►
From YouTube: Jupyter Community Call - September 28, 2021
Description
Recording from the Jupyter Community Call in September 2021.
The notes from this call can be found here: https://jupyter.readthedocs.io/en/latest/community/community-call-notes/2021-september.html
Read more about these calls on Discourse: https://discourse.jupyter.org/t/jupyter-community-calls/668
B
Okay,
I
need
to
actually
change
the
view
to
speaker
there.
We
go.
I
always
forget
to
do
that
hello
and
welcome
to
our
september.
2021
jupiter
community
call
whoo,
I
love
autumn.
I'm
super
happy
to
be
here.
I
hope
you
all
are
enjoying
whatever
season
it
is
where
you
are,
I'm
isabella
I'll,
be
your
host
today
and
yeah.
Welcome.
B
If
you
haven't
been
here
before
you
don't
know,
jupiter
community
calls
are
a
place
that
we
gather
across
jupiter
projects
and
sometimes
outside
to
celebrate
what
people
are
doing
in
the
community
show
off
resources
that
have
helped
one
another,
just
generally
kind
of
get
excited.
Remember
you
know
why
we're
all
working
hard
on
this
project
and
learn
from
each
other
in
the
process.
B
There
are
a
few
things
I
do
have
to
point
out
before
we
start
one.
This
call
is
recorded
two.
This
is
a
jupiter
community
event,
so
we
are
held
to
the
jupiter
code
of
conduct
right,
and
that
includes
me.
You
can
find
this
at
any
time.
Jupiter.Org
conduct
and
I
think
it
also
is
linked
at
the
top
of
the
agenda.
B
I
haven't
seen
the
number
of
participants
go
up,
so
I
won't
link
the
agenda
again
yet
how
we're
going
to
run
this
is
we're
basically
following
that
document.
So
the
first
thing
we're
going
to
do
today
is
talk
about
the
quick
reports,
but
we're
just
going
to
go
in
order
of
the
agenda.
People
have
signed
up
ahead
of
time.
Hopefully
I
think
we
still
have
a
little
bit
of
room,
though
so,
if
you
are
so
inspired
by
the
absolutely
amazing
work
that
I'm
sure
we're
gonna
see
today,
we
may
still
have
a
little
time.
B
C
Yeah
I
had
to,
I
had
to
remove
my
grey
screen
filter
just
to
be
like.
Oh,
are
things
still
weird
but
yeah?
That's
a
shout
out
so
shout
out
to
you
for
painting
your
wall
and
the
satisfying
feeling
of
pulling
away
the
painters
tape
as
like.
Isn't
it
is
it
like?
I
just
want
a
wall
where
I
can
put
paint
or
tape
on
and
peel
it
off.
A
B
C
Well,
that's
that's
because
isabella
had
to
go
and
find
the
the
pigments
for
yellow.
You
know
she
had
to
travel
the
world
like.
D
B
D
B
Okay,
well,
thank
you
all
for
that
entertaining
note
on
my
freshly
striped
wall.
We're
gonna
go
into
the
agenda
items
and
I
am
super
excited.
We
have
angus
first
on
the
list,
are
you
ready
to
present.
A
A
Okay,
okay,
good,
so
I
haven't
got
too
much
for
structured
talk
today,
which
is
quite
freeing
so
hi.
I'm
angus.
I've
worked
on
a
few
different
jupiter
lab
extensions,
but
the
sort
of
the
one
that's
been
sort
of
the
most
popular,
I
think
has
been
jupiter
lab,
markup
and
nick
who's
here
today
did
a
lot
of
work
recently
in
making
that
some
extensible,
so
that
other
people
can
kind
of
hook
into
it
and
the
general
sort
of
premise
there
initially
was
just
that.
A
I
wanted
to
add
a
little
bit
more
intel
interesting
markup
to
markdown
than
just
you
know,
bold
text,
images
and
so
on.
So
it
started
off
with
diagrams
and
then
from
there
there's
been
sort
of
a
few
other
extensions
that
have
been
added.
I
think,
but
there
is
actually
a
deeper
kind
of
question
embedded
in
the
idea
of
duplicate
markup,
which
is
like
what
else
can
we
do
with
markdown
rendering?
A
How
else
can
the
notebook
itself
be
a
bit
more
interesting,
and
I
think
I
hadn't
really
grasped
that
until
recently-
and
I
think
nick
had
been
sort
of
commenting
on
the
github
issues
and
I
came
to
appreciate,
there's
actually
sort
of
there's
a
whole
way
to
go
there
with
what
else
we
could
be
doing,
and
it
does
raise
some
interesting
questions
but
related
to
dublin.
Markup
is
this
extension
that
we've
been
prototyping
based
on
the
discourse
for
rich
sort
of
mark
down,
rendering
with
kernel,
outputs
and
so
jupiter
lab
eye.
A
Markdown
is
an
extension
that
sort
of
integrates
with
jupiter
lab
by
markup,
and
its
purpose
is
to
kind
of
realize
one
particular
implementation
of
this
concept.
So
I've
got
the
discourse
open
here,
just
to
kind
of
go
the
way
back
to
august,
which
I
guess
isn't
that
long.
But
the
idea
was:
could
we
embed
things,
sort
of
variable
declarations
in
our
markdown
and
have
them
render,
and
so
this
extension.
E
A
Yeah,
I
know
I
was
thinking.
Oh
yes,
that's
definitely
a
problem.
Yes,
so
the
this
extension
has
like
it's
been
realized
in
a
particular
version,
and
this
is
not
really
supposed
to
be
something
that
will
end
up
in
jupiter
lab
core.
It's
more
of
just
an
idea
of
what
this
could
look
like.
So
I've
got
a
couple
of
notebooks
here
that
make
this
maybe
a
bit
more
apparent.
So
here's
a
nice
mermaid
diagram,
which
is
quite
cool
and
I'll
reset
the
kernel
and
switch.
A
Yeah,
of
course,
so
yeah.
This
is
a
markdown,
well
a
mermaid
diagram
and
the
plugin
that
implements
this
has
chosen
to
use
a
code
block
and
repurpose
it
with
the
mermaid
syntax
to
actually
render
it
as
a
diagram
and
there's
also
another
syntax
for
svg
bob,
if
anyone's
seen
that
which
is
like
ascii
diagrams
but
they're,
sort
of
more
self-explanatory.
