►
From YouTube: CNCF Reference Architecture Meeting - 2018-08-08
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
A
Okay,
we
unfortunately
don't
have
ten
on
the
phone.
I
hope
you
can
join
us
soon
and
equally.
Unfortunately,
I
did
not
do
my
homework
over
the
last
month
and
published
the
revised
version
of
this
online,
but
I
did
want
to
just
walk
people
through
for
a
second,
where
we
had
rough
consensus
on
this
material
last
week
and
then
it
came
up
in
the
TSU
meeting.
Yesterday,
pasta
talked
about
potentially
doing
some
additional
work
in
his
space
and
how
we
might
allocate
that
out.
A
Yeah
I
just
defined
this
email
now
of
what
we
were
talking
about
last
time,
okay,
so
the
changes
that
I
am
good
at
made
as
a
really
secure
images
to
security
and
I.
Think
that's
going
to
be
a
major
topic
for
us
to
come
back
and
engage
on
based
on
but
be
back
from
the
from
the
TOC
and
then
the
three
categories
that
are
all
at
the
bottom
of
the
landscape,
which
have
been
a
little
arbitrary
in
the
past.
So
those
are
host
management,
tooling
in
destructor
automation
and
private
clouds.
A
A
Derek
collision
from
Nass
and
that's
the
question
of
do:
we
need
how
we
separate
out
naps
and
what's
called
screaming,
on
the
landscape
right
now
from
this
RPC
category,
and
he
made
the
proposal
that
we
remade
screaming
to
be
screaming
and
messaging
and
have
it
included
everything.
This
is
screaming
now,
plus
RPC
on
the
argument
that
they
that
distinction
between
these
just
is
never
clear,
and
if
you
heard
the
call
yesterday,
we
would
also
put
the
new
our
socket
in
there
as
well.
Okay.
A
A
A
B
Last
that
time,
I
know
that
I
kind
of
meeting
yesterday
in
the
CLC
meeting
there's
some
interest
in
us.
Looking
at
you
know,
a
couple
of
different
like
Caddy
was
a
canoe,
a
silver,
so
I
have
I
cold
security,
one
of
them.
The
other
idea
was
us.
Take
a
couple
of
the
CAF
project.
It's
kind
of
like
try
to
map
out
all
the
things
we
put
together.
So
I
think
it's
you
know
interesting,
but
at
the
same
time
we
can
kind
of
look
boo.
A
couple
of
different
options.
I
think.
A
Yeah
wondering
anybody
else
would
like
to
comment
on
the
to
see
Co
yesterday,
lonely
I,
don't
feeling
Alexis's
proposal
is
pretty
totally
clear
and
so
I
think
I
might
just
give
him
a
call
and
understand
what
he's
getting
at
I
will
point
out
to
you.
That's
the
interactive
landscape
right
now
can
show
you
just
the
CMC
I
projects
and
where
they
fit
into
the
landscape,
and
so
I
just
pasted
it
into
the
chat
window
of
view
of
that.
A
B
A
Did
that
live
on
the
interactive
last
year
in
the
next
day
or
so,
and
then
we
are
managed
to
publish
a
new,
fixed
landscape
satellite.
We
ask
you
next
week
with
this
and
you
know,
and
he
thinks
go.
We
often
just
have
a
ton
of
new
projects
where
I
think
we're
up
to
now
and
we're
gonna
have
about
15
more
projects.
Since
the
last
time
we
publish
yeah
quicken.
C
Yeah
yeah
I,
think
I
understood
at
Lexuses,
motivation
and
and
Ryan
Cantrell
echoed
some
of
those,
which
is
that
I
think
in
general.
It's
it's
difficult
to
navigate
this
stuff
and
one
solution.
One
potential
solution
and
I
don't
think
this
is
mandated.
But
but
one
way
of
doing
it
is
more
in
a
hierarchical
fashion.
