►
Description
On May 27th, the Clojure visual-tools group had its 18th meeting.
https://scicloj.github.io/docs/community/groups/visual-tools/
Agenda:
On the main part, Jérémy Vuillermet presented Mosaic, the prototype of a data-oriented design tool built in clojurescript and inspired by Bret Victor’s Dynamic visualisations.
Kira McLean shared an update about the Clojure data cookbook.
Daniel Slutsky shared an update about the Clojure data scrapbook using Kindly-Clay-Quarto.
Moderator: Teodor Heggelund
Summary at Clojureverse:
https://clojureverse.org/t/visual-tools-meeting-18-mosaic-and-other-updates-summary-recording/
A
Welcome
to
to
YouTube
this
is
the
beginning
of
cycloser
visual
tools,
meeting
18,
and
we
will
start
with
a
round
of
introductions
so
Daniel.
Why
don't
you
go
ahead
and
introduce
yourselves.
B
A
So
I
guess
on
my
screen:
Elena
you're,
the
next
one.
We
just
follow
the
order.
Would
you
introduce
yourself.
C
Hi
I'm
Juliana
I
teach
computer
science
at
the
University
of
Minnesota
Morris
and
I've
been
working
for
quite
a
while
on
exploring
beginner
friendly
tools
for
teaching
closure,
in
particular
error
messages,
so
modifying
error
messages,
but
also
other
things
and
several
of
my
students,
former
students
became
members
of
closure.
Community
worked
in
various
grocery
companies.
A
D
Hello,
my
name
is
Brian
I
am
a
software
developer.
I
use
closure,
primarily
I
work
at
a
company
called
Bright,
Health
and
I
love
closure.
A
Okay,
perfect
I
aim,
if
I
pronounce
your
name
correctly,
would
you
like
to
introduce
yourself?
A
Okay,
no,
no
introduction
from
Iron
to
data
Kira
who
who
is
Kira.
E
Hey
yeah,
I'm
Kira
I,
write
closure
code
by
day
for
publishing
data
on
the
internet
with
a
team
based
out
of
the
UK
so
mostly
like
back-end
web
services
and
that
kind
of
stuff,
but
I'm
interested
in
this
whole
data
science,
world
and
and
closure
stack
for
data
science
and
am
trying
to
help
build
some
like
educational
resources
for
that
for
kind
of
helping
people
get
started.
E
I
think
there's
like
a
lot
of
really
awesome
tools
that
are
pretty
much
ready
to
go
and-
and
you
know,
maybe
need
some
polishing,
but
I-
think
one
of
the
big
gaps
in
the
ecosystem
is
like
just
kind
of
like
guides
and
tutorials
and
books
and
courses
and
stuff
like
that.
E
So
I've
got
too
many
projects
underway
in
that
kind
of
vein
and
then,
like
part
of
it,
is
a
little
bit
selfish
too
I'm,
like
kind
of
using
it
as
an
opportunity
to
simultaneously
like
learn
all
of
these
tools
and
and
things
because
I
I'm,
not
a
data
scientist
but
I,
find
that
stuff
really
cool
so
yeah
anyway.
That's
me:
I'm
I'm,
based
on
the
east
coast
of
Canada
too,
so
joining
you
from
a
lovely
sunny
day
in
Nova
Scotia.
A
Yeah
I
think
at
some
point
we're
all
learning
together.
F
I
am
Jeremy
from
France
first
time
I
attend
this
meeting
and
I've
been
doing
closure
mostly
closure
script
as
a
front-end
developer
for
I,
don't
know
eight
years,
maybe
and
I'm
walking
at
pitch.
No
for
so
representation
software
building,
closure,
enclosure
script
and
where
I've
been
a
bit
for
five
years.
G
G
Hi
yeah
I'm
Tim
I
build
a
diagram,
app
called
Hami
and
mainly
just
because
I,
like
graphs
and
diagrams,
so
I
started
the
past
episodes
that
you
put
up
online
for
this
group
and
it
had
a
lot
of
really
interesting
topics.
So
yeah
I'm
really
excited
to
hear,
hear
what
gets
discussed
today
and
meet
some
people.
A
Okay
yeah,
my
name
is
theater,
so
the
brain
and
Power
behind
this
is
Daniel.
He
just
asked
me
to
to
talk
today,
so
we
could
not
say
anything.
I
worked
as
a
developer,
a
civil
engineer
and
as
a
product
manager
and
I
want
to
use
closure
for
data
science.
So
I'm
really
happy
to
see
the
ecosystem
evolve.
A
I'm,
presenting
a
talk
on
the
upcoming
babashikov,
which
is
called
build
your
own
little
manics
with
Babushka,
where
I'm
going
to
talk
about
building
websites
and
building
knowledge
together
and
I
live
in
in
Norway
yeah.
So
today,
on
the
schedule
we
are
first
going
to
hear
about
Jeremy
presenting
his
tool
in
protra
Mosaic.
A
A
So
without
further
Ado
Jeremy.
Would
you
like
to
present
your
prototype
yeah.
F
A
A
Yeah,
so
my
notes
say
about
30
minutes,
so
I
think
it's
good
to
not
spend
too
much
time
because
we
have
different
things
but
but
take
your
time
and
if
there
are
questions,
I
think
it's
good
to
trust.
F
F
So
this
is
a
pitch
and
I'll
just
show
pitch
to
just
to
kind
of
show
the
what
kind
of
work
I'm
doing
in
the
daily
basis,
so
I
joined
earlier
early.
So
as
a
front-end
developers,
I
worked
on
initially
on
the
canvas
and
then
moving
things
around
resizing.
All
that
the
editor
part
then
I
spent
quite
a
lot
of
time,
thinking
about
layouts
and
how
to
have
some
layouting
capabilities
like
swapping
or
aligning
or
Distributing
Without
Really,
adding
too
much
complexity.
F
So
it
stays
like
a
presentation
tool
and
doesn't
become
the
design
tool
or
something
else,
and
then
I
walked
on
the
recording
part
where
you
can
record
yourself
and
I
have
your
face
on
a
slide
and,
lastly,
working
on
animations,
so
yeah,
just
just
to
show
and
I.
Like
really
I,
don't
know
at
pitch
sometime,
we
say
the
front
end
of
the
front
end
and
yeah
this
my
other
project,
it's
here.
F
So
this
is
a
mosaic
prototype
I'm
working
on
a
lot
of
it
is
inspired
by
Brad.
Victor
I've
been
watching
his
videos,
I,
don't
know
every
every
year,
every
two
years,
I
watched
again
to
a
point
where
I
don't
really
know
anymore,
which
ideas
I'm
calling
from
me
originally
or
him.
F
So
you
might,
if
you're
familiar,
you
might
see
a
lot
of
similitude,
but
the
way
it
works
is
there
is
a
canvas
on
the
right
part
when
I
can
draw
a
rectangle
circles
and
a
row
for
now,
so
it's
still
a
prototype
because
I'm
not
planning
to
do
or
two
ship
a
product
at
this
point,
I'm
just
really
curious
to
see
with
this
concept
how
far
I
can
go.
So
that's
why
this
limit
limited.
The
number
of
shapes
is
enough.
F
Then,
on
the
left
left
we
have
the
data
browser,
so
it's
really
all
kind
of
data.
It
can
be
any
any
type
of
data
and
for
convenience,
it's
splitted
in
two,
although
they
work
the
same
way
but
at
the
top.
Those
are
the
elements
of
the
designs
and
the
rest
is
all
the
data.
So
I
have
the
the
sheer
curl
and
the
rectangle
here,
and
this
is
the
input.
