►
From YouTube: OCI Weekly Discussion - 2021-03-31
Description
Recording of the OCI developer's weekly call from 31 Mar 2021. Agenda/notes here: https://hackmd.io/El8Dd2xrTlCaCG59ns5cwg#March-31-2021
B
B
I
actually
did
a
few
other
things
today,
but
yeah
I'd,
love
to
just
kind
of
say,
a
few
words
and
kind
of
frame
I
think,
try
and
frame.
I'm
sure
I
won't
collect
everyone's
perception
of
where
we
are
with
the
discussions,
but
I
will
attempt
to
because
I
have
to
somehow
get
my
paycheck
for
being
the
oci
tob
chair.
So
chris
won't
pay
me
if
I
don't
do
this
once
in
a
while.
No.
C
C
C
B
Have
a
raiding
window
pop
up.
H
B
So
yeah,
I
think,
we've
topped
our
highest
participation
in
quite
a
while.
So
I'd
rather
than
wait
longer,
I
know.
Usually
we
have
people
trickling
after
the
hour
but
but
yeah.
I
assume
you're
all
here,
because
you're
excited
about
oci,
which
is
great.
B
Now
really
I
mean
there's
been
lots
of
activity
lately
and
you
know
I
think,
for
any
regular
participant.
You
know
that
you
know
the
agenda
has
been
more
and
more
packed
lately
and
you
know
we
have
lots
of
interesting
ideas.
I
think
we're
you
know
not
to
overload
the
term
inflection
point,
but
you
know
there.
B
There's
we've
had
images
containers
registries
for
for
a
good
while
now
and
we
have
people
using
them
for
new,
interesting
things,
and
I
think
we're
kind
of
at
that
point
where
we're
trying
to
figure
out
how
that
fits
into
specs
and
the
components
that
effectively
were
developed.
You
know
and
published
now
you
know
a
couple
years
back
so
out
of
that
it
seems
like
there
are
kind
of
a
couple
key
things
that
I
think
are
being
struggled
with
on
a
general
level.
B
One
is
that
similar
to
you
know,
someone
can
correct
me
if
I'm
wrong,
but
similar
to
to
run
time
spec
there.
There
was
kind
of
this
relief
of
you
know
getting
that
across
the
finish
line
and
then
kind
of
a
loss
of
of
kind
of
strong.
You
know
regular
activity
and
I
think
image
spec
we're
in
a
similar
spot,
where
you
know
you
look
at
that
maintainers
list
and
I
posted
that
in
steve's
tob
issue
earlier
today.
B
You
know
that
list
of
people
do
not
show
up
to
to
this
call
and
again
I'm
not.
This
is
not
about
shaming
anyone
or
naming
you
know,
participants,
and
some
of
that
is
time
zone
pain.
You
know
we
all
we
all
know.
These
calls
are
a
struggle
to
figure
out
how
to
where
to
place
them
on
the
24
hours
we
each
have,
but
we're
at
a
point
where
a
lot
of
these
ideas
really
need
the
input.
B
Ascent
agreement
challenge
whatever
of
people
who
have
the
commit
authority
to
make
changes
to
these
specifications-
and
you
know
maybe
we
find
out
that-
and
I
think
steve
has
brought
this
up
a
few
times
and
he
brought
it
up
with
me
today.
There's
also
the
complexity
that
these
things
don't
live
by
themselves.
So
they're,
not
the
image.
B
Spec
does
not
stand
alone,
it
it
interacts
with
distribution,
spec
and
how
registries
operate,
and
so
we're
dealing
with
you
know,
essentially
the
complexity
of
not
just
getting
something
quote:
unquote
into
image:
spec
we're
dealing
with
the
complexity
of
how
that
interoperates
with
distribution,
spec
changes
or
registry
operations,
and
so
you
know,
I
think,
what
I'd
like
to
do
or
early
suggest
and
again
I
I
hold
no
authority
here
other
than
you
know.
This
weekly
call
has
been
going
on
for
a
very
long
time.
It's
morphed
into
many
different
things.
B
Over
the
years,
different
people
have
led
it
kind
of
orchestrated
it,
and
so
you
know
I
have.
I
have
this
chair
role
and
I
feel,
like
it's
important
to
kind
of
you
know
clear,
clear,
the
air
so
to
speak,
that
here's
where
we
are,
I
don't
have
solutions.
I
don't
have
perfect
ideas,
but
but
I
have
at
least
a
voice
to
say:
I
think
we
need
a
different
model
right
now,
because
what
we're
doing
is
not
leading
to
actual
progress
for
the
things
that
people
care
about.
B
Other
people
have
other
ideas.
You
know
justin's
idea,
I
think,
is
very
interesting
and
valuable,
but
is
kind
of
an
another
take
on
these
topics,
and
so
we
could
spend
weeks
kind
of
having
people
present
and
go
around
on
these
things,
but
I
feel
like
unless
we
have
the
set
of
people
who
will
actually
collaborate
on
these
outside
of
these
60
minutes
that
we
spend
here.