Where
you
know
each
level
of
the
hierarchy,
you
know,
has
a
small
number
of
branches,
so
you
know:
we've
got
CN
CF
at
the
top
I
think
we
can
hopefully
agree
that
there
are,
you
know
somewhere
between
five
and
seven.
C
You
know
major
areas
under
that
and
then
within
each
of
those
there
are
yet
more.
You
know
more
specific
areas.
So
if
I
just
look
at
your
category
changes
there
Dan
all
those
kind
of
fit
into
one
bigger
category,
our
pcs
load,
balancers,
API
gateways,
service
measures,
they're
all
part
of
a
thing,
and
and
as
we
grow
the
number
you
know
it
looks
like
we
have
a
very
large
influx
of
new
projects
coming
into
the
CNC
F,
and
so,
as
these
things
grow,
it's
going
to
become
increasingly
difficult
to
to
kind
of
put
it
all.
C
Together
in
a
logical
mental
model,
so
the
sooner
we
we
sort
of
break
it
down
into
a
hierarchy
that
the
easier
it
will
be
to
incorporate
new
projects
into
that
hierarchy
rather
than
having
what
is
if
I
understand
correctly
Anna
I
haven't
been
at
all
in
development
of
what
we've
got
up
to
now.
So
please
correct
me:
if
my
perception
is
wrong,
but
we've
we've
got
a
reasonably
flat
structure.
Where
we
have
these
categories
and
each
category
we
have.
You
know
projects
and
those
are
either
CNCs
projects
or
nonslip
c
CN,
CF
projects
and
I.
A
C
A
A
You
had
observability
and
analysis
that
kind
of
went
across
all
of
those,
but
all
we're
talking
about
right
now
is
really
swapping
around
subcategories
inside
of
them,
but
I'm
wondering
if
maybe
what
Alison
and
Brian-
and
perhaps
you
are
looking
for-
is
more
trying
to
highlight
those
top-level
categories
or
focus
on
the
more.
C
A
C
Yeah
I
mean
it
does
become
problematic
as
to
how
to
represent
that
on
something
people
can
understand,
but
I.
Think
logically,
there
probably
are
subdivisions
in
some
of
these
areas.
Now
whether
we
want
to
expose
those
or
not
is
perhaps
a
different
question,
but
you
know
just
looking
at
a
few
of
them
cloud
native
storage.
You
know
there's
so
many
of
them
and
they're
not
all
the
same.
Some
of
them
sort
of
operate
at
different
levels
than
others,
and
the
same
can
probably
be
said
for
host
management
and
tooling.
A
What
we
were
trying
to
do
with
it
and
I
mean
I'm
so
varied
opportunity,
changes
which
is
hate
any
subcategory
that
seemed
too
large,
and
service
management
in
particular,
seems
like
a
grab
bag
or
teaching
utility
for
kind
of
subcategory
and
break
it
into
multiple
subcategories
and
I
would
love
to
do
the
same
for
storage,
but
what
the
two
sub
networks
would
be
didn't
jump
out
at
me.
I
want
to
fancy
that
we
don't
want
to
do
open
source
versus
closed
source
because
that's
represented
by
the
gray
background
versus
transparent
backgrounds,
yeah
and.
C
A
We're
looking
at
some
more
architectural
that
but
storage
I
mean
the
other
one
that
jumped
out
was
understand,
database
and
data
warehouse,
which
is
now
the
biggest
category,
but
this
sort
of
obvious
ones
that
came
to
mind
way
over
snows
people
aren't
actually
quarry
categories
anymore.
There's
a
lot
of
things,
sort
of
spinning
certain
videos.
Oh.
C
Yeah,
yeah
and
and
I
absolutely
recognize
the
challenges
and
and
I,
don't
think
anyone's
got
a
you
know
very
good
solution
to
the
problem.
The
problem
is,
this
isn't
overwhelming
and
we
almost
need
a
step-by-step
way
and
then
we
have
a
two-step
mechanism.