A
So
anyway,
this
is
just
the
idea
of
the
premise
here,
and
so
I
can
run
this
ipython
kernel
and
if
I
create
an
html
rich
object,
which
has
some
basic
styling
setting
some
text
to
green-
and
this
here
is
a
markdown
cell,
that's
sort
of
been
hacked
a
little
bit,
don't
tell
anyone
to
enable
this
rich
rendering
and
using
these
curly
brace
syntax,
which
is
an
arbitrary
choice
of
of
tokens.
But
we've
chosen
these
curly
braces
for
this
example
example,
and
we
can
now
embed
these
kernel
outputs
here.
A
So
this
gets
a
little
bit
more
interesting
when
you
do
things
with,
for
example,
sliders
so
or
widgets.
So
I've
created
a
little
slider
widget
here
and
then,
if
I
actually
run
the
markdown
cell,
which
again
is
just
the
same
syntax.
Referring
to
the
slider
variable,
it
will
actually
render
the
slider.
A
Now
if
I've
been
really
clever-
and
I
thought
about
this
beforehand-
I
would
have
also
installed
the
zeus
kernel
and
looked
at
the
x
widgets
to
demonstrate
that
with
a
different
kernel,
but
I'm
not
that
clever,
so
I
forgot
to
but
take
it
as
read
that
this
is
not
supposed
to
be
kernel
specific.
So
all
of
the
kernels
should
implement
a
particular
part
of
the
protocol,
which
enables
us
to
say
what
do
the
results
of
these
expressions.
A
Look
like
it's
not
a
mandatory
part
of
the
specification,
so
some
kernels
might
not
do
this,
and
but
it's
possible
for
all
kernels
to
do
so
and
so,
like.
Lastly,
here's
a
little
matrix
in
and
then
a
list
rather
than
a
numpy
array
and
again
you
can
sort
of
see
that
it's
embedded,
the
the
text
here
and
the
formatting
actually
is
a
little
bit
off
here.
A
I've
changed
it
a
lot
recently,
so
maybe
I
forgot
to
change
it
back
and
some
of
the
trickier
parts
is
sort
of
working
out
the
appropriate
styling.
But
you
can
see
the
general
premise
here.
This
is
what
we're
trying
to
achieve,
and
one
last
example
is.
A
I
can
remember
the
shortcut
there.
We
go
here's
senpai,
which,
if
anyone
hasn't
used,
it
is
sort
of
symbolic
expression.
So
you
can
work
with
maths
and
here's
the
result
of
an
expression
embedded
in
some
markdown
here.
So
this
is
the
actual
markdown
and
then
this
is
the
result
which
takes
a
little
while
to
load
so
yeah.
This
is
the
the
basic
tool
and
I
haven't
really
got
a
lot
more
to
say
on.
You
know
what
it
does.
A
This
is
sort
of
you've
seen
everything
there
is
to
see,
but
it
does
for
me
raise
some
interesting
questions
about
what
else
could
we
do
here?
So
there
are
some
questions
around
you
know.
A
Should
this
be
a
new
cell
type,
so
we
have
a
marked
down
cell
and
then
a
an
x
mark
down
or
an
I
marked
down
cell,
so
that
we
have
back
backwards
compatibility
and
so
that
we
maintain
a
distinction
between
the
codes,
a
cell
that
executes
code,
which
is
what
these
markdown
cells
now
do,
and
the
cells
which
don't,
which
was
the
old
markdown
cell
and
there's
also
other
questions
about
to
me
at
least
like.
What's
the
next
evolution?
A
Do
we
just
add
a
new
markdown
cell,
or
do
we
think,
could
we
go
even
further
and
rethink
other
parts
of
the
notebook,
because
these
are
all
sort
of
different
interpretations
of
rich
marked
at
markup?
They're
rich
objects,
but
we
could
do
the
same
thing
potentially
with
diagrams
or
with
geojson
or
something,
I
think,
there's
a
lot
of
opportunities
here.
So
that's
sort
of
it
from
from
my
side.
So
I
don't
know
what
the
protocol
is
now,
so
I
stopped
sharing
and
and
go
from
there
or.
B
F
I
have
a
question
first,
this
is
fantastic.
Let
me
just
just
start
with
that.
So
the
question
that
so
getting
back
to
your
question
of
whether
or
not
there
should
be
a
new
kind
of
markdown
cell
or
a
new
kind
of
cell
or
whether
they
should
just
fit
into
the
old
markdown
cells.
One
question
that
I
have
is
is
a
is
just
standard
markdown
completely
within
the
scope
of
this
thing.
So
is
this
a
pure
extension
to
mark
down
or
does
or
does
it
change
some
differences?
A
It's
a
good
question
actually
because
part
of
the
the
questions
I
alluded
to
at
the
start
is
that
markdown
itself
isn't
sort
of
one
thing
sort
of
does
now:
a
common
mark
varian,
a
lot
of
people
agree
on,
is
sort
of
what
markdown
represents,
but
you
don't
have
to
use
that
and
when
we've
this
syntax
here
that
we've
got
that
enables
these
curly
braces
to
inject
mark
down
sorry
to
inject
kernel,
outputs
that
technically
isn't
part
of
the
common
mark
specification.
A
So
we've
now
got
kind
of
got
a
common
mark
plus
going
on
here,
which
is,
in
theory
a
superset
of
markdown.
So
it's
everything
you
would
normally
do
and
it's
a
bit
extra.
But
it's
not
it's
not
out
of
the
question
that
paul
introduced
stuff
that
breaks
things,
so
you
can
put
normal
mark
down
in
here
and
it
will
just
work
normally.
So
if
I
were
to
do
that
in
the
cell
below
I
could
say,
this
is
bold
text.
A
What's
that
I
was
ho.
I
remember
where
there's
actually
keyboard
shortcuts,
but
there
aren't
oh
okay.
This
is
a.
This
is
bold
text
so
yeah.
This
is
it's
all
normal!
Markdown,
but
you
can
also
use
these
curly
braces
to
actually
execute
code
and
the
way
that
this
this
particular
extension
works.
Is
it
scans
the
markdown
for
these
tokens?
If
it
finds
them,
it
will
ask
the
kernel
to
give
it
some
extra
information
so
that
it
can
then
inject
the
results
into
the
markdown.
G
A
A
good
question
so
that
we
were
talking
about
as
if
we,
the
people
on
the
discord
side,
were
talking
about
one
of
the
ways
you
could
do
this
and
another
way
a
different
way
would
be
to
use
magic.
So
the
this.
The
answer
to
the
question
is:
if
you
use
a
markdown
display
object
and
you
present
that
to
jupiter
lab.
