►
From YouTube: OCI Weekly Discussion - 2021-11-11
Description
Recording of the OCI weekly developer's call from 11 Nov 2021; agenda/notes here: https://hackmd.io/El8Dd2xrTlCaCG59ns5cwg?view#November-11-2021
A
A
A
B
B
So
can
you
see
my
browser
right?
B
Yes,
so
I
set
out
to
kind
of
scrub
the
repositories
in
the
github
organization
to
look
for
who
wasn't
on
github
actions,
because
we
had
a
bunch
of
failures
related
to
travis
and
things
like
that
and
other
projects
were
using
github
actions.
B
Beyond
that,
I
started
looking
into
like
other
types
of
discrepancies
between
the
repos
and
one
of
them
seemed
to
be
using
pool
approved
versus
code
owners.
So
if
anyone
saw
a
bunch
of
prs
come
in
yesterday,
it's
a
proposal
that
we
basically
move
completely
to
code
owners
from
pool
approve
but
yeah.
Basically,
in
doing
this,
I
put
together
this
simple,
like
script,
repo,
oci
stats
and
you
can
it'll
clone
in
all
the
repos
and
actually
spit
out
this
little
report,
so
red
x
is
supposed
to
be
like
bad.
B
So
I
wanted
to
like
show
and
tell
this,
but
beyond
that
I
just
want
to
have
an
overall
discussion
on
like
do.
We
want
to
standardize
on
governance,
slash
automation,
type
files
across
the
repos.
B
Is
there
any
value
in
doing
that?
I
know
I
I
chris
had
mentioned
that,
like
the
organization
was
paying
for
pool
approved,
so
there's
like
another
reason
there
beyond
just
standardization
anyway,
any
thoughts
on
that
somebody
would
pay
for
that.
A
But
it
used
it
used
to
be
a
good
thing
and
it's
it's
not
done
for
a
little
while
same
with
travis
I
mean.
Obviously
I
can't
I'm
not
I'm
not
a
maintainer
on
all
the
repos,
but
I
think
in
general
travis
and
bull
approved
need
to
go.
It's
interesting.
I
didn't
even
think
about
like
glide
and
some
of
those
I
do
think
one
of
the
projects
was
meaningfully
using
circle
ci.
So,
like
there's
kind
of
a
space
here
to
say,
we
want
to
standardize
on
some
stuff.
B
Something
like
I
want
to
say:
runtime
tools
was
using
cirrus
because
of
it
was
related
to
vagrant.
I
think,
maybe
not
run
time
tools,
one
of
the
repos,
so
that
one
I
kind
of
let
go
because
akihiro
had
opened
a
pr
about
like
how
actions
was
way
too
slow
for
laborers,
so
that
I
think,
is
a
one-off.
But
besides
that,
like
circle
is
really
not
used.
Travis
is
really
not
used.
B
Circle
is
used
on
this
repo
called
web,
which
I
don't
actually
know
what
we're
doing
with
this.
But
vanessa
had
done
a
ton
of.
A
Yeah
but
it's
like
archived,
you
know
maybe
something.
B
B
A
A
But
a
script
like
this
could
actually
make
sure
that
at
least
the
github
handles
that
are
called
out
and
that
they
match
some
aspect
of
code
owners
yeah
matches
what's
in
the
maintainers
file
or
you
know
whatever,
because
I
would
like
to
only
have
one
file
there
to
maintain,
but
there
might
be
places
where
we
start
breaking
out
additional
roles.
You
know
so
called
on
code
owners
could
could
be
different,
not
exactly
the
same
because
it
might
have
different
roles.
A
But
I'm
just
saying
hey
like
there's
things
like
that
and
then
even
the
topic
of
like
the
governance
document
of
like.
A
B
Anyways
yeah
there's
there's
kind
of
a
bigger
question
too
of
like
who
owns
some
of
these
repos,
so
if
it
doesn't
have
a
maintainers
file
like-
and
I
think
it's
just
I
mean
overall
permissions
wise
in
actuality,
it's
managed
by
github
teams
correct
as
far
as
I
understand,
but
it's
it's
like
there's
strange
things
like,
for
example,
conformance
which
I'm
gonna
bring
up
sometime
after
this,
like
this,
has
people
who
have
been
kind
of
not
around
in
maintainers
and
meanwhile,
like
my
name
is
not
in
here
and
so
like
who?
What
file?
B
B
A
There's
two
issues:
okay,
click
on.
I
think.
B
A
I
think
I
think
some
of
this,
you
know
a
few
of
these
things.
We've
either
like
somebody
takes
it
upon
themselves
to
fix
or
clean
up
some
aspect
of
it
and,
however
far
they
get,
and
then
some
of
these
are
like
one-offs
that
we
ping
amy
and
chris
about,
but
it
probably
just
needs
to
be
almost
like
a
checklist
or
almost
like
this,
like
a
spreadsheet
of
like,
like
things,
things
that
need
to
be
done
to
actually
clean
this
up.
A
You
know,
and
some
of
this
even
like
a
a
script
like
this
that
could
run
and
just
you
know,
be
kind
of
a
report
of
things
that
don't
look
like
consistent
across
the
repos
anything
that
can
help
like
point
out
or
call
out
things
that
are
out
of
out
of
sync,
it's
kind
of
like
the
the
the
mundane
aspect
of
watching
a.