So
you
you
know
if
you're
looking
for
some
particular
technology.
Oh
you
have
a
vague
idea
of
way.
You
know
problem
space.
This
is
you.
You
didn't
mean
on
that
layer,
you.
What
is
it
seven
of
them
to
choose
from
at
or
eight
to
choose
from
at
the
moment?
C
C
Was
gonna
say
and
and
and
if
we
look
not
too
far
into
the
future,
I
presume
it's
not
inconceivable
that
we
we
have
a
hundred
CN
CF
projects,
and
maybe
you
know
two
or
three
times
as
many
non
CN
CF
projects
as
we
have
today
and
then
you
know
this
becomes
even
more
overwhelming
and
even
less
navigable.
So
the
question
is:
how
do
we?
How
do
we
resolve
that
problem?.
A
Okay,
but
you
wanna
reiterate
the
primary
document,
what
we
think
of
it:
the
front
side,
darkness
other
thing
that
I
just
pasted
in
of
the
trail
map,
and
so
that
is
trying
to
tell
a
story
that
there's
an
order
in
which
you
want
to
consider
this
technologies,
and
we
particularly
want
you
just
to
think
about
the
CMC
at
once.
As
you
go
through
that
process,.
A
D
A
Mean
on
the
on
the
trail
map
I
mean
on
the
landscape
itself.
You
know
it
is
simplest
level.
If
we
just
want
to
get
a
bunch
of
icons
off
here,
we
could
move
the
platforms
and
the
at
the
bottom
right
to
keep
it
kubernetes
certified
service
providers
and
it
can
raise
training
partners
and
make
that
their
own
sheet.
My
hesitation
on
doing
that.
It's
just
that
these
are
a
lot
of
the
biggest
supporters
of
Sancia
and
so
I
sort
of
hate
to
kick
them
off
when
I
feel
like
they.
A
That's
actually
doing
a
ton
of
work.
You
know,
but
I
guess
I
felt,
like
Alexis,
had
a
some
other
idea
that
you
sort
of
been
American
double
click
into
security
and
and
get
like
a
full
wave
of
that
and
I
guess.
I'm
not
I
mean
I.
Of
course
the
interactive
landscape
has
way
more
information
on
these
projects.
So
it's
very
good
for
doing
pollen
research,
and
it
also
has
the
less
than
tweets,
which
has
to
be
extremely
helpful,
but
I'm
still
a
little
unclear
on
what
that
for
security.
How
bout
with
look
like.
C
C
So
I
guess
one
of
just
we
have
here
is
that
that
there
are
so
many
different
angles
that
you
can
look
at
this
from
you
know.
Are
you
do
you
care
about
seeing
all
the
container
registries
together
or
do
you
care
about
it,
I'll,
seeing
all
the
security,
its
and
stuff
together,
which
may
incorporate
secure
container
registry,
secure
networks,
authentication,
etc?.
C
And
I
I,
don't
I,
don't
claim
to
know
the
answer
to
the
question.
I
just
know
there
is
a
common
request
for
making
this
stuff
more
navigable,
and
maybe
the
you
know
the
and
I
think
this
is
a
topic
Brian
Cantrell
brought
up
is
if
we
can
just
verify
just
be
they're
very
clear
on
what
the
goal
is.
The
goal
is
to
make
consumers
of
the
CN
CF
technology
able
to
come
in
and
understand,
what's
going
on
as
quickly
and
easily
as
possible,
and
that
may
require
in
a
multiple
views
of
the
problem.
A
E
A
Interactive
landscape,
for
me,
always
has
a
different
goal,
which
is
that
it's
not
about
yes,
yes,
it's
saying:
hey,
there's
so
many
projects,
and
so
many
products
in
this
space.
How
do
they
play
out
and
when
I
do
decide
that
I
want
to?