A
So,
for
example,
if
you,
if
you
use
python,
for
example,
if
you
were
to
create
a
markdown
object
from
ipython.display
that
gets
rendered
by
the
front
end
in
this
case,
jupiter
lab
as
a
markdown
object,
but
we
don't
expect
those
things
to
embed
code
that
can
be
executed.
So
this
is
all
done
at
the
notebook
level
rather
than
by.
A
So
it's
done
in
the
by
the
code.
That's
running
the
cell!
So
if
you
don't
have
markdown
in
a
markdown
cell,
it's
sort
of
like
a
normal
markdown,
but
if
it's
in
a
markdown
cell,
then
it
does
it
does
this
extra
feature
that
enables
us
to
sort
of
do
stuff
with
the
the
outputs
it
actually
wouldn't
be
impossible
to
modify
the
markdown
renderer,
and
so,
instead
of
doing
this
on
the
cell,
we
could
say
anytime,
we
see
any
markdown
that
needs
to
be
rendered.
A
We
could
ask
a
kernel
for
cell
outputs,
but
then
we,
when
we
start
to
do
that,
we
get
into
sort
of
some
tricky
questions
like
if
we
had
no
kernel
and
we
were
just
rendering
mark
down
somewhere
else
in
tube
to
lab.
What
would
we
do
so?
The
short
answer
is
magic.
Won't
work
to
give
you
any
of
this
extra
functionality,
but
also
render
it
as
normal
markdown.
D
And
so
I
was
actually
prototyping
this
a
little
bit
yesterday
and
the
magic
can
work
if
the
if
it
embeds
both
the
plane,
markdown
and
the
markdown,
like
with
the
curly
brackets
and
also
saves
into
the
notebook
file,
the
output
items
which
are
going
to
be
rendered
in
line
which
is
possible.
I
I
don't
think
that
these
are.
You
know
the
two
approaches
you
know
it
can
be
one
or
the
other.
I
do
think
it'd
be
sort
of
nifty
to
have
an
eye
markdown
like
output,
mime
type,
but.
A
Yeah
and
if
I,
if
I
could
speak
to
that,
I
have
a
habit
to
talk
so
if
anyone
else
wants
to
interrupt,
I
don't
know
whether
it's
just
my
zoom.
I
can't
actually
easily
change
the
the
view,
so
I'm
just
sort
of
looking
at
like
four
people,
so
interrupt
me.
If
I
need
to
stop
but
yeah
that
we
that's
it
there's
loads
of
different
ways.
A
We
could
do
this
and
I
think
I
I
we
were
actually
just
sort
of
talking
on
the
discourse
just
before
the
meeting,
but
the
the
kind
of
it
would
be
quite
exciting
if
you
could
kind
of
move
away
from
it
has
to
be
a
code
cell
or
a
markdown
cell,
or
it
has
to
be
a
cell
output.
There
are
lots
of
different
ways.
A
We
could
have
rich
markup
in
a
jupiter
notebook
and
currently
the
only
way
we
can
do
that
is
cell
outputs
or
if
you
have
one
of
the
three
three
cell
types,
so
yeah
there's
so
many
different
ways.
We
could
do
this
and
it
would
be
quite
nice
to
imagine
in
a
world
where,
okay,
I
have
this
idea
in
my
head,
where
the
notebook
is
rather
than
having
cell
types,
it's
just
a
load
of
mime
types
and
then
the
front
end
knows
how
to
either
render
it
or
edit
it.
A
We
get
into
other
problems
if
we
start
doing
things
like
that,
but
yeah
there's
a
lot
of
scope
here.
To
do
this
and
to
be
really
clear.
This
is
this
is
an
extension
that
I've
worked
on
to
kind
of
deliver
a
prototype,
but
this
could
be
binned
tomorrow
and
something
else
could
be
put
in
this
place.
If
we
decided
this
was
going
to
be
something
that
should
be
in
jupiter
lab
itself.
C
H
I
just
wanted
to
toss
out
one
idea,
so
one
thing
that
that
I
had
thrown
out-
and
I
didn't
pick
it
back
up
because
it's
been
busy
one
thing-
that
every
markdown
representation
does
have
to
agree
on
are
links,
and
so
links
and
images
are
strongly
represented
across
every
flavor,
because
you
can't
break
those
yeah
and
they
bring
with
them
a
lot
of
nice
features
of
of
hyper
media.
So
you
know
if,
instead
of
curly
braces,
we
had
the
eight
different
forms
of
image,
references,
right
images
and
links.
H
They
give
you
the
the
handy
feature
of
when
they
break.
You
still
can
find
out
something
about
them.
They
have,
they
have
all
texts
they
have.
They
have
titles
things
like
that,
and
that's
the
thing
that
I
most
worry
about
is
as
we
get
fancier
and
fancier
markup.
H
The
next
client
won't
be
able
to
render
it,
and
now
it
looks
like
absolute.
You
know
manure
right
like
that.
These
are
the
things
that
I
I
really
dislike.
Is
I
get
this
cool
notebook
from
a
couple
years
ago
and
it
worked
with
some
specific
way
that
someone
had
manipulated
their
jquery
notebook
environment
and
then
it
doesn't
render
anymore.
So
that
would
be
another
syntax
that
wouldn't
require
too
much
special
casing
of
the
token
stuff,
and
I
differ.
C
Which
is
great,
I
love
all
these
things.
I've
built
something
similar
to
this
in
pipe
in
a
pure
python
kernel
and
I
would
not
do
away
with
inputs
and
outputs
and
just
go
straight
to
wrappers,
but
rather
my
approach
has
been
to
say.
I
have
a
markdown
cell
with
some
templates
in
it,
I'm
going
to
show
that
output
and
then
the
outcome
of
my
document
is
going
to
be
all
the
outputs
put
together.
C
So
I've
gotten
to
a
place
where
I
haven't
really
been
separating
the
inputs
and
the
outputs,
because
the
next
point
in
a
few
years
when
I
get
to
this
thing,
we're
not
looking
at
it
at
the
cell
level
anymore
and
we're
looking
at
it
at
the
document
level.
So
you
know
this
is
great
stuff
angus.
As
you
move
forward.
You
know,
there's
this
really
fun
high
velocity
stuff
that
happens
at
the
cell
level,
but
remember
to
question
that
stuff
against,
like
the
enduring
document.
C
So
like
I
like
the
ginger-like
syntax,
because
I'm
like
hey,
I
can
use
that
everywhere,
but
you
know
you
can
do
a
lot
with
this
stuff.
