►
From YouTube: How We Built This // Slate
Description
In our first How We Built This, the @_slate team will talk through the evolution of Slate, going from a simple Filecoin client to a robust media sharing product and walking through the development and innovation processes!
Keep up with events for the Filecoin community by heading over to the Filecoin project on GitHub:
https://github.com/filecoin-project
Check out the Filecoin community resources:
https://github.com/filecoin-project/community
And stay connected on Filecoin Slack:
https://app.slack.com/client/TEHTVS1L6
A
So
be
sure
to
check
out
the
slingshot
website
on
the
events
page
and
you
can
find
all
the
details
there
with
registration
links,
but
for
today
I'm
happy
to
introduce
everyone
to
jim
lee
who's,
also
known
as
cake
around
here.
So
jim's
going
to
be
taking
us
through
today
of
how
slate
was
built
so
jim.
If
you're
all
set
I'm
going
to
pass
it
over
to
you.
B
B
So
teamwork
makes
the
dreams
work
so
a
lot
of
what
is
possible
today.
What
you
see
when
you
visit
slave
host
is
possible
because
horus
really
killed
it
on
design.
Martina
has
been
killing
engineering
and
jason
built
like
some
of
the
coolest
examples
and
like
it
has.
I
think
the
chrome
extension
is
probably
one
of
the
coolest
things
that
we
offer
outside
the
product
and
william
tara
have.
B
Basically,
I
mean,
create
the
marketing
page
that
you
see
here,
and
so
just
want
to
remind
everyone
that
I
guess,
like
you
know
it,
it
takes
a
team
to
build
really
cool
things.
We
have
we've
been
lucky
to
have
contributions
from
pretty
much
all
around
the
world
so
and
we're
getting
more
contributors.
So,
if
you're
interested
in
contributing
and
helping
out
with
slate,
you
know
you're
more
than
welcome
and
I'm
always
around.
B
If
you
have
any
questions
textile
as
I'll
get
into
later
in
this
presentation
has
been,
I
mean
instrumental
in
like
making
all
this
possible.
B
Puja
sha
and
juan
banay
really
set
the
groundwork
for
any
of
this
to
like
kind
of
sit
in
my
head
and
kind
of
like
give
me
the
ability
to
kind
of
come
up
with
the
foundation
for
the
work
of
slate
which
I'll
get
into
as
well,
and
it's
been
an
insane
privilege
lucky
to
have
anyone's
attention
dedicated
time
extremely
grateful
for
all
of
it,
and
so
slate
is
completely
open
source,
and
you
can
ask
me
if
you
need
anything,
we
have
a
github
issues
tab.
If
you
file
an
issue,
I
will
be
there.
B
I
will
help
you
with
getting
buckets
working
I'll
help
you
get
powergate
working.
My
time
is
your
time.
So
whatever
you
guys
need
today
is
a
really
special
day,
because
it
marks
seven
months
of
searching
for
the
optimal
client
experience
for
file
coin,
and
so
we
get
a
lot.
That
slate
is
an
ipfs
client
too.
We
get
a
lot.
That's
like
can
do
other
things,
but
we're
really
looking
for
the
optimal
file
coin.
Experience
and
I'll.
B
Take
you
back
to
march
1st
before
slate
even
existed
before
we
even
called
it
slate,
and
it
was
called
the
filecoin
client
I'll.
Take
you
to
transition
point
in
june
when
the
filecoin
client
became
sort
of
slate
and
we
rebrand
it
and
there's
other
moments
such
as
when
we
were
only
using
powergate.
B
The
first
day
we
supported
pdfs
and
videos
which
started
showing
our
range
as
a
product,
not
just
for
visuals,
and
also
separate
four
teams
a
special
day
for
me,
because
martino
long
basically
saved
the
day
and
made
a
lot
of
the
engineering
work,
just
really
killed
it,
and
so
I
think
the
first
step
for
anyone
who
really
wants
to
build
for
file
coin
is
kind
of
understanding
the
romantic
vision
behind
foulcoin.
Like
I
think
in
engineering,
it's
really
easy
to
become
a
fatalist
and
think
about
things
and
kind
of
a
deterministic
way.