C
So
I'm
I'm
still
wondering
whether
oci
has
a
like
an
overarching
process
for
just
like
regular
housekeep
housekeeping
such
as
something
like
this.
Looking
at
maintainers
files
seeing
who's
active
who's,
not
active,
then.
B
Decisions,
I
think
it
doesn't.
I
think
that
probably
linux
foundation
has
some
level
of
compliance
stuff,
but
I
think
like
specific,
like
ci
tools
and
things
like
that,
I
think
it
then
just
left
up
to
the
repo
people.
A
A
I
agree
I
don't
there's
times
when
I
like,
don't
want
to
ask
for
any
more
privileges
just
because,
but
but
how
does
it
break
break
down
into
different
responsibilities?
Because
even
like
the
the
some
of
those
repos
like
you
said
that
don't
even
have
maintainers
files
like
what
should
be
the
bucket
that
they
fall
into?
Because
if
it's
just
like
a
website,
you
know
or
like
a
bot
or
something
like
that?
It's
just
like
organizational.
A
A
A
union
of
of
all
the
current
maintainers,
or
something
like
that
like
if,
if,
if
other
maintainers,
can
come
together
and
not
have
to
have
like
two-thirds
of
them,
I'm
just
saying
like
something
to
where
folks
could
you
know
if
you're
is
that?
Okay?
So
let
me
let
me
step
back,
I'm
not
I'm
not
a
fearless
leader,
I'm
full
of
fear.
In
the
days
of
your
when
it
was
github.com
open
containers,
I
think
for
a
while
it
was.
A
There
was
literally
like
one
or
two
repos
for
a
little
while
that
was
just
kind
of
like
you
know,
foundational
tob
charter
stuffs,
maybe
like
the
websites.
Maybe
there
are
three
repos
but
pretty
much.
It
was
like
foundational
stuff
and
like
website,
which,
like
were
some
members,
and
then
linux
foundation
like
chris
and
the
only
technical
repo
for
a
while
were
the
specs
and
lib
container
and
the
maintainers
of
those
two
two
things.
A
It
was
just
saying
the
developers
of
these
kind
of
things
not
necessarily
github
and
administrative
breakdown,
of
like
maintaining
specific
repos,
but
just
you're,
either
like
a
a
member
who
has
potential
like
legal
patent,
whatever
rights
that
you
negotiate
on
like
the
foundation
side
or
you
are
a
developer
of
the
specs
or
flip
container,
and
that
was
a
those
two
maintainers
files
were
in
sync,
and
so
some
of
the
wording
was
kind
of
meant
for
that
and
that's
completely
broken
down
over
the
years
to
where,
like
effectively
the
tdc
is
the
union
of
every
maintainer's
file
across
the
repos.
A
A
If
there's
a
bot,
you
know
kick
kick
kick
a
ci
job
if
it
needs
to
be
kicked
but,
like
you
know,
like
kind
of
like
grading
of
who
could
who
could
actually
have
like
admin
rights
of
a
particular
repo,
even
like
an
owner
admin
of
a
repo
and
kind
of
grade
that
out
into
different
buckets.
The
part
of
this
that's
even
more
confusing.
B
Is
the
actual
accesses
like
point-and-click
and
github
teams,
so
code
owners
doesn't
actually
like
code
owners
is
gonna
say
I
believe
like
who
should
be
asked
for
review
on
prs.
A
B
A
Interestingly,
no
that's
that
is
kind
of
like
a
workflow
thing.
Code
owners
is
like
a
workflow
thing
and
then
members
you're
you're
correct
that
some
of
the
members
and
otherwise
comes
down
to
the
github
teams
and
looking
at
this,
like
you
know,
there's
groups
like
core
that
still
have
patrick
shenzhon
and
ben
freshman,
who
you
know
haven't
been
involved
in
years
like
like
a
lot
even
just
looking
at
these
numbers.
It's
pretty
stale
and
I
don't
think
that
there's
a
team
for
every
repo,
even
or
project
or
focus
anisha.
C
So
I
would
like
to
make
a
proposal
here
how
about
how
about
I
go
off
and
map
all
of
the
resources
that
need
that.
That
is,
that
is,
under
this
whole
open
containers
of
github
organization,
and
you
know,
map
out
who
has
who
has
ownership
and
who
needs
to
like
what
level
of
ownership
is
required
for
each
one
of
them.
C
I
do
have
another
proposal,
which
is
you
know,
maybe
maybe
take
a
look
at
this
tob
github,
because
I
took
a
look
at
it
and
it
looks
like
there
are
a
number
of
proposals
that
are
like
create
project
criteria,
guide
for
proposals
and
governance
and
no
confidence
monitors.
I
mean
it
seems
like
there's
a
lot
of
documentation,
work
that
needs
to
happen
within
this
project.
C
I
don't
have
a
proposal
over
there,
but
it's
just
something
that
you
know.
It's
probably
just
like
administrative
work.
