
The
review.
E
Yeah,
the
community
issue
would
be
the
best
thing
too
I
think
thanks
for
bringing
it
up
Tyler,
you
can
also
Define
the
this
scope
of
what
you.
What
you
would
like
to
have
reviewed
there
in
such
makes
it
easier
to
discuss
it
afterwards.
D
Oh
actually
yeah,
that's
a
good
point,
whether
it's
like
the
API
or
the
SDK
2.
Sometimes
we
want
to
review
separate.
C
A
Okay,
moving
down
JavaScript
SDK
1.11
was
released
with
exponential
histogram
support,
go
or
preparing
a
third
metric
API
release
candidate
this
week.
So,
let's
see
plus
plus
no
new
updates,
Community
demo
we're
cutting
a
1.4
release
within
the
next
week
focused
on
some
core
bug
fixes
very,
very
cool
website.
You've
kind
of
your
maintainers
track
talk
was
moved
to
Thursday,
which
is
great
because
if
it
stayed
on
Friday,
basically
no
one
would
have
been
presenting
or
attending
the
end
user
working
group.
A
Oh
sorry,
that's
not
an
update
for
functions
of
service,
we're
almost
at
the
point
to
start
releasing
Community
Lambda
layers,
rather
than
relying
on
the
ADOT
release
yep,
and
we
got
those
permissions
from
the
CNC
a
few
weeks
ago.
So
that's
fully
unblocked
SDK
configuration
Otep
has
been
merged.
Next
steps
will
be
pouring
this
into
the
core
specification.
A
Anyone
from
the
SDK
work
group
want
to
chat
about
this
because
I'm
guessing
most
maintainers
are
not
super
familiar
with
this.
It
might
be
good
to
give
them
a
refresher.
F
H
Yeah
sure
I
mean
the
the
high
level
overview
is,
if
you
haven't
been
paying
attention
to
the
previous
Otep
that
was
merged
is
that
we
want
to
provide
a
configuration
format
that
will
configure
the
SDK
and
potentially
other
things,
but
really
focusing
on
the
SDK
right
now
across
all
language
implementations
of
Otep
or
of
a
hotel,
and
so
what
this
is
doing
is
it's
going
to
require
these
language
implementations,
to
support
a
particular
format
of
that
configuration
file
and
how
that's
defined
has
been
kind
of
you
know
done
via
a
schema.
H
So
what
to
expect
next
is
the
specification
is
going
to
have
an
update
as
to
the
language
of
you
know
what
sigs
are
required
or
recommended
to
do,
and
then
likely
we're
also
going
to
build
tooling
to
help
in
the
process
of
validation
and
ensure
that
you
know
it's
as
easy,
easy
as
possible
to
start
using
this,
but
yeah
I
mean
I.
H
Think
we're
really
excited
to
to
get
this
forward,
because
providing
Cloud
configuration
is
something
we've
been
looking
at
for
a
while
check
anything
you
want
to
add
to
that.
F
No,
unless
folks
have
questions.
A
A
Okay,
we'll
keep
on
going
discussion
topics.
First,
one
up
is
from
Tyler
spec
States,
some
emerging
Resources
with
different
schema.
It
is
an
error.
Is
this
how
it
Go
implements
the
behavior,
but
it's
causing
user
frustration
goes
looking
to
add
resource
translation
based
on
schema
files
and
we're
wondering
if
other
sigs
have
the
same
problem
and
how
they
address
it.
H
Yeah
I
just
wanted
to
ask,
because
this
is
something
that,
like
we've,
had
a
few
different
issues
come
up
on
and
go.
Oh,
it
may
be
also
how
we
handle
our
cementing
convention
versioning
as
well.
That's
specific
to
go,
but
I
wanted
to
ask
around
just
for
the
community,
because
one
of
the
issues
is
the
specification
says
that
if
you
try
to
merge
to
resources
from
different
schemas,
it
is
considered
an
error
if
they
don't
share
the
same
schema
URL,
which
is
problematic.
H
We
implement
this
behavior
authentically
and
go,
but
a
lot
of
users
will
use,
say,
detectors
from
third
parties
or
old
versions
of
detectors
that
don't
implement
the
same
semantic
convention
versioning
and
when
they
try
to
merge
these
resources
together,
like
it's
as
specified
in
error,
that's
returned
technically,
it
doesn't
I
mean
it
shouldn't
it
like
it's
solvable
right
like
that's.