I
look
forward
to
seeing
where
it
goes,
but
the
question
that
I'm
curious
about
is
like
in
the
document
that
you're
running
now.
Are
you
effectively
using
two
different
kernels
and
does
that
change
like?
Is
there
a
javascript
kernel
effectively
and
a
python
kernel?
C
A
Yeah,
I
think
well
I
mean
I
was
thinking
about
this.
The
other
day.
There's
two
questions
here:
there's
one
about
link,
syntax,
oh
well,
ideas
and
the
other
one
is
yeah
about
this.
Multiple
kernel
answer
nick's
first,
which
is
yeah.
I
I
definitely
agree
at
the
moment
if
you,
if
you
try
and
render
this
in
in
jupiter
notebook,
for
example,
where
we
don't
have
a
rent,
a
front
end
renderer,
it
will
look
terrible.
A
I
actually
haven't
tried
it
so
and
in
some
cases
terrible
might
be
okay,
but
it
would
be
useful
to
have
a
fullback
and
yeah.
I
think,
in
an
ideal
world
the
fullback
would
be
we'd,
have
multiple
different
display
types
in
the
mind.
A
Bundle
and
we'd
fall
back
on
something
that
had
made
it
clear
that
this
wasn't
rendering
properly,
because
although
we
could
do
some
with
links,
I
think
as
well
for
all
of
the
other
kinds
of
markdown
extensions
we
can
do
besides
this
particular
one
we
sort
of
have
the
question
of
how
do
we
make
this
look
work
on
other
clients?
How
do
we
tell
other
clients?
A
We've
got
some
special
markup
here
and
I
think
it's
a
really
big
question
and
that
was
sort
of
I
don't
know
if
nick
would
say
when
I
was
talking
about
this,
but
yes,
that's
what
I
was
alluding
to
earlier.
There
are
some
big
questions
here
with
markdown.
Isn't
actually
this
very
highly
specified
thing
for
a
lot
of
clients.
We
need
to
maybe
work
on
specifying
it
more
carefully
in
the
notebook
itself.
At
the
moment,
the
notebook
specification,
I
think,
is
marked.js,
but
it's
not
actually
written
down
in
a
an
rfc
or
anything.
A
A
Tune
doing
the
nb
combat
stuff.
I
actually
do
like
the
idea
of
the
link
syntax
from
just
an
ergonomics
perspective,
because
adding
an
alt
text
is
quite
useful
and
we
can
sort
of
move
on
in
the
same
way
we
have
attachments
in
notebooks
use
this
attachments
resolver
in
jupiter
lab.
We
could
have
a
similar
kind
of
protocol
for
for
this
kind
of
stuff
for
the
kernel.
So
that's
the
first
thing
with
links.
I
definitely
agree.
A
We
can't
just
sort
of
move
forward
with
this
as
like
as
a
blessed
extension
or
anything,
because
those
are
the
big
problems
and
I
think
we
do
want
to
support
this
and
yeah
stupid
lab
is
one
of
the
front
ends,
but
there's
loads
of
different
ways:
people
use
jupiter,
including
mb,
convert
and
all
that
stuff.
A
The
second
point
is
about
the
kernel,
so
yeah
I
mean
in
absolute
terms,
there's
only
one
kernel
here,
it's
the
the
ipad
kernel,
but
I
definitely
agree
that
you
can
think
about
this
in,
like
what's
in
a
high
level
sense.
What's
going
on
here
and
and
yeah,
we
are
obviously
doing
stuff
matt
some
kind
of
magic
at
the
front.
A
End
the
front,
and
it's
not
a
dumb,
just
rendering
what
we
see
in
the
markdown
we're
also
kind
of
weaving
things
in
and
that's
maybe
a
bit
cleverer
than
we
used
to
have
things
happening
in
jupiter,
but
I
think
at
the
moment
this
is
a
reasonable
level
of
spooky
magic
and
I
think
the
difficulty
is
in
making
sure
that
any
front-end
can
render
this
and
that's
about
standardizing,
which
is
what
I'm
sort
of
interested
in
just
maybe
having
conversations
around
like.
A
What's
the
how's
this
going
to
look,
if
we
tried
to
make
this
a
part
of
jupiter
lab
or
jupiter,
sorry
as
a
whole
and
in
particular,
when
you
talk
about
documents,
tony
have
you
been
working
on
the
death
beds,
project
yeah,
that's
nick,
and
I
yes,
okay,
so
that
makes
sense
yeah
I
mean
when
I
was
thinking
about
how
you
could
like
notebook
2.0
but
tongue-in-cheek.
What
could
you
do
with
the
notebook
next,
if
you
were
to
like,
if
you
could
start
again
today?
How
do
we
design
it.
H
I
A
No,
I
think
this
is
a
community
called,
isn't
it
we
all
put
our
thoughts
in
yeah
I
mean
I
think
prizes
are
actually
really
useful
for
the
markdown,
let
alone
everything
else.
My
rhetorical
question
was
gonna,
be
I
wonder
whether
in
my
head,
the
the
notebook
is
just
a
collection
of
objects
and
I've
used
the
jupiter
flex
for
voila
extension,
where
which
it
basically
takes
the
notebook
and
restructures
it
into
the
sort
of
front
end,
which
has
no.
It
doesn't
have
the
notebook
structure
anymore.
A
It's
all
over
the
place
and
that's
one
front
end,
but
there's
loads
of
different
ways
that
people
use
notebooks
and
so
obviously
the
more
general
we
get
the
less
we
can
specify
in
the
notebook
itself.
We
can't
say
a
notebook
is
a
top-down
document
anymore
and
you
can
even
go
further,
and
so
I
think
that's
what
I
think
of
the
bare
minimum
abstraction
for
a
notebook.
A
Is
it's
a
collection
of
things
which
have
important
semantic
meaning,
so
the
front
end
should
be
able
to
understand
at
least
a
subset
of
so
front-end
is
currently
understand,
marked
down
in
code.
But
could
we
generalize
that
even
further,
and
so
they
can
understand
anything
as
long
as
it
has
a
mime
type,
that's
sort
of
my
way.
My
thoughts
have
been
going
at
the
moment.
D
I'm
going
to
hop
in
really
quick
and
say
that
my
my
sort
of
desire
is
that
notebooks
can
actually
be
a
distribution
format,
and
for
that
you
want
to
be
able
to
send
notebooks
to
people
without
having
to
have
a
list
of
here
are
the
extensions
that
you
need
to
view
this
notebook
or
instructions
on
how
to
use
it,
and
that
that
requires
a
more
concrete
specification
for
how
the
notebook
is
displayed.
