►
From YouTube: Kubernetes Sig Docs 20171121
Description
Meeting notes: https://docs.google.com/document/d/1Ds87eRiNZeXwRBEbFr6Z7ukjbTow5RQcNZLaSvWWQsE/
The Kubernetes special interest group for documentation (SIG Docs) meets weekly to discuss improving Kubernetes documentation. This video is the meeting for 21 November 2017.
https://github.com/kubernetes/kubernetes.github.io
A
A
Let's,
let's
start
with
a
follow-up
on
versioning
for
documentation.
I
realized
that,
after
our
meeting
on
Halloween
that
I
completely
forgot
to
follow
up
on
November
7th
to
have
a
further
talk
about
versioning.
So
we
we
had
talked
on
October
31st
about
Steve,
putting
together
a
prototype
of
what
versioning
might
look
like.
I
know,
Steve
you've
been
really
busy
with
auto-generated
reference
Docs.
Do
you
want
to
do
want
to
say
a
little
bit
about
where
we're
at
with
versioning
and
about
what
are
you
still?
Are
you
still
looking
at
doing
a
prototype?
B
B
B
Yeah
so
say:
if
you
go
to
the
1.7
docks,
this
is
what
they
look
like.
If
every
page
has
this
big
warning
that
says
this
is
not
the
latest
docks
added,
it's
not
maintained
and
link
back
to
the
latest
set,
and
this
works
pretty
well.
If
you,
if
you
follow
links
within
this
dock,
set,
it
stays
in
the
in
the
1.7
set.
B
C
B
C
B
B
A
B
See
we
have
yeah
I
just
have
this
one
one
extra
sentence
this
this
is
part
of
them
include,
so
that
this
gets
included
and
it
says
different
things,
depending
on
your
page
version
variable,
so
we
at
least
tell
the
person
that
to
do
what
what
you're
about
to
see,
you
must
have
kubernetes
server
version.
One
point
eight
so
that
I
had
forgotten
about
that
was.
Was
that
the
discussion
you
were
thinking
of
when
you
started?
Okay,
that.
A
B
Someone
else
added
and
that
anyway,
so
we've
we've
done
a
little
bit
to
help
people
understand
I
thought
we
had
a
link,
maybe
back
to
the
one
point:
seven
Docs
or
something
like
that.
You
know
some
statement.
That
said,
if
you
don't
happen
to
have
one
point
eight
you
could
look
here,
but
maybe
we
haven't
gotten
to
that.
Yet,
okay,
like.
A
We
need
to
over
engineer
like
a
link.
A
link
back
would
be
cool,
but
at
this
point,
I
feel
like
I
agree
with
Andrea
I
think
the
the
code
common.
This
is
just
going
to
be
sufficient
there
and
Steve.
The
change
that
you
made
to
the
to
be
include
and
before
you
begin
I
think
that's
that's
awesome
and
that
will
work
so
yeah.
Then
we
just
need
to
make
sure
that
there
are
issues
open
for
support,
preserving
version
and
search
functionality,
specifically,
okay.
Well,
what
was
last
point
again?
A
B
D
Problematic,
because,
right
now,
we
specifically
having
in
the
headers,
like
the
note
in
the
note
indexing,
is
that,
like
these,
like
the
different
versions,
don't
get
put
in
the
general
search
results
right,
so
I'm
not
quite
sure
how
to
separate
them
out.
So
you
can
still
just
search
within
there,
but
they
don't
end
up
in
the
regular
school
yeah.
B
And
we
really
don't
want
them
in
the
general
search
index.
You
know
we
don't
want
people
just
out
on
the
internet
somewhere
searching
for
something
and
ending
up
in
the
1.7
dodge,
so
yeah
I,
don't
I
mean
yeah.
Let's
keep
thinking
about
it,
but
right
now,
I,
don't
know
how
we
would
do
it
now.
That's
totally
fair.
A
Yeah,
no,
that's
a
stumper
I
have
no
idea
either.
You
know
I
mean
it's
nothing
else.
We
can
I
mean
we
don't
have
to
open
a
PR
against
it,
but
you
can
at
least
open
an
issue
saying
we're
aware
of
that.
This
is
how
things
are.
It
looks
super
problematic,
we're
not
sure
it's
shrug
emoji
we're
not
sure
how
any
of
this
goes.
C
E
Just
looking
at
the
docs
project
that
Steve,
curry
and
github-
and
there
are
more
issues
in
there
so
I'm
just
you
know,
I
can
give
you
the
numbers.
If
you
want
against
the
versioning
problem,
I
guess
some
of
them
have
been
closed.
A
A
We
are
at
capacity
for
the
event,
which
is
awesome.
