►
From YouTube: Octant Community Meeting - November 18th, 2020
Description
Octant community meeting is held weekly. We discuss and talk about the current state and future of Octant, demo upcoming features and releases, and preview new ideas we are considering for Octant.
Meeting agenda: https://hackmd.io/CzaPxtmXT_SW8nEpdwvGzw?view
A
B
A
All
right
thanks
everyone
for
joining
the
november
18th
octant
community
meeting
today,
we've
got
some
fun,
some
fun
stuff.
So
let
me
just
get
the
notes
shared
and
we
will
get
started.
Sorry
for
the
late
start.
We
we
had
a
little
bit
of
technical
difficulty.
A
Well,
we
I
had
some
technical
difficulty
with
the
with
the
stream
getting
started
so
got
it
got
it
sorted
out
now,
meeting
notes.
Where
are
they.
A
A
Zoom
in
a
little,
and
that
looks
big
enough,
people
can
see
that
okay.
A
All
right
so,
let's
see
first
things.
Oh
sixteen
two
we
released
it
I
put
in
I
yeah.
I
filled
this
out
on.
I
filled
this
out
on
monday
and
put
in
there.
It
was
released
and
then
today
I
went
out
of
the
maybe,
but
I
think
it
is
released.
Well
is
it?
Is
it
published.
A
Yeah,
so
I
think
that's
where
we're
I'm
not
gonna.
Do
it
right
now
I'll
do
it
after
the
the
community
meeting's
over
but
yeah,
we
we
have
a
draft
published.
It
is
here.
It's
it's
ready
to
go.
We
will
be
publishing
this
release
right
after
the
meeting,
and
it
contains
some
really
great
fixes,
a
host
of
just
like
usability
things
and
some
fixes
for
the
form
groups
for
plugin
authors.
A
What
else
there's
some
yeah
there's
some
things
that
we've
been
talking
about
it
for
a
while,
so
it
fixes
that
npm
registry
issue
that
made
it
its
way
into
master
at
some
point
with
the
artifactory
builds.
So
this
this
would
be
a
good
one.
This
this
will
fix
some
homebrew
recipes
and
and
installations
that
are
building
from
source,
as
well
as
get
some
people
making
progress
towards
on
plugins,
and
things
like
that.
So
that's
coming.
A
Thank
you,
sam
and
milan,
for
kind
of
dealing
with
all
of
the
cherry
picking
of
of
open
issues
and
and
getting
things
sorted
out
for
the
release
and
pulling
the
changelog
together
and
stuff
like
that,
I
really
really
appreciate
it.
So
yeah,
that's
that's
it!
There
anything
I
missed
on
that
one.
A
No
cool,
I
thought
pio
is
here,
I
swear
I
saw
pia
earlier
and
then
she
left
is
she
not
here?
I
think
she's
still
here,
I'm
here,
oh
great,
fantastic.
Okay,
so
I
just
wanted
to
introduce
you.
I
think
this
is
the
first.
Is
this
your
first
community
or
your
second
one?
I
don't
remember
this.
D
A
D
A
Great
yeah,
that's
what
I
was
saying.
This
is
the
first
one,
and
so
everyone
pia
is
the
new
product
manager
for
octant.
A
The
the
the
program
that
octant
open
source
octant
is
a
part
of
within
vmware
has
been
kind
of
restructured
a
little
bit,
and
so
now
we
we
got
access
to
esho
a
while
back,
which
has
been
super
great,
and
now
we
also
have
access
to
a
product
manager,
which
is
pretty
amazing,
so
pia
will
be
helping
with
like
issue
management
and
prioritization,
as
well
as
doing
some
community
management
stuff
kind
of
helping
the
rest
of
the
team
out.
A
So
we
don't
have
to
do
that
all
by
ourselves
and
so
pia
go
ahead
and
say
hi
and
add
anything
you'd
like
to
add
that
I
forgot.
D
Sure,
yes,
I'm
excited
to
be
working
with
the
team.
Before
this
I
was
initially
at
pivotal
and
I
came
in
through
the
acquisition
so
excited
to
continue
supporting
the
open
source
community
now
at
vmware.
A
Awesome
yeah
we're
looking
forward
to
it,
so
you'll
be
seeing
her
around
so
yeah
and
I
believe
she'll
she'll,
probably
start
popping
up
in
github
issues
and
and
the
suck
channel.