You
know,
make
it
invested
in
a
team
and
commit
or
storage
or
anything
else.
So
what
are
the
options
that
I
could
be
considered?
C
B
A
In
that
way,
I'm
just
I'm
just
trying
to
do
something
rigorous
and
kind
of
maintainable
longers
lives,
but
I
mean
it's
definitely
the
keep
now
that
you
know
just
trying
to
I
mean
one
of
the
most
common
things.
I
guess
for
people
coming
to
it
is:
oh,
we
shouldn't
have
to
only
be
in
one
category.
Our
project
for
our
official
product
works
in
multiple
categories.
E
A
F
Don't
really
know
concretely
what
he's
trying
to
accomplish
I
know
he
just
wants
to
simplify
the
landscape,
so
you
could
basically
take
like
a
category
like
security
and
only
see
those
those
projects
that
are
kind
of
associated
with
that
he's
going
for
a
more
simple,
simple
view.
He
just
doesn't
want
to
be
presented
with
here.
300
things.
A
F
Mean
I'm
committed
to
whatever
solution
we
come
up
with.
Won't,
please
everyone,
so
I,
don't
I
think
it's
impossible
to
do
that.
So
I
think
we
just
need
to
decide.
As
a
group,
you
know
an
improvement
based
on
currently
what
we
have
and
just
make
a
decision
and
move
forward.
We
could
bike
shed
on
this
forever.
To
be
honest,
yeah.
C
But
it
is
worth
making
sure
that
we
understand
what
the
problem
is,
that
we're
trying
to
solve
and
and
maybe
bike
shed
a
couple
of
alternative
approaches
to
solving
it
and
then
pick
the
least
bad
of
them
and
and
drive
it
to
to
an
actual
complete
thing.
Rather
than
just
being
an
idea
that
that
hasn't
been
worked
into
a
proper
plan.
A
A
Oh
well,
let's
put
an
extra
piece
of
metadata
on
it
that
on
each
of
these
cards
and
it
could
be,
is
it
security,
related
or
I?
Don't
know?
Is
it
observation
related
or
maybe
something
else
and
then
allow
people
to
only
to
zoom
in
or
to
filter,
so
that
they're
only
seeing
the
items?
That
is
true
for
that
metadata.
C
Yes,
I
think
I,
think
that
is
one
of
the
tentative
proposals
and
and
with
the
assumption
that
each
one
of
these
tiles
items
products
projects
can
can
have
multiple
of
those
labels
and
and
maybe
that
even
the
high
right
yeah.
So
maybe
maybe
you
zooming
to
security
first
and
you
get
all
the
all
the
things
with
label.
You
know
security,
slash
star
and
then
once
you
look
at
that
you,
you
then
have
more
subcategories,
which
is
key
management
or
this
or
that
or
the
next
thing.
C
A
Well,
let's
agree
upfront
that
there
is.
There
is
a
single
piece
of
metadata
right
now,
there's
at
least
two
pieces
of
metadata
associated
with
it.
So
there's
a
category,
slash
sub
category
and
you
exclusively
can
choose
on
the
interactive
landscape
to
just
show
either
one
but
either
the
whole
category
or
subcategory.
A
C
B
Know
I
think
the
only
thing
I
can
think
of
Dan
is
maybe
some
of
the
application
and
the
CITV
things
we
should
put
off
in
a
separate
area
because
they're
not
anything
about
what
you
know
I
mean
it
was
about
they're,
not
early
technologies,
that
we're
going
to
like
bring
into
a
CNC
FM.
In
most
cases,
right
I.
B
Decide
for
the
T
or
C,
but
it
seems
like
at
least
they
could,
but
I
am
bending
it
more
like.
No
category,
though,
seems
like
like
spinning
on
his
name.
Like
a
you
know,
the
kubernetes
version
is
more
of
a
specific
technology
version,
which
makes
sense,
but
the
entire
category
me
me
doesn't
make
sense.