H
The
whole
point
of
these
schemas
is
to
provides
translation,
both
upgrades
and
tab
grades,
so
this
PR
here
is
just
for
go
and
it
actually
does
that
it
doesn't
upgrade
your
downgrade
based
on
the
how
the
user
wants
to
merge
it.
The
problem
is
that,
like
it
adds
this
functionality
instead
of
replaces
the
existing
merge
functionality
and
I'm.
H
Just
kind
of
wondering,
like
do
other
sigs,
have
this
problem
because,
like
it
seems
like
a
common
problem,
I,
don't
understand
why
it
wouldn't
affect
other
cities,
but
maybe
maybe
they
handled
it
differently,
and
if
that's
the
case,
then
maybe
changing
it
at
the
spec
level.
Is
the
preferable
way
to
address
this
as
well.
I
H
E
B
K
Can
you
open
an
issue
somewhere
because
that
I
think
other
languages,
probably
including
Java,
need
a
moment
to
go
and
see
how
we're
actually
doing
this?
Whether
we're
emitting
schema
URL
for
one
and
whether
we're
merging
throwing
an
error
is
the
second
thing
I
think
we
need
to?
Probably
most
of
us
need
to
go
and
look
at
the
code
to
know
yeah.
H
Okay
I
can
I
can
do
that
I'll.
Take
that
as
an
action
item.
F
C
K
H
Okay,
I'll
I'll
be
sure
to
tag
you
in
the
issue
Trask
later
this
morning,
when
I
create
the
issue.
Okay,
perfect.
A
Okay,
next
is
Dan
contrib
ownership
tooling,
and
you
want
to
take
this
way.
J
Yeah
so
over
the
last
week
or
so,
I've
just
been
looking
at
ways
that
we
can
refine
the
the
contrib
ownership
experience
in
JavaScript
I
had
written
this
GitHub
action
a
long
time
ago,
which
doesn't
do
very
much
other
than
assigning
owners
or
assigning
like
maintainers
to
PRS
and
requesting
reviews
of
component
owners,
but
I
had
sort
of
written
this
and
then
we're
using
a
JavaScript.
It's
largely
unmaintained
I
didn't
realize
until
like
a
week
ago,
that
seems
like
half
of
hotel
is
using
it.
J
J
Is
it
working
for
you
and
can
it
be
improved
and
then
also
I
kind
of
wanted
to
just
ask
if
anyone
is
willing
to
help
me
maintain
this
I
am
planning
on
being
more
active
and
maintaining
it
moving
forward,
but
don't
have
a
ton
of
time
to
do
it
completely
by
myself
and
having
a
second
set
of
eyes
for
PRS
and
stuff
would
be
helpful
anyways.
J
I
J
I
did
I
talked
to
him
on
Friday
for
a
significant
period
of
time.
The
Collector,
contrib
repo,
indeed
does
have
a
bunch
of
tooling
around
similar
things,
they're
all
written
in
bash,
which,
according
to
him,
has
been
a
pain
and
he's
been
interested
in
rewriting
some
of
their
functionality
in
JavaScript,
because
that's
the
the.
G
J
Sdk
for
GitHub,
apparently
is
in
JavaScript.
I
wrote
this
in
JavaScript
just
because
that's
what
I
was
familiar
with,
but
he's
been
interested
in
doing
that
anyway.
So
we've
been
talking
about
consolidating
some
of
the
the
features,
and
that
was
part
of
the
reason
that
I
wanted
to
bring
it
up
in
this
meeting
is
because,
if
we're
going
to
do
significant,
Feature
work
on
this
finding
out.
H
J
Features
are
important
to
the
maintainers
using
it
and
if
there
are
any
shortcomings
that
they're
really
feeling
at
the
moment,
we'd
like
to
address
them,
foreign.
G
For
people
using
component
owners,
the
issue
is
not
really
that
it's
not
networking
as
intended
or
it's
lack
of
features.
But
the
main
thing
is
I,
don't
know
if
other
things
are
having
the
same
issue.
But
contributors
don't
really
seem
to
respect
the
tagging
and
assignment
of
PR
reviewers,
especially
contributors
who
aren't
very
active
or
not
frequently
joining
the
sigs
so
much
so
that
the
component
owners,
at
least
for
python,
is
kind
of
useless.