B
B
Yeah,
thanks
vincent,
so
you
know
this
meeting
is
a
touch
point
for
all
of
us
to
kind
of,
say:
okay,
what's
stuck
or
you
know
how
come
nobody's
commenting
on
this
thing
or
you
know
I
came
up
with
this
new
idea.
What
are
your
thoughts,
but
instead
it
seems
like
we're
trying
to
jam
everything
through
this
meeting
like
this
is
the
place
for
approval
or
disapproval
of
ideas,
and
it's
really
not
because
at
least
at
this
moment
most
of
us
here
are
not
maintainers
on
the
projects
being
discussed.
B
So
it's
a
great
you
know
it's
a
great
place
for
discussion.
It's
a
great
place,
for
you
know
bringing
new
ideas.
It's
a
great
place
for
mentioning
that
that
you
need
help.
You
know
figuring
something
out,
but
to
me
I
think
we
need
to
reorganize
around
that
principle,
and
you
know
thanks
to
chris,
for
for
opening
up
that
issue
to
go.
Do
a
survey
on
maintainer
activity
or
or
actives.
G
G
Another
thing
so
I
think
doing
an
act
of
just
reaching
out
to
the
folks
who
have
been
active,
saying:
hey:
do
you
want
to
step
down
and
then
doing
a
public
call
for
folks,
based
on
kind
of
tob
guidance
of
who
could
kind
of
step
in
and
fill
phil,
because
there's
been
some
people
that
have
been
active
that,
I
think,
should
be
maintainers
and
we've
done
an
okay
job
for
some
other
things
like
distribution,
spec,
but
image
spec
for
sure
has
been
as
big.
B
Yeah-
and
you
know
again,
none
of
this
is
to
say
that
that
you
know
good
work
hasn't
continued
to
go
on
distribution.
Spec
is
right
there
on
the
verge
of
release,
which
is
great,
so
I
I've
spent
more
than
enough
minutes
of
the
call
on
this
vincent.
J
Yep,
I
think
that's
where
traditionally
topics
were
raised
if
they
couldn't
get
results
on
the
gap
issue
and
unfortunately
I
mean
like
even
the
list
has
gone
gotten
pretty
quiet
the
mailing
list,
even
the.
J
J
I
think
we
tried
that
a
couple
of
times
we
just
added
like
yet
another
dimensional
discussion,
but
the
calls
were
just
basically
a
coordinated
effort
of
who's
who's
involved
that
we
couldn't
arrive
at
a
consensus
in
an
ac
pattern
that
was
effectively
used
and
there
were
plenty
of
times
that
we
were
like
would
start
the
call
and
not
have
enough
topics
and
adjourn
for
the
week.
Even
during
the
hot.
You
know
like
getting
close
to
the
release
for
runtime.
J
But
like,
but
like
what
chris
said
awesome,
but
I
like
that
approach
to
paul
from
containers,
because
we
do
I
mean
like
geez
louise.
We
knew
that
after
the
run
times
back
an
image
that
got
to
v1
that
a
lot
of
people
are
going
to
like
take
a
cognitive
break.
You
know
that
there
would
be
some
some
a
period
of
cooldown.
Some
faults
would
probably
like
step
away
permanently.
J
That
was
partially.
Why?
J
Even
though
it
is
a
little
confusing,
I'm
glad
that
we're
working
on
the
charter
also,
but
of
like
that
there
was
kind
of
a
technical
community
that
could
basically
come
in
and
breathe
life
back
into
it
once
we've
had
a
period
for
it
to
be
out
there
and
adopted
folks
see
how
it
looks
in
production,
then
come
back
and
iterate
on
it
in
a
future
time,
and
if
those,
if
the
maintainers
had
cooled
down,
if
the
maintainership
had
cooled
down
that
we
could
have
enough
mechanics
in
place
by
the
tob
or
whoever
else,
it
is
to
breathe
life
back
in
and
you
know
get
it
fired
back
up
again.
A
I
think
obviously
the
maintainership
has
been
a
challenge.
I
think
one
of
the
things
that
we
need
to
figure
out
is
what
process
do
we
want
to
follow
for
changes,
but
I
think
a
bunch
of
these
aren't
as
simple
as
just
image:
spec
and
distribution,
like
we've
kind
of
acknowledged
that
registries
have
been
storing
things,
storing
additional
things
for
a
while.
A
How
do
we
decouple
these
two
things
that
were
initially
you
know
built
together,
like
it
made
sense
at
the
time
when
registry's
only
sorted
images,
but
as
we
get
into
signing
things
and
I'm
not
even
trying
to
say
one
signing
format
should
win
versus
another?
The
approach
has
been.