F
So
if,
when
I
want
to
change
a
value,
let's
say
the
text
here,
I
can
type
here,
because
usually
there
are
long,
URL
or
formulas
and
similar
to
a
spreadsheet
at
the
top
it's
easier
to
to
manipulate,
although
the
ux
or
the
UI
I'm
not
sure
yet
so
so
far
it
works
like
maybe
figma
or
a
normal
design
tool.
You
would
know,
or
you
might
know,
but
the
specificity
is
the
the
in
the
the
value.
Sorry
can
be
any
formula
so
I
can
say
hello
world.
F
And
I
have
yeah
my
text
here,
so
there
are
a
limited
set
of
formulas
right
now
that
are
function,
that
it's
supported,
I,
just
hand-picked
sum
and
created
some,
but
also
we
can
reference
other
values
by
using
a
pass.
So
the
time
here,
for
example,
I,
could
reference
that
with
the
path
directly
in
a
vector
so
time
that
seconds
and
I
would
get
the
the
seconds
here
and
it's
updating
I
could
also
to
go
faster.
F
I
could
click
the
text
here
to
link
and
then
I
link
to
the
minute.
For
example,
it
does
the
same,
but
it's
a
little
bit
faster
because
that's
something
we
do
a
lot
so
at
a
high
level,
the
easy
West,
the
easiest
way
to
think
about
it.
F
In
my
opinion,
is
this:
although
it
doesn't
look
like
a
spreadsheet,
it
is
like
a
spreadsheet
with
two
key
differences,
one
that
they
really
have
names
instead
of
just
coordinates
and
a
2d
two-dimension
table,
so
things
have
names
which
have
some
trade-offs,
sometimes
better,
sometimes
not,
and
it's
nested
Infinity
nested,
like
a
Json
data
structure
or
a
map
enclosure.
F
It's
not
as
easy
to
enter
new
data
as
in
a
spreadsheet,
you
have
to
come
up
with
a
name,
but
for
this
particular
use
case,
we'll
see
why
I
think
it's
quite
nice,
my
notes
till
so
here
I
will
again
connect
with
the
time.
So
the
text
will
connect
with
the
seconds
and
yeah.
Something
else
I
want
to
show
is
when
I
update
the
or
when
I
use
direct
manipulation.
So
this
prototype
really
has
to
focus
data
and
direct
manipulation
and
all
the
two
interact
together.
F
So,
for
example,
if
I
do
a
line
from
this
rectangle
to
this
circle,
what
I
just
did
is
created
a
new,
a
new
line,
but
the
X
Y
X1,
sorry
to
one
point
of
the
line
is
the
center
of
the
rectangle.
The
Y
one
is
the
center
vertical
center
of
the
rectangle
and
so
on.
So
just
by
directly
manipulating
things
on
the
canvas
I
actually
create
those
formulas
which
has
very
interesting
properties.
So,
for
example,
let's
say
that
I
want
to
do
a
a
clock.
F
First
thing
it
could
be
interesting
also
is
to
I
can
connect
the
height
with
the
width.
This
way
so
I
make
sure
this
is
a
circle
and
now,
when
I
change
right
so
I
can
create
create
some
constraints.
This
way,
although
try
to
avoid
saying
constraints
because
kind
of
just
the
connotation
is
bit
negative
or
too
complex.
F
So
let's,
let's
try
to
do
a
clock,
so
I
can
make
a
line
like
this,
and
if
we
look
at
the
property,
we
have
a
rotation
which
is
the
angle
right
now
between
the
two
points,
but
I
can
update
that.
So,
if
I
do
zero,
it's
horizontal
but
I
can
connect
it
to
the
the
time
again
so
the
seconds.
So
this
will
move
this
way.
I
need
to
multiply
by
six,
so
I
get
the
correct
offset
every
time
and
then
I
subtract
90
because
of
the
initial
rotation.
G
F
Yeah
I
get
that
one.
You
know
yeah
better.
F
If,
if
you
have
a
clock
that
runs
like
that,
maybe
say
something
wrong,
but
yeah
I
think
what
does
it
do
wrong.
F
F
Correct
I'm
in
a
normal
Universe
again,
so
that's
really
the
the
basics
and
my
hypothesis
is
that,
just
with
that,
like
a
slightly
improved
or
different
spreadsheet
connected
to
a
canvas,
we
can
actually
do
a
lot
and
do
a
lot
of
what
I
would
consider
is
programming
as
a
front-end
developer.
That's
mostly
what
I
do
all
day
long
getting
the
data
in
and
then
Translating
that
to
visuals.
F
So
the
the
data
is
very
important
and
I
think
in
a
in
a
closure,
Meetup
or
group.
It
can
be
nice
to
to
talk
about
that,
because,
when
I
presented
that
in
a
in
a
London
Meetup
a
few
weeks
ago,
I
had
a
lot
of
I
forgot
that
doing
something
in
closure
was
maybe
not
not
expected,
because
I've
been
in
a
closure
bubble
and
at
pitch
for
so
long
that
I
didn't
expect
to
have
so
many
questions
of.
Why
are
you
using
closure?
F
Because
it
was
like
the
future
of
programming
Meetup
in
London,
so
people
who
are
mostly
using
JavaScript
I
think-
and
it
made
me
think
because
initially
I
said
yeah
just
because
that's
the
programming
language
I
know
but
I,
don't
think
it
giving
it's
giving
me
an
edge
for
this
particular
prototype.
I
could
have
done
it
with
JavaScript,
but
it
got
me
thinking
more
also
because
I've
seen
other
tools
similar
to
this
one
and
they
were
all
made
with
the
elect
of
lisp
or
some
existing
one
or
some
made
up
one.
F
So
I
thought
they
were
common,
a
common
point
and
I.
Think
that's
probably
data
right,
we
say
like
closure
is
a
data
first
language
or
we
like
to
manipulate
the
data
and
now
that
I
think
about
it.
I
think
this
is
playing
the
same
principles
about
data
to
a
UI
or
a
product.
So
the
first
thing
is
that
the
data
is
always
visible.
We
don't
necessarily
need
an
abstraction.
G
F
Can
even
hurt
to
have
a
different
abstraction
and
try
to
have
the
data
behind
different
panels
or
different
type
or
different
different
settings.
All
the
data
here
is
exposed
and
it
has
some
Traders,
but
we
can
try
to
see
what
it's
good
at.
So,
let's
try
not
to
bury
the
data
and
similar
to
spreadsheet,
like
everything
is
visible.
So
you
already
know
you
always
know
how
the
system
works,
because
everything
is
there.
F
There
is
nothing
hidden
and
then,
but
did
I,
and
from
that
there
are
quite
a
few
benefits,
so
it
composed
well.
So,
for
example,
I
can
create
a
new
viable
here
and
if
I
connect
that
to
the
time,
for
example,
because
time
point
to
another
map,
you
can
just
expand
this
way
and
it
compose
nicely
and
we'll
see
how
and
another
benefit
is
that
I
think
I
really
like
from
closure,
script
and
I.
F
Think
that's
what
sold
me
when
I
discovered
it
or
closure
in
general,
like
if
you
want
to
talk
to
an
API
and
you're
getting
Json
back.
I
was
coming
from
Scala
and
I
was
learning
ascal
at
that
point
and
you
would
have
to
do
some
crazy,
in
my
opinion,
sometimes
mapping
or
like
not
the
exact
shapes
of
all
the
Json
to
get
the
correct
mapping
where
enclosure
you
just
do.
Js2
CLG,
like
I,
think
that's
what
you
get
when
you
have
a
data
in
that
program.
F
So
it's
just
a
normal
call
and
returning
just
Json
I'm,
not
processing
it
at
all
and
I
can
create
a
weather
variable.
There
is
a
API
function.
I
will
do
that
for.