G
At
this
point,
and
all
of
the
reviews
are
kind
of
falling
on
the
maintainers
and
the
active
approvers
anyways
I'm
wondering
if
anyone
else
is
experiencing
this.
J
G
The
the
letter-
usually
it's
from
contributors
who
just
drop
like
a
certain
instrumentation,
usually
ones
that
the
current
maintainers
and
approvers
are
not
experienced
in
and
then
when
issues
do
come
up
with
those
instrumentations
like
the
contributors
never
respond
or
anything,
which
is
why
I
python
has
kind
of
put
a
hard
limit
on
instrumentations.
Now
we're
not
really
accepting
any
new
ones
anymore.
It's
just
a
too
much
of
a
maintenance
burden
for
us.
J
We
definitely
have
the
same
issue.
We
have
not
put
the
same
limitation
of
no
new
instrumentations,
but
we
we
definitely
do
have
the
same
issue.
I,
don't
have
a
good
solution
for
it.
We
have
been
talking
about
the
possibility
of
more
aggressively
marking
those
components
as
unmaintained
and
possibly
even
marking
them
as
deprecated
in
npm,
which
is
the
primary
distribution
channel
in
JavaScript,
which
would
for
sure
get
people's
attention,
but
I,
don't
know
whether
it
would
actually
solve
the
problem
or
whether
it
would
just
make
people
angry
and
also
unresponsive.
J
But
it's
something
we're
definitely
thinking
about,
because
right
now
end
users
have
no
idea,
but
these
components
are
effectively
unmaintained.
For
the
same
reason,
somebody
opens
an
issue
or
creates
PR.
The
component
owners
never
look
at
it.
The
maintainers
just
frankly,
don't
have
time.
There's
three
JS
maintainers
and
there's
I
the
last
count
like
75
packages
or
something
like
that.
J
H
Daniel
under
that
last
point,
I
think
that
you
know
and
go
we've
seen
similar
situation
as
both
Python
and
JS
here
I.
Would
you
know
one
of
the
things
that
I've
heard
before,
especially
in
like
the
collector
contrib
is,
if
it's
unmaintained,
then
they
aggressively
deprecate
it
and
I.
Think
that
that's
probably
I
mean
that's
what
we've
decided
to
do.
H
H
Another
thing
that
we've
also
looked
at
is
ownership
models
and
similar
to
the
electrical
contrib.
So
this
is
based
on
them.
They
would
have
I
think
better
understanding,
there's
ability
to
add
code
owners
to
the
repository
in
like
subsections
of
any
sort
of
instrumentation
and
anything
without
a
code
owner.
Our
plan
is
to
do
just
this
is
to
deprecate
it
and
if
code
owners
aren't
responsive
to
then
go
through
the
deprecation
cycle
as
well
and
I.
Think
that's
just
like
it
adding
the
code
owners
I
think
helps
just
to
get.
H
You
know
what
your
tooling
does
here
and
that's
just
some
sort
of
notification
to
the
code
owners
that
they
have
a
PR
or
some
sort
of
issue.
They
need
to
track
and
I.
Think
that's
I,
don't
know
it's
useful,
but
it
if
you
don't
have
a
commitment
from
somebody
to
actually
maintain
it.
Then
I
think
that
it
needs
to
live
somewhere
else
or
you
know
it
needs
to
get
deprecated.
J
Yeah
code
owners
has
severe
limitations,
the
most
obvious
of
which
is
that
you
have
to
have
right
access
to
the
repo
in
order
to
get
notifications
from
a
code
owners
file,
which
makes
it
essentially
useless
because
I'm
not
going
to
give
right
access
to
the
whole
contribury
bow
to
people
who
are
just
contributing
a
single
instrumentation.
That
was
the
original
motivation
to
even
write
this
action
yeah.
J
It
does
a
handful
of
other
things,
but
mostly
it's
just
code
owners
without
right
access.
But
if
the
deprecation
and
unmaintained
status
is
important
thing,
that's
coming
up
in
multiple
sigs,
then
maybe
this
tooling
can
also
help
to.
H
J
J
J
Yes,
so
I'm
I'm,
not
sure
I
I
mean
I
I,
know
for
sure
that
you
could
run
an
action
out
of
the
build
tools,
repo,
any
repo
can
be
an
action,
but
I
think
you
can
only
run
one
action
out
of
it
and
it
has
to
be
the
name
of
the
repo,
so
I
mean
if
we're
fine,
with
the
action
being
called
open,
Telemetry
build
tools,
then.