You
can
store
other
things
in
a
registry
and
we
want
to
be
able
to
now
link
these
things.
A
K
I'd
say
I'll
say
that
I
mean
you
know
oco,
you
know
we
at
darker
originally
created
it
largely
to
standardize
things
that
were
already
there,
and
you
know.
I
think
that
that
that
role
has
kind
of
you
know
as
as,
as
it
kind
of
said,
it's
kind
of
I
mean,
as
you
know,
it's
kind
of
come
to
an
end
ish,
because
we've
standardized
the
things
that
were
there
and
that's
been
great
and
very
successful.
But
you
know
one
of
the
reasons
I
wrote
the
dog
I
did.
K
The
other
day
was
that
you
know
there
are
some
fundamental
design
problems
that
the
formats
that
we
standardize
are
simply
not
upgradable,
and
we,
you
know,
we've
been
struggling
with
making
changes
because
of
those
fundamental
design,
things
that
were
made
a
very
long
time
ago,
at
docker
for
entirely
other
reasons
than
providing
a
long-term
upgradeable,
future-proof
path.
That
people
can
build
new
and
better
things,
and
I
think
a
lot
of
you
know
I
think,
there's
a
there's
a
there's,
a
huge
amount
of
really
exciting
work
around
new
image
specs.
K
K
K
You
know,
bubbling
and
like
with
ideas,
but
recently,
like
people
have
really
compelling
reasons
why
they
want
to
implement
things
like
new
image
formats.
New
compression
formats
like
there
are
things
that
have
as
well
as
things
like
signing,
but
like
even
things
with
images
and
things
where
we
have
now,
where
we're
just
you
know,
I
think,
there's
a
there's
some
things
I
would
love
to
see
in
production
that
we
just
having
great
difficulty
getting
getting
over
the
line.
B
So
yeah,
I
was
thinking
something
about
that
today
I
mean
so
I
think,
there's
kind
of
this
tactical.
Can
we
solve
the
image
spec
maintainers
issue,
which
I
think
we
can't
we
know
we
can
do
that.
It's
you
know
some
work
to
figure
that
out
it's
almost
like
there.
There
are
going
to
be
a
set
of
things
that
can
fit
into
the
steady
state,
incremental
improvements
of
image
spec
for
some
from
general
things,
but
I
always
saw
alexa's
presentations.
B
You
know
here
in
this
meeting
and
others
who
got
involved
as
being
like
kind
of
thinking
about
the
next
generation,
which
I
think
justin
is
what
you're
you
know.
The
proposal
you
put
together
is
essentially
you
know
if
we
knew
everything
we
knew
now
here,
here's
how
we
might
have
laid
out.
You
know
these
concepts
and
it's
almost
like
that's
a
separate
like
that,
isn't
something
that
that
you
need
two
lg
gtms
and
emerge
an
image
spec
like
that
is
a
new
kind
of
idea.
B
Maybe
maybe
it
fits
in
this
call,
but
maybe
it
doesn't
maybe
it's
something
that
gets
iterated
on
you
know
and
I
think
steve,
maybe
that's
where
you're
hitting
into
this.
Like
you
know,
what's
the
what's
the
decision
points
for
this,
how
do
we
decide
what
happens?
What
doesn't
happen
like
that?
That
seems
separate?
J
We're
because
we're
trying
to
figure
out
like
some
of
the
silly
challenges
of
just
like
how
to
handle
simver
for
some
of
these
specifications,
which
is
not
not
a
clean
match,
but
you
know
like
even
when
it
comes
down
to
the
api
and
justin
you're,
saying
of
like
some
of
the
pieces
that
were
in
that
design
process,
where
we
even
arrived
at
the
v2,
schema,
2
and
or
what's
now,
oci
distribution.
J
If
somebody
were
to
want
to
rewrite
that
api
to
experiment,
it's
wholly
possible
for
them.
To
do
that,
I
think
there's
there's
we
find
this
interesting
spot
where
there
has
been
a
whole
a
hold
on
to
not
breaking
everybody
in
the
community
or
existing
clients
versus
enabling
ourselves
the
future
and
there's
actually
a
lot
that
we
can
do
in
that
space
with
what
we
already
have
have
worked.
I
J
Oh
sorry,
but
like
finding
effectively
the
middle
path,
all
these
different
things,
I'm
just
having
so
much
discussion
that
it's
not
actually
solving
anything
yet
or
having
something
that
just
goes
off
and
effectively
wants
to
throw
it
all
away
and
start
over
which
people
are
welcome
to
do
both
of
those
it's
just
not
at
that
point,
actually
being
a
group
consensus
division.
I
I
like
the
idea
of
forming
working
groups.