B
So
what
is
its?
What
it?
What
is
the
fundamental
truth
about
the
function?
Calls
and
what
are
the
preconditions
and
post
conditions,
and
understandably,
falcoin
you
know,
will
run
to
performance
issues
we'll
run
to
bugs
and
snags,
but
the
romantic
vision
is
really
really
wonderful.
So
I
got
lucky
like
in
february
2020
I
was
able
to
work
on
the
documentation
project
for
lotus
and
I
had
a
ton
of
help
from
their
team,
so
it
had
a
pretty
unfair
advantage.
I
could
talk
to
magic,
who
was
like
basically
one
of
the
smartest
people.
B
I've
ever
had
the
chance
of
meeting,
and
you
know
I
re.
I
actually
completed
a
small
storage
and
retrieval
deal
using
lotus
at
the
time,
and
I
got
a
taste
of
the
file
coin
stack
and
that's
when
I
became
a
true
believer.
I
think
I
actually
thought
that
this
thing
could
really
work
and
be
amazing,
and
we
just
have
to
make
it
available
to
everyone
and
why
I
put
together
this
team
that
just
I
think,
like
always,
was
willing
to
share
information
openly
like
travis
person.
B
Whenever
you
ask
some
questions,
you
always
share
what
he
knows
and
it's
just
really
easy
to
kind
of
understand.
So
I
I
started
to
read
a
lot
of
the
the
materials
and
they're
they're
really
really
dense,
like
there
are
a
lot
of
wonderful
technologies
that
are
being
built
to
get
filecoin
to
work,
and
so
I
had
a
lot
of
assistance
and
the
filecoin
spec.
Is
they
updated
the
the
image
so
or
the
webpage?
B
So
right
now
you
see
like
a
this
is
a
newer
one,
but
the
older
one
was
super,
informative
and
kind
of
captured
a
lot
of
the
original
vision
for
file
coin,
and
so
I
I
naively
I
I
have
some
design
experience,
so
I
didn't
naively
approach
documentation
project.
You
know
with
my
typical
individual
contributor
style,
so
you
know
gotta
produce
a
hasty
vision,
so
determine
themes.
Determine
insights,
determine
where
you
gamble.
Where
are
the
opportunities
beyond
existing
solutions
and
and
to
be
brave
right?
B
Because
you
know
that
you
could
be
wrong,
but
you
got
to
give
it
your
best
shot,
and
so
I
prototyped
I
mood,
boarded
pull
up,
155
samples
of
different
documentation,
websites
and
then
iterating
quickly
to
like
test
division
test.
The
test
to
test
the
solution
and
in
retrospect
designing
documentation,
was
fairly
simple
compared
to
the
file
coin,
client,
which
is
significantly
harder.
So
you
know,
if
you
look
right
here,
like
you,
just
look
at
a
command
line
tool.
B
B
It's
like
there's
a
lot
of
stuff
that
you
have
to
kind
of
keep
into
context
before
you
produce
any
forms
and
puja
I
gotta
give
a
huge
shout
out
to
puja.
Pujo
was
the
first
person
that
really
gave
me
a
huge
lift
and
helped
me
realize
the
detailed
complexity
required
to
solve
this
problem.
So
we
were.
We
were
in
a
mode
where
she
really
she
has
a
really
deep
understanding
of
follow
coins,
so
we
were
able
to
kind
of
like
with
her
help.
B
She
before
I
got
into
like
mocks
or
prototypes,
really
just
draft
out
all
the
features
and
flows,
and
this
made
a
ton
of
work
possible
like
I,
there
are
ideas
from
ipfs,
companion
and
client,
client
and
other
apps
in
the
ecosystem
that
I
wasn't
even
aware
of
so
seeing
this
was
extremely
enlightening
and
I
was
unprepared
for
topics
like
data
transfer
and
payment
channels
from
you
know
previously,
before
working
on
the
filecoin
client,
I
was
working
in
games
and
you
know
we
make
up
rules.
We
just
have
fun.
B
So
there's
a
lot
of
things
that
you
know.
I
needed
a
really
good
primer
on
in
and
juan
and
puja.
You
know
we.