Yes,
that
would
work.
But
there
is
some
weird
limitations.
Like
the
action
itself,
you
have
to
actually
check
in
the
compiled
JavaScript.
J
You
can't
check
in
the
typescript
source
and
some
things
like
that
that
I
I
think
people
who
are
working
on
build
tools
not
related
to
this
would
be
annoyed
at
how
invasive
it
has
to
be
in
the
repo
in
order
to
make
it
an
action.
K
That
makes
sense
yeah
I
I
would,
if
you
want
to
add
me
over
in
the
open,
Telemetry
copy
I,
think
that
would
make
sense
to
me
as
far
as
maintaining
it
over
there.
J
A
All
right,
let's
move
on
so
I
added
this
next
topic,
apparently
at
the
governance
committee
meeting
last
week,
which
I
didn't
attend
because
we
had
an
internal
customer
fire
drill
happening,
so
I
could
join
but
I
believe
there
was
a
discussion
around
public
video
recordings.
So
today,
all
of
the
open,
Telemetry
Sig
calls,
including
this
one,
are
recorded
and
then
automatically
posted
to
YouTube.
A
There's
some
concern
raised
by
I
think
by
T
Gran,
as
well
as
a
few
other
community
members
that
with
AI
tools
online
advancing
as
quickly
as
they
can.
You
could
easily
use
all
these.
Our
own
talking
faces
and
voices
to
generate
fake
versions
of
ourselves,
and
there
was
concern
about
this,
leading
to
just
a
lot
of
source
material
for
impersonation
and
so
I
think
there
was
a
discussion
about
this
at
the
GC
meeting,
so
I
was
wondering
Trask
or
Daniel
or
others
on
the
GC
who
were
there
last
week.
J
Yeah
there
was
significant
discussion
about
this
I
think
all
the
Talking
Points
you
would
expect
came
up.
Obviously.
Yes,
it
is
a
large
volume
of
source
material
for
these
types
of
tools.
These
tools
are
becoming
more
common
and
more
powerful,
they're
they're,
really
honestly
becoming
quite
good
with
a
shockingly
small
amount
of
training
data.
J
The
question
that
we
had
for
maintainers
specifically,
is:
how
useful
do
you
find
in
the
recordings
we
we
have
the
recordings
now,
so
that
people
who
Miss
Sig
meetings
can
watch
them
later
and
you
know
sort
of
a
transparency
effort,
which
is
why
we
publish
the
governance
meetings
and
some
things
like
that.
But
one
of
the
questions
we
had
is:
does
anyone
actually
watch
them?
There
are
any
of
the
Sig
maintainers
aware
of
any
regular
contributors
in
their
Sig.
J
That
depends
on
these
recordings,
particularly
we
were
thinking
about
sigs
that
have
regular
contributors
or
maintainers
in
Asian
time
zones,
which
don't
typically
play
well
with
our
meeting
schedules.
H
So
I
could
say
for
go.
We
do
I,
know
I
I
can't
say
actually
as
of
today
but
I
know
in
the
past.
We
definitely
have
had
people
that
are
actively
and
like
when
the
upload
broke
to
YouTube.
They
were
on
slack
paying
us
because
they
wanted
to
get
a
a
view
of
the
previous
Sig
meeting.
A
I
I
can
definitely
also
attest
to
like
I
personally.
Do
not
use
the
recordings
just
because
I
tend
to
be
able
to
join
the
meetings,
but
I
happen
to
know
because
I'm
often
the
person
who's
asked
to
fixing
them
when
the
recording
pipeline
breaks,
usually
within
a
day
of
it,
going
down
I'm
getting
pinged
by
usually
about
two
people
about
hey
the
recordings
are
broken.
Can
you
please
fix
it.
J
Okay,
if
we
have
a
reasonable
idea
of
who
these
people
are,
can
we
find
out
whether
the
zoom
transcript
feature
would
be
an
acceptable
substitute
and
then
also
we
just
wanted
to
ask
generally
how
worried
are
people
about
these
types
of
tools?
J
Is
this
something
that
that
we
should
just
accept
as
a
risk
and
say?