B
B
B
C
A
C
It's
it's
all
right
like
at
the
end.
At
the
end
of
the
day,
all
of
us
are
being
repetitive
and
redundant
because
donald
knuth,
like
wrote
about
literate
programming
a
long
time
ago.
I
really
think
that,
like
in
some
way
or
another,
we're
just
like
delving
back
into
old
research
and
there's
like
some
good
foundations
to
build
off
of
so
I
hope
we
can
bring
the
old
and
the
new
together
and
design
something
radical
again.
B
B
I
Yeah,
I
am
I'm
gonna
share
my
screen
here
and
I
don't
think
this
will
take
up
all
the
rest
of
the
time.
I
know
that
there's
some
some
other
demos
coming
that
will
be
more
exciting,
probably
than
talking
about
security
stuff.
But
what
I
wanted
to
do
wait.
Can
you
still
see.
I
You
can
okay,
that's
weird
because
it
says
my
screen
sharing
is
paused,
but
that's
weird.
Well,
I
don't
want
to
quite
do
it
this
way.
I
guess,
but
anyway,
if
people
can
specific.
J
Tab,
rather
than
your
window,
that's
why
it
says:
sharing
is
paused
because
you've
moved
away.
I
I
So
what
I
wanted
to
share
with
everybody
was
that
we
have
this
engagement
with
trusted
ci,
which
is
the
national
science
foundation
center
of
excellence
for
cyber
security,
and
that
runs
through
the
end
of
the
year
and
so
rick
wagner,
and
I
rick
at
ucsd,
we've
been
working
a
little
bit
on
this
on
the
jupiter
security
subproject
in
collaboration
with
trusted
ci
and
one
of
the
first
things
that
they
they've
done
for
us
is
actually
try
to
do
a
full
census
of
all
of
the
security
related
documentation
from
across
the
project
and
k.
I
Avila
did
this
and,
as
you
can
see,
it's
about
seven
pages
worth
of
links
and
in
in
among
these
links
are
some
some
comments
about
how
clear
the
documentation
is
and
whether
or
not
things
are
being
repeated
across
different
sub-projects
or
whether
or
not
actually
the
documentation
says
one
thing
in
one
place
but
says
something
else
in
another
and
trusted
ci
is
going
to
work
on
this
census
and
try
to
put
together
kind
of
a
high
level
document
about
jupiter
security,
best
practices
and
then
you
know
hand
it
share
it
with
the
community
as
they've
shared
this
document
here.
E
D
B
I
I
got
it
all
right
so
back
on
track
this
document's
public,
so
anybody
can
see
it
and
everybody's
welcome
to
place
comments
on
it,
of
course,
so
I've
left
that
link
there,
there's
also
a
link
to
it
from
the
discourse
that
that
rick
very
kindly
put
up
so
we're
gonna
be
discussing
this
document
at
the
jupiter
security
subproject
meetings
which
are
bi-weekly
on
fridays.
I
You
can
find
the
invitation
to
the
to
the
bi-weekly
security
meeting
here
in
the
document
is
part
of
the
calendar,
and
everybody
is
welcome
to
join
and
talk
to
us
about
jupiter
and
security
stuff.
I
I
Team
already,
but
you
know
the
the
rick
and
rolling
show
as
you
call
it,
yeah
we're
we're
still
we're
the
hosts.
So
I
guess
that's
kind
of
how
it
works.
I
J
A
bit
of
a
question
about
the
security
meetup
like
is
there
a
kind
of
a
like
a
like
a
report,
reporting
program
process
or
anything
like
that?
I
don't
know.
If
people
saw
a
few
weeks
ago
on
twitter,
there
was
like
some
notices
of
travis
not
being
super
cool,
and
that
was
kind
of
a
thing
that
I
wanted
to
like
flag
to
someone
at
jupiter.
I
So
yeah
this
is
part
of
why
we've
started
a
jupiter
security
subproject,
because
it
is
actually
frankly
figuring
out
how
to
share
that
kind
of
conversation
with
people
or
you
know,
get
the
word
out
about
things
isn't
entirely
clear
and
so
we're
trying
to
figure
out
how
to
make
that
process
of.
Where
do
you
go
when
you
want
to
share
something
or
where,
when
you
need
to
go,
get
help?
How
do
we
make
that
more
transparent
for
for
vulnerability
reporting?
I
We
want
to
make
it
clear
to
everybody
that
there
is
basically
one
place
to
be
doing
that,
and
that
is
security
at
ipython.org,
and
that's
that's
on
every
bi-weekly
agenda
and
it's
also
on
the
new
jupiter.org
security
page,
which
is
at
jupiter.org
security,
which
we
didn't
have
until
a
couple
weeks
ago,
when
rick
put
that
together.
So
we
have
a
high
level
page
for
that
yeah.
I
There's
every
so
often
there
are
these
nice
little
things
that
we
have
to
react
to,
and
the
travis
thing
that
you're
mentioning
is
one
of
those
but
yeah
in
terms
of
informing
informing
all
the
various
sub
projects,
the
things
that
they
need
to
be
aware
of.
That's
one
of
the
things
we're
going
to
try
to
work
on
and
that's
part
of
why
we
have
these
little
updates
from
the
security
subproject.
I
So
sarah,
what
I'm
going
to
do
with
that
is
I'm
going
to
go
back
to
our
our
github
and
maybe
make
an
issue
out
of
that
out
of
that
kind
of
dissemination
of
things
to
sub
projects.
What
kind
of
mechanisms
can
we
put
in
place
for
that?
I
think
that's
a
that's
a
good
issue
to
raise.
F
Roland,
do
you
know
steph
savage
by
any
chance?
I
do
not
okay
well
or
I
would
suggest
that
you
and
rick,
or
just
some
time
when
you've
got
some
time
head
over
he's
head
over
to
the
cs,
to
the
cse
department
there
at
you
know
it
gives
you
a
chance
to
visit
the
granite,
fair
and
so
steph
and
his
group
is
definitely
now.
I
know
who
you're
talking
about.
F
Okay,
yeah
he's
he's
he's
a
he's
a
friend,
so
I
can
you
know,
but
he's
also
a
very
welcoming
and
warm
guy.
So
you
won't
need
any
intros,
and
you
know
he
and,
and
his
colleagues
like
jeff
bulker,
would
be
extremely
interested
in
this
project
and
would
be
very-
and
I
think,
might
have
some
suggestions
and
help
for
you
all
right.