We
worked
in
a
figma
file,
I'm
happy
to
share
that
fixmo
file
with
people
later
if
they
like,
but
it
was
just
a
lot
of
ideation
and
it
was
really
fun
like.
I
could
see
basically
an
ideation
that
we
came
up
together.
B
We
actually
saw-
I
actually
saw
that,
like
almost
the
entire
history
of
computing
and
it's
kind
of
fun
like
to
trace
things
back
to
like
the
90s
2000
and
how
software
has
evolved
and
it
kind
of
the
weight
started
to
set
in.
I
started
to
realize.
Oh,
my
goodness,
like
this
is
kind
of
a
homage
to
software
that
we've
used
all
our
lives
and
maybe
just
being
really
cool
here,
and
so
I
went
and
did
what
I
what
I
do
best.
B
I
don't
know
if
it's
the
best
design
process
whatever,
but
I
went
and
produced
like
a
ton
of
mood
boards
so
put
like
346
references
that
the
pro
labs
team
liked.
I
would
share
some
of
the
stuff
I
posted
in
slack
too,
but
that's
you
know
like
a
lot
of
it's
just
a
continuation
of
the
same
thing.
So
I
really
wanted
to
get
a
sense
of
how
people
felt
when
they
looked
at
visuals.
B
I
wanted
to
be
there
to
talk
to
them
and
and
kind
of
get
a
feel
for
like
oh,
what
do
you
like,
and
what
do
you
like
about
these
things-
and
you
know
intrigue
is
an
important
part
of
when
you're
designing
application
and
the
file
coin
network
when
it
hits
the
mainstream,
won't
be
any
different
than
any
type
of
other
tool
that
we've
used
in
the
history
of
human
civilization.
It
has
to
be
interesting
has
to
be
intriguing.
B
It
has
to
feel
like
it
gives
you
superpowers,
so
aesthetic
matters,
visual
representation
matters,
and
so
that
led
to
figment
sign
party
all
march-
and
this
is
all
I'm
really
comfortable
with
showing
from
that
that
figma
party,
we,
I
I
made
a
ton
of
things-
I
crossed
out
a
ton
of
things
and
then
countless
prototypes
through
april
written
and
react,
so
people
could
interact
with
the
flows.
So
it
was
super
important
for
me
that
you
could
actually
use
an
experience.
There
was
no
back
end.
B
You
could
just
basically
go
through
every
single
view.
That
was
in
the
fileclone
client
and
get
a
feel
for
if
the
data
was
there,
if
you
could
actually
make
storage
and
retrieval
deals,
what
would
it
feel
like?
And
so
I
I
love
going
to
museums
and
looking
at
art,
so
I,
I
kind
of
you
know
selfishly
put
that
story
into
the
falcon
client
and
you
know
made
it
so
everyone
can
see
this
full
vision.
B
I
think
there's
about
like
20
or
30
or
so
scenes
that
you
could
experience
going
through
the
falcon
client,
and
I
was
actually
pretty
excited
about
this
path.
I
thought
you
know
this
would
be
really
cool
as
a
desktop
client.
We
were
obviously
planning
to
build
this
electron
app
and
I
built
it
starting
with
this,
this
github
repo,
so
this
get
a
repo,
is
basically
the
the
starting
point
for
a
lot
of
my
projects.
I
just
use
like
a
few
libraries
and
nexus
is
probably
my
favorite
framework
for
the
web.
B
After
all,
that
kind
of
heads
down
design
duration
and
help
from
juan
and
puja,
I
was
able
to
kind
of
get
a
tweet
out
there.
Finally
and
say:
hey
here's,
what
I've
been
working
on
and
hope
you
guys
like
it
like
a
first
stab
the
first
genuine
stab
at
the
filecoin
network
and
so
yeah.
A
B
Turned
out
to
be
received
pretty
well,
I
have
to
give
a
huge
shout
out
to
render
render
is
one
of
my
favorite
services
for
hosting
they
they're
back
in
this
gcp
and
kubernetes,
so
it's
pretty
pretty
powerful
stuff
and
worked
really
well
to
just
show
the
whole
world.
This
follicle
client
example,
and
you
know
some
ideas
I
also
took
you
know.