B
C
I
mean
it
seems
like
t.o.b
is,
has
opinions
about
it,
but
it's
not
written
down
anywhere.
D
D
D
B
Right
because,
like
I
like,
I
don't
think
we
should
prohibit,
for
example
no,
but
I
like,
I,
don't
want
to
make
the
overhead
for
making
a
reap
like.
We
had
to
have
a
repo
for
container
images
recently,
and
it
was
kind
of
it's
kind
of
just
like
thrown
together.
B
D
Well,
is
that
documented
for
each
project
in
the
tob
project?
Should
that
not
so
we
don't
have
a
list
right
of
like
it
should
be
like
here's
all
the
projects
and
here's
the
proposal
that
created
the
project
and
then
here
is
you
know,
a
link
to
that
proposal
and
I'll
link
to
the
repo
right.
We
don't
have
that.
C
D
C
A
I
mean
I
guess,
if
it's
not
clear,
then
it
needs
to
be
more
clear,
but
the
the
tv
usually
is
just
for
resolving
conflicts,
and
so
I
think
it's
kind
of
morphing
into
a
general.
You
know
not
that
I
want
it
to
or
otherwise
just
it
seems
like
it's
in
in
the
lack
of
clarity
that
we're
describing
it's
kind
of
assumed.
A
Like
the
t
this
all
this
would
roll
up
to
the
tob
if
we
can't
figure
it
out,
and
I
don't
think
that
that
was
originally
the
intention
for
the
tob,
but
so
I
think
it's
probably
best
just
to
kind
of
iron
out.
A
You
know
here's
here's
categories
for
these
projects
and
if
these
projects
don't
have
clear
categories-
and
this
is
almost
like
recommendation
for
the
other
week
of
like
if
we
went
through
and
made
some
kind
of
clear
delineation
like
this
project-
is
a
spec.
You
know
therefore,
here's
the
guidance
on
like
how
voting
breaks
down.
You
know
whatever
for
things
that
are
specs.
A
You
know
and
here's
this
other
project.
It
is
code,
you
know,
go
digest
or
se
linux
or
whatever
it
is.
Here's
how
votes
work
for
code
projects
you
know,
and
so,
if
every
repo
that
we
have
doesn't
fall
into
a
defined
kind
of
bucket,
you
know,
and
even
if
that
bucket
is
like
marked
down
or
whatever
like
these
things
have
to
be
handled
in
a
certain
way,
whether
it's
the
github
repo
or
you
know,
whatever
some
kind
of
like
website
generator.
A
Like
the
you
know,
every
every
repo
should
kind
of
fall
into
those
buckets
and
at
least
guidance
on
how
those
things
get
maintained.
Who
do
you
ping,
if
you
need
repo
level
or
if
it's
just
an
lgtm,
that
you're
looking
for
that
kind
of
guidance?
I
think
all
that
is
is
basically
what
we're
we're.
Looking
at
and
you're
kind
of
fine
you're
kicking
up
against
with
like
the
inconsistencies
against
these
repos
in
this,
the
scripts
repo,
the
scripts
stats
earlier
nisha,
when
you
said.
A
A
I
just
don't
want
you
know
like
waste
your
time
if,
if
there's
a
good
way
to
view
it
or
generate
it,
whether
it's
like
what
what
repos
exist,
what
teams
exist?
What
you
know
who
who
has
what
rights,
if
it's
just
like
on
a
github
level
or
what
or
code
on
this.
C
Level,
I
don't
know
stephen
and
stephen
and
jason
have
their
hands
up.
Is
it
okay
for
me
to
go
forward
and
answer
that
question
or
do
you
want
to
wait
for?
I
guess
he
has
his
hand
down
now.
This.
C
Okay,
well
so
I'll
answer
the
question,
so
I
I've
been
on
these
calls
for
a
while
now
and
what
I've
seen
is
that,
as
far
as
like
governor
general
maintenance
or
like
statistics
on
all
of
the
the
state
of
all
of
the
repos
is
usually
something
that
josh
does
and
he's
the
one
that
brings
brings
it
up
in
these
meetings
and
then
there
is
no
follow-up.
It's
like
okay.
I
found
this
issue.
C
What
do
we
do
and
the
the
reason
for
that
is
because
the
governance
is
there,
there's
no
documentation
in
the
governance
or
the
t.o.b
I
mean,
maybe
maybe
that's
the
problem.
Maybe
open
containers
needs
a
governance
repo
that
talks
about
what
each
of
the
processes
are,
what
the
tob
is
supposed
to
do.
C
E
Don't
think
it's
not
even
not
documented,
there
isn't
a
process
to
document
right
like
there
isn't.
We
all
keep
saying
someone
should
fix
this.
Someone
should
do
this.
Someone
should
do
a
survey
of
repos
and
put
them
in
the
right
order,
but
nobody
says
who
that
is.
I,
I
think,
vincent
your
comment
that
that
this
isn't
exactly
tob's
responsibility,
kind
of
scared
me
because
I
don't
know
what
higher
authority
there
is
than
t
ob
in
this
case.
E
A
That
the
the
thing
was
just
like
in
the
days
when
there
was
just
like
specs,
which
was
really
ambiguous
and
that's
why
it
later
broke
up.