I
think,
as
phil
was
saying
because
yeah
previously,
if
you
look
at
all
the
specs
and
how
they
came
along,
they
effectively
had
working
groups
at
some
point
that
generated
and
pushed
these
specs
out
to
users
in
production
and
all
the
specs
have
been
accepted
today
have
had
these
production
use
cases
before
even
coming
to
oci
and
becoming
standardized.
I
So
today,
we're
in
kind
of
a
weird
state,
especially
with
the
image
stack,
and
we
saw
this
happen
a
little
bit
with
b2,
where
kind
of
an
informal
working
group
formed
went
out
to
find
something
that
that
could
be
actually
used
and
and
proven
out.
I
But
I
think
if
we
had
some
sort
of
formalization
of
that
process,
then
we
could
actually
see
some
new
and
innovative
stuff
come
from
oci.
So
if
we
had
a
way
to
form
working
groups
that
were
actually
like,
voted
on
by
the
top
scoped
and
had
some
defined
owners-
and
maybe
we
could
move
forward
on
some
of
these
ideas
that
aren't
necessarily
specs
but
could
come
out
with
drafts
that
could
go
and
be
implemented.
I
So
I
see
that's
kind
of
the
problem
today
is
like
nobody
wants
to
take,
take
on
the
risk
of
defining
a
new
spec
that
may
or
may
not
be
accepted
by
oci
in
the
future
or
become
the
standard,
but
at
the
same
time
oci
isn't
really
the
place
to
take
on
that
risk,
either
of
something
that's
completely
unproven.
J
J
Some
of
these
proof
of
concepts
or
otherwise
introduced
are
not
breaking
changes,
because
I
know
that
was
even
some
of
the
some
of
the
recent
discussion
points
were
around
this
also,
and
it
kind
of
allows
at
least
a
conversation
place
to
take
a
conversation
to
take
place
and
it
it
least
mimic.
Maybe
the
working
groups,
the
sigs
kind
of
like
kubernetes
cncf
style.
It's
because
they
have
the
comfort
of
the
you
know.
J
B
B
You
know
is
it:
is
it
time
to
come
up
with
that
list
of
like
what
are
these
kind
of
concrete
areas
that
people
are
are
struggling
with
to
get
changes
in
and
can
we,
you
know,
form
something
around
each
one
of
those
that
that
you
know
by
next
week
or
two
weeks
from
now?
Have
you
know
some
more
specific?
You
know
a
paragraph
what
it
is
we're
trying
to
achieve
and
who's
owning
who's,
because
I
feel
like
then
this
meeting
could
become
more
of
hopefully
something.
B
You
know
practical
that
you
know.
If
you
look
at
the
agenda
and
say
the
thing
I
care
about,
isn't
even
on
this
week's
agenda,
I
that
you
know
I
may
listen
in,
but
maybe
not
that
way
you
know,
maybe
there's
even
a
flow
to
kind
of
the
month
of
oci
weekly,
calls
that
you
know.
There's
distribution
focus,
there's
v2
new
idea,
focus
et
cetera,
et
cetera,
you
know,
but
it
sounds
like
maybe
that
that
helps
and
then
this
meeting
is
a
bring
back.
You
know,
what
did
we
do?
Where
are
we
stuck
is?
J
They
had
effectively
a
working
dock
where
they
tracked
all
those
different
working
groups
and,
had
you
know
the
names
of
people
involved,
kind
of
rolled
up
various
pieces,
each
one
of
those
became
incremental
conversations.
J
J
J
I
think
some
amount
of
that
also
fill
is
you
know,
as
a
lot
of
this
is
also
as
we
breathe,
breathe
enough
life
or
with
enough
championing
to
even
see
those
see,
see
the
status
of
those
working
groups
through
because
if
we,
if
it
adds
too
much
process,
you
know
and
we're
in
the
process
of
breathing
life
into
all
the
different
areas
that
would
need
traction
or
review
that's.
It
would
be
difficult,
but
I
think
it's
still.
F
Feasible
once
once,
you
figure
out
what
those
working
groups
are
it'd
be
nice.
If,
on
this
meeting,
you
could
do
kind
of
a
stand-up
style
like
no
more
than
five
minutes
per
working
group,
just
say
that's
what
happened
this
week
and
let
the
let
the
kind
of
discussions
or
like
the
discussions
right
now
are
free
for
all.
You
know
it's
like
whoever
gets
to
that
hack.
B
I
think
you
know
having
some
specific
focus
areas
where
you're
actually
talking
to
the
people
who
can
what's
the
right
to
make
decisions
is
a
strong
word,
but
you
know
talking
talking
with
a
group
of
people
who
all
have
a
vested
interest
in
that
area,
and
then
this
meeting
you
can
bring
back
those
five
minutes
that
lets.
You
know
everyone
be
aware
of,
like
kind
of
where
that's
heading,
and
that,
and
that
may
be
a
vehicle
to
get
more
interest
for
a
specific
area,
because
then
it
becomes
a
much
more
practical.