I
think
we
stand
on
the
shoulders
of
giants.
My
peer
ben
gave
me
the
idea
for
pre-brand.
B
I
remember
seeing
his
presentation
a
long
time
ago
and
being
like
yeah.
B
That's
really
smart,
like
a
brand,
that's
adaptable
and
can
change
over
time,
and
so
I
use
sites
like
logo
archive
with
the
team
to
kind
of
come
up
with
what
the
like
rv
one,
what
people
like
generally
so
I
was
asking
people,
you
know
hey
like
we're,
we're
naming
the
filecoin
client,
and
I
give
I
I
had
this
page
and
this
page
had
a
place
where
you
could
vote
or
you
couldn't
vote
on
the
page
at
a
dm
me,
your
vote,
but
I
keep
track
of
everything
and
here
I'm
going
to
hide
the
person
who
said
it,
but
we
had
a
lot
of
comments.
B
People
liking
the
name
slate,
and
I
was
I
was
kind
of
like
oh
slate's,
like
a
cool
name,
and
I
hadn't
really
like
before
in
the
past,
had
a
group
kind
of
come
together
and
name
a
product
together
like
usually
it's
like
one
person
decides
because
it's
like
an
endless
debate,
but
I
I
felt,
like
you
know
like:
let's
ask
everyone,
so
we
we
pulled
everyone
together
and
it
was
cool
that
a
lot
of
people
kind
of
gravitate
around
the
same
name,
and
I
was
like
okay.
Well,
that's
cool,
so
we
we.
B
Actually.
I
put
that
on
the
back
burner.
Naming
naming
the
pro
the
the
falcon
client
wasn't
as
important
as
getting
this
relationship
right.
So
textile
is
the
the
ecosystem,
the
platform,
the
the
tooling.
That
really
makes
this
all
possible
and
it's
a
joy
to
work
with
textile
like
they
are
they
really
care,
they're,
really
responsive,
and
you
can
see
that
even
at
like
high
duress
moments
when
there's
just
like
a
ton
of
stuff
going
coming
their
way,
I
a
huge
joy
to
work
with
them,
and
so
by
june.
B
After
doing
all
that
design
work
that
you
see
or
you
had
saw
in
the
previous
slides,
the
plan
was
eventually
used
textile
services
and
at
the
time
we
were
still
the
file
point
client.
So
here's
an
internal
dock
I
wrote
about.
Like
my
you
know,
timeline
I
actually
missed
the
launch
date
that
we
were
gonna
if
we're
just
gonna
be
a
powergate
client.
B
I
missed
that
launch
date
by
like
a
month,
but
we
actually
expand
the
product
and
I'll
get
into
that
more
of
those
details
soon,
and
so
you
can
see
here
very
simple
model.
Textiles
powergate
is
such
a
useful
tool.
All
you
need
is
a
token
and
you
have
to
protect
that
token
from
bad
actors
for
your
user,
and
so
you
might
want
a
way
to
manage
that
or
if
you're,
just
a
desktop,
app
or
offline
app.
It
doesn't
really.
B
You
don't
have
if
you
don't
connect
to
internet
or
you
don't
share
that
information
now
on
the
internet,
it's
okay,
so
I
just
came
up
with
a
simple
model
based
on
all
the
design
work
I
had
done
so
just
you
know.
We
have
a
token.
We
have
offline
storage
directory
that
token
can
be
used
to
you
know,
make
deals
interact
with
the
falcon
file
system.
B
B
I
didn't
understand
fully
the
full
ipf
as
except
the
time
I
was
just
kind
of
hyper
focused
on
powergate,
and
so
you
know
I
I
started
my
brain
started
going
into
like
galaxy
brain
mode
like
okay.
Well,
you
know
this
is
how
it's
going
to
work
at
a
multiple
client
level,
so
you
have
client
a
client
b.
They
do
roughly
the
same
thing:
they're
different
users.
They
all
communicate
with
powergate
they're,
all
gonna
send
follow
coin
between
each
other.
There's
gonna
be
a
deal
registry.
Whenever
you
make
a
storage
deal,
you
can.