You
know
and
broke
it
into
runtime
and
image,
but
in
the
days
of
like
specs
and
lib
container,
that
was
the
that
was
all
the
developer.
You
know
maintainers
like
really.
A
That
role
has
changed.
That's
what
I
was
saying
it's
like
it
was
never
really
intended
on
that,
but
it
is
kind
of
the
bucket
that
a
lot
of
these
things
fall
all
up
to
but,
like
you
said,
there's
no
process
defined
because
that's
originally
it
was
like.
I.
I
just
expect
that
we'll
be
doing
nothing
as
a
t.o.b
we're
just
there.
If
somebody
needs
us
to
break
the
logic
yeah,
not
not
to
like,
say
not
to
come
in
and
say
this
is
how
things
should
be
done
like.
E
E
We're
like
you're
right
at
the
time
that
the
tob
was
defined.
That
was
their
main
job,
and
so,
if
you
expect
never
to
have
conflicts,
then
it's
a
pretty
cushy
job.
You
don't
have
to
do
a
lot,
but
I
think
now
you
have.
We
have
a
conflict
against
inertia.
We
have
a
conflict
against
like
lack
of
clarity,
and
you
know
nobody
knows
who
owns
half
of
these
things.
Like
that's,
that's
the
conflict
now.
E
D
The
chat,
I
think
I
think
it's
important,
though,
to
frame
rather
than
saying
like.
Oh,
we
have
a
conflict
of
inertia.
I
think
I
find
it
best
to
be
very
specific
about
what
you're
trying
to
do
and
you
can't
do
because
otherwise
they're
just
kind
of
like
I
mean
I
think
when
we
created
this
project.
It
was
very
much
like
your
bias
for
action
right,
so
you
know
you
do
something
you
figure
out
how
to
do
it.
If,
if
it
really
is
against
the
rules,
don't
do
it
and
then
go
figure
out.
D
You
know
why
not
and
then
have
a
discussion
around
it.
What
concerns
me
about
this
conversation
a
bit
right
now
is
like
we're
saying:
okay,
we
need
these
governance,
but
we
don't
really
understand
what
we're
not
doing
and
the
areas
that
we've
we're
so
so
like
take
the
artifact
spec
or
the
some
of
the
stuff
around
the
working
groups
like
we
created
the
working
groups,
I
think
a
lot
of
people
lgtm
did
it.
Why
aren't
we
driving
that
like
it
like,
so
so
when
we
say
hey
we're
fighting
against
inertia?
D
What
what
do
you
mean?
What
are
you
trying
to
do
that?
You
can't
do
be
specific,
and
then
we
can
go
figure
out
whether
more
work
needs
to
be
done,
but,
like
I
think,
like
creating
a
bunch
of
governance
structure
without
having
something
technical
or
something
for
our
users.
That's
driving
that
governance
structure
is
just
going
to
be
a
practice
in
rulemaking
and
not
necessarily.
E
Yeah,
to
be
clear,
I
don't
think
I
don't
think
I'm
advocating
for
more
more
formal,
documented
governance
and
this
you
know
this
group
shall
be
responsible
for
whatever.
E
I
think
the
the
specific
tactical
issue
I
think
we
are
unable
to
resolve
is
that
some
repos
are
clearly
just
abandoned
and
nobody
knows
how
to
nobody
knows
whether
they
are
functionally
actually
should
be,
abandoned
or
or
should
be
removed,
and
even
just
like
getting
travis
removed
from
repos
is
not
people
don't
know
who
to
go
to
to
ask
whether
we
can
get
rid
of
who
to
ask
to
approve
that
pr
and-
and
that's
like
basic,
you
know,
cleaning,
you
know
sweeping
the
cobbs
out
of
the
attic-
it's
not
possible
to
get
done
because
nobody
knows
who's
responsible
for
that
addict.
D
If
we
really
need
to
change
the
the
charter
or
something
like
that
great,
but
I
think
like
the
tob
can
handle
that,
like
that's
their
in
their
role,
their
job
is
to
in
that
in
that
role
they
are
to
resolve
these
conflicts,
and
I
think,
like
going
further
with
governance.
That
would
be
great,
but
I
I
don't.
I
think
we
just
need
to
solve
the
I
mean
what
is
there
five
five
repos?
D
Maybe
we
need
to
go
in
and
figure
out
the
future
of
right,
like
going
through
governance
and
rule
making
to
solve
that
seems
heavyweight
when
we
could
just
go
and
figure
out
which
which
of
these
repos,
I
would
just
say
you
know,
send
it
to
dev,
send
it
to
the
tob
mailing
list.
Let's
have
a
discussion
there
get
it
worked
out,
change
the
you
know,
figure
out
what
votes
we
need
to
do,
and
you
know
I
think
I
think
I
think,
for
for
a
lot
of
these.
D
D
F
F
In
there
the
the
tob
has
the
already
has
the
role
to
deal
with
what
what
it
calls
high
level
issues
that
cannot
be
resolved
in
the
tdc,
and
so
I
think
that
right
now
we're
having
high
level
discussions
around
management
of
the
things
that
exist
in
oci,
and
that
is
the
role
of
the
tob
to
deal
with.