B
Oh
I
see
what
we're
trying
to
accomplish
and
yeah.
We
have
a
product
we're
trying
to
develop
and
we
should
get
involved
in
that
so
yeah
and
I
agree
with
vince,
and
I
the
danger
here
is
to
make
it.
You
know
another
thing
that
people
have
to
figure
out
following
a
set
of
processes
and
posting
statuses
somewhere,
but
I
I
hope
and
think
we
can
make
that
fairly
lightweight
and
maybe
even
steve's
kind
of
nv2.
B
A
Well,
I
look
at
some
other
projects,
like
you
know,
kubernetes
and
some
of
the
others,
and
I
think,
there's
here's
some
cool
things
we're
doing
with
what
exists
and
then
there's
because
it
just
there's
enough
flexibility
built
into
the
system.
I
wonder
how
much
we're
at
a
point
now
where
we
need
to
enable
enough
flexibility,
so
people
don't
have
to
come
back
here
to
get
approval.
They
basically
can
just
show
off
something,
or
do
I
think,
that's
part
of
the
bottleneck.
We're
starting
to
see
is
the
only
way
I
can
get.
A
My
thing
in
is,
if
I
make
this
change
and
that
you
know,
and
does
that
change
specific
to
that,
does
it
have
other
implications
as
opposed
to
can
we
get
some
changes
in
that
enable
people
to
have
extensibility
without
having
to
come
back
to
get
approval,
as
opposed
to
hey?
Let
me
show
off
what
I
pulled
off.
B
Yeah,
I
agree,
and
I
think
that's
that's
why
I
started
you
know
calling
it
some
sort
of
inflection
point
and
derek
sort
of
clarified
that
you
know
that
it's
almost
as
if
phase
one
of
oci
kind
of
closed
and
we
have
to
figure
out
how
phase
two
works,
because
it's
very
different,
it's
not
I'm
bringing
something
that
a
million
people
already
use
and
let's
figure
out
how
to
like,
put
that
in
a
nice
standardized
format.
You
know
it's
yeah,
so
there's
a
tension
there,
because
we
we
have.
B
B
J
Yep,
that's,
I
feel,
like
that's
some
of
those
conversations
that
I
think
or
even
hypotheticals
we
had
four
years
ago
now.
You
know
various
folks
have
experimented
with
and
said.
Actually
this
improves
x,
y
and
z,
and
I
think
it
would
do
the
same
for
everybody.
It
wouldn't
break
anybody
and
like
we're
right,
we're
right
and
ready
for
those
kind
of
changes
as
well,
and
sometimes
those
don't
even
need
whole
work.
That's
just
that's!
J
J
Kind
of
like
carry
that
conversation
forward.
That
means
it
worked
great
if
it
means
just
like
having
active
maintainers
engaged.
That's
that's
great.
B
Yeah,
I
think
chris
brought
up
a
good
point
in
the
chat
I
mean
there.
I
think
we
still
have
this
decision
tree.
That
I
think,
is
hard
to
figure
out
with.
Can
my
thing
go?
Go
into
a
existing
spec,
hey,
you
know
I'll
fill
out
your
form
that
I
promise
not
to
break.
You
know
current
clients
and
servers
and
all
this-
and
I
guess
yeah-
I
don't
know-
maybe
part
of
that-
can
be
solved
with
templating
and
github
to
kind
of
answer,
a
bunch
of
questions
but
yeah
that
that's
a
tricky
one.
H
G
In
some
ways,
but
you
still
have
this
issue
where
you
have
a
fundamental
you
know,
specs
that
are,
should
be
stable
and
not
necessarily
break
people
with
the
desire
to
do
fundamentally
new
interesting.
You
know
things
that
kind
of
the
ecosystem
demands
and
we
have
to
have
to
somehow
bridge
these
two
things
and
kind
of
enable
innovation,
while
also
ensuring
stability.
H
H
If
we
roll
out,
we
imagine
like
we
meet.
We
group
rolling
groups
come
up
with
the
plans
right
at
the
end
of
the
day,
if
it's
not
something
where
the
majority
of
the
clients
are
either
forward
compatible
with
changes
that
we've
got
or
are
self-updating
to
be
into
the
new
thing
and
there's
a
whole
set
of
questions
around
that
it
doesn't
matter
we're
just
masturbating
in
spec,
right
or
making
up
incompatibilities
or
balconizing
the
space
that
we're
in
right.
None
of
those
are
good
stories,
and
so
maybe
this
is
a
meta
problem.
H
H
Some
version
of
something
and
not
keeping
up
with
regard,
because
that's
another
way
of
getting
the
same
balkanization
only
now
version
oriented
list
whatever,
and
so
to
me
that,
like
there's
a
lot
of
technical
details,
we
have
to
figure
out
and
a
lot
of
clarity
around
what
problems
trying
to
solve.