Yes,
we
know
that
being
putting
your
your
face
and
and
voice
publicly
on
the
internet
regularly
introduces
risks
that
you
wouldn't
otherwise
have
and
we're
okay
with
that,
or
is
it
something
that
people
are
worried
enough
about
that
we
should
consider
discontinuing
the
YouTube
videos
in
favor
of
something
else
like
transcripts.
I
J
Another
option
that
we
had
discussed
was
to
upload
the
meetings
and
leave
them
unlisted
by
default,
and
then
people
would
specifically
request
individual
meetings.
It
sounds
like
from
the
response
from
Tyler
and
Morgan,
though,
that
that
would
not
work
for
at
least
some
of
the
segs,
because
every
meeting
would
have
at
least
one
request.
I.
A
Yeah
there
might
be
other
ways
to
do
that,
like
I
suppose
in
theory
we
could
have
them
unlisted
and
keep
a
I.
Don't
know.
Have
this
the
scripted
or
whatever
tool
we
use
that
publishes
them.
A
Perhaps
it
has
a
way
of
just
like
plopping
the
links
down
in
a
document
somewhere
or
in
some
place
where
it's
not
crawlable,
I
I,
don't
know
if
I
I
don't
really
publish
videos
to
YouTube,
so
I,
don't
know
if
what
I'm
about
to
describe
as
possible
like
I,
don't
know
if
you
can
just
publish
YouTube
videos
that
are
only
accessible
to
certain
people
or
unless
it
is
the
main
way
to
do
that,
like
there
may
be
sort
of
a
middle
ground
here
we
can
still
keep
the
recordings,
but
you
know
like
unless
you'd
have
to
be
in
the
community
to
access
them.
E
I
think
you
can
upload
them
as
unlisted
and
then
you
would
need
to
share
individual
links,
but
you
can
also
edit
to
a
playlist
and
I
believe
if
you
have
an
unlisted
playlist
and,
as
you
said
in
an
in
an
unfoldable
location,
you
post
a
link
to
that
playlist
and
everyone
should
have
access
to
it
still.
I
think.
A
Sure
I
mean
the
challenge
of
that,
though,
is
they
could
just
join
the
calls
and
record
them
locally
right
like
like
I,
don't
know
I
just
I'm,
not
really
an
expert
on
this,
but
my
my
gut
is
like
if
someone
really
says
I
want
to
impersonate
Morgan,
McLean
and
I'm
going
to
create
a
fake
version
of
him.
I
mean
they
could
just
join
us.
Maintainer's
call
right
now
and
be
recording
locally
and
I
mean
short
of
not
having
a
maintainer's
call,
there's
no
way
to
avoid
that
yep.
J
Alex
just
asked
in
the
chat,
expiring
videos.
That's
definitely
something
that
we
discussed
in
the
GC
meeting
is
having
some
sort
of
retention
policy.
Multiple
people
in
the
past
have
brought
up
the
fact
that
these
videos
live
forever
and
they
would
just
be
more
comfortable
if
they
had
some
expiration
period,
mostly
because
you
know,
if
you
accidentally
slip
up
and
say
something
about
I,
don't
know
your
internal
company
or
a
customer,
or
something
like
that.
J
Then
it's
on
the
internet
forever,
unless
you
specifically
remove
it,
which
is
kind
of
a
bummer
Anthony
responded
in
the
chat
by
saying
he's,
found
it
useful
in
the
past
to
refer
to
old
videos.
Another
argument
for
keeping
the
videos
but
also
I,
guess
an
argument
against
a
retention
policy
Anthony.
J
I
Yeah
I
think
that
would
be
a
reasonable
compromise.
The
the
situations
I've
had
to
do
that
in
are
not
super
frequent,
but
you
know
like
I,
want
to
go
back
and
and
refer
to
a
prior
discussion
that
I
know
was
had
and
communicate.
What
happened
there
to
someone
while
making
sure
that
I
get
clear
what
what
actually
was
said.
A
Daniel,
can
you
just
repeat
the
your
suggestion.
J
Yeah
the
suggestion
would
be
after
like
30
days
or
so
the
video
would
be
made
either
unlisted
or
private,
whatever
we
decided
and
then,
if
you
needed
to
refer
to
a
video
older
than
30
days,
you
would
have
to
reach
out
to
somebody
with
access
to
the
channel
in
order
to
specifically
request
a
meeting.
J
I
guess
I
also
wanted
to
get
the
temperature
of
the
room,
since
we
have
all
the
maintainers
here.