F
So,
if
I
I
think
what
what
steven
day
was
just
saying
is
right,
like
this
is
the
kind
of
thing
to
post
on
the
dev
mailing
list
and
come
to
consensus
or
the
tob
mailing
list
and
come
to
consensus,
and
then
we
can
act
on
it.
I
don't
know
that
we
need
a
governance
change
for
it,
but
we
already
have
the
the
structure
that
says:
here's
who
is
responsible
for
dealing
with
it,
it's
the
t.o.b.
F
Yeah
so
that,
like
these
weekly
meetings
are
not
really
part
of,
like
not,
everyone
comes
to
them,
they're,
not
part
of
how
we
necessarily
do
decision
making
they're
a
good
chance
to
like
talk
about
things.
But
you
know
the.
If,
if
you
read
the
charter-
and
you
read
the
the
stuff,
that's
in
the
tob
repo,
it
mostly
talks
about.
We
do
discussions
in
the
open
on
the
mailing
list.
D
Yeah,
it
would
be
great
to
hear,
like
maybe
chris
anderson
has
he's
like
yeah
here's,
what
the
what
was
the
one
that
started
with
f
the
foundation
right
like
you
know
he
could
say,
hey
the
foundation.
Repo
is
here
for
this
reason
and
we're
going
to
go
document
that
reason
we
can
set
up
vr
right
like
this.
We
should
just
do
this
on
the
mailing
list
and
and
get
through
it,
and
I
think
we
can
move
on
from
it.
C
C
So,
if,
if
some
now
that
you
have
several
projects
going,
where
can
someone
okay,
someone
like
me
who
has
not
been
involved
for
that
long
and
is
wanting
to
it-
wants
to
ask
a
question:
where
can
they
go
to
do
that
because.
C
D
C
Okay,
well.
C
Okay,
so
what
I'm
noticing
over
here
and-
and
this
is
my
experience-
your
mileage
may
vary,
but
what
I'm
noticing
is
that
I
get
the
most
immediate
feedback
when
I
come
to
these
meetings.
C
The
first
thing
I
started
off
with
was
the
mailing
list
and
there
wasn't
really
much
activity
going
on
the
mailing
list.
So
what
I
want
to
say
is
that,
yes,
all
of
this
is
documented,
but
there
isn't
really
any.
You
know
solid
follow-up.
Stephen
is
asking
for
specifics
and
that's
what
everyone
else
is
asking
for
too.
But
it's
not
a
two-way
street
like
it.
The
tob
is
asking
for
specifics.
We
don't
get
any
specifics
from
the
tob.
C
C
It's
bounced
around
from
you
know,
submitting
a
proposal
to
the
distribution
spec,
and
then
it
moved
to
its
own
repo,
and
now
it's
sitting
in
a
proposal
stuff
that
to
me
at
a
higher
level,
makes
me
think
that
the
the
tob
is
well
it
it.
It
has
some.
It
has
some
trouble
making
decisions.
C
C
Like
yes
or
no,
is
this
something
that
the
community
wants?
Yes
or
no,
what
we
say
is
we'll
look
at
it
and,
oh,
I
have
some
ideas
and,
oh
I
I
think
it
would
be
good
if
it
goes
over
here.
There's,
no,
yes
or
no.
G
Well,
I
mean
it.
The
the
artifact
stuff
is
the
interesting
example
of
this,
and
I
think
you
know
we
got
far
enough
to
create
an
artifacts
repo
which
was
scoped
at
documenting
what
the
existing
specs
does,
and
we
did
that
about
two
years
ago.
Whatever
it
was,
and
then
you
know,
we've
had
a
lot
of
success
with
it
and
then
we're
trying
to
go
to
the
next
step.
What
I
think
what's
happened
is
when
we've
gone
to
the
next
step.
G
Is
we
kind
of
hit
a
head
wall
of
how
the
original
specs
were
set
up
and
that's
why
we're
like
we
try
to
just
operate
within
the
artifacts
repo
right?
That's
that's
what
the
charter
says,
but
the
charter
did
not
approve
us
to
actually
create
new
specs.
It
was
going
to
be
documentation
how
to
use
the
existing
specs.
That's
fine!
So
now
we're
faced
with
this
question
of.
G
We
are
talking
about
refactoring
some
of
the
specs
and
how
do
they
split
up
from
one
particular
type
to
container
image
to
supporting
more
generic
types,
which
was
a
conversation
we
had
when
we
created
the
artifacts
repo?
Did
we
want
to
do
it
because
we
talked
about?
We
can
go
to
cncf.
That
was
a
choice
at
the
time
and
we
as
a
group,
the
tob,
decided
to
keep
it
under
oci.
G
What
I
think
we're
talking
about
here
is
it's
not
just
the
original
container
image
format
and
how
it's
run
and
stored
as
a
single
thing
in
a
registry,
it's.
How
does
oci
evolve
to
support
a
broader
set
of
spaces
and
scenarios
and
like
the
artifacts
artifacts
project,
is
very
active.
It's
just.
We
hit
a
wall
of
what
we
can
do
in
that
repo,
so
we're
creating
a
spec
under
cncf
to
hopefully
have
this
adopted
by
oci.