But
there's
a
really
fundamental
one
of
are
we
actually
stuck
in
a
situation
where
we're
frozen
on
a
spec
that
isn't
going
to
make
a
difference
if
we
update
it
because
nobody's
going
to
move
anyway.
G
I
I
I
I
mean
my
view
is
if
the
oci
gets
together
and
everyone
says
hey,
this
is
v2.
We're
all
going
to.
You
know,
do
this
and
support
this
then
mission.
You
know
potentially
mission
accomplished
if,
if
people
go
out
on
their
own
and
kind
of
build
their
own
things,
obviously
you're
going
to
have
that
balkanization,
but
I
think
the
group
has
to
come
together
and
decide
if
they're
going
to
do
something
forward.
A
H
A
H
But
like
let's
say
we
all
agree
and
have
a
v2
spec
great.
What's
the
rollout
plan,
what's
the
go-to
market
with
like
the
bazillions
of
people
using
the
current
v1
stuff,
how
many
years
will
be
before
we
can
count
on
v2
being
there
like,
because
otherwise
it's
like
great
we've
solved
this
problem,
but
you
have
to
upgrade
to
v2
and
nobody
wants
to
for
another
thousand
stupid
reasons
or
whatever.
A
Well,
I
think
this
is
where
the
the
use
case
and
the
you
know,
customers
will
opt
in
for
what's
valid
and,
what's
not
I
mean
things
get
delayed,
not
because
it's
just
the
process
of
delay
that
things
get
delayed,
because
it
just
wasn't
considered
to
be
that
useful,
and
I
think,
if
there's
nothing,
stopping
us
from
like
what
happened
to
the
time
where
what
justin
was
saying
is
that
we
go
off.
We
incubate
some
ideas
based
on.
A
What's
there
we
get
feedback
and
we
continue
on
that
design
and,
as
it
looks
more
generally
useful,
we
bring
into
the
group
hey
here's
some
ideas
and
you
know
as
that
adoption
drives.
You
know
the
the
various
providers
of
these
tools
will
adopt
things
that
customers
are
asking
customers
and
users
are
asking
for.
So
I
don't
think
it's
necessarily
time
latency
built
in
it's
more
on.
K
My
feeling
is
the
time
scale
for
change
on
existing
things.
We've
got
some
past
history,
we
added
multi-arch
images
which
it
took
maybe,
and
we
added
I
mean
we
changed
the
image
format
from
v1
to
you
know
the
registry
5
from
v1
to
v2
in
the
first
place,
like
those
things
they
took
a
few
years
for
to
roll
through.
K
I
think
some
of
the
things
that
people
are
working
on
have
more
of
a
push
for
people
to
upgrade,
because
there's
actually
like
bigger
advantages,
probably,
but
I
think
we're
on
that
sort
of
time
scale.
So
we
don't
want
to
like
make
things
that
are
we
don't
want
to
change
things
all
the
time
for
existing
use
cases
there
are.
We
are
trying
to
enable
a
whole
lot
of
new
use
cases.
You
just
can't
do
now,
and
I
think
those
are
I
mean
I
think
you
know.
K
Artifacts
have
been
driving
a
lot
of
this,
because
those
are
just
things
you
couldn't.
You
can't
do
and
they're
not
breaking
anyone
they're
just
enabling
new
things,
and
so
I
think
those
can
move
at
a
slightly.
You
know
a
slightly
faster
pace,
but
even
then
they've
got
issues
like
you
know,
helm
moving
into
registry.
Is
it's
been
a
slogan?
You
know
that's
going
on
the
order
of
a
few
years
change
because
it's
an
existing
thing,
but
so
you
know
so
I
think
we
would.
K
We
need
to
plan
on
that
those
kind
of
time
scales
and
they
look
at
the
rate
of
people
which
people
update
cuban
s's.
It's
not
that
quick.
So
we
can't
we
can't
plan
for
things
to
be
short,
but
they
do
mostly
happen
over
time.
A
To
the
community
stuff
is
they
did
build
in
some
extensively,
so
you
know
all
the
emission
controllers
and
I
don't
even
try
to
list
all
the
ones
that
I'm
not
sure
of
all
the
extensibility
points.
I
think
that's
that
kind
of
focus
some
of
them
came
late,
fair,
but
they
don't
all
come
to
the
aka
working
group,
because
once
the
extensibility
was
in
there's
just
a
playing
field
for
them
to
use
what's
already
there,
but
I
think
you
know
we
can
innovate.
A
K
A
A
K
Yeah
I
mean
I
think
yeah
multi-arc
didn't
have
as
high
any
like,
so
it
took
longer.
It
eventually
happened
when
you
know
again,
it
was
need
driven
like
people
needed
new
feature.
New
feature
was
specific.
It
was
in
the
spark
if
there
were
implementations
there
was
code
when
you
need
it.