Great.
F
And
they're
really
good-
and
this
is
an
incredibly
important
project
that
you
have.
B
K
B
I
K
F
Okay,
thanks
thanks,
so
so.
First
of
all,
this
is
just
to
let
everybody
know
this
is
very
early
and-
and
I'm
really
looking
principally
for
feedback
and
which
you
my
email,
is
there
and
you
can
and
you
can
always
be,
give
me
feedback
offline
or
here
offline
or
by
or
on
discourse
whatever
you
like,
I'm,
I
I'm
really
any
comments
at
all
are
appreciated,
so
you
can
see
here.
I've
got
a
dashboard
which
is
u.s
presidential
elections
from
1828
to
2000..
F
With
here
we
were
looking
at
1960.
The
states
are
color-coded
in
the
usual
way,
which
is
red
and
red
for
republican,
do
blew
for
democrat
and
white
and
white
in
this
case
for
third
party,
and
if
I
click
on
a
state
like
nevada.
For
some
reason,
I
can
see
how
nevada
voted
in
the
1960
election.
I
also
see
the
popular
vote
for
the
the
electoral
college
vote
for
1960.
F
F
So
back
in
1892,
the
south
was
blue
in
the
north
was
red
and
in
2000
2008
things
looked
very
different.
So
that's
so
that's
a
dashboard,
and
I
produced
this
strictly
from
a
jupiter
notebook
and
I'd
like
to
show
you
a
little
bit
about
how
we
about
how
that
was
done.
This
extension,
which
is
which
we're
which
we're
going
to
make
available
when
we
make
it,
get
it
just
a
little
bit
more
mature.
F
So
here's
the
database
that
we
start
and
I'd
like
to
just
spend
a
few
minutes
doing
a
couple
of
minutes,
building
the
first
couple
of
charts
on
that
and
you
can
and
show
you
how
it's
published.
So
we
started
off
with
this
with
this
database.
It
comes
from
the
cook
political
report
and,
as
you
can
see
here,
we
have
presidential
elections
with
the
year
in
the
first
row,
the
candidates
in
the
second
row
and
the
values
for
each
state
and
nationwide
in
the
in
the
subsequent
rows.
F
And
now
I'm
going
to
go
to
a
jupiter
notebook
right
here,
and
this
is
simply
going
to
take
that
data
and
turn
it
into
a
turn
it
into
a
table.
Format
which
would
have
been
had
I
used
had
I
used
pandas
in
this-
would
have
been
a
lot
shorter
and
so
stupid
of
me
not
to
anyway.
So
I'm
going
to
start
off,
but
before
I
start
to
execute
the
notebook
I'm
going
to
fire
up
a
galileo
dashboard
as
it's
called,
and
this
is
an
editor
which
will
which
will.
F
Let
me
just
construct
the
dashboard
that
you
saw
from
the
data
in
the
notebook
using
drag
and
drop
operations.
F
As
I
execute
this,
we
can
now
see
the
columns.
I've
got
year
state
name
of
the
candidate
party,
the
candidate,
the
votes
he
got
and
the
percentage.
So
we
just
have
a
single
table
like
that
and
which
is
sitting
in
what
is
what
amounts
to
a
pandas
data
frame,
what
we
call
a
galileo
table
and
we're
sending
it,
and
once
we
now
we've
computed
it.
I'm
going
to
go
back
to
my
dashboard
file
that
you
can
see.
F
I'm
editing
here
and
I
currently
have
no
tables
and
now
the
kernel
using
the
standard
standard
mechanism,
that's
familiar
from
my
five
widgets
and
certainly
was
only
from
angus's
demo
earlier,
just
sends
the
data
off
to
the
off
to
the
to
the
dashboard.
So
now
you
can
see
we
have
presidential
vote
table
which
we've
just
computed
and
once
we
have
that
I
can
filter
it.
F
So
I'd
like
to
I'll
just
create
that
and
drag
it
down
here
and
go
back
to
my
notebook
and
create
some
more
tables,
I'm
just
creating
a
pivot
record,
a
nationwide
vote
table
and
I'm
now
going
to
turn
the
pivot
table
into
a
margin
table
where
plus
10
is
repo
is
a
democratic,
landslide
plus
five
is
a
narrow,
democratic
victory.
Minus
five
is
a
republican
landslide.
F
Minus
ten
is
a
narrow
republican
victory
and
zero
is
a
third
party
victory
so
that
computes
up
from
the
pivot
table-
and
I
send
it
over
to
the
dashboard-
and
now
I,
as
you
can
see.
If
I
go
back
to
tables,
I've
got
a
bunch
of
them
all
the
tables
that
I've
computed
and
just
set
over
and
now
that
I've
got
a
table,
I
can
create
what
I
call
a
view
on
it
and
a
view
is
just
a
subset
of
a
table.
F
So
in
this
case
I'm
going
to
take
the
presidential
margins
table,
I'm
going
to
give
it
just
the
margin
map,
the
name
margin
map
create
it
and
I'm
going
to
be
asked.
Well,
what
columns
do
I
want?
Well,
in
this
case,
I
want
state
and
margin
from
this.
The
the
table
has
three
columns.
I
only
want
two,
the
state
and
margin,
and
I'm
going
to
use
this
thing
to
filter
the
rows
that
I
look
at,
so
I
only
get
the
the
views
that
are
that
are.
F
I
only
got
the
I
only
get
the
years
that
are
compatible
with
this
with
this
filter.
So
I
update
it.
Then
I
create
a
chart
based
on
which
I'm
also
going
to
call
margin
math
and
I'm
going
to
base
it
on
the
view.
F
I
get
all
the
tables
and
views
here
margin
map,
so
I
create
that
and
now
this
appears
the
chart
itself
appears
in
the
top
left
corner
which,
and
it
always
appears
as
a
table
because
that's
universally
rendered.
It's
also
not
typically
what
I
want.
So
I
get.
This
chart
editor
popped
up.
I
want
a
map
chart
over
the
united
states
region.
F
The
democrats
are
blue
and
we'll
make
third
parties
white
and
no
value
black,
say
that's
okay
and
it
changes
and
now
indeed,
as
I
move
the
slider
as
I
move
the
slider,
you
can
see
the
the
map
changes
and
with
respect
to
what
I
get
I'll.
Do
one
more
very
quickly
be
just
to
just
to
illustrate
one
more
aspect
of
this
and
then
take
it.