G
I
think
that's
the
question
for
our
tob
is:
what
do
we
want
to
see
going
forward
related
to
these
specs
is?
Are
these
specs
very
narrowly
focused
on
a
container
image?
If
that's
the
case,
that's
okay,
but
but
that
I
think
that's
part
of
the
guidance
that
nisha
is
talking
about.
Is
we
don't
have
clarity?
We
do
keep
on
getting
you
know
now.
Actually
this
is
interesting.
We
do
want
to
consider
it
here
as
opposed
to
no,
it
doesn't
fit.
Go
over
cncf,
that's
fine!.
D
D
I
mean
I'm
I'm
not
a
little
cheap,
I'm
not
on
the
tob,
but
I'm
getting
a
lot
of
feedback
on
the
positioning
of
the
artifact
spec
and
you
know
I
think,
there's
some
issues
that
I'm
kind
of
seeing
with
it
and
you
know,
maybe
others
don't
feel
as
comfortable
to
give
that
feedback.
D
So
you
know
that
it
could
be
that,
but,
like
I
don't
know
what
the
like,
I'm
looking
through
my
mailing
list
box
for
dev,
open
containers
and
stuff
and
I'm
not
seeing
the
proposals
laid
out
in
those
mailing
lists,
but
if
that
could
just
because
I'm
poorly
searching,
so
it
would
be
good
to
show
us
what
those
proposals
are
and
show
why
they're
getting
direct,
getting
rejected
right
or
just
sitting
there
right.
Because
again,
it's
really
easy
to
say
like
oh,
the
tob
doesn't
do
anything
right,
but
that's
not
actual
actionable
feedback.
D
Better
feedback
would
be.
They
didn't
do
anything
on
this
specific
item
and
I
think
that
like
if
we
talk
about
these,
the
very
specific
proposals
like
I
made
the
proposal
for
xyz
and
it
just
sat
there
so
that
we
can
specifically
address
that.
I
think
we'll
get
much
further
than
if
we
like
say:
oh,
the
tob
is
not
doing
anything.
G
Yeah,
no,
it
makes
perfect
sense,
but
not
everything
has
happened
on
the
dev
mailing
list
that
you
know
the
dev
mailing
list
is
not
a
great
forum
for
conversations
a
great
it's
a
great
way
to
tease
something
in
and
say
hey
by
the
way.
There's
this
pr.
Can
we
discuss
this
so
there
there
is
lots
of
feedback
on
the
artifacts
prs,
there's
27,
there's
29,
there's
37,
and
then
we
went
for
the
ocitob
96
that
prompted
the
working
group.
G
G
So
I
because
I
don't
want
to
without
for
the
feedback.
I'd
love
the
feedback,
any
feedback
that
we've
had
we've
addressed,
but
I
think
your
question
is
from
a
process
perspective.
G
We've
been
trying
to
follow
as
much
as
we
can
and
then,
like
even
the
vote
for
a
working
group,
it's
been
going
on
for
a
while
there's
various
conversations
around
why
we're
waiting
for
the
next
step,
but,
like
all
of
that,
has
been
extremely
transparent,
open
and
waiting
for
action.
D
D
G
I
see
vince
got
his
hand
raised.
I
mean
this
has
been
the
conversation
we've
had
for
months.
There
was
a
first.
G
Group
definition,
so
that's
how
that
got
spawned.
So
then
the
group
went
off
and
created
a
definition
for
working
groups.
While
that
definition
was
recreated.
I
made
a
proposal
so
that
hey
can
we
satisfy.
We
will
make
changes
to
the
working
group
proposal
based
on
how
the
definitions,
but
this
is
the
way
we're
thinking
about
it-
derek-
did
a
great
job
getting
that
structured.
The
lawyers
took
over
and
now
we've
been
sitting
there
for
a
couple
months
to
figure
out.
How
do
we
vote
and
move
forward.
C
So
josh
has
a
conformance
topic
and
vincent
has
his
hand
up.
I
think
I
think
we
can
ask
vincent
for
his
opinion.
A
It
was,
I
did,
did
not
okay,
let
me
say
something
to
say
it
was
very,
very
frankly,
I
didn't
think
it
was
my
responsibility
to
kick
off
the
vote
for
for
the
tob
like.
Is
this
completely
ready
to
be
voted
on
yes
or
no,
and
I
thought
that
there
was
still
conversation
going
around
and
we're
waiting
to
say
are
we
are
we
ready
to
vote
on
this?
Yes
or
no?
I
know
I
think
we
were
waiting
on.
A
I
was
waiting
on
that
if,
if
it's
in
my
lap
to
kick
off
that
vote-
and
I'm
like
like-
yes-
we're
ready
start
the
vote
now
because
I've
been,
I
was
thinking
even
when
we
talked
about
this
at
the
summit.
Like
is
this
ready
to
vote
on?
Now?
Yes
or
no,
can
we
get
this
closed
within
a
week
or
two
of
like
the
voting
time
period
and
there
was
no
definitive
like
yes,
we're
ready
to
vote
now,
so
I
did
nothing.