You
know
oh
yeah,
we
have
one
landscape.
That
kind
of
has
everything.
Then
we
have
another
landscape
and
it's
like
CMT.
Yet
projects.
E
E
B
E
A
Okay,
so
I
mean
what's
hard
from
here,
is
if
we
just
come
up
with
new
categories
and
sub-categories
that's
like
super
easy
for
me
to
do
and
fix
and
such
but
I,
don't
I
haven't
quite
heard
a
proposal
yet
for
a
new
set
of
labels.
If
we
call
it
that
that
could
apply
to
essentially
all
the
items
on
the
landscape
right
now,.
A
Taylor,
do
you
have
any
view
pointer?
Would
you
like
to
jump
in
on
owner
or
anyone
else
on
the
call.
E
A
E
All
right,
hi
Dan
now
Michael
P,
is
Michael
Payne
here
from
JPMorgan
Chase,
so
I'm
also
just
listening
in
and
understand
their
concerns
about
the
complexity
and
I
landscape.
But
you
know
I
think
that
the
discussion
so
far
sort
of
makes
sense.
I
I
agree
with
Quinton,
though
I
think
that
you
know
trying
to
understand
exactly
what
the
purpose
is
can
help
with
the
directions,
but
could
tell
us
along
look
yeah.
A
Along
those
lines,
Michael
I
mean,
let
me
read
you
the
phrasing
from
the
very
bottom
of
the
static
landscape
that
we
say
this
landscape
is
intended
as
a
map
through
the
previously
uncharted
terrain
of
cloud
native
native
technologies.
There
are
many
routes
to
deploying
a
cloud
native
application
with
CNCs
projects
representing
a
particularly
well-traveled
path
so
and
so
yeah.
We
were
exclusively
trying
to
say
that
there's
not
only
one
way
of
doing
it,
I
mean
I.
A
Think
it's
generally
the
case
that
most
organizations
will
only
pick
one
or
two
options
from
each
of
these,
so
I
mean
I,
think
we
can
say
aspirationally
the
point
of
our
of
a
taxonomy.
Is
that
you
know
you're,
probably
not
you
probably
don't
want
to
run
nomads
and
kubernetes
and
mazes
and
swarm
all
in
the
same
organization.
A
Now
you
know
you
can
say
well
if
it's
a
big
enough
organization,
it
probably
is
but
aspirationally
you'd
probably
like
to
pick
one
of
those
or
at
most
and
and
selected
and
invested
that
and
so
again,
aspirationally
I
think
we
had
a
similar
thought
for
all
of
our
other
categories.
Here
that
you
know
you
can
do
multiple
container
registries,
but
you
probably
don't
want
to.
E
Sure
that
makes
sense,
look
I,
think
that
I
do
find
a
landscape
useful
and
we've
certainly
shared
it
at
work,
and
you
know:
pay
attention
to
sort
of
categories
like
the
service,
one,
for
example,
but
really
it's
sort
of
I,
think
it's
useful
at
a
category
level
for
us
or
for
me
at
least
because
the
the
30,000
foot
view
of
everything
is
certainly
daunting,
but
no
I
understand
the
idea
there
and
and
I
do
think,
there's
a
value
in
in
making
this
more
inclusive
than
just
this
ncf
stuff.
A
Yeah
the
solution
is
to
have
that
be
the
default
view,
but
it
would
sort
of
hide
everything
else.
Was
there
and
available
for
it?
I
guess
I
would
be
curious
if
anybody
maybe
could
just
speak
for
a
second
about
how
they've
used
the
landscape
I
mean
if
we
say
look,
the
trail
map
is
fine
for
what
it
is,
we're
not
really
interested
in
iterating
on
that
right
now,
but
the
landscape
itself,
I
mean
I'll
just
say
first.
Obviously
this
started
out
as
a
much
much
smaller
project.