G
F
And
again,
because
this
resolve
into
more
data
I
can
just
inline
it
and
now
use
this
data
as
part
of
by
drawing.
So,
if
I
remove
that,
for
example,
the
example
I
like
to
do
is
I
won't
do
the
the
full
weather
which
the
one
I
also
I
published,
but
I'll
just
show
the
the
basics.
So,
for
example,
if
we
look
at
currents
whether
now,
let's
touch
here
in
the
API,
they
show
how
to
show
a
the
pictogram
for
the
current
weather.
F
G
F
No,
it's
updated
with
the
current
weather
in
France.
Let's
remove
the
Border
like
that
and
now,
let's
quickly
add
the
actual
temperature
so
because
it's
so
common
to
want
to
show
the
text
of
some
probabilities.
I
can
click
here
directly
link
and
then
start
drawing
a
rectangle
which
then
would
be
the
the
value
font
size,
20.
G
F
Drawing
from
this
point,
so
I
use
some
elements
of
the
the
design
to
help
me
create
those
relationship,
and
one
was
an
important
part-
is
the
what
I
call
the
collections
so
in
in
some
other
demo.
I
I
made
this
one,
but
I
think
this
one
is
also
interesting.
It
shows
the
the
pretty
precipitation
and
we
can
do
that
quickly,
so
we
have
the
data.
F
So
sorry,
let's
check
again
so
here
what
we
have
is
the
hourly
precipitations.
Here
we
have
a
level
every
three
hours,
I
think
it
was,
and
then
one
rectangle
for
every
hour.
One
data
point,
but
I
don't
have
that.
G
E
F
I
have
only-
and
this
is
a
list
and
if
I
just
want
to
connect
to
one
precise
like
the
first
one
I
would
go
to
zero
and
then
start
connecting
thing
from
there.
But
if
I
want
to
do
things
at
repeat,
I
use
the
index
because
which
you
then
why?
But
then
the
formula
will
use
the
index.
So
let's
do,
for
example,
humidity
and
I
can
start
to
draw
from.
G
F
F
So
now
I'm
connected
to
all
the
all
the
humidity.
It's
not
exactly
precipitations,
but
precipitation
is
zero
Always
Somewhere
over
here
and
what
I
actually
want
to
do
is
to
connect
the
the
height
right
with
the
the
humidity
like
that.
So
now,
I
have
this
value.
Let's
move
that
to
the
top
beat.
F
Sorry
and
to
make
it
smaller
I
can,
if
I
make
one
smaller
like
it's,
it
will
scale
all
of
them.
Let's
move
that
down.
So
if
I
change
the
the
first
one
I
can
draw
it
like
I
change,
the
bass
and
everything
move,
but
if
I
change
this
one,
for
example,
I
can
control
the
the
spacing
and
yeah
the
the
size.
F
Hopefully
I
didn't
maybe
edit
crash
it.
Let's
see,
but
now
it's
in
the
in
the
wrong
direction.
I
can
click
here
and
unfortunately,
I
change,
something
and
then
I
get
not
the
number.
But
since
the
data
is
always
there
and
sometimes
I
get
it
wrong.
I
can
always
change
that
and
I
don't
want
the
center
X.
That's
the
issue.
Shouldn't
change
I
want
the
left
to
be
Shield,
say
so
now
it's
back
here
so
ideally
yeah
I
think.
Let
me
start
again
so
again.
Humidity.
F
F
G
F
And
the
last
part
was
yeah.
What
I
find
interesting
as
well
is
here
in
in
another
tool
on
maybe
HTML
or
CSH
would
have
to
to
make
the
border
and
think
about
that.
But
I
don't
think
if
that
I
think
a
border
may
be
the
concept
from
from
developers
or
people
used
to
CSS
or
you
could
think
as
well
as
actually
I
just
want
to
create
a
line
from
here
to
here.
G
F
You
can
see
how
it's
slow,
because
it's
calculating
all
the
dependencies
while
dragging
there
is
no
caching,
so
it
has
to
calculate
everything
and
that
I
worked
again,
not
number.
F
Like
that,
and
the
last
thing
we
can
do
for
this
demo
is
adding
the
the
number
again,
but
let's
say
below,
as
a
text
like
here
like
that,
and
we
can
see
that
the
the
text
that
I
didn't
show,
that's
why
we
need
the
index.
So
it's
the
weather,
hourly
index
like
a
closure,
pass
and
then
I
get
the
correct
value
and
if
I,
for
example,
want
to
show
once
every
three
hour
I
can
do
that.
And
then
the
center
of
the
text
is
connected
to
the
center
of
the
rectangle.
F
F
So
yeah,
that's
what
I
wanted
to
show
Mostly
you
about
yeah
what
happened
or
what
I
want
to
to
see
what
happened
when
we
expose
the
data.
We
just
have
a
small
formula
and,
very
importantly,
that's
the
difference
between
the
two
live
scene
and
Brad
Victor
presentations
really
connecting
to
actual
API
and
be
able
to
to
handle
Json
to
or
any
other
things
so.
I
can
also,
for
example,
connect
to
chat,
TPT
or
open
Ai
and
ask
for
data.
F
A
For
me,
I
think
that
was
that
was
quite
amazing.
According
to
my
my
schedule,
we
have
eight
minutes
for
for
questions,
for
sticking
to
to
half
an
hour,
I'm
just
going
to
go
ahead
and
ask
one:
how
do
you
handle
cyclic
dependencies?
What
happens
if
I
make
two
things
that
depends
on
each
other
yeah.
F
Right
now,
with
trash
I
mean
the
the
library
that
I
use
would
have
an
exception,
circular
dependency
and
blank
page,
so
yeah,
probably
one
of
the
better,
not
that
like
tell
the
user
like
in
a
spreadsheet
but
yeah
yeah
right
now,
it
would
be
an
exception.
So
I
know
it
I
try
to
avoid,
but
sometimes
I
create
those
Myself
by
clicking
on
the
right
wrong.
Dots,
yeah.
A
Nice
are
there
any
other
questions,
just
unmute
your
mic
and
ask
if
you
haven't.
B
Yeah,
so
wonderful,
could
you
tell
maybe
a
bit
more
about
the
data
model
behind
the
scenes
about
how
things
are
represented.
F
So
it's
using
reagent
at
the
moment
so
I
have
a
big
map,
which
is
the
entire
state
of
the
yeah
I
can
show
that
the
entire
state
of
the
the
app
which
you
can
see
so
it
does
all
the
data,
but
I
also
exposed
the
data
about
the
editor.
So,
for
example,
the
mouse
position
everything
so
right
now,
I'll
pick
me
debug
things,
but
I
think
it
can
be
further.
F
F
When
I
want
to
add
something,
I
have
this
spacing,
because
usually
you
want
to
to
connect
things
like
that.
This
is
using
this
variable.
So
if
I
want
to
change
the
spacing,
you
can
create
a
design
system.
This
way,
for
example,
so
this
will
update
the
spacing
and
now
I
will
have
consistent
spacing
and
you
could
have
a
colors
or.
G
F
So
that
there
is
a
this
big
map
of
data,
which
is
both
the
data
from
the
user,
the
data
from
the
API,
the
data
from
in
this
case,
the
editor,
what
I
use
then
I
use
one
Library
I,
don't
remember
the
name
to
get
a
topic,
topological
sort
of
all
the
the
dependencies
between
the
yeah,
the
values.
F
So
what
is
important
as
well
is
that
the
dependencies
are
value
level,
not
as
a
entity
or
so
it's
more
fine
grain
and
have
less
sheikhilar
dependencies
as
well,
and
once
we
have
the
the
topological
thoughts
again.
Yeah
resolve
all
the
the
values
one
by
one.