A
A
Strong
there's
a
strong,
yes
hoping
that
somebody
else
will
do
it
like.
You
know
whether
it's
falls
in
your
lap
or
not,
and
because
we're
working
across
you
know
different
teams
and
companies
and
people
and
like
saying
like
assigning
somebody
something
to
do
or
whatever
and
knowing
that
this
might
not
be
some
people's
day
job.
I
think
that's
why
some
of
these
tasks,
you
know
getting
the
flurry
of
activity
and
then
they
don't
they
fall
by
the
wayside.
A
That's
that's!
That's
kind
of
natural
there
is
like
there's.
A
lot
of
like,
like
was
just
said.
It
was
a
can
of.
Worms
has
been
opened,
the
the
the
the
thing
the
actionable
things
here
are
from
where
this
conversation
started
and
then
kind
of
you
know.
Obviously,
testing
has
very
viable
things
to
address
here.
A
There's
organization
that
needs
to
be
clarified
and
like
just
tidying
and
rolls
and
whatnot.
I
think
we
have
good
ideas
on
that.
A
The
uniformity
stuff
that
the
stats
repo
is
out
there
is
great
because
that's
been
bothering,
I
think,
a
lot
of
us
that
things
seem
out
of
seat
out
of
sync,
with
each
other
and
kind
of
stale
or
otherwise,
and
that's
great.
But.
A
All
the
stuff
about
mailing
lists
and
all
like
that
could
all
be
you
know
a
sinkhole
conversation.
I
think
the
big
thing
is
like
actually
probably
people
getting
comfortable
with
like
either
appointing
somebody
or
standing
up
and
taking
action,
and
if
somebody,
if
something
doesn't
have
like
somebody
who
says
that
they're
going
to
do
it
and
then
follow
up
to
say,
hey,
you
said
you're
going
to
do
this.
A
Did
you
do
it,
then
not
really
getting
upset
about
it,
because
I
think
everybody's
busy
and
everybody
has
important
stuff
to
talk
about,
but
it
falls
by
the
wayside.
So
there's
also
just
some
amount
of
patience
that
or
patients
are
understanding
or
otherwise
you're.
Just
saying
like.
Oh
it
didn't
have
an
owner,
that's
why
it
got
dropped
and
so
what's
the
process
of
making
sure
that
owners,
you
know
get
tracked
all
the
way
through
just
it's
hard
to
dredge
up.
A
You
know,
years
of
years
of
minut
interactions
and
say
this
is
a
single
point
or
something
went
off
the
rails.
So
I
was
trying
to
why
what
I
was
trying
to
like
actually
type
out.
What's
what's
a
gen
actionable
items
and
then
like?
How
can
we
actually
just
maybe
set
a
part
of
a
rolling
agenda
to
say
here's,
the
to-do's
that
somebody
said
in
the
past
what
happened
on
them
and
when
do
you
think
you'll
have
them
done
josh?
A
B
G
I
was
just
scared.
I
wasn't
trying
to
redirect
it.
I
was
trying
to
use
as
an
example.
I
really
wasn't
trying
to
turn
the
system
to
hey.
Let's
do
the
tob
vote.
I
think
that
started
my
por.
My
point
was
more
of
testing
some
of
what
we're
discussing
that.
I
think
we
are
trying
to
do
some
of
that,
but
we
are
hitting
some.
I
don't
know
what's
process,
if
it's
time
or
what
on?
G
How
do
we
proceed
with
things
that
aren't
I
mean
even
some
of
the
easy
peasy
simple
things
take
forever
to
get
done,
and
I
think
it's
because
we
don't
have
as
much
engagement,
because
people
are
busy
the
things
that
are
more
complicated,
don't
fit
within
one
replay.
Those
have
been
all
about
pause.
So
that's
that's.
The
question
I
think
goes
up
to
what
you're
trying
to
do
here
with
some
of
the
government
stocks.
B
I
want
to
point
your
attention
to
this
image
and
I
don't
know
if
anyone's
heard
of
this
mean,
but
this
could
be
us,
but
you
plan,
I
don't
recommend
that
you
actually
google
that,
but
I
want
to
point
your
attention
to
this
repository,
kate's
conformance
and
just
how
active
this
is
465
contributors
last
commit
six
days
ago
and
just
to
you
know,
give
some
background
on
how
this
works.
This
is
what
the
oci
conformance
was
supposed
to
be
and
we
haven't
had
people
submitting
so
beyond
me
pinging
people
on
the
side.
B
I
don't
really
know
how
else
to
encourage
folks
to
submit,
but
basically
the
way
that
the
kubernetes
one
works
is
for
each
new
version
of
kubernetes.
B
There's
a
folder
in
here
for
a
specific
distribution
of
kubernetes,
and
you
all
you
have
to
do
is
just
submit
a
readme
file
and
a
xml
like
a
junit
xml
that
says
how
to
run
conformance
against
this
distribution.
B
C
B
Instead
of
having
top
level
versions
folders
in
the
repo,
the
idea
was
to
nest
it
one
level
deeper
under
the
name
of
the
spec,
so
we
have
a
distribution,
spec,
folder
and
under
here
I'm
hoping
v100
and
then
azure
container
registry.