Let's
see
if
there
is
anyone
else
who
is
not
if
you
are
going
to.
If
you
are
interested
in
participating
but
you're
not
going
to
be
a
coop
con,
we
can
definitely
use
more
reviewers
for
PRS
that
that
sprint
participants
open
up
during
those
times.
So
if
you
are
going
to
be
on
line
between
what
is
it
between
1:00
and
5:00
Central
Time,
that
is
an
excellent
opportunity
to
participate
and
and
get
some
get
some
quality,
PRS
reviewed
and
pushed
against
the
documentation.
A
A
D
Exact,
so
for
that
I'll,
let
you
sent
out
like
I
noticed
you
talking
with
them
Jennifer
about
it,
and
we
know
you've
linked
the
user
journey
stuff
to
issue
and
I
thought
we
had
discussed,
not
making
it
so
broad,
like
indeed
giving
them
more
specific.
You
know
pieces
you
know
related
to
used
to
her
journeys
but
like
to
not
open
it
up
to
like
sort
of
like
the
whole
kind
of
framework,
but
just
like
give
them
specific
tests
within
that,
so
I
don't
want
to.
D
E
I
I
had
a
couple
of
thoughts
about
this
too,
because
I
saw
the
email
yesterday,
and
this
is
partly
Zak
I
feel
like
I
owe
you
an
apology.
It
occurred
to
me
after
Andrew
and
I
were
talking
about
this-
that
we
should
have
looked
you
in
too,
but
I
think
so.
One
of
the
things
that
I
think
might
be
useful
to
do.
I've
been
looking
again
at
Steve's
project,
and
you
know:
we've
got
that
column
in
it
now
so
many
get
up.
E
Debs
we've
got
that
columns
or
the
docs
print
and
I
was
wondering
whether
we
maybe
wanted
to
make
a
Doc's
print
project
and
sort
those
issues
out
a
little
more
clearly.
That
would
let
us
create
some
sprint
specific
issues:
around
user
journeys
we
could
even
make
a
user
journeys
column
and
then,
if
we
send
out
a
questionnaire,
we
can
send
it
out
with
an
email
saying
you
know
sort
of
please
go
here
and
take
a
look
at
this
link.
E
Instead,
you
know
we're
refining
the
work
that
we're
doing
you
know,
I,
don't
think
it
does
any
harm
for
people
to
look
at
the
larger
context
of
the
issue,
and
you
know
if
they
poke
around
some
of
the
discussions
about
user
journeys
anyway.
But
if
we
do
it
like
that,
then
we
can
invite
them
to
think
in
a
more
focused
way.
Without
saying
we're
sending
you
guys
mixed
messages
and
I
think
it
might
help
us
plan
the
work
a
little
more
easily
too
I'm.
E
That
is
I
should
at
least
provide
better
ideas
about
how
to
get
started
there
so
that
we
can
give
people
something
that
they
can
get
done
in
the
space
of
a
few
hours,
because
that's
definitely
going
to
be
something
that
we
need
to
iterate
through
we're,
not
gonna
get
it
all
done
on
the
first
pass,
no
matter
who
does
it
or
when
they
do
it
right
so
I
understand
you
know:
we've
tried
to
make
the
docs
print
stuff.
You
know
easily
bite
off
and
finished
chewy,
that's
of
things.
E
I've
got
a
vested
interest
in
specific,
potentially
bigger
things,
I'm
still
trying
to
break
them
down,
to
make
them
more
adjustable
and
I
will
stop
motoring
metaphor
now,
anyway.
Those
were
my
thoughts
about
how
to
both
how
to
organize
the
work
for
the
sprint
generally
and
how
to
communicate
that
out
ahead
of
time.
B
B
A
E
A
E
A
A
A
What
are
you
most
interested
in
pursuing
and
I
think
we
can
iterate
without
causing
without
causing
or
worsening
confusion,
as
the
case
can
be
so
I
tell
you
what
I
will
schedule
a
short
meeting
with
you,
Andrew
and
Jennifer,
and
if
anybody
else
is
interested
in
this
as
well
happy
to
invite
you
as
well
is
anybody
else
interested?
Does
anybody
else
want
to
come
talk
about
chunking
out
the
work
of
a
baduk,
sprint
yeah.
E
That
would
be
good
I
started,
able
form
for
a
questionnaire
just
based
on
the
injuries
of
my
conversation
last
week,
I'll
share
it
with
you
guys:
okay,
cool
I,.
A
Will
I
will
set
up
a
meeting
where
we
can
really
just
sync
up
about
all
of
this
and
have
a
have
a
have
a
more
coherent
plan?