So
if
you
have
questions
and
around
well
really
anything,
you
can
ask
her
as
well.
She
she'll
be
there
so.
A
Yeah,
so
I
put
this
in
there's
like
new
designer
new
product
manager.
Now
what
or
what
now
so
this
is.
This
is
what
now
we
like.
A
We
have
some
really
cool
work
that
milan
and
esha
have
been
doing
that
they're
gonna,
share
out
and
but
part
of
part
of
that
new,
getting
a
new
designer
and
getting
a
new
product
manager
is
also
working
to
getting
some
more
dedicated
engineering
time
for
octant
and
that's
going
to
be
coming
in
kind
of
two
two
there'll
be
kind
of
two
pieces
to
that
there'll
be
just
some
like
carry-over
work.
Some
of
the
the
internal
work
that
our
program
does
around
developer.
A
Enablement
octant
will
get
some
shared
resources
out
of
that
people,
not
resources
we'll
get
some
shared
engineering
effort
from
people
out
of
that,
and
then
we
may
be
able
to
add
another
kind
of
dedicated
engineering
person
to
the
acting
team.
We're
still
trying
to
figure
that
out,
but
I
think,
would
be
really
great
for
us.
We've
we've
been
a
little
short-handed
recently
with,
like
brian
transitioning
over
to
do
other
work
within
vmware,
and
so
we've
lost
some
of
his.
A
You
know
contributions,
and
so
now
it's
just
me,
sam
and
milan,
on
the
engineering
side.
So
hopefully
we
can
get
some
more
help
there
and
I
and
that'll
help
the
community
at
large
and
kind
of
drive
us
towards
that.
1.0
we've
been
talking
about
which
you
can
learn
more
about
in
our
roadmap.
A
E
Sure
I
can
start
by
sharing
my
screen
so
give
me
one
second.
E
So
before
we
dig
into
this,
I
just
want
to
preface
that
this
exploit.
Can
everyone
see
my
screen
yeah
good?
Okay?
I
just
want
to
preface
that
this
is
only
a
first
pass
of
like
what
the
resource
sphere
might
look
like,
so
these
designs
are
not
set
in
stone.
It's
more
of
an
exploration
to
understand
what
are
the
open
questions
that
we
have,
and
we
definitely
want
to
get
some
feedback
on
this
so
I'll
be
going
through
like
the
designs.
E
What
are
some
of
the
open
questions
that
we
have
and
if
you
have
feedback,
definitely
leave
feedback
in
the
octane
channel
I'll
leave
it
here.
You
can
ping
me
I'm
in
the
kubernetes
channel,
so
right
now,
so
we're
looking
to
redesign
the
resource
viewer.
If
anything,
we
think
it
needs
some
visual
polish.
E
Now
we
want
to
solve
the
problem
of
the
resource
viewer,
getting
really
really
wide
and
having
these
connections
that
are
kind
of
hard
to
see,
because
you
can't
really
pan
in
or
out-
and
it's
just
visually
hard
to
comprehend
when
you,
when
it
gets
really
large
and
also
we
want
to
explore
what
what
might
it
look
like
to
pan
in
and
out
of
the
resource
viewer
like
have
a
really
high
level
overview
and
then
be
able
to
drill
down
deeper
into
like
the
specific
information
that
you're
looking
for
and
and
so
milan
had
actually
started.
E
This
exploration
with
scott,
who
was
another
designer
on
the
team,
and
this
is
the
direction
they
were
going
in,
where
you
can
see
the
overall
deployment
you
can
see
the
replica
set
and
the
pods
and
the
adjacent
services
and
resources.
But
the
thing
we
want
to
think
about
is
scale.
So
if
there's
multiple
replica
sets
multiple
pods,
they
might
not
be
able
to
be
represented
in
this
model
over
here,
and
so
we
did
a
little
bit
more
exploration
about
how
we
might
format
that
and
and
thinking
about.
E
But
we
also
looked
at
what,
if
we
were
able
to
keep
that
hierarchy
model
that
the
resource
viewer
has
now
but
still
be
able
to
make
it
easier
to
navigate
to
some
of
those
relationships
and
view
what
is
errored
out
and
go
deeper
for
troubleshooting
and,
like
I
mentioned
before,
we
also
wanted
to
think
about
panning
in
or
out
so,
for
example,
say
this
is
your
view
of
the
resource
viewer.