All
of
all
of
us
in
this
meeting
are
on
more
of
these
meetings
than
anyone
else
and
most
likely
to
be
talking.
All
worried
generally.
Are
people
about
these
types
of
tools?
J
H
D
H
Maybe
clear
I
think
that
it's
not
necessarily
the
concern
today,
but
what
concerns
me
is
tomorrow
and
the
next
day
after
that,
because,
as
you
pointed
out
like
these,
are
historical
records
and
like
they
may
already
have
been
downloaded
I
guess
like
I
I,
guess
that
whole
issue
of
like,
if
you
capture
encrypted
communication
and
then
you
know
in
the
future,
are
able
to
decrypt
it.
You
have
a
problem,
it's
kind
of
the
same
thing
here,
it's
like.
J
Yeah
so
Aaron
and
Pablo
both
said
in
the
chat
I.
Guess
we
something
we
that
we
both
already
brought
up
at
the
GC
meeting,
but
the
two
main
camps
of
thought
that
we
had
in
the
GC
meeting
Aaron
said:
there's
already
enough
video
of
me
online
that
it's
a
drop
in
the
bucket
I
think
that's
the
case
for
a
lot
of
us,
particularly
the
particularly
the
ones
who
speak
at
conferences
and
Pablo,
said
he's
worried
enough
that
he
doesn't
show
his
face
on
Zoom
meetings.
J
I
mean
these
are
that's
generally.
The
two
cats
have
thought
that
we've
seen.
A
I,
don't
I
say
it's
like
myself,
I'm
in
the
first
Camp
self,
like
just
selfishly,
for
myself,
like
I,
like
I,
do
press
interviews
and
things
like
oh
I'll,
like
my
face,
is
already
out
there,
but
I
suspect
if
I
had
to
guess
that
most
of
the
people
in
our
community
are
probably
more
in
the
camp
of
what
Pablo
mentioned
of
like.
Well,
no
I
don't
have
a
big
Persona
like
I'm,
not
on
a
lot
of
online
videos.
If
someone
copies
this
I'm
very
uncomfortable
with
them,.
E
J
And
then
I
guess
one
final
point,
and
this
is
not
yet
decided
in
any
way.
But
it's
probably
worth
mentioning
here,
since
you
are
all
the
maintainers
one
thing
that
we
talked
about
is
the
process
for
specifically
requesting
videos
be
taken
down
right
now,
it's
sort
of
an
informal
process,
but
if
somebody
says
I'm
in
this,
video
I
have
the
right
to
be
forgotten.
I'd
like
you
to
take
it
down.
That's
where
we
get
into
really
sort
of
tricky
water
and.
G
E
J
J
J
Yes,
you
know
something
along
those
lines.
Obviously,
I
don't
know
what
the
legal
implications
are
and
whether
we
even
could
have
a
policy
like
that,
but
it
was
one
of
the
things
that
was
discussed
as
an
option.
J
Council,
we
definitely
will
be
talking
with
cncf
Council
about
the
right
to
be
forgotten
in
general,
since
it
was
already
brought
up-
and
none
of
us
is
a
legal
expert.
So
well.
A
Certainly
Daniel,
like
you
mentioned
it,
would
cause
us
to
have
to
remove
most
recordings
if
we
end
up
following
this
strategy
here
of
making
the
recording
expire
on
some
time
schedule.
But
at
that
point
it's
like
the
decimation
is
far
less
relevant
than.
I
A
But
I
think
then
the
the
the
challenge
that
Daniel
was
discussing
is
no
longer
relevant.
F
Is
okay,
so
we
we
record
these
these
meetings
and
we
upload
them
to
YouTube
and
presumably
in
the
future,
somebody
could
crawl
them.
Could
we
just
and
this
this
I
don't
know
this
sounds
incorrect,
saying
it
out
loud,
but
I
just
want
to
say
it
anyways.
Can
we
in
some
way
say
that
it's
not
fair
use
of
this
content
to
use
it
for
training
these
models
or
to
you
know
anyway,
incorporate
into
you
know
a
corpus
of
data
for
training
purposes,
regardless.
J
Of
whether
we
did
that
or
not
that
was
not
the
specific
concern
brought
up.
The
concern
was
something
more
along
the
lines
of,
and.
J
Example
today
there
are
banks,
for
instance,
that
use
your
voice
as
like.
A
second
factor
of
authentication.