This
way
and
and
get
the
final
thing
that
I
draw
and
because
it's
a
reagent
atom
as
soon
as
I
update
it,
it's
updated,
I
can
also
save
it.
We
should
do
a
locally
on
local
storage.
F
Just
edn
do
a
read
string
and
that's
that's
back
and
then
the
last
point.
Maybe
for
the
the
formulas
there
is
no,
that's
the
the
part
that
I
have
the
list.
Experience
with
so
I'm,
not
sure
I
can
probably
do
other
things,
but
right
now
it's
just
passing
the
Indian
dress.
With
string
and
then
I
get
a
vector
for
the
formulas
and
then
I
pass
the
vector
and
then,
when
I
find
a
symbol
that
I
that
I
know
I
call
the
function.
F
A
Yeah
nice
are
there
any
more
questions.
D
Is
this
currently
like
available
on
GitHub?
Is
this
something
I
can
run
locally.
F
Yeah,
thank
you.
Yeah
yeah.
If
you
have
use
cases,
I
would
also
be
curious
because
right
now,
I
consider
it
as
a
technology
like
because
of
what
I've
been
doing
at
pitch
and
what
I
was
doing
before
I
found
a
gap.
My
technology
got
bit
before
finding
a
use
case
for
myself.
I
have
kind
of
an
initial
use
case,
but
yeah
if
in
some
kind
of
data,
visualization
data
trying
to
get
in
contact
with
some
data,
visualization
data
designer
or
even
information
designer,
sometimes
it's
even
closer
to
what
I
need
to
see.
F
If
something
like
that
exists,
I
would
help
but
yeah.
So
anything
in
that
area
would
be
welcome.
A
Okay,
other
more
questions.
B
G
F
Yeah,
so
the
if
I
was
to
to
publish
it
to
publish
it
summer
right
now,
with
what
I
would
I
think
he
would
be
hosted
and
then
what
you?
The
output
that
you
have
at
the
end,
can
be
directly
published
in
a
newer
in
a
URL
and
then
that's
what
you
would
send
to
people
and
then
trying
to
also
keep
the
you
can
hide
the
data
panel.
If
you
just
want
to
to
ship
something,
I
think
it
could
even
be
used
to
do
some
kind
of
website
with
very
light.
F
Layouting
wouldn't
be
responsive,
but
it
can
go
in
that
direction
as
well.
A
small
website,
for
example,
budget
or
a
small,
something
that
you
can
embed
in
a
in
a
blog
or
something
like
that.
You
would
just
get
the
URL
from
this
hosted
and
then
display
it
in
place
and
yeah.
F
G
A
A
Half
an
hour
so
I
unfortunately
have
to
leave
a
bit
early
because
I
missed
the
I
I
made
a
mistake
on
the
on
the
time
when
this
started
so
Daniel
will
will
take
over
as
facilitator
in
the
middle
of
Kira's
presentation.
E
Sure
yeah
thanks
so
much
yeah,
so
there's
not
like
a
ton
of
stuff
to
present,
but
I
am
curious.
I'll
show
what
basically,
what's
what's
there
so
far,
but
I'm
I'm,
mostly
curious
to
like
just
kind
of
get
like
feedback
from
people
in
the
community.
E
A
couple
like
big
things
or
like
whatever
high
level
things
I'm
like
thinking
about
these
days
are
like,
first
of
all,
how
to
like
publish
the
book
and
how
to
kind
of
distribute
it
which
I
think
Daniel's
work
around
like
all
the
quarto
stuff
is
like
really
interesting
and
exciting
and
then
I'm
also
I,
don't
know
I'm
going
back
and
forth
a
lot
on
like
the
sort
of
organization
of
the
chapters
like
because,
like
it's
a
classic
thing
where,
like
you,
you
write
something
or
make
something
and
then
compare
it
to
other
similar
things
and
you're
like
mine
is
different,
it's
worse
but
anyway.
E
So
okay
I'll
share
my
screen.
First
I'll
pull
it
up
somewhere,
so
so
far,
I've
just
been
playing
with
Clerks
like
capable
excuse
me
capabilities.
So
this
is
like
a
sample
of
this
is
what's
kind
of
on
the
main
branch.
So
far,
and
obviously
eventually
the
idea
is
to
get
oh
I,
don't
have
it
up,
maybe
I
don't
have
so
many
things
open
anyway.
E
Eventually,
obviously,
this
will
be
like
a
nicer
like
landing
page
or
whatever,
but
right
now,
basically,
clerk
does
have
like
a
a,
not
half
bad
way,
to
publish
like
multiple,
multiple
chapters
of
like
one
document.
So
so
these
are
the
namespaces.
Basically
excuse
me,
oh
the
headings,
and
this
one
are
broken,
so
there's
no
like
table
of
contents,
but
there's
obviously
like
stuff
missing
so
like
if
you
compare
this
to
like,
for
example,
this
style
book.
E
So
this
this
is
the
r
for
data
science
book,
which
is
generated
with
Cordo,
which
I
I
really
like,
because
then
there's
just
like
a
lot
of
little
things
that
make
it
really
nice
for
navigating
a
lot
of
content.
E
So
like
there's,
these
page
turn
things,
for
example,
which
is
like
a
small
thing,
but
it's
like
non-trivial
to
hack
this
in
after
the
fact
there
is
this
like
little
internal
table
of
contents
here
on
the
side
which
I
really
like,
and
it
kind
of
like
updates
itself
as
you
scroll
and
then
there's
really
good
search
capability.
E
So
if
you
search
for
like
a
bar
chart
or
whatever
it'll
oops
it'll
bring
up
these
like
relevant
and
useful
results,
and
so
this
is
obviously
like
much
more
pleasant
to
navigate
like
as
a
user
compared
to
this.
This
is
pretty
good,
and
this
is
a
start,
but
you
know
there's
there's
a
lot
of
a
lot
of
work
to
do
so
anyway.
E
I
guess
yeah
in
terms
of
the
like
high
level
and
and
that's
the
other
thing
these
right
now.
The
namespaces
basically
are
not
linked.
So
the
the
way
to
link
the
way
to
sort
of
navigate
is
to
go
back
to
the
index
scroll
to
the
top
go
back
to
the
index
and
then
click
on
the
next
one
that
you
want
to
look
at
so
anyway.
E
There's
that
so
one
thing
I
guess
like
I,
don't
know
I,
guess,
I'm
curious,
I'm,
curious,
first
of
all,
like
if
there's
like
I
find
I
find
Clark,
really
really
cool,
but
it's
I
find
it
complex.
A
lot
of
things
like
there's,
there's.
Obviously
the
really
nice
like
notebook
like
auto,
updating,
like
caching
mechanism
or
whatever,
that
they've
implemented,
which
is
like
a
really
cool
piece
of
software.
E
Excuse
me
that
gives
you
this
like
illusion
of
like
re-running,
the
entire
name
space
from
top
to
bottom.
Every
time
you
change
anything.
So
it's!
It's
really
nice
because,
like
your
code
always
executes
in
the
right
order,
like
everything
always
runs
as
expected
in
the
right
order,
I
mean
assuming
the
the
caching
works
like,
except
for
the
times
where
the
caching
breaks
down.
E
But
then
you
can
just
blow
it
away
and
restart,
but
there's
that
so
there's
like
the
caching
stuff
and
the
namespace
rendering
stuff,
but
then
there's
this
whole
other
world
of
like
I,
mean
there's
a
whole
separate
concern
about
like
interactive
stuff,
which
I
guess
I
haven't
even
touched
on,
but
then
there's
also
like
this
publishing
stuff.