Maybe
it
starts
with
this.
E
Maybe
it
starts
with
the
statuses
and
the
coloring
that
tells
you
whether
or
not
there's
an
error,
and
you
can
pan
in
and
see
more
details
about
the
replica
again
there's
the
issue
of
scale
with
pods
so
right
now.
This
is
how
you
can
view
the
pods.
But
what?
If
there
was
more
information
explicitly
about
the
pods
or
what,
if
there's
an
easier
way
to
drill
down,
so
I
looked
at
what,
if
you
could
go
into
the
deployment
replica
set,
see
how
many
pods
there
are
and
drill
down
from
there.
E
So
one
way
of
doing
this
might
be
repurposing
the
sidebar,
so
you
can
see
the
replica
set,
you
can
click
it
and
then
you
can
see
the
pods
which
ones
are
aired
out.
Maybe
there's
a
high
level
now
about
what's
going
on
with
the
pod
and
you
can
click
it
and
go
to
the
specific
page
for
that
pod
or
maybe
that
information
is
accessed
through
this
fly
out
where
all
the
pods
are
listed,
and
then
you
could
also
click
it
and
maybe
there's
a
fly
out
that
gives
you
that
information.
E
So
that's
another
way
of
looking
at
it,
and
one
thing
to
note
is
we're
for
now
we're
focusing
on
the
specific
hierarchy
of
deployment
pods
and
then
the
adjacent
resources,
such
as
services
secrets
service
accounts,
but
there
might
be
other
types
of
information
that
we
would
have
to
retrofit
into
this
model.
E
So,
for
example,
here,
if
you
looked
at
this
specific
resource,
keypad
controller
admin,
there's
not
really
much
information
or
hierarchy,
so
we'd
have
to
think
about
how
what
might
we
use
the
framework
that
we're
using
right
now
to
show
deployment
replica,
sets
and
pods
and
apply
that
to
different
types
of
information,
but
right
now
for
the
mvp
and
for
the
sake
of
exploration,
we've
just
narrowed
down
to
this
specific,
this
specific
set
of
information
and
then
another
thing
we
want
to
think
about
is:
where
does
this
view
start
so
right
now,
when
you
want
to
go
into
the
resource
viewer,
you
can
go
into
deployments
or
whatever
you
click
on
the
resource
and
then
you're
in
the
resource
viewer
for
that
specific
resource.
E
E
If
you
know
what
you're
looking
for
or
maybe
you're
just
checking
the
health
of
some
of
your
deployments
and
to
take
that
a
step
back
what
if
there
was
a
high
level
overview
of
the
entire
cluster,
and
so
you
get
pretty
much
a
map
of
the
world
for
that
specific
cluster
and
you
can
drill
down
from
there
so
we're
thinking
about.
Where
do
we
start
with
this?
E
It's
a
matter
of
like
what
are
user
behaviors
around
drilling
down
like
do
they
come
to
the
dashboard
and
find
just
to
check
if
they're
errors
or
do
they
know,
there's
a
specific
error
and
they
navigate
towards
that
error,
and
it's
also
a
matter
of
scope.
What
we're
able
to
bite
off?
This
is
probably
like
the
ideal,
but
we
might
want
to
start
with
just
reimagining
what
it
looks
like
at
this
level.
B
Now
this
this,
this
is
perfect
yeah.
Thank
you
so
much.
This
is
really
good
to
see.
You
know
that
we
are
approaching
this
a
little
differently
than
a
little
more
upfront
design
and
a
little
more
balanced
approach.
So
I
I
really
like
that
and
hope
that
it
will
get
to
the
really
good
solution.
This
is
a
really
hard
problem,
and
this
is
probably
the
only
way
to
to
get
to
the
solution
that
that
that's
going
to
work
better
than
the
current
one.
A
Yeah,
I
I
think
these
look
awesome.
I
really
like
the
the
idea
of
kind
of
being
able
to
drill
down
into
these
things
and,
like
each
layer,
uncovers
more
detail,
I
think
that's
really
important,
especially
for
like,
because
in
the
success
case,
most
of
the
time
you
don't
care
about
those
details
right,
you
just
care
that
everything
is
successful
and
so,
but
even
in
the
success
case
still
being
able
to
get
to
them.