Throw
it
open
for
questions
which
is
I'll
go
to
a
I'll,
create
another
view.
F
I
want
this
on
presidential
vote
actually
so
here,
so
here
are
the
various
things
I've
got
in
this
case.
I
want
my
x-axis
to
be
party,
my
y-axis,
to
be
percentage
of
the
vote
and
you'll
notice.
F
I
have
two
filters
here:
the
year
filter
which
we
see
here,
but
also
this
map
is
filtered,
and
this
is
how
clicking
on
a
state
changes,
selects
a
selects,
a
particular
state
and
any
chart
has
this
property
that
clicking
on
clicking
on
it,
selects
that
x
selects
the
appropriate
x-axis
value,
which,
in
this
case,
states
charts
create
a
chart
based
on
this.
This
view
that
I
just
created
you
don't
have
to.
I
don't
have
to
make
to
use
the
same
name
for
charts
and
views
I
tend
to
because
it's
easy.
F
So
it's
easier
for
me
to
remember
what
input
is
to
what
chart
create
this
guy,
what
the
heck
I'll
take
a
bar
chart
or
a
column
chart,
and
there
we
go
and
as
now
as
I
click
on
the
state,
let's
say
texas,
I
see
how
texas
voted
in
1948
and
I
could
continue,
but
it's
just
a
repetition
of
these
operations,
and
you
see
plenty.
F
Oh
one
thing
I
do
want
to
point
out
is
that
this
thing
it
does
have
a
is
that
this
thing
is
a
is
a
json
file
which
should
update
aside,
and
I
can
also
look
at
it
and
look
at
it
as
a
json
file,
and
you
can
see
this
thing
pops
up
this
this
way,
so
this
could
also
be
generated
if,
if
I
wanted
as
a
just
as
just
a
straight
jason
file,
and
then
the
dashboard
here,
as
you
can
see,
takes
the
name
of
the
json
file
to
render
as
a
right
here
as
the
as
the
dashboard,
so
basically
just
moving
this
to
any
joining
public
url.
C
I
think
mike
this
looks
super
cool
with
these
tools.
There
are
a
few
out
there
like
holoviews
does
stuff
like
this.
One
of
the
things
that
always
seems
like
a
toss
up
to
me
is
being
able
to
use
the
ui
and
then
also
being
able
to
programmatically
work
on
things.
Does
this
tool
have
that
ability.
F
So
a
couple
of
so
first
of
all,
I
should
make
some
the
the
answer
is
that's
a
really
good
question
and,
as
with
many
really
great
questions,
the
answer
is:
I'm
not
really
sure
we'd
have
to
dig
into
it
deeper,
which
I'd
love
to
do.
The
answer
is
there
are
several
way
the
answer
is:
there
are
several
ways
we
can
cut
this,
so
one
is
yeah.
You
can
certainly
you
don't
even
need
to
use
this
drag
and
drop
editor.
F
You
can
just
push
you
could
just
you
could
just
push
out
the
data
structure,
but
two.
F
So
let
me
just
so
the
reason
we
built
this
thing
is
there's
actually
three
chunks
of
technology
here,
of
which
one
is
just
pure
jupiter
and
the
other
two
are
things
that
that
we
would
like
to
integrate
into
integrate
into
jupiter
in
in
a
in
a
very
deep
way,
one
of
which
just
goes
to
that
extension
that
you
talked
about
tony,
so
the
first
is
just
is
obviously
jupiter
itself.
F
The
second
is
that
there's
a
is,
there's
a
little
chunk
of
code
so
that
this,
the
editor
is
the
editor
here
is
this
guy
is
a
web
page
sitting
within
an
iframe
and
so
there's
an
extension
to
jupiter,
which
is
currently
a
closed
repo,
but
is
under
the
bsd
license.
F
I'm
going
to
open
it
up
when
I
get
the
bugs
out
and
the
other
thing
I'd
really
like
to
do
is
genericize
this
thing
so
that
any
any
web
applications
sitting
in
an
iframe
can
talk
to
the
jupiter
can
talk
to
the
jupiter
file
system
directly,
just
using
the
same
extension
and
all
you
have
and
you
don't
have
to
write
code.
All
you
have
to
do
is
say:
here's
the
iframe.
F
Here's
the
here's,
the
file
type
go,
and
it
just
and,
and
it
happens,
fine,
and
so
anybody
can
use
this
for
their
for
their
for
their
web-based
editor
extensions
and
then
the
third
is
this
web-based
editor
itself,
which
is
what
I
which
is,
and
this
is
so
the
web-based
editor.
F
The
dashboard
editors
is
an
application
that
well
largely,
I
wrote
on
top
of
an
open
source
project
called
lively
next
and
what
I
would
really
really
and
the
and
the
limitations
on
this
right
now
in
terms
of
the
dashboard
application,
is
what
I
can
expose
as
a
in
a
widget-based
ui
on
top
of
lively.
F
What
I'd
love
to
do
is
is
is
really
in
that
so
people
could
put
together
simulations
is
so.
The
answer
is,
I
would
the
answer.
Tony
is,
I
would
love
to
do
it.
F
The
community
could
use
this,
so
so
so
so
the
answer
is,
thank
you
and
I'm
sorry
to
spend
so
long.
Answering
that
such
a
short
question.
A
Mic
switches:
okay!
Yes,
it
was
an
intentional
hand,
raise
I
put
it
on
late,
so
thanks
for
noticing
rick.
This
looks
really
interesting
and
I
was
wondering
so
is
what's
happening
here-
that
the
notebook
itself
is
pushing
sort
of
what
becomes
static
content
to
this
front-end
sort
of
renderer.
So
is
everything
embedded
that
needs
to
be
embedded
inside
this
json
object,
for
example
the
table
data
and
whatever
else
you've
pushed
from
the
notebook.
F
Yes,
that's
yes,
that
is
yes,
it
is
nick
and
here's
or
or
adam
sorry,
pardon
me,
and
there
are
two
things
I'd
like
to
do
about
that.
So
the
first
is
the
the
first,
which
is
just
straight
plumbing,
is
to
have
the
widgets
on
the
is
to
have
the
widgets
on
the
dashboard
ask
for
data
from
the
ask
for
data
from
the
notebook
and
get
it
back,
and
that's
that
isn't
really
very
hard.
That's
just
making
the
well!
You
know
this
better
than
I
do
how
easy
that
is.