So
like
publishing
these
artifacts
and
I
don't
know
I'm
torn
because,
like
I
like
I
like
it,
but
I,
also
I,
don't
like
how
it's
like
all
all
kind
of
in
one.
E
If
you
know
what
I
mean
so
anyway,
like
like
it's
crossed
my
mind
to,
for
example
like
like
contribute
stuff
to
clerk
to
like
make
this
kind
of
UI
like
this
is
obviously
possible
to
have
this
kind
of
like
master
table
of
contents
here
and
then,
basically,
in
clerk
like
these,
these
little
side
table
of
contents
are
what
this
is.
So
this
is
what's
over
here
and
then
this
one
this
one
on
the
left
is
like
the
the
index.
What
what
is
this?
E
So
this
is
really
this
and
then
the
search
is
is
equal
to
so
anyway,
so
yeah.
So
that's
that
sort
of
stuff
I'm
thinking
about
in
terms
of
like
how
to
actually
like
disseminate
the
content,
because
it's
at
a
point
now
so
this
is
not
everything
is-
is
pushed
to
the
main
branch,
but
so
far,
there's
a
bunch
of
stuff
written
now
for
basically
getting
data
like
in
and
out
of
notebooks
and
like
tidying
it
up
a
little
bit.
E
So,
for
example,
like
databases,
random
sequences,
there's
a
little
bit
of
text
to
fill
in
here,
maybe
explaining
things
a
little
bit
better,
but
I
am
trying
to
keep
it
like
fairly
concise
and
focus
more
on
the
code,
as
opposed
to
like
really
verbose
like
like
the
goal
is
not
to
be
like
a
really
verbose
collection
of
kind
of
blog
posts
or
like
tutorials.
E
The
goal
is
to
be
more
of
like
a
like
a
reference
guide
kind
of
thing
for
people
like
if
you
already
kind
of
know
what
you're
looking
for-
and
you
just
want
to
know
like
the
syntax.
The
ideas
you
can
go
be
like
how
do
I
get
data
from
like
a
like
a
database
into
my
notebook,
and
then
it'll
just
be
like
this?
Really
short
kind
of
collection
of
examples
or
whatever,
so
that's
the
idea,
so
these
are
the
ones
that
are
oops
I,
keep
see,
I
keep
expecting
that
to
go
back
already.
E
But
yeah
so
anyway,
so
that's
that
this
one
needs
better
headings,
but
this
is
a
basically
the
chapter
about
kind
of
tidying
up,
misformatted
data
or
as
far
as
data
or
whatever
that
kind
of
thing,
and
then
this
is
for
getting
data
out.
So
there's
other
things
like
probably
I'll
hide.
E
So
so
that's
that's
basically
where
it's
at
so
I'm
curious.
If
anyone
has
ideas
about
like
publishing
like
multi,
multiple
pages
of
documents,
as
one
resource
is
one
kind
of
topic
and
then
the
other
topic
that
I'm,
like
waffling
about
a
lot
in
my
mind,
is,
is
how
the
stuff
is
laid
out
so
like
you'll
see
so
this
is.
This
is
just
one
similar
kind
of
book
that
I've
been
like
using
as
a
reference
the
data,
the
the
content
here
is
sorted
a
little
bit
differently.
E
E
Excuse
me,
but
yeah,
like
I
I,
find
it
a
little
counterintuitive
because,
like
obviously
you
can't
like
visualize
your
data
until
you
like
import
it
and
clean
it
up
and
and
transform
it,
and
so
the
way
I
had
laid
out
the
content
so
far
is.
E
This
is
bad
example,
because
it's
not
all
here
but
like
it
goes
basically
like
there's
a
little
intro
and
then
loading
data
and
like
getting
data
just
in
and
out
of
of
a
closure,
notebook
or
namespace
and
then
cleaning
it
up,
hiding
it
kind
of
like
yeah
like
manipulating
the
data
and
then
dataviz
and
the
database
chapter
I.
Don't
think
it'll
be
done
for
a
long
time.
E
There's
like
a
million
questions
about
how
to
do
that
best,
but
yeah
anyway,
I'm
just
curious
like
if
you
were,
if
you
were
the
kind
of
person
who
would
use
a
resource
like
this,
like
what
like
what
would
be
useful
because
I
can
see
where
he's
coming
from
like
this
is
probably
like.
E
Basically,
database
is
really
interesting
and
fun,
and
so
it's
like,
if
you
want
to
get
people
into
something,
it's
like
you,
just
Dive,
Right,
In
and
start
making
graphs,
and
so
it's
it's
like
useful
to
have
that
like
upfront,
but
it
assumes
a
lot
of
stuff
like
you
need.
You
need
like
a
loaded
cleaned
up
like
tidy
data
set
before
you
can
even
think
about
visualizing
it
so
anyway,
I
think.
That's
that's
mostly
what
I
wanted
to
say:
there's
not
a
whole
lot
more
than
that
other
than
yeah
I.
E
Guess,
like
a
brief
status,
update
about
just
kind
of
how
it's
going,
which
is
slowly
but
yeah,
just
kind
of
curious
I,
don't
know
like
I
I
would
like
for
this
to
be
a
resource
like
kind
of
buying
for
the
community,
and
so,
if
there's
like
input
or
whatever,
like
I'm,
happy
to
happy
to
hear
it,
especially
on
these
on
these
particular
topics,
but
in
general,
I'm
always
open
to
any
any
other
thoughts
too.
So
yeah
I
think
I.
Think.
A
Yeah,
so
I
I
don't
have
any
thoughts
about
outline
and
structure.
I.
Think
I'll
have
to
try
to
read
the
r
book
and
read
your
stuff
before
I
have
an
opinion
on
that
yeah
on
on
clerk.
Specifically,
your
experience
matches
mine
it's
great
for
single
documents
and
for
multiple
documents.
A
I've
considered
trying
to
make
my
whole
website
with
clerk,
but
the
linking
part
is
just
iffy,
so
I
kind
of
like
having
control
of
that
for
myself
and
my
idea
over
the
long
term
for
trying
to
make
or
publish
my
own
documents
is
to
be
able
to
just
use
clerk
to
create
the
single
documents
either
just
as
data
or
an
edge
demand,
an
HTML
file
and
then
being
able
to
do
the
rest
myself
and
and
have
control
over
that
I
think
it's
possible
to
get
pandok
Json
out
in
the
middle
from
the
clerk.
A
Yeah,
rather
than
yeah,
buying
the
whole
cake
and
the
presentation
and
the
delivery
and
everything.
E
Yeah,
that's
the
thing
I
find
it's
like.
You
said
it's
a
lot
of
a
lot
of
concerns
like
all
kind
of
like
mixed
up
in
like
one
library,
and
so
that's
really
interesting.
I
hadn't
thought
about
using
pandok
because,
like
one
option
is
of
course
to
like,
because
this
is
obviously
just
a
list
of
HTML
pages,
so
you
can
take
these
pages
and
like
collect
them
with
like
just
like
a
babashka
script
and
like
make
your
own
thing
and
that's
one
option.
E
This
style
of
thing
like
a
book
with
pagination
and
like
tables
of
contents
and
master
tables
of
contents
and
search
there's
a
lot
of
types
of
content
that
fit
this
kind
of
like
desired,
like
artifact
and
anyway,
so
I'm
just
curious
how
other
people
are
doing
it.
So
that's
interesting,
I
hadn't
even
thought
about
panda,
but
that
obviously
would
be
a
good
one.
That's
what
Corda
uses
if
I'm,
if
I'm
not
mistaken,.
A
E
A
Ask
them
this
this
question,
because
I've
I've
heard
them
kind
of
say
about
this:
is
just
data
all
the
way,
but
then
I'm
looking
into
it
myself
and
I
kind
of
don't
understand
the
data
performance
so.