B
You
know,
pull
retrieval
deal
from
there
and
to
kind
of
like
understand
that
the
best
way
I
know
how
to
I
created
a
design
system
around
it,
and
so
this
was
actually
the
opportunity
that
began.
I
think,
like
external
contributions
to
the
project,
because
what
I
was
able
to
do
here
was
kind
of
like
get.
You
know,
other
people
work
on
small
bits.
B
So,
like
you
know,
you
don't
have
to
be
a
blockchain
expert
to
help
me
come
up
with
a
really
great
button
or
icons
or
colors,
and
so
I
really
wanted
to
make
it
accessible
for
everyone.
So
I
was
like
okay,
I'm
gonna
start
introducing
some
new
concepts
like
powergate.
B
What
are
some
things
that
I
can
get
the
community
involved
in
that
aren't
as
intimidating,
so
it
was
really
great
to
kind
of
start
this
out,
and
so
the
the
the
framework
for
design
system
like
what
we
wanted
to
teach
was
something
as
simple
as
this
you
had.
You
have
a
user,
they
have
a
token.
How
do
they
make
a
deal?
How
to
interact
with
the
file
coin
file
system?
B
How
to
interact
with
a
wallet
we
didn't
want
to
get
any
crazier
than
that,
whatever
architecture
people
are
using,
you
know
it's
fine
if
they
use
those
things.
We
just
want
to
present
this
in
a
really
simple
way,
and
so,
if
you
guys
remember
from,
I
think
this
was
a
long
time
ago
now,
you
know
I
I
did
a
I
did.
I
did
a
hack
fest
session
before
textile
supported
tons
of
teams
during
hackfest.
B
I
I
produced
like
a
quick
tutorial
of
how
you
could
build
an
app
using
this
design
system
and
and
use
powergate
and
and
make
a
follicle
and
storage
deal,
and
so
here's
an
example
of
that,
like
you
basically
have
these
components
written
in
react,
their
style
cross
browser
the
tool.
Emotion
is
a
tool
that
has
auto
prefixing.
So
it
gave
me
that
confidence
that
okay,
like
we
can
use
this
on
any
browser
and
then
the
react
code
so
showing
that
in
under
23
lines
you
know
hey
like
we
can.
B
B
You
know,
use
the
filecoin
file
system
to
create
a
new
address,
console
log,
the
response
and
then
to
have
a
lot
of
the
ui
bindings
that
are
intimidating
to
certain
engineers
kind
of
figured
out,
and
this
is
when,
like
shout
out
to
one
of
my
mentees
daniel,
who
basically
just
kind
of
like,
went
in
and
started
having
ton
of
fun
like
he
put
odometer
on
this
token
thing,
it's
still
live
on
our
design
system.
B
If
you
look
at
it
and
we
started
getting
help
from
the
community
to
like
kind
of
help
make
this
powergate
thing
have
a
visual
system,
and
you
can
see
here
all
the
code
for
all
the
components
are
roughly
the
same.
It's
they're
just
kind
of
you
know
it
starts
to
become
repetitive
enough.
We're
like
okay,
I
get
it.
I
just
gotta
get
this
token
and
I
can
get
my
follicle
imbalances.
B
I
said
I'll
get
this
token,
and
then
I
can
get
my
you
know
all
of
the
the
the
each
balance
the
address
type.
You
know,
if
I
get
the
token-
and
I
call
some
of
the
options
that
file
filecoin
file
system
provides.
I
can
you
know,
learn
about
my
storage
deals,
my
retrieval
deals
and
then
you
know
martina
stepped
in
and
really
helped
with
some
of
the
more
robust
components
so
like
we,
we
had
a
conversation
where
I
was
like
you
know,
maybe
kind
of
cool.
B
If,
when
you
use
a
component,
you
actually
learn
more
about
the
file
coin
network
too,
and
so
the
settings
component
she
kind
of
just
like
killed.
It
was
like
just
did
an
amazing
job
where
you
kind
of
just
once
you're
using
it
you're
like
okay,
I
get
it.
There's
this
notion
of
hot
storage,
cold
storage
and,
of
course,
like
definitions
for
the
network
change
over
time
as
we
implement
new
things,
but
it
was
just
really
cool
to
see.