F
But
the
other
issue
that
I
have
in
doing
is
that
is
that,
okay,
once
we
do
that,
what
we'd
really
like
to
do
is
let
people
package
up
what
they've
got
and
throw
it
into
and
just
throw
it
into
a
json
file
which
can
be
rendered
as
a
which
can
we've
we've
built
a
renderer,
but
we
can
also
just
aesthetically
freeze
this
and
throw
throw
it
into
a
into
a
zip
file
that
people
can
can
pull
anywhere.
F
Anyhow,
I'd
like
to
be
able
to
do
that,
but
the
trouble
is
that
if
the
data
source
is
sitting
externally,
then
we
have
to
find
some
way
to
to
co-publish
the
notebook
as
a
data
as
it
as
a
data
server
and
that's
going
to
take
that's
a
little
bit
bigger
left.
F
B
G
Yeah
hi,
look
I'm
really
new
to
this.
The
the
jupiter
lab
development.
I've
been
the
user
of
jupiter
jupiter
software
for
the
past.
You
know,
like
maybe
four
four
or
five
years,
it's
a
piece
of
software
that
really
revolutionized
the
way.
I
do
research
and
programming
and
I'm
really
grateful
to
the
community
for
for
for
doing
this.
About
a
month
ago,
I
started
looking
into
developing
jupiter
lab
extensions
and
customizing
jupiter
lab
to
fit
my
own
needs
and
I've
went
through.
G
You
know
the
fantastic
tutorial
from
last
year's
jupiter
con
that
was
done
by
alex
carla
and
clara
or
carla.
There
was
like
someone
else
like
fantastic.
It
was
like
a
three
hour
stream,
but
you
know
the
the
one
and
a
half
hours
of
the
content
that
that
is
there
is
is
really
nice
and
the
jupiter
lab
extensions
extension
examples
repository.
G
I
was
really
I
like
popping
into
open
source
community
meetings
to
see
the
people,
and
maybe
you
know
if
someone
can
point
me
to
good
development
tutorials
that
teach
how
to
utilize
the
jupiter
lab
front
end
itself.
You
know
like
like,
for
example,
there
is
this
plus
icon
in
the
file
browser,
and
I
want
to
have
this
file
like
this
plus
icon,
like
in
the
in
the
left
panel,
when
the
file
browser
is
closed,
so
that
I
can
create
a
launcher
panel
without
opening
the
file
browser.
G
C
It
does
just
I
just
I
don't
have
any
answers.
Angus
might
have
better
answers,
but
the
jupiter
lab
dev
meeting
might
be
a
good
place
to
go.
That
happens
every
wednesday,
so
that
might
be
a
place
to
bring
this
up.
There's
a
more
focused.
G
Yeah
I
was
at
the
deaf
meeting
a
couple
of
weeks
ago
and
the
level
of
issues
being
discussed
there
is
like
so
super
important.
I
think
that
you
know,
like
a
question
of
pointing
to
tutorials.
Just
does
not
belong
to
it.
At
the
meeting
of
that
magnitude,
I
mean
like
that
serious
meeting
there.
I
That's
that's
unfortunate
as
a
committer
on
that
project.
I'd
like
to
argue
that
that's
not
the
case,
and
I'm
sorry
that
we
gave
you
that
impression.
It
might
have
been
one
of
those
weeks
where
a
couple
of
the
high
tier
guys
both
decided
to
have
a
really
in-depth
discussion.
That
does
happen
in
that
meeting
every
once
in
a
blue
moon,
but
it
is
actually
not
how
most
of
those
meetings
go.
We
actually
do
want
new
people
to
come
in
and
ask
questions
like
this.
I
In
that
meeting,
I
would
recommend
going
into
that
meeting
again
tomorrow
and
trying
again
and
just.
I
On
the
agenda,
like
you
did
here,
will
always
give
you
time,
we'll
always
make
sure
to
have
time
to
discuss
those
topics.
I
G
The
magnitude
of
this
project,
the
impact
this
project
has
in
my
life
is
like,
is
it's
impossible
to
overestimate
this,
but.
F
Theodore,
imagine
thousands
hundreds,
tens
or
tens
or
hundreds
or
thousands
of
people
like
you
who
have
seen
the
same
problem
who
have
had
the
same
issues
you
have
but
didn't
have
and
by
the
way.
Thank
you.
So
much
for
coming
didn't
speak
up.
Documentation
is
amazingly
important.
I
mean
saying
you
know
this
is
important
enough
for
the
jupiter
lab.
Sorry,
I've
been
an
open
source
contributor
for
30
years.
Let
me
tell
you
there
is
nothing
more
important.
No
programming
tool
is
more
important
than
a
pen.
G
A
Please
I
think,
looking
for
straws
is
always
a
good
one,
because
people
go,
oh
I'm
sure
I've
seen
something
somewhere,
but
I
haven't
written
it
down
and
I
would
also
recommend
I
don't
also
have
any
good
links
right
now
to
recommend
to
you,
but
I
would
recommend
if
you
haven't,
if
you
use
any
of
these
sort
of
the
asynchronous
platform.
So
if
you
use
like
gita
or
discourse,
I'm
not
so.
I
A
On
discourse,
I
don't
use
it
for
certain
conversations
as
much
myself,
but
the
gita
discussion
is
usually
really
good.
If
you
have
a
specific
question,
you
can
go
on
there
and
ask
it.
G
Yeah
I
had
one
of
those
and
I
had
a
very
good
link
to
a
github
repo
or
that
came
to
me
through
through
getter,
and
it
actually
like
it,
opened
up
a
whole
set
of
like
opportunities,
because
it
shows
you
how
you
need
to
structure
your
search
query
for
github
to
get
very
good.
You
know
very.
G
And
so
this
yeah
okay
I'll,
ask
like
probably
the
same
question
no
at
the
deaf
meeting
tomorrow,
and
thank
you
very
much
for
this
advice
and
thank
you
all
for
you
know,
being
a
part
of
this
community
and
they're
doing
it's,
it's
absolutely
fantastic.
G
B
B
B
Okay,
then,
for
that's
it
we
did
it.
That
was
awesome.
Everyone.
Thank
you
for
so
many
great
discussions.
It
was
awesome
to
hear
what
everyone
was
thinking
and
such
a
great
range.
Today,
too,
I'm
gonna
be
super
excited
to
re-watch
this
recording
myself,
but
two
things
also,
you
can
always
give
feedback
on
this
call
on
that
google
form
if
you
want.
If
there's
anything
you
want
to
say,
I
also
won't
know
who
you
are
if
you're
like
this
was
terrible
fix
things
or
if
you're
like
this
is
great.