A
The
the
interactive
landscape
in
particular,
was
you
know
very
much
the
tool
that
I
wanted
to
use,
just
as
I,
was
playing
around
and
learning
the
stuff
and
investigating.
But
of
course,
the
tool
that's
ideal
for
me
is
not
necessarily
the
one
that
makes
sense
for
the
community,
so
I
guess
I
I'd
love
to
hear
other
people's
perspectives
on
how
you
are
using
it
and
then
related
to
that
and
to
Quentin's
question
just
how
is
it
falling
down?
A
I
mean,
besides
that
basic
overwhelming
question:
how
else
is
it
not
doing
good
enough
job
yeah.
D
Hi
Dan,
this
is
dims,
so
one
thing
that
helps
me
is
like
trying
to
figure
out:
okay,
what
are
the
different
container
runtimes
that
are
there
right
and
what
are
the
options?
So
I
am
able
to
use
the
map
to
say
that
okay
rocket
run,
we
CRI,
you
are
kind
of
the
same
similar
things
right,
so
it
helps
from
that
regard.
But
then
what
doesn't
help
is
what
do
I
use
these
along
with
right.
I
can
use
these
with
kubernetes,
but
not
me
source
right.
D
So
if
we
can
like,
if
I
click
on
rocket,
for
example
right
if
we
can
highlight
some
of
the
things
that
rocket
works
with,
maybe
that
will
be
useful,
I
think
so,
and
I
also
like
the
tag
hierarchy.
That
Quinton
was
mentioning.
That
kind
of
like
tells
people
that
you
can
tag
yourself
each
project
when,
when
they
people
are
proposing
projects
they
can
select,
which
set
of
tags
that
they
want
to
apply
to
their
own
project.
So
it
helps
in
like
showing
reduced
views.
D
C
A
A
C
It's
gonna
say:
I
mean
we
have
everyone
here
and
it
seems
like
we're
having
a
useful
productive
talk
as
long
as
everyone
can
tolerate
my
sort
of
unstructured
rambling
I'm
happy
to
talk
a
little
bit
more.
So
so
in
that
one
idea
that
a
couple
of
ideas
just
came
to
mind.
While
we
were
having
this
discussion,
the
first
one
is
I.
Think
this
landscape
diagram
is
very
useful,
but
it's
too
much
for
people
to
take
in
and
it
needs
to
be
interactive.
Then
I
think
the
interactive
landscape.
C
Luck
lacks
the
structuring
that
the
non
interactive
one
has
so
and
I
haven't
used
it
extensively.
But
the
impression
I
got
that
it
was
that
it
kind
of
felt
more
like
a
database
in
the
sense
that
you
know,
I
could
specify
a
couple
of
filter
criteria
and
then
I
would
only
get
those
things
that
fulfill
that
instead
of
criteria
but
I
lose
the
mental
structure
that
I
have
in
the
non
interactive
landscape.
So
I
don't
know
if
that's
true
and
they.
A
A
C
C
A
C
Yeah
now
what
I'm
referring
to
is,
if
I
look
at
the
non
static
landscape,
but
this
may
be
going
into
too
much
detail.
So
at
the
bottom
of
this,
of
the
stack
I
have
a
cloud
but
that
I
have
provisioning
than
I,
have
runtime
or
castration
and
application
definition,
and
that's
a
logical
structure
and
when
I
do
what
I
just
did
on
the
internet,
landscape
and
I
choose
group
by
category
the
category.
A
E
A
C
C
And
within
that,
and
so
I've
lost
the
cloud
category
actually
I've
got
a
cloud
public
category
and
under
that
a
cloud
private
category
and
then
at
the
same
regulan
hierarchy.
I've
got
provisioning
non-interactive
when
I
have
a
cloud
category
and
subcategories
within
that
of
private
and
public.
So
I
mean
these.