A
In
the
case
that
you
do
is
really
nice
and
then
the
I
really
like
that
when
you
were
talking
about
where,
where
does
the
view
start,
the
idea
that
was
kind
of
surfaced
in
that
overview
of
everything
that
to
me
is
like
the
the
the
kind
of.
A
New,
like
the
we
have
that
applications
view
which
shows
like
the
donut
chart
and
like
how
things
are,
are
shaping
up
for
the
name
space,
and
I
could
see
this
view
kind
of
being
the
the
new
applications
view
where
it
incorporates
some
metrics
and
incorporates
just
the
the
what
we
determined
to
be
like
the
application
or
the
workload
and-
and
we
create
this
like
overall
idea
of
of
healthy
workloads
and
or
non-healthy
ones,
and
expose
them.
And
then
right
from,
I
don't
know
in
my
head.
A
B
Yeah,
that's
really
powerful
and
you
know
the
the
content.
The
way
the
the
boxes
look
might
be
a
little
different
too
right.
We
may
yeah
indicate
in
some
way
that
this
is
maybe
an
abstract
group.
Like
you
know,
deployment
together
with
some
things.
Maybe
those
can
be
circles
or
some
some
other
symbols,
as
I
think
the
paradigm
I
I
like
to
use
is
a
map.
B
Figma
is
a
good
one
too,
but
you
know
when
you
zoom
in
on
the
map
or
figma
you,
you
start
seeing
more
details
and
when
you
zoom
out
you
those
things
may
look
differently.
You
definitely
don't
see
every
single
street
or
every
every
single
item.
You
just
see
the
contours
or
maybe
the
very
important
things
so
yeah.
I
think
this.
If
you
can
make
that
that
would
be
awesome.
I
feel.
E
Yeah
yeah,
I
think,
yeah
there's
some
open
questions,
though,
that
I
have
before
I
move
forward
and
I'll
post
this
in
the
slack
channel
as
well.
But
one
is
like
what
informations
do
do
you?
E
What
information
do
users
actually
want
to
see
up
front
like
right
now
we're
hypothesizing,
but
this
is
the
information
we
want
to
surface
like
deployment
pods
deployment
replica
sets
pods
and
those
adjacent
services
and
secrets,
but
would
there
be
something
more
useful
that
you
would
rather
see
when
you
open
the
resource
viewer
or
when
you
come
to
not
the
dashboard
page,
but
the
page
that
lists
the
status
of
all
your
like
resources
and
objects?
E
So
that's
one:
what
information
do
you
want
to
see?
Number
two
is:
do
we
want
to
preserve
this
like
hierarchical
view
that
gets
like
wider
and
wider,
or
do
we
want
to
have
a
better
view
of
like
all
the
relationships
in
one
screen?
What
would
be
more
helpful?
What
would
you
lose
if
we
were
to
break
the
hierarchical
deal
and
then
are
you
using
the
resource?
Would
you
be
using
the
resource
we
are
more
for
observability
like
oh.
E
Let
me
come
in
here
and
check
the
status
of
something
versus
like
I
know
something
is
wrong,
and
I
know
that
like.
I
just
want
to
find
that
specific
thing,
that's
wrong,
because
that
might
give
us
more
information
of
like
how
much
information
display
what
what
information
to
display
as
well.
A
A
Because
the
one
of
the
big
uses
of
the
resource
viewer
right
now,
is
you
install
a
new
crd
into
your
cluster?
You
deploy
some
of
those
new
custom
resources.
How
are
how?
What
are
those
things?
How
do
they
look
in
the
cluster?
What
are
they
connected
to?
What's
using
them?
What
and
and
that
tree
that
can
be
very
big
like
that.
A
That
graph
can
be
quite
large
and
so
being
able
to
ensure
that
on
the
engineering
side,
we
architect
it
in
a
way
that
it
is
configurable
and
and
that
we
can
create
those
relationships,
maybe
not
even
directly
through
like
owner
refs,
but
through
labels
or
annotations
or,
however,
we
we
end
up
deciding
on
it,
but
ensuring
that
when
people
deploy
things
into
kubernetes,
there
is
a
way
for
us
to
visually
represent
not
only
the
hierarchy,
but
the
relationships
like
it's
important
that
we
do
both.