I
guess
we're
where
we
are
all
aware
of
what
we
are
all
doing
and
since
we'll
suppose
we'll
be
facilitating
the
sessions
itself,
the
sessions
themselves,
you
get
a
chance
to
sync
up
about
what
that
might
actually
look
like
in
advance.
A
A
B
A
B
So
yeah
I
I've
been
looking
into
this
and
for
a
while
I
was
concentrated
on
just
the
working
on
the
code
that
generates
the
kubernetes
api
reference,
Docs
and
I've.
Now
I've
set
that
aside
to
look
at
the
bigger
picture
and
just
for
q4
at
least
for
what
I,
what
I
think
I
can
get
done.
It's
to
just
understand
what
all
the
different
generators
are
and
get
the
process
documented.
So
we
understand
it
and
anyone
else
who
wants
to
understand
it
has
a
place
to
to
read
about
it
for
us.
B
You
know
we
have
these
documents
that
we're
using
for
our
discussion.
Joe
has
this
document
and
I
see
this
as
the
them
where
we
have
our
main
discussion?
We
have
our
long-term
vision
and
we
just
keep
talking
about
things
in
this.
This
document,
the
one
after
that
that's
called
proposal
for
q1.
That
was
just
my
thoughts
about
what
I
think
we
can
realistically
get
done
in
q4
and
q1
and
I
I
think
what
that
is.
Is
we
can
see
what
all
the
existing
processes
are
understand?
B
How
to
do
them
ourselves,
get
that
documented,
more
and
more,
it
seems
like
people
feel,
there's
no
need
to
keep
any
of
that
a
secret
so
that
it
can
just
be
documented
to
the
general
public
and
our
kubernetes
dot
IO
site
that
if
anybody
still
thinks
that's,
not
a
good
idea
speak
up,
but
that
seems
to
be.
The
thinking
is
that
we
can
just
make
this
public
information
how
you
auto-generate
documents
come.
B
So
then
we
have,
we
probably
don't
need
to
click
into
each
of
these,
but
we
have
an
issue
where
we're
tracking
this
work
on
getting
all
auto-generated
reference
Docs
into
this
folder
called
docs
reference
generated
and
that
that's
going
well
I
think.
Well
what
I'd
like
to
take
a
look
at,
though,
is
this
this
work-in-progress
PR
for
documenting
the
generation
process.
If
everybody
can
see
what
I'm
in
the
middle
of
here.
B
So
if
any
of
you
want
to,
let
you
know
just
follow
along
with
what
what
I'm
doing
and
comment
that
would
be
nice.
If
we
go
to
the
the
back
on
the
meeting,
though
it's
the
link
right
below
this
that'll
take
us
to
one
of
the
pages.
That's
in
the
preview,
whether
the
I'm,
a
link,
Rick,
you
know
see
where
it
says:
work-in-progress
PR
for
documenting
the
generations.
B
So
what
I've
done
here
is
is
to
get
started.
I've
documented
everything
I
can
find
just
because
I
felt,
like
I
needed
to
understand
it.
I
found
this
kind
of
confusing
and
with
the
idea
that
we
might
not
keep
this
all
as
public
information,
but
right
now
this
is
just
the
place
where
I
I
figure.
We
can
all
see
what
what
we
have
and
talk
about
what
what
we
want
eventually.
So
there's
this.
This
script
called
gen,
cubed
access,
h
and
it
calls
for
different
golang
programs,
each
of
which
does
a
different
thing.
B
B
You
know
these
components
like
cube,
API,
server,
controller
management,
each
of
those
has
a
Microsoft
markdown
document
generated,
and
then
we
have
to
you,
know
manually
copy
those
over
to
our
kubernetes
website
a
repository.
But
what
I
didn't
realize
is
that
if
you
go
down,
we
generated
these
files
in
other
formats.
B
B
B
C
You
arey
2015
I
was
never
certain
how
they
got
absorbed
into
rpms
there
Debs
and
then
Bush
through
the
various
providers.
I
knew
that
they
were
there,
but
I
I
I
don't.
I
think
when
fabrizio
was
actually
asking
a
couple
weeks
ago.
Nobody
answered
the
question
of
how
they
go
from
in
the
documentation
repo
out
into
the
internet,
where
they're
clearly
being
found,
but
it
I
don't
know
what
the
process
is.
You.
B
Well,
yeah,
I
just
I
put
it
in
the
in
the
topic
for
now,
just
because
for
now,
I
want
this
topic
to
be
our
our
place
for
seeing
what
we
have
and
and
talking
about
what
what
we
want
to
do,
but
III
think
it's
bad
to
have
multiple
copies
of
a
forgiving
page
all
over
the
place,
because
it's
it's
just
harder
to
find
the
one
we
want.
You
know
for
search
to
find
the
one
we
want
them
to
find
absolutely
and
then
there
are
also
ya,
know
files.