B
You
know,
members
of
our
community
kind
of
just
like
create
informative
components,
even
not
just
like
you
know,
like
a
button,
but
like
something
that
teaches
you
how
to
use
the
network,
and
so
this
was
the
kind
of
the
mental
model,
the
hope
where
we
would
get
to
where
it
was
just
like.
Okay,
a
ton
of
people
understand
this
paradigm.
You
got
a
user,
you
have
a
token,
and
now
you
got
deals.
B
There
were
some
caveats
here,
of
course,
like
I
think
you
know
we're
on
nerpa
defnet,
and
you
know
today
you
know
if
you're
in
slingshot
right
now
you
know
you're
trying
to
work
with
the
test
net,
which
is
the
live
environment,
and
you
know
we
one
thing
that
in
retrospect,
looking
back
on
it,
I
think
you
know
it
was
awesome
that
nerf
was
so
performant
and
it
made
everything
feel
so
fast.
But
you
know
we
I
it
was
it's
kind
of
on
me
to
like
I.
B
I
didn't
get
the
testing
against
a
test
net
soon
enough.
So
a
lot
of
the
exploration
there.
You
know
if
powergate
was
awesome,
but
it
was
kind
of
in
optimal
situations.
So
I
kind
of
wish
that
I
had
more
opportunity
to
kind
of
test
in
the
test:
net
environment-
and
you
know
people
who
are
self-hosting
their
power
gates
kind
of,
I
think,
had
the
ability
to
do
that,
and
I
think
we
could
have
learned
a
lot
more
sooner.
B
So
I
I
definitely
needed
the
help
and
I
needed
a
I
needed
a
team
that
I
could
work
with
and
we
were
getting
contributors
and
partners,
but
I
started
to
realize
that
okay
building
this
entire
thing,
this
falcon
client
on
my
own,
you
know
I'm
not
like
the
sharpest
person
in
the
world,
so
it's
it
would
be
really
awesome
to
have
a
group
of
people
to
work
with,
and
you
know,
like
my
friend
rich
says,
teamwork
makes
the
dreams
work
and
I
believe
this
to
my
like
with
all
my
heart.
B
I
think
you
do
things
with
other
people,
and
you
know
you
learn
from
them.
You
learn
where
you're
wrong
too
you're,
not
always
right,
and
I
I
thought
it
was
a
really
awesome
opportunity
to
build
a
diverse
team
from
start.
So
you
know,
let's,
let's
get
a
range
of
developer
backgrounds
or
no
backgrounds
and
start
with
a
fresh
slate
and
kind
of
see
how
you
know
how
to
how
to
you
know
what
that,
what
kind
of
impact
that
would
have
on
our
product
process.
B
You
know
we're
not
all
just
mono,
cultured
and
we're
all
thinking
about
things
in
different
ways:
july
10th.
It
felt
right
to
remove
the
file
coin
branding
and
become
slate,
and
so
I
talked
about
this
a
little
bit
earlier.
You
know,
after
everyone
had
a
chance
to
vote
protocol
labs
and
22
plus
names.
The
second
runner-up
was
argo,
so
that
could
have
been
a
fun.
B
If
you
guys
have
been
following
slate,
it
could
have
been
fun
to
imagine
what
it
would
have
been
like
if
we
were
named
argo-
and
you
know
during
this
time,
working
with
textile.
I
think
when
you
work
with
like
a
really
engaged
group,
you
start
discovering
things
on
the
fringes
of
your
your
offering
you
start
learning
more
about
what
your
tool
can
do
and
I'm
sharing
a
really
bad
diagram.
B
I
made
because
you
know
I
we
were
just
getting
all
this
new
information
of
like
oh
there's,
buckets,
there's,
redcb
and-
and
I
I
wanted
to
like-
I
want
to
make
use
of
all
those
tools
you
know
and,
and
I'm
still
learning
kind
of
you
know
still
wrapping
my
head
around
powergate,
but
I
also
had
you
know
this
buckets
tool
that
just
seemed
really
interesting
and
and
when
andrew
proposed
it
I
was
like
okay,
we
got
to
try
this.
It's
like
you
know
this.