B
A
One
of
the
one
of
the
examples
that
that
milan
and
I
worked
on
when
we
took
a
pass
at
doing
this-
was
the
custom
resource
for
what
was
it.
It
was,
was
it
tkg
or
it
was
one
of
the?
I
think
it
was?
Maybe
it
was
vmware's
sas
offering,
but
but
the
you
know
there
was,
I
don't
recall
exactly
how
many
objects
were
in
there,
but
it
was
a
combination
of
maybe
about
50.
yeah.
A
I
was
going
to
say
it
was
like
eight
or
nine
different
custom
resources
and
then
like
50
relationships
within
that
one
thing,
and
so
yeah
scale
is
definitely
something
we'll
run
up
against
as
well,
because
in
comparison
like
that
was
relatively
small
scale,
we
were
just
using
a
small
test,
bed
instance
of
it,
but
these
things
can
can
really
fan
out
quite
wide.
So
that's
also
something
we
want
to
think
about
in
the
design
is
how
to
capture
those
really
large
scale
things
and
and
again
I
think
it
comes
down
to
like
drilling
in.
E
A
B
There's
a
there
is
that
the
diagram
on
the
right
side,
if
you
can
scroll
there,
I
think
the
idea
yeah
see
the
frame
19..
B
So
the
idea
there
is,
I
think
I
mean
there's
obviously
clustering
and
actually
I
went
through
one
of
those
diagrams
that
had
like
50
or
60
elements,
and
I
was
able
to
divide
it
in.
Like
eight
groups,
there
were
quite
a
few
connections
between
them
still,
but
if
you,
if
you
do
it
that
way,
you
know
if
you
can
do
the
cluster.
The
clustering
like
that
you
can
replace
that
hairy
chart
at
the
top
of
that
terry
graph
at
the
top
with,
like.
E
Yeah
and
and
also
there's
different
ways
of
showing
like
the
hierarchy
like,
for
example,
this
we
can
show
hierarchy
through
size,
and
maybe
this
is
the
replica
set
and
then,
if
you
want
to
see
pods
you
go
to
a
different
page
or
maybe,
like
the
view,
changes
where
it's
like
you
click
pods,
and
now
it
actually
shows
those
50
pods,
so
yeah
just
want
to
iterate.
This
isn't
set
in
stone
with
these
boxes.
There's
so
many
other
ways
to
do
this,
but
yeah
collecting
those
use.
E
Cases
like
you
mentioned
wayne
like
this
is
the
scale
that
we
could
see
or
like
those
specific
numbers
like
these
many
replica
sets
these
mini
pods
or
whatever
types
of
information
like
that
would
be
helpful
to
stretch
test
how
far
these
designs
can
go
and
like
what
the
shape
and
form
should
be.
A
Yeah
yeah,
I
think
I
think
I
can
find
out
find
some
really
good
examples
out
in
the.
A
Serverless
and
and
some
other
ci
cd
resources
and
things
like
that.
That
would
be
helpful
in
shaping
like
the
direction
of
this.
A
B
Yeah
absolutely
agreed.
The
other
thing
we
should
not
forget
is
how
we
gonna
expose
this
to
the
plugin
api,
because
we
still
have
to
provide
the
plugin
author's
ability
to
modify
the
chart
to
add
their
own
stuff.
So,
while,
while
we
do
this
that
that
should
be
pretty
close
to
top
of
our
concerns,
how.
A
B
Extensibility
and
how
we
can
allow
the
users
to
configure
it
the
the
best
for
their
needs.
E
E
A
A
All
right,
so
moving
on
to
the
next
item
in
the
agenda,
keep
going
is
on
right
now,
as
a
octane
as
a
project
doesn't
have
any
big,
like
you
know,
make
a
big
splash
for
this
kubecon,
probably
the
next
one,
but
not
this
one.
The
we
do
have
a
ama
happening
on
the
19th,
so
that's
tomorrow,
and
that
is
from
4
to
4
45
eastern
time.
So
you
can.
You
can
find
that
a
link
right
from
our
notes
here.
A
The
community
meeting
notes
we've
also
shared
it
on
our
slack
channel,
but
we'll
share
it
again
tomorrow
before
the
event
and
we'll
tweet
out
about
it
as
well,
but
we
encourage
folks
to
stop
by
and
ask
any
questions
if
you're
on
this
meeting,
you
probably
already
have
asked
them,
but
otherwise
it's
there.