Are
these
are
silly
little
usability
into
you
know
tribes,
but
but
this
is
exactly
why
people
are
finding
the
stuff
difficult
to
use.
Oh.
A
And
I
and
I'm
very
willing
to
change
it.
It's
just
I.
My
hesitancy
on
adding
categories
versus
subcategories
here
is
that
I
was
worried.
It
would
be
more
confusing
to
deal
with
it,
but
I
do
want
to
just
point
out
to
you
that
if
you
scroll
down,
you
are
seeing
hope
Rory's
at
a
time
from
the
bottom
and
of
course
it
would
also
be
easy
to
change
it.
A
Yes,
so
I
guess
my
I
think
my
asked
of
the
people
on
the
call
would
be
to
play
around
this
this
a
little
bit
more
and
then
throw
some
ideas
onto
the
mailing
list.
I
mean
I.
If
you
feel
like
a
two-level
visual
is
going
to
make
a
make
it
clearer
for
people
I'm
very
open
to
doing
that.
The
other
option
would
be
you
know,
that's
an
overwhelming
way
of
looking
at
it.
Michael.
E
A
A
Just
want
to
give
a
quick
shout
out
to
what
a
big
believer
I
am
in
SVG's,
where
we
actually
have
high-res
versions
of
every
single
one
of
these
logos.
So
you
can
zoom
out
this
far
and
it
looks
good
and
also
if
you
zoom
it
in
hugely
each
of
the
logos
looks
nice,
but
no,
please
go
ahead.
I
think
you
were
going
to
the
more
critical
point.
Yeah.
E
I
also
believe
I
agree.
The
SVG's
are
the
best
way
to
do
this,
but
I
guess
I
may
be.
The
default
is
something
to
think
about
because
I
grows
you
down,
like
you,
are
obviously
very
familiar
with
navigating
the
interactive
landscape
and
I
sort
of
feel
like
I,
somewhat
M
as
well,
but
I
guess
you
know
defaults
matter
and
I.
Just
wonder.
Is
there
a
different
default?
E
We
can
select
that
would
sort
of
make
make
this
more
intuitive
on
how
to
navigate
or
maybe
we
need
to
spell
out
how
to
navigate
I,
don't
know,
but
just
a
thought,
because
obviously
you
wouldn't
see
this.
What
I'm
sharing
you
now,
because
it's
this
zoomed
out
view,
but
it
is
overwhelming,
even
if
it's
off
the
screen
so.
E
I
also
like
the
idea
of
of
tags
like
I'm,
a
long
time,
user
of
pin,
board
and
stuff
like
that
and
I
know,
a
lot
of
the
people
on
this
list
want
to
be
in
multiple
categories
and
and
tag.
Tagging
could
be
a
way
of
doing
that
and
then
allow
a
waiter
to
sort
of
view
by
tag
how
to
do
that
in
concrete
practices.
Different
story,
but
I
do
like
the
idea
of
multiple
tags
potentially
to
address
that
complaint.
A
D
E
C
C
Okay,
that's
the
one
I'm,
the
one,
that's
his
current
version
there.
So
so
that
seems
like
very
useful.
It
is
one
particular
structuring
of
these
things,
which
is
basically
architectural
with
you
know,
clouds
at
the
bottom
and
applications
that
we'll
talk
right
and
I
think
that's
very
useful.
So
the
two
objections
I
can
think
of
off
the
top
of
my
head,
but
I've
heard
or
one
is
there's
too
much
stuff
on
there
and
two
that
isn't
the
view
that
I
want
I
want
the
view
that
says
security
versus
this
versus
that
versus
the
other
thing.
C
C
Oh,
these
are
the
similar
products,
and
this
is
how
my
one
that
I'm
about
to
buy,
compares
with
the
other
ones,
something
like
that
and
and
that
you
know
intuitively
to
me,
I,
just
like
going
to
click
on
container
runtime,
section
and
I
get
that
information
that
seems
to
me
like
it
would
address
most
of
the
concerns
I've
heard
from
Michael
from
Chris
from
Alex
and
others
and
I'm.