Eventually,
you
get
around
to
doing
it.
C
I
think
there's
another
interesting
point,
michael
to
answer
your
question
in
terms
of
the
go
to
market
and
the
end
users
of
this,
I
think
it
really
depends
on
where
we
defined
our
end
users,
because
if
the
end
users
of
these
specifications
are
the
large
platforms
that
are
developing
tooling
around
this,
then
what
we
are
doing
is
going
to
market
by
defining
this
and
talking
about
it
with
those
people
in
the
room
and
moving
the
industry
with
the
specification
and
the
you
know,
test
tooling
and
reference
implementations
and
such
as
we
go
ahead
and
implement
these
things
into
our
own
tooling.
C
All
that
said,
if
you're
talking
about
an
end
user
who
may
consume
it
good
lord,
I
hope
they
never
need
to
know
whether
or
not
this
isn't.
You
know
container
image,
spec,
something
or
another
versus
versus
others,
but
generally
the
open
source
projects
don't
work
on
a
go
to
market,
particularly
because
this
is
about
open
source.
This,
isn't
this,
isn't
the
product
it's
the
project
which
supports
each
of
our
products
around
it.
F
C
F
To
sorry,
it's
much
easier
to
ask
for
resources
if
there's
an
actual
spec
defined
and
it's
less
easy,
but
still
easier
to
ask
for
resources.
If
there
is
a
working
group
proposing
a
spec
that
has
been
like
empowered
as
a
working
group
and
it's
almost
impossible
to
ask
for
resources.
If
it's.
If
it
looks
like
it's
one
guy
asking
filing
a
pull
request
against
a
repository
for
a
spec.
B
So
something
tells
me
we're
not
doing
a
whole
lot
of
presentations
today.
B
J
Okay,
so
what
but?
What
are
the
big
takeaways?
I
mean,
there's,
there's
good
stuff
here,
I
think,
breathing
breathing
life
into
like
literally
we
we
discussed
this
and
knew
that
this
was
going
to
be
a
thing
that
there
would
be
a
slowdown
period.
There'd
be
a
rekindling
at
some
point
and
apparently
we're
there.
Even
though
there's
been
like
constant
activity,
it's
been
on
all
things
around
it,
and
so
now.
F
J
Point
and
we've
even
almost
anticipated
it,
because
I'm
leaving
like
scrolling
through
a
lot
of
these
different
issues
that
are
open
like
how
to
clean
up
the
charter.
How
to
make
this
more
even
some
drafts
about
this
stuff,
whether
what
to
do
with
some
of
the
different
projects
or
repos,
that
we
have
that
have
arguably
slowed
to
a
crawl
and
they're
not
used.
That's
fine,
so
I
think
just
kind
of
beginning.
The
flurry
of
activity
and
kind
of
champion
is
something.
That's
really.
J
J
Well
me
personally,
I
I
I
I
intentionally
tried
to
you
know
be
less
involved
for
right
after
1.0
just
to
give
myself
a
cognitive
break,
but
then,
particularly
last
year
have
been
a
little
bit
more
occupied
but
happy
to
jump
in
and
be
more
involved.
B
Yeah
thanks
vincent
so
yeah,
I
think
takeaways.
We
we
need
to
hopefully
not
belabor
this
to
death
but
find
a
simple
kind
of
straightforward
here's.
What
working
groups
look
like
here's
a
framework
for
how
they
operate
the
we
can
use
the
tob
mailing
list
and
issues
to
hash
that
out.
So
it's
public
other
people
can
comment
and
get
involved
as
as
necessary.
B
I
think
the
other
takeaway
is
is
just
well.
I
think
we're
all
we'll
all
be
interested
and
help
in
any
way
with
chris
and
amy
with.
You
know,
call
for
kind
of
maintainership
some
more
activity
around
that,
and
then
I
think.
B
Hopefully
we
can
turn
that
around
you
know
without
letting
a
lot
of
other
time
go
off
the
calendar
so
that
you
know
we
can
restructure
this
meeting
so
that
people
who
attend
here
feel
like
they're.
Actually,
you
know
getting
value
for
their
time.
I
know
you
know.
I
definitely
want
to
hear
more
from
justin
on
his
dock
and
these
other
changes-
and
I
know
sargon-
has
brought
some
good
stuff.
So
you
know,
let's,
let's
try
and
figure
out
a
way
to
to
make
all
those
constituents
have
a
have
a
place
here.
C
That
one's
for
legit
hand
raise
so
one
of
the
things
we
didn't
address
here-
and
I
know
is
maybe
I'm
maybe
I'm
overstating,
but
I'm
pretty
sure
is
a
goal
of
at
least
some
of
the
groups
is
also
speed.
It's
I
mean
we're.
We,
microsoft
are
much
more
okay
with
the
let's
call
it
solid
pace
that
oci
continues
to
move
at.