A
And
I
also
wanted
to
just
call
out
quickly
time
off
so
there'll
be
no
community
meeting
next
week.
I
think
I
think
everyone
on
the
octane
team
is
is
either
off
for
most
of
or
all
of
next
week.
So
just
keep
that
mind
we'll
not
have
one
next
week
and
there
will
be
just
some
general
like
more
time
off
throughout
the
end
of
this
year.
A
A
So
if,
if
responses
are
a
little
slower
or
things
seem
like
they're
moving
a
little
slower
apologize
in
advance,
but
it's
just
people
kind
of
taking
a
break,
but
we
I'll
be
around
and
doing
upstream
stuff
in
in,
like
on
issues
and
triaging,
and
things
like
that,
but
but
probably
not
as
active
in
the
slack
channels
and
we'll
still
have
community
meetings,
but
they
there
just
won't
be
one
next
week
since
literally
everyone's
gone
next
week
and
open
q
a
is
what's
next.
A
A
All
right,
so
what
I'm
gonna
do
is
I
have
a
build
of
octant
that
I
that
I've
been
working
on
trying
to
speed
up
and
now
this
might
not
come
over
well
over
the
the
stream
and
the
zoom,
but
I
I
do
think
it's
I
think
it
will.
I
think
I
think
it'll
be
pretty
pretty
immediate
you'll
see
the
difference,
so
I'm
going
to
stop,
sharing
and
explain
what
I'm
going
to
do.
A
Is
I'm
going
to
I'm
going
to
launch
an
instance
of
octant
as
we
currently
have
it
built
out
and
and
then
I
will
pull
up
the
browser
to
that
to
that
instance
of
octane
and
people
can
see
how
painfully
slow
it
is
to
load-
and
that's
that's
my
opinion,
but
it's
it's
true.
It
is
painfully
slow
to
load.
So
let
me
share
my
screen.
A
Share
share.
Okay,
so
octane
is
running.
This
is
my
first
time
going
to
it.
You
can
see
it's
just
spinning,
spinning
spinning.
Oh
it's
so
bad!
Oh,
that's
horrible!
I
do
not
like
this,
so
this
is
the
thing
that
I
hate
most
in
octan
that
okay,
so
there
it
is
that
is
like
that.
Took
an
extremely
long
time
so
now
what
I'm
going
to
stop
octant,
I'm
going
to
start
a
new
instance
of
octant
that
has
my
speed
up
changes
in
it
and
then
I'm
going
to
to
load
it
so,
okay!
A
A
It
happens
so
much
faster,
so
I'm
still
working
through
the
tests
and
still
working
through,
like
the
just
general
refactoring
of
it,
but
that
is
that
will
be
part
of
0.17,
which
will
be
the
the
first
electron
release.
So
anyway,
I
just
wanted
to
preview
that
and
show
it.
It's
it'll
there'll
be
an
octan
speed
up
branch
that
will
be
coming.
A
It's
all
around
access
control
and
has
access
and
the
problems
there
and
and
the
hundreds
of
access
checks
that
octant
currently
does
versus
other
methods
of
checking
access
and,
failing
so
I'll,
probably
do
a
technical,
deep
dive
on
the
actual
code
changes,
maybe
in
the
first
community
meeting
in
december
on.
A
Oh,
maybe
not
I
december
2nd
is
my
birthday
and
I'm
pretty
sure,
I'm
pretty
sure
I
have
that
off,
but
I
might
still
do
the
community
meeting
because
I'm
really
excited
about
showing
this
so
we'll
we'll
see,
but
anyway,
that
that's
it
that's
all
I
wanted
to
show
real
quick
was
like
a
very
short
demo
of
how
much
faster
it
is.
B
Yeah,
that's
so
exciting
that
startup
time
is
just
killer
right
now,
yeah,
it
really
can
fix.
That
would
be
awesome.
A
And
this,
and
this
you
know
just
to
like
it,
this
is
a
speed
up
anytime
you're,
like
switching
clusters
as
well
like
it
it's
the
startup
time,
it's
cluster,
switching
time,
it's
namespace,
switching
every
every
anything
we
do
that
causes
a
reload
of
the
entire
window
is
sped
up
by
this
set
of
changes.