Not.
Let
me
see
if
I
can
turn.
A
That
into
something
a
kind
of
concrete
solution,
but
then
we're
at
the
end
of
our
call,
which
is
that
the
interactive
landscape
is
an
open
source
project
that
you
know
has
this
y
llamo
file
and
anyone
can
edit
it
and
play
with
it
and
generated,
and
such
the
static
landscape
is
done
by
our
grass,
our
in-house
graphic
designer
Alex
Contini,
who
copies
and
pastes
all
of
these
images
and
then
resizes
them
and
puts
them
on
a
little
grey
background
and
such
and
spends
tons
of
time
getting
everything
to
fit
in
and
having
the
data
match.
A
We
have
had
the
aspiration
in
the
past
of
automatically
generating
this
diagram.
Now,
obviously,
we
have
all
the
data
for
it,
but
it's
just
tricky
to
dynamically
render
this,
but
my
contractor
in
in
Moscow
who's
worked
with
me
on
this
whole
project
is
really
ready
to
go,
and
so
the
sort
of
natural
follow-up
here
would
be
a
automatically
generate
this
version
of
the
diagram
and
then
be
to
bring
back
that
filter
box
on
the
left
and
start.
A
Allow
you
to
start
filtering
it
by
things
like
CNCs
projects,
open
source,
locate,
you
know
Europe
or
China
or
other
kinds
of
things
and
allow
you
to
see
different
perspectives
on
the
landscape,
or
it's
really
different
filters
on
the
landscape.
But
maybe
we
could
think
about
different
perspectives
as
a
second
step.
I
think.
A
Okay,
I
I
need
more
concrete
proposals
on
the
mailing
list,
then
of
at
least
two
meaningful
perspectives
that
are
different.
Then
I'm
not
arguing
with
the
concept
in
principle.
I
mean
I,
understand
in
principle.
You
can
always
blow
things
inside
out
and
in
such
but
I'm.
Looking
for
semi
practical
things
that
people
would
actually
want
to
do.
A
C
E
E
Yeah,
yes,
sure
thing:
the
categories:
I
guess
the
difference
being
that
over
here,
as
as
Quinton
pointed
out,
you've
got
like
the
clown
at
the
bottom
sort
of
moving
up
the
stack
to
to
rain
which
which
is
not
represented
here.
But
what
you
saying
there
that
would
be
hard
to
do
automatically,
but
certainly
something
to
think
about.
A
E
Now
I
think
it's
great
I
think
something
we
can
definitely
improve
upon.
I
just
think
you
know,
I'll
definitely
give
some
thoughts
to
to
defaults
and
yes,
some
other
improvements,
but
it's
good.
As
I
said:
we've
shared
it
occasionally
etre
yeah
at
JPMorgan.
Secondly,
with
interest
in
the
similar
stuff,
so
there
are
people
with
you
know
within
you
know
our
organization.
A
A
Them
up
again
in
the
future,
but
when
I
work
with
Derek
on
it,
we
couldn't
come
up
with
a
bright-line
rule
but
separated
and
that's
92
kinds
of
projects.
So
as
long
as
we
understand
that
this
next
generation
will
be
different,
but
is
not
by
any
means
the
final
version
of
it
right,
we're
gonna
that
would
just
keep.
A
If
this
would
be
useful
for
this
reason
and
not
having
to
fill
in
every
detail
about
it,
but
just
it
once
I
think
we
have
some
concrete
examples
in
mind,
then,
hopefully
the
labels
for
the
metadata
that
would
implement
it
will
pop
out
from
it.
So,
okay
thanks
everyone
for
the
call,
so
we'll
do
this
again
to
month,
but
hopefully
we'll
have
much
more
activity
on
the
mailing
list.
In
the
meantime,
yeah.