You
know
or
stayed
pace
how
about
that,
but
I
know
that
there
are
other
groups
that
are
being
challenged
by
by
the
slowness.
C
Is
there
something
in
the
short
term
that
we
can
do
with
that?
Or
do
we,
because
I
I
fear,
with
this
group
and
with
some
others
that
where,
when
we
say
okay,
we
need
to
go
back
to
governance.
We
need
to
go
back
to
like
how
do
we
do
this
or
write
if
you
know,
update
the
chart
or
that
kind
of
thing
that
that
feels
like
delay
tactics
as
opposed
to
progress,
so
I
just
want
to
make
sure
we
address
that
in
some
way,
instead
of
just
saying.
C
G
Yeah,
I
think
I
think,
a
quick
response
to
that.
One.
Sarah
is
like
having
this
meeting
currently
rolling
of
howard
kind
of
how
we're
doing
it
could
kind
of
continue
while
orthogonally
the
oci
tob,
who
basically
has
the
power
to
change
rules
anyway.
It
goes
out
and
kind
of
comes
up
with
a
quick
mechanism
for
working
groups
and
cleans
up
the
maintainer
list,
but
I
do
think
like
any
time
you
get
a
meeting
with
you
know:
15
20
20
people
in
the
industry
that
kind
of
basically
control
how
containers
works.
G
You
keep
that
meeting
going
and
then
we
could
go
clean
up
some
of
the
governance
stuff.
That's
just
it's
basically
governance
debt
from
from
over
the
years.
That
would
be
my
thing
like
keep
this
going,
yeah.
C
It's
like
every
non-profit,
I've
been
a
part
of
you
join
the
board,
you
go,
are
we
doing
the
things
our
bylaws
say?
We're
doing
or
our
governance
rules
say
we're
doing.
Yeah.
B
B
We
at
least
have
time
for
josh's
announcement,
so
we
don't
have
to
carry
that
over
to
april.
F
Sure
yeah,
so
the
distribution,
spec,
rc2,
1.0
rc2,
just
went
out
an
hour
two
ago.
So
that's
pretty
much.
That's
pretty
much
it
for
1.0,
there's
a
few
like
cleanup
tasks
and
I'm
hoping
to
cut
a
rc3
or
put
a
vote
out
for
rc3
by
the
end
of
the
week
and
if
there's
no
rejections
on
that,
that
should
be
the
1.0.
F
I
I
we
could
certainly
release
rc2.
There
is
just
a
few
missing
links
and
old
code
and
I
looked
up
sorry.
I
looked
at
the
governance
somewhere
in
the
org
and
it's
for
major
releases.
It
has
to
have
three
rc's,
I
believe
so
we
would
need
a
rc3
anyway.
It
could
be
the
same
commit
as
rc2,
but
there
are
a
few
things
to
clean
up.
So
congratulations.
F
But
yeah,
I
I
guess,
like
the
final
thing
to
say
on
that,
is
it's
mostly
typo
cleanup
type
work.
If
you
feel
very
strongly,
you
know
it's
up
to
the
I
guess
maintainers
to
decide
if
it's
worth
holding
up
a
release,
but
if
you
feel
so
strongly
that
something
should
get
in
there
prior
now
and
tomorrow
and
friday
is
like
a
great
time
to
do
that.
So
any
last
minute
reviews
I
appreciate
it.
F
B
L
I
just
say:
don't
don't
feel
the
need
to
do
too
much,
it
seems
like
getting
active,
maintainers
would
solve
like
90
of
the
stuff
and
you
risk
over
correcting
and
overfitting.
If
you
try
to
design
something
too
elaborate
at
this
point
to
address
some
problems,
it
might
just
be
solved
with
more
interesting
activity.
B
Yeah
yeah
good
point
and
yeah
we're
using
a
term.
That's
used
all
over
the
place,
so
working
groups
mean
a
lot
of
different
things
to
a
lot
of
us.
I'm.
I
hope
when
people
hear
that
here
we're
talking
about
something
very
simple
and
lightweight
that
just
kind
of
shifts
some
ownership
on
things
that
don't
fit
neatly
into
this
is
a
pr
on
this
repo
and
here's
the
maintainers,
because
I
I
think
that's
what
we're
trying
to
understand
is
there
are
things
that
are
broader
than
that.
H
This
is
definitely
like
getting
together
a
whole
bunch
of
people
I
haven't
seen
in
a
while,
and
you
know.
Hopefully,
we
can
go
from
a
crisis
model
to
a
like.
I
think
this
working
groups
will
lead
to
a
better
path.
We
have
a
lot
of
things
to
talk
about,
including
you
know,
just
things
that
have
come
up
recently,
but
also
you
know
the
last
time
we
met
for
a
crisis
meeting
which
wasn't
directly
oci
related
but
was
related
to
the
docker
hub
situation.