E
E
A
Interesting
anyway,
yeah
it
might
even
be
possible
to
do
do
something
less
I
I
think
they
have
their
own
markdown
or
sorry
yeah,
markdown,
parser.
A
Puts
but
there's
been
some
thought
about.
A
Some
other
stuff
I
think
that
maged
made
this
essay
about
Clerk
and
they've.
E
E
Like
some
sort
of
format,
that
would
be
like
easier
to
read
on
like
like
a
mobile
device
than
like
just
a
website.
So
that's
really
interesting.
Yeah
I
mean
I'm.
Sure
like
Martin,
has
a
lot
of
ideas
about
this.
It's
probably
a
really
common
thing,
so
it
makes
sense
to
talk
to
him
about
it,
but
anyway
that's
cool
thanks
for
your
yeah.
Thanks
for
your
thoughts,
yeah.
B
G
G
B
For
a
moment,
yeah
any
other
questions
to
Kira
or
any
other
comments
by
Kira.
C
C
But
for
beginners
I
think
the
ability
to
have
the
entire
code
base
and
see
the
entire
code
base
is
important
and
I'm
wondering
if
that
could
be
some
sort
of
toggle
feature
where
you
show
this
to
that,
and
that
or
some
way
of
sort
of
getting
the
entire
code
base,
because
I
know
that
students
often
Google
for
something,
and
they
would,
you
know,
see
some
fragment.
And
actually
that
was
my
experience
with
r,
often
where
I'm
reading
something
and
it
assumes
a
gazillion
of
different
things
before.
D
C
E
E
Would
love
I,
don't
know
if
it's
possible,
it
might
be
possible
to
do
something
with
like
Psy
or
ski
or,
however,
they
pronounce
it
like
I
think
it
would
be
super
cool
to
be
to
have
like
so
yeah
be
able.
You
know
how
you
see
those
like
little
code,
pens
and
stuff
where,
like,
if
you
see
a
JavaScript
snippet
on
the
internet,
you
can
click
like
try
this
out.
That
would
be
super
cool
to
have
a
little
button.
C
A
F
F
Yeah
click
I
never
tried,
yet
it's
not
there
yet
for
some
of
the
reason
you
mentioned
and
during
during
my
research
like
the
just,
maybe
as
inspiration
the
it's
called
dokusaries
I
think
it's
the
most
flexible,
maybe
modern,
that
I
found
not
sure
if
it's
possible
to
to
plug
it
noisy
but
I
think
yeah
and
they
have
the
the
same
UI.
As
you
mentioned,
like
the
search
on
the
left
and
eye
level
and
on.
D
A
E
E
G
E
It
should
be
pretty
straightforward,
so
yeah
the
goal
is
definitely
for
that
to
be
like
possible.
So
there's
a
like
it's
just
all
of
the
the
whole
project.
Right
now
is
just
a
collection
of
namespaces,
so
it's
here
in
the
cyclose
organization,
I
should
probably
like
write
a
little
bit
more
there,
but
there's
I
guess
this
anyway
yeah
these
are
just
like.
E
So
it
should
be
like
self-contained
there's
a
couple
external
things
that
that
could
possibly
break
down
like
we're,
using
like
dbpedia
for
some
of
the
examples
for
the
sparkle
stuff,
but
like
yeah,
it's
it's.
The
goal
is
for
that
to
be
to
remain
possible
and
I
would
love
to
add
that's
another,
like
long-standing
issue,
that,
in
the
back
of
my
mind,
there's
just
no
time
it's
enough
hours
in
the
day,
but
two
like
I
want
to
like
add
some
like
tests,
at
least
like
some
some
sort
of
basic
test
around
this
stuff.
E
So
that,
like
just
to
check
that
it's
working
because,
like
the
goal
is
for
this
to
be
like
useful
for
more
than
just
the
first
six
months
after
it's
released
like
as
libraries,
get
updated
and
changed
and
and
the
ecosystem
evolves.
I
would
like
to
keep
this
up
to
date
and
just
be
able
to
like
update
the
dependencies,
update
the
examples
and
then
have
like
a
test
Suite
to
just
double
check
that
they're
still
working
and
then
the
idea
is
like.
E
If
the
test
pass
and
the
book
renders
then
it's
it's
correct
like
because
it's
it's
just
closure
code
is
namespaces.
So
it's
not
like.
That's
that's
one.
The
one
thing,
maybe
not
the
one
thing,
but
one
of
the
main
things
keeping
me
from
switching
away
from
clerk
completely
is
because,
like
I
would
I
like
that,
it's
just
closure
name
spaces,
because
the
problem
is
like
if
I
switch
to
just
like
plain
markdown
files,
they'll
they'll
they'll
get
stale.
E
The
website
that
it
produces
is
not
super
polished,
but
anyway,
I'm
really
excited
to
hear
from
Daniel,
because
I
know
like
you've,
done
a
lot
of
work
on
this
like
exact
thing
so
yeah,
maybe
if
I
mean
I'm,
certainly
having
to
talk
with
this
forever,
but
I
also
don't
want
to
take
any
more
time
away
from
you.
B
B
D
E
I
should,
like
probably
add,
a
contributing
guide
or
something
to
the
thing,
but
yeah
I
think
there's
there
there's
like
a
range
of
like
a
spectrum
of
like
ways
to
contribute
so
like
the
thing
is
like,
like
most
people,
just
don't
have
time
to
like
write
a
chapter
or
like
write,
a
section
and
I
think
that's
one,
one
of
the
things
that
we
were
talking
about
solving
with
like
this
other,
like
little
scrapbook
project,
but
certainly
like.
E
E
That's
it's
like
a
big
commitment,
but
but
there's
anyway,
there's
there's
that's
like
the
far
end
of
like,
like
the
sort
of
extreme
end,
but
they're
like
there's
that
from
everything
all
the
way
back
to
just
like
reading,
through,
like
one
namespace
and
pointing
out
like
typos
and
mistakes
and
stuff,
like
that,
that's
another
really
helpful
thing
or
sort
of
like
kind
of
being
like
a
like
an
editor,
almost
and
and
just
like,
leaving
little
comments.
So
you
can
like
leave
comments.
You
can
make
pull
requests
against
the
repo.
E
Probably
the
best
thing
is
to
just
like
Ping
me
like,
and
we
can
talk
about
what,
like
whatever
you
want
to
do.
If
there's
something
in
particular
you
like
or
if
there's
something
they
particularly,
you
want
to
learn
like
I'm
talking
with
one
one
woman
right
now,
who's
like
like
interested
in
learning.
You
know
certain:
do
it
how
to
do
certain
things
with
closure
and
data.
So
it's
like.
We
can
kind
of
work
together.
E
So
it's
like
really
helpful
that
way,
so
yeah
I
think
like
if
I
guess
for
now
it's
it's
mostly
just
like
if
you're
interested
in
this,
like,
let
me
know
and
we'll
find
something
we'll
find
a
way
to
get
those
contributions
in
or
just
like,
like
a
couple,
people
have
made
like
small
pull
requests,
just
with
like
little
like
small,
Corrections
and
stuff,
like
that,
that's
always
welcome
to
and
yeah
the
goal
is
like
I'll
I'll
make
for
now.
E
I'm
really
I'm,
just
like
really
hung
up
on
the
whole
like
publishing
process,
I.
Think
for
now,
I'm
just
gonna
try
to
make
like
a
slightly
nicer,
landing,
page
and
I'll
put
that
on
the
internet
as
like
a
work
in
progress
as
I
keep
like
you
know,
adding
new
chapters
and
whatever
and
yeah
it's
just
it's
going
so
much
slower
than
I
wanted
to.