B
So
if
we
go
back
to
that
the
topic
where
we
get
markdown
files,
we
get
man,
page
files
and
down
at
the
bottom
of
this.
We
also
get
llamó
files
and
I,
don't
know
if
these
are
published
anywhere
or
if
they're
used
for
anything
daniel
Smith
proposed
this
in
an
issue
some
time
ago.
He
said
it
would
be
helpful
to
have
the
docs,
not
only
in
markdown
and
man,
page
format,
but
also
yamo
format,
and
so
that
that
got
done.
But
I
don't
know
if
anybody
actually
uses
these.
B
B
The
only
thing
we
actually
publish
is
cube
control,
MD
and
you
can
probably,
if
you
see
we're
in
the
sentence
right
after
the
bulleted
list,
it
says
cube,
control,
dot.
Md
is
published
here.
Let's
click
that
so
this
of
all
that
this
is
the
only
one
that
we
actually
still
use.
The
rest
of
them
are
generated
by
a
different
tool
and
they're
published
as
a
single
giant
file
instead
of
a
bunch
of
individual
files
for
commands,
and
if
you
click
on
any
one
of
these
commands
will
see
that
act.
So.
A
B
Tool
that
has
all
the
cube
control
commands
so
part
of
what
I
found
confusing
is
that
you
know
we
have.
We
have
some
duplication
and
we
have
you
know
in
the
way
if
we
got
man
pages
yamo
files
and
markdown
files,
and
we
also
have
some
duplication
in
the
sense
that
we
have
two
different
ways
of
generating
cube
control,
Docs
and
only
one
of
them
is
published
on
our
site.
B
C
E
C
A
B
C
B
Yeah,
stop
sharing.
Okay,
so
back
in
the
meeting
notes.
Is
there
anything
did
it
direct?
Let
me
look
here
if
I
put
anything
about
there
about
cube
admin,
that's
another
story.
So,
let's
see
where
I
said
three
separate
areas
of
concern.
Okay,
so
one
is
the
API
ref
can
cube
control
command,
docks,
that's
filled
with
rocks
tool,
then
the
all
other
ref
docks
down
at
the
bottom
we've
just
been
talking
about
and
in
the
middle
we
have
cube
admin
ref
dachshund.
B
This
is
Fabrizio's
work.
There's
a
PR
and
I've
done
an
edit
pass
on
this.
So
that's
all
in
progress,
but
what
that
does
is
it?
There
are
a
bunch
of
automatically
generated
files
that
don't
get
published
as
those
files,
but
they
get
blended
in
as
includes
two
manually
authored
files.
So
he's
he's
like
greatly
expanding
the
cube,
control,
docks
and
I.
Think
it's
a
good
good
effort.
Okay,.
A
C
Quick
question
on
that
yeah
yeah,
just
one
little
tidbit
Fabrizio
has
asked:
should
he
change
where
he
drops
those
files,
cuz
I,
think
he
was
putting
him
in
a
specific
location
and
we've
decided.
Maybe
it
went,
we
wanted
under
edmonds
less
generated
or
something
if
we
want
him
to
update
that
now
is
the
time
to
let
him
know
yeah.
B
B
B
Okay,
so
if
we
go
down
a
little
bit,
we
see
all
these
underscore
so
see.
This
thing
called
underscore.
You
know,
cube
admin
dot.
You
know
they're
there,
a
zillion
of
those
those
are
auto-generated
and
they
get
they
have
to
live
somewhere
in
our
repository,
but
they
just
serve,
as
includes
into
the
like
three
or
four
manually
generated
files
and
I
can
see
that
he
hasn't
moved
them.
Yet
we've
we've
been
discussing
where
to
put
these
so
we're
gonna
put
them
under
Doc's
reference.
B
And
I
read
it
not
quite
because,
because
he
can't
the
fact
that
these
are
used
as
includes
meanings,
they
can't
be,
they
have
to
be
in
a
child's
directory.
So
anyway,
if
if
we
look
anybody
who's
interested,
can
look
through
the
comments
in
this
PR
and
we
were
having
this
discussion
about
where
to
put
these
things,
but
they
won't
be
an
admin
because
we're
trying
to
make
admin
go
away.
A
A
A
A
All
right,
Steve
I
would
love
to
pick
your
brain
about
about
what
to
do
with
all
of
this
for
one
night,
because
that's
coming
up
here
pretty
quickly:
okay,
okay!
Is
there
anything
else.
Paris
was
here
earlier
and
she
left
I
would
love
to
get
her
input
on
all
of
our
stuff,
but
is
there?
Is
there
anything
else
that
we
need
to
discuss
today?
Anything
else?
If
anybody
wants
to
talk
about.