Google
container
registry.
B
Open
source
registries,
but
it
has
to
be
submitted
by
the
people
who
are
responsible
for
the
registry.
So
if
there's
anything,
I
can
do
to
help
you.
B
I
know
many
people
are
running
the
test,
so
it's
a
matter
of
if
you're
passing
the
tests,
even
just
the
pool
tests,
which
I
know
is
controversial,
but
even
if
you're
just
passing
the
pool
test,
I
would
like
for
you
to
come
and
open
a
pr
with
these
files
and
there's
a
there's,
a
there's,
a
file,
there's
a
markdown
file
instructions
in
here
and
when
you
run
the
conformance,
the
distribution
conformance
suite
against
your
registry.
B
For
example,
so,
basically
run
that
get
it
working,
you
don't
even
need
to.
You
can
like
scrub
details
about
the
register,
you're
testing
against
and
submitted
and
yeah.
That's
that's.
My
sales
pitch.
B
A
B
And
maybe
I
ought
to
just
send
a
mailing
list
thing
to
remind
people,
but
the
idea
was
after
distribution,
spec
hit
1-0
yeah,
which
has
been
several
months
now
that
people
submit
vrs.
So
there's
like
zot
was
the
first
people
to
get
on
this,
but
this
was
back
when
we
were
starting,
and
so
it
had
like
a
v10
pre
folder.
B
B
C
D
No,
no,
it's
it's!
The
reference
implementation
I
mean
like
like
the
original
spec,
was
written
at
the
same
time
as
that
as
distribution
and
then
most
of
the
oci
stuff
we've
prototyped
inside
of
the
distribution,
so
not
all
of
it,
but
a
good
part
of
it.
F
C
Well,
I
mean
I
do
have
a
follow-up
question
if
I
may,
since
since
distribution
is
a
de
facto
implementation
and
if
that
is
not
conformant
to
the
distribution
spec,
then
what
incentive
is
there
for
other
registry
providers
to
also
be
conforming.
H
Hey,
I
put
the
completion
of
the
proposal
for
the
extension
apa.
I
think.
Last
week,
mike
brown
had
support
kind
of
like
spoken
about
it.
Two
weeks
ago
and
last
week,
steve
steven
had
kind
of
like
spoken
about
the
reference
implementation
to
just
show
in
distribution
project.
Also,
is
there
anything
pending
from
the
merge
of
that
one,
or
do
we
need
to
do
something
more?
Just
kind
of
a
follow-up
would
be
really
helpful
if
at
all,
that's
111
for
reference.
I
D
Over
there
so,
but
I
am
a
reviewer
and
I
have
looked
at
it,
do
you
I
don't
know
if
I
can
get
anonymous
camera
off.
Is
that
what
you're?
Looking
for
so
jay.
H
Yeah
we
can
follow
through
with
the
distribution
pr.
If
the
proposal
for
respect
is
what
you
say
acceptable
in
some
draft
form,
that's
what
michael
brown
was
helping
with
like
saying
that
this
is
subject
to
change
and
has
no
has
no,
what
you
say:
support
that
could
possibly
come
through.
H
And
regarding
the
alignment
of
which
projects
would
help
it,
I
think,
if
there's
a
if
the
direction
is
positive
in
that
form,
then
we
can
submit
the
follow-up
implementation
of
how
references
can
be
stored
and
how
the
notary
project
can
actually
use
that
as
well.
So
that
will
be
the
alignment
for
the
project
question.
D
H
Yeah,
so
do
we
so
we
have,
I
think,
brandon
had
put
a
proposal
to
kind
of
like
change
the
output
of
the
extensions.
I
don't
have
any
strong,
yes
or
no
reasons
for
that.
So
we
can
take
that.
D
Oh
yeah,
so
so
brandon
brandon's
reason
is
very
good
and
I
I
just
you
know
I
linked
this
yeah,
okay,
that
one's
another
one
we
so
he's
saying
put
that
put
it
in
in
a
name
field
and
the
reason
to
do
that
is,
if
you
add
other
fields
or
change
it
in
the
future.
It's
a
lot
easier
to
add
fields
if
you
ever
have
an
object,
whereas
it's
just
strings,
it's
really
hard
to
extend
in
the
future.
So
that's.
A
Very
good
I'll
actually
try
and
get
the
recordings
up
sooner
than
later
see
that
the
one
from
last
week
hadn't
been
posted
yet
so
it
probably
is
even
useful
if
being
able
to
you
know,
have
referenceable
clips
a
point
in
time
when
we
talked
about
things.
So
thanks
everybody
for
joining
the
conversation,
even
if
it
is
you
know,
spilling
a
discussion
out
is
good
because
it
shows
at
least
symptoms
of
things
that
need
to
be
addressed.
So
there's
a
few
I
would.
A
I
would
encourage
everybody
to
like
actually
put
you
know
in
in
action
items
on
the
on
the
dock.
A
Put
your
name
down
on
specific
things
like,
I
think
it's
great
the
the
stats
there
and
I
think,
looks
like
nisha
had
to
drop,
but
those
those
kind
of
items
to
address
or
like
map
out
are
useful.
But
let's,
let's
actually.