But
this
I'm
realizing
like
evenings
and
weekends,
is
just
like
it.
E
B
B
And
yeah
and
this
part
will
be
really
very
much
a
continuation
of
what
Kira
was
telling
about
I
think
you
can
see
my
screen
now
and
yeah
so
so.
First,
maybe
that
the
context
is
that
this
is
something
we
have
tried
in
the
past,
and
we've
had
actually
failed
in
creating
the
book
writing
process
a
few
times,
and
the
point
was
that
one
of
the
things
needed
was
this
kind
of
editorial.
B
Attention
somebody
who
is
continually
thinking
about
what
the
book
needs
and
kind
of
making
it
tidy
and
then
then
that
is
why
it's
so
exciting
to
hear
Kira
telling
this
story,
which
is
evolving
continually
and
in
actually
thinking
about
these
dilemmas
of
of
layout
and
all
usability
questions.
And
so
that
is
one
thing
and
other.
B
And
you
know
people
had
to
to
leave
for
a
while
and
it
just
didn't
haven't
been
working
for
us
and,
and
so
the
current
idea
that
we've
been
discussing
recently
is
that
yeah
there
is
the
closure
that
data
cookbook,
that
Kira
is
creating
and
facilitating,
and
maybe
there
is
another
piece
of
the
story
alongside
it
and
in
a
great
conversation
with
Kira
and
with
Blaine
moose.
It.
B
B
It's
the
entire
thing
that
is
evolving
as
the
the
nice
book
and
the
other
thing
is
more
wide
and
more
maybe
open
to
contributions
and
is
never
so
tidy
as
the
cookbook
and
we
may
need
both
and
so
now
I'll
kind
of
describe
our
efforts
in
creating
a
very
basic
structure
for
the
Scrapbook
that
will
encourage
contributions
and
gradually
allow
contents
to
flow
to
the
cookbook.
B
Hopefully,
and
I'll
just
describe
a
few
of
the
tools
we're
using
for
that,
and
is
it
making
sense
so
far,
any
any
comments
about
it,
but
this
Duality
between
two
projects
and
the
flow
between
them
right
and
yeah,
and
everything
I'm
talking
about-
is
very
much
kind
of
driven
by
the
guidance
of
Kira
in
the
last
few
weeks
and
and
I
happy
to
have
this
conversation
that
allows
to
kind
of
gradually
create
it,
and
so
yeah
of
the
structure
of
this
scrapbook
is
that
it
will
be
one
repo
that
is
very
much
welcoming,
pull
requests
and
is
so
open
that
it
is
almost
immediate
to.
G
B
To
it-
and
it
is
great
from
many
closure
projects,
because
you
know
teaching
a
certain
technique
or
a
certain
library
or
topic-
may
need
a
separate
project
for
it
for
the
dependencies
that
set
up
and
so
on.
So
it
will
have
many
projects
inside
and
it
will
also
have
a
book
which
is
the
rendering
of
these
projects
Again
by
this
approach
that
he
represented
of
the
namespace
as
a
notebook.
B
So
ideally,
every
name
space
of
code
will
be
a
page
in
the
book
in
the
resulting
book
right
and
then
the
projects
will
have
one
main
project,
which
is
the
easiest
to
contribute
to,
because
it
has
some
default
dependencies
which
are
okay
for
most
common
data
science
tasks,
but
it
have,
it
will
have
other
projects
also,
for
example,
this
hdf
project
will
be
about
handling
this
data
format
called
hdf
that
some
people
care
about.
So
it's
made
additional
dependencies
which
are
not
in
the
main
project.
B
B
B
So
this
annotation
of
kind
is
part
of
this
kindly
Library
and
the
idea
of
kindly
is
to
State
about
certain
values
that
they
should
be
rendered
in
a
certain
way.
So,
for
example,
this
thing
is
of
the
Vega
light
specification,
so
it
should
be
rendered
this
way
and
the
idea
is
that
we
may
have.
We
need
a
convention
that
different
tools
will
know
how
to
treat.
B
B
B
As
this
plot
of
this
vehicle,
like
specification,
but
I
could
also
send
it
to
be
visualized
in
Porta,
so
now
I
send
it
to
portal
this
amazing
tool
for
data
navigation
and
it
handles
it
well.
But
there
are
some
layout
problems
to
be
to
be
improved
in
the
Adaptive
portal,
but
the
idea
is
that
using
this
kindly
annotation,
we
can
have
different
tools,
respect
the
notion
that
something
should
be
rendered
in
a
certain
way.
So
we
we
saw
that
play
this
tiny,
simple
tool
for
data.
B
B
This
other
amazing
two
amazing
notebook
tool,
so
I
have
another
key,
binding
and
I
have
this
clear
notebook
that
is
updated
by
another
keybinding
right,
so
different
tools,
hopefully
in
harmony,
at
least
in
common
use
cases
of
common
things
like
a
Vega
light
plot,
and
so
we
saw
the
clay
a
single
value.
B
B
B
D
G
B
B
Is
maybe
one
of
the
editors
emacs
or
vs
code
and
Clay,
because
you
just
press
key
binding
and
you
get
it
visualized,
but
there
will
be
better
tools
and
one
of
the
exciting
ones
or
maybe
more
beginner.
Friendly
tools
is
the
Calvin
oakles,
which
is
not
supported?
Yes
in
the
current
setup.
But
that's
the
idea
that
we
will
create
things
in
a
compatible
way
and
that
we
will
encourage
contributions.
B
And
the
only
thing
we'll
ask
is
for
people
to
try
to
use
this
kindly
convention
when
they
can
so
that
things
will
be
friendly
for
Future
tools
like
album
books
and
flow
current
tools
like
the
other
ones,
and
that's
the
idea
of
the
Scrapbook.
And
the
idea
is
that,
hopefully,
after
some
things
mature
in
the
process
of
the
Scrapbook,
they
can
hopefully
move
to
the
cookbook.
If
the
cookbook
editor
accept
that
more
carefully,
if
it
makes
sense.
G
B
A
moment
Kira
will
have
some
comments
about
about
the
vision
of
that
side.
So
maybe
we
see
just
last
one
last
thing,
then
that
is
the
book
part
of
the
story.
So,
oh
sorry,
so
now
in
the
Scrapbook
we
have
this
book
directory
and
I
now
run
a
bus
script.
I
wrote
a
tiny
script
that
what
does
is
collecting
or
the
quote
of
documents
that
have
been
created
by
the
user
in
the
book,
so
I'll
just
run.
It
then
you'll
see
it's
collecting
those
q
and
d
files.
B
Qnd
is
the
markdown
live
format
of
of
the
quarter,
and
and
now
it
is
creating
the
book
of
all
these
quotes
or
documents
which
have
been
created
from
closure
and
expenses.
So
we
are
connecting
the
namespace
as
a
notebook
approach
to
a
whole
quarter
book
that
will
be
rendered
in
a
moment
and
at
the
moment
these
are
two
separate
processes.
Though
you
see,
we
did
have
one
stage
where
we
explore
things
dynamically
in
closure
and
render
them
Page
by
Page,
and
then
afterwards
we
can
collect
everything
to
into
one
book.
B
B
Contribute
namespaces
and
projects,
and
we
will
collect
them
to
this
chaotic
book
called
scrapbook
that
will
be
a
source
of
sun
content
that
would
gradually
move
to
the
crypto
codefully.
So
that's
the
situation
at
the
moment
and-
and
there
are
some
technical
troubles
and
and
many
questions
about
these
compatibility
hopes
which
are
not
solved
yet
because
those
amazing
tools
are
not
always
compatible
with
each
other
and
yeah.
So
that's
the
situation.
We
are
now
around
the
official
End
of
Time.