J
And
there
are
other
there
was
a
tool
that
was
announced
on
Hacker
News
recently
for
this
and
then
in
the
comment
thread
a
bunch
of
people.
It
was
a.
It
was
a
a
demo
website
and
all
of
the
demos
were
public
and
then
in
the
comment
thread,
people
were
already
finding
scams
where
somebody
was
faking
voices
to
say
you
know.
B
J
Is
your
son,
can
you
please
send
money
to
you
know
I'm
in
jail
or
whatever
also
another
attack
Vector?
That
is
not
explicitly
related
to
our
voices
being
included
in
the
training
Corpus,
but
more
of
a
a
targeted
attack
type
of
deal
I.
A
Or
they're
going
to
be
again,
I
wasn't
on
the
the
GC
call,
but
but
are
there
going
to
be
like
follow-ups
or
further
conversations,
understanding
like
what
was
the
actual
outcome.
J
Well,
the
the
main
outcome
was
that
we
wanted
to
gather
as
much
information
from
the
maintainers
as
possible,
since
they
are
the
most
public
face
of
the
project
and
in
most
of
these
YouTube
videos,
any
decision
we
make
related
to
the
YouTube
channel.
The
YouTube
channel
affects
the
maintainers
more
than
anyone
else,
so
the
initial
outcome
was
to
gather
this
feedback
and
to
reach
out
to
cncf
Legal
team
to
find
out
what
our
obligations
are
with
with
regard
to.
J
If
somebody
requests
their
video
be
taken
down,
how
is
that
generally
handled
in
the
cncf?
We're
probably
not
unique
there
I
know
at
least
kubernetes
is
also
publishing
their
meetings.
I
think
right
so
finding
out
what
the
what
the
policies
are
from
the
cncf
perspective.
J
Yes,
we
will
definitely
have
follow-up
meetings,
I
suspect
more
than
one
more
than
two.
This
seems
to
be
a
topic
that
a
lot
of
people
have
a
lot
of
differing
opinions
on
and
there's
no
obvious
solution.
A
Yeah
I
definitely
I
think
from
the
comments
there.
There
were
several
people
calling
this
one
here
out
as
making
the
recordings
expire,
but
I
I
mean
that's
just
me
from
looking
at
the
meeting
chat.
A
Maybe
we
can
get
like
a
straw
poll
amongst
people
but
which
these
prefer.
A
Actually,
maybe
everyone
look
on
the
call
like
just
opened
this
Doc
and
just
put
if
you
have
a
presence
for
these,
just
like
put
an
X
next
to
each
one,
if
you
prefer
it-
and
this
is
not
binding
but
like
this
is
just
so
like
we
can
just
get
on
the
GC.
A
vague
sense
of
like
which
of
these
people
think
are
acceptable,
feel
free
to
put
x's
on
multiple
ones.
If
you
like,
multiple,
oh
Alex,
so
can
we
make
this
community
issue?
That
sounds
like
a
thousand
times
better
than
what
I
just
suggest.
K
J
I
don't
think
there
was
anybody
arguing
against
some
sort
of
retention
policy.
We
had
not
really
thought
of
I
mean
we
had
thought
of
the
idea
of
historic
videos,
but
we
had
a
little
bit
dismissed
it,
because
these
these
the
nature
of
these
videos,
means
that
most
of
the
time
they're
not
really
relevant
after
some
period
of
time.
J
A
A
Just
FYI
I
was
going
to
skip
the
Otep
triage
this
week.
Unless
there
are
any
objections,
I
think
we
usually
have
Ted
for
that
he's.
Not
here
and
I
just
took
a
look
while
we
were
discussing
this
at
the
stack
Overflow
issues,
there's
plenty
of
new
questions,
but
they're
all
language
collector
or
vendor
specific.
So
we
do
not
need
to
review
those
today
because
I
just
took
care
of
it.
A
Okay,
then
we
can
wrap
this
up.
This
was
a
long
meeting.
Lots
of
content,
we'll
wrap
this
up
we'll
file,
a
community
issue
on
this
people
can
weigh
in
I.
Don't
think
we're
looking
to
make
a
snap
decision
here,
but
certainly
something
in
the
next
few
weeks
is
probably
ideal
and
then
we'll
go
from
there
alrighty.
Thank
you
very
much.
Everyone
we'll
see
you
throughout
our
call
string
a
week.