Some
people
may
need
to
go
I
see
some
people
that
did
need
to
go.
B
Maybe
what
we
should
do
is
have
some
short
comments.
Questions
by
anybody
who
may
need
to
leave.
B
B
So
maybe
I
wish
to
Germany
to
ask
Jeremy
one
more
questions,
one
more
question,
and
that
is
it.
You
know:
how
could
the
community
possibly
help
with
your
Amazing
Project
what
you
may
like
to
see
happening
in
terms
of
community
interest,
and
you
know
how
what
what
kind
of
of
feedback
or
or
questions
you
would
like
to
see
coming
from
people
regarding
your
project
and
but
if
it
is
at
this
stage
where
it
makes
sense.
D
F
Yeah,
it's
still
a
very
much
an
inspiration.
So
it's
hard
to
say,
I
think
the
what
I'm
most
looking
forward
is
I,
think
I've
I've
been
far
enough
or
what
I
wanted
to
achieve.
What
I
had
in
mind.
F
I
could
achieve
it
like
verify
that
okay,
you
could
really
visualize
and
create
information
software
like
with
the
system
in
a
flexible
way,
so
I
think
it's
mostly
again
about
the
the
use
cases
where,
if
those
kind
of
tools
exist,
like
the
simplest
way
like
spreadsheet,
plus
visualization
connected
what
and
a
way
to
privilege
that
easily
on
the
web,
what
use
cases
does
it
solve
and
I'm
still
not
sure
the
the
niche
use
case?
F
I
was
talking
about
because
I
was
I
used
to
be
a
professional
video
game
player
I
couldn't
strike
so
I
spent
a
lot
of
time
on
Twitch
and
I
developed
a
lot
of
tools
for
twitch
and
streaming
in
general,
and
that's
where
that's
I
know
where
it
would
be
a
use
case
already,
because
when
you
stream
usually
on
a
production
or
even
on
TV,
you
add
some
overlays
on
top,
so
those
are
non-dynamic
just
information.
On
top
of
your
overlays
of
your
stream,
maybe
the
player
that
are
playing
the
current
game.
F
If
it's
things
like
that-
and
this
would
work
well
for
that-
that's
where
I
was
coming
from
initially,
but
that's
one
use
case,
but
I
think
it
could
be
more
General
like
how
Brad
Victor
said,
like
really
information
software.
B
Yeah,
fantastic
and,
of
course,
Jeremy,
you
know,
I
think
we
all
have
this
feeling
that
there
is
so
much
more
to
hear
about
it
and
and
to
learn
about
your
implementation,
about
about
the
the
things
that
can
be
created.
So
of
course,
if
one
day
you
wish
to
present
the
game,
a
short
update
or
a
big
update
or
a
whole
meet
up
about
about
it.
Then
please
use
this
space
as
much
as
you
find
it
helpful,
because
people
will
be
enjoying
that
again
and
yeah.
E
G
Just
have
a
question:
I
I
think
it's
for
both
Kira
and
Daniel
about
the
book.
Publishing
process
specifically
like.
G
It
seems
like
you
can
have
code
that
contains
markdown
or
you
can
have
markdown
that
contains
code
and
I'm
a
little
bit
confused
like
which
is
better
or
like
have
you
tried
both?
It
sounds
like
currently
you're
erring
on
the
side
of
code
that
contains
markdown,
but
I'm,
not
I'm,
not
sure
I
understood
that
correctly.
So
could
you
maybe
just
talk
about
that
dialectic?
A
little
bit.
E
E
So
it's
like
yeah,
you
can
either
have
text
that
has
code
Snippets
in
it
or
you
can
have
code
that
has
like
text
Snippets
in
it
and
I
think
the
trade-off
is
like
the
the
pro
the
common
problem
with
the
sort
of
notebook
Style
with
with
like
a
text
document
that
has
code
Snippets
in
it
is
that
the
code
sentence
goes
stale
like
you.
Can
you
can
they
execute?
You
can
get
these
like?
E
You
can
get
systems
that
can
actually
execute
the
code
and
and
like
produce
little
Snippets
of
results,
kind
of
like
Jupiter
notebooks,
but
there's
there's
a
really
good
paper.
It's
in
the
clerk
documentation
about
I'm,
just
gonna
bring
it
up
about,
like
all
the
problems
with
notebooks,
which
explains
all
of
this
like
much
better
than
I
will
I'll
share
it
in
oh,
the
link
is
broken.
E
E
So
like
this
is
like
this
is
like
a
lot
of
the
common
complaints
like
summarized
in
a
really
elegant
way,
and
so
like
one
of
the
one
of
the
answers
to
that
like
set
of
complaints
or
whatever
is
to
like
do
like
basically
invert.
It,
like
you
said,
like
write,
just
write
code
and
then
like
render
the
comments
as
text,
and
so
that's
kind
of
what
clerk
does,
but
then,
like
the
flip
side
of
that,
is
that
it's
like
it's
like
annoying
to
write
a
book
as
comments
in
a
closure
file.
E
Obviously,
like
you,
don't
want
to
write
a
lot
of
text
that
way,
but
there's
there's
work.
Arounds,
like
you
can
write
for
the
the
chunky
text
Parts,
you
can
write
like
markdown
files
and
like
slurp,
those
into
your
namespace
or
something
like
that,
but
the
like
I
think
the
advantages
outweigh
the
downsides,
because
the
upside
is
that,
like
it's
just
code
so
like
you
can
execute
it.
E
So
if
you're,
if
your
book
renders,
then
you
know
it's
like
up
to
date,
because
you
know
it
actually
works
like
if
the
code,
because
the
thing
is
like
you
could
publish
a
collection
of
markdown
files
with
code
Snippets
in
them
and
the
code
sentence
may
or
may
not
work,
and
you
don't
really
know
unless
you
test
them
separately
somehow
but
like
if
you're,
if
you're
whole.
If
all
of
your
content
is
just
like
namespaces,
then
if
it
renders,
then
you
know
it,
it
works
and
it's
it
opens
up.
E
Possibilities
for
like
testing
like
I
would
like
to
like
automate
see
you
later
Jeremy
thanks
so
much
to
automate
some
of
the
testing
for
this
kind
of
stuff,
so
that
with
an
eye
to
like
the
future,
so
it's
like.
If
we
can
like,
have
a
test
Suite,
then
you
can
like
update
dependencies
and
then,
if
the
test
pass,
then
you
can
like
have
some
confidence
that
your
book
is
still
like,
giving
accurate
information,
instead
of
like
out
of
date
out
of
date,
apis
and
stuff.
E
So
anyway,
I
don't
know,
there's,
but
then
it's
it's
complicated.
It's
hard
to
find
like
a
sweet
spot
where
it's
like
easy
to
write
and
publish,
but
also
getting
the
benefits
of
the
code.
So
it's
like,
if
you're
interested
in
that
there's,
like
a
million
there's
a
whole
Literature
Like
literate
programming
is
like
often
what
it's
called.
E
So
it's
this
style
of
like
writing
code,
that
is
like
produces
readable
documents
but
yeah.
It's
like
a
hard.
It's
a
hard
problem.
A
E
B
B
Maybe
one
more
comment
is
that
we
are
still
organizing
the
real
world
data
group
that
will
begin,
probably
at
the
end
of
June,
and
this
group
will
be
a
space
for
people
who
wish
to
practice
their
data
and
science
projects
and
possibly
do
it
in
a
way
that
is
also
contributing
to
the
this
content
of
of
the
Scrapbook
and
cookbook
or
other
tutorials,
and
and
to
have
this
process
of
a
group
different
individuals
and
group
them
friends
can
share
their
process
of
exploration
and
study
together.
So
this
thing
will
begin
in
very
few
weeks.