B
This
is
a
really
cool
piece
of
technology,
and
this
I
think
july
19th
is
when
slate
was
no
longer
a
pure
filecoin
client
and
it
wasn't
fair
to
call
it
a
falcon
client
anymore.
Of
course,
like
you
know,
officially
branding
something
is,
is
another
discussion
all
in
itself,
but
you
know
we
started.
B
We
started
to
have
some
function
in
this
code
base,
as
we
were
evolving
and
growing
it
that
it
seemed
to
make
sense
to
start
moving
to
different
brand,
and
I
was
looking
through
textile
diagrams
and
puja
and
andrew
have
presented
about
this
a
ton
of
times
before
and
this
tech
is
super
solid.
I
mean
like
they
say
it's
beta
stack
or
alpha
stack
or
whatever,
but
in
my
experience
it's
like
it's
been
very,
very
fun
to
work
with,
and
I
think
it's
it's
beyond
hobbyist
stage
where
you
know
it
you
can.
B
You
can
kind
of
once
you
have
a
stable
enough
like
infrastructure
to
run
this
code.
It
actually
works
really
well,
and
it's
only
on
the
fringe
cases,
where
you're
like
testing
the
bounds
that
you
kind
of
experience,
some
issues,
but
you
know
a
ton
of
fun
to
work
with,
so
they
really
broke
out
like
there's
a
white
paper
as
well.
I
can
share
with
people
later
if
they're
interested,
but
they
really
broke
out
like
how
their
tooling
works
and
you
kind
of
see
what
all
the
considerations
are
putting
into
things.
B
And
so
buckets
was
this
dynamic
directory
storage,
and
so
you
have
like
powergate,
that's
probably
really
great,
at
storing
like
single
c
ids.
But
then
you
have
with
buckets
you
know
conceptually.
It
would
be
possible
maybe
to
get
a
unity
game
onto
the
file
code
network
like
a
whole
folder
once
you
have
the
root
cid
for
that
directory,
and
so
it
made
total
sense
for
us
to
use
it,
and
this
was
my
first
modeling
of
that,
and
I
I
tried
my
best
here.
I'm
actually
like.
B
I
don't
know
if
I
got
it
right,
but
you
know
you
have
a
local
react
application,
which
is
what
we
have
been
working
on
for
months
on
what
I've
been
working
on
for
a
while
as
well-
and
you
know
you
have
your
identity,
your
lip
pdp
token
and
your
little
pdb
token
is
basically
your
gateway
into
all
of
the
hub
and
something
cool
about
the
token
like
this
could
be
gateway
to
other
things
as
well.
B
So
you
know
if
you,
if
you
leave
our
product,
you
can
use
the
sleep
ptp
token
to
kind
of
go
anywhere,
and
so
you
know
you,
we
have
a
storage
reference
on
slate,
so
you
know
we
don't
have
to
always
go
to
the
bucket
to
get
the
information
about.
What's
on
there.
If
we
keep
a
storage
reference,
we
can
have
a
little
bit.
It's
like
acting
as
a
cache,
and
then
we
can
use
the
gateways
to
pull.
B
So
you
know
we're
not
making
total
round
trips
when
it
comes
to
checking
if
something's
in
the
bucket
and
then
pulling,
we
can
just
hit
up
the
gateway,
url
and
get
the
asset,
and
you
know
this
directory
aspect
has
always
been
something
that
I've
really
wanted
in
slates.
I
just
I
think
software
is
better
with
friends
like
when
you
can
just
kind
of
hang
out
with
other
people,
and
so
I
refined
it
over
time-
and
I
think
slate
is
now
kind
of
like
this.
Where
you
visit
a
web
page,
you
open
a
program.
B
The
open
program
full
doesn't
really
work
right
now,
but
you
have
a
library
of
cd
references.
You
have
your
slates,
which
is
our
user
experience
and
slates
I'll
get
into
more
later.
But
slates
are
important
because,
when
you're
working
with
data
it
has
to
be
enjoyable,
it's
not
just
like
visually
enjoyable.
You
got
to
want
to
feel
like
you,
you're
collecting
things
and
having
fun,
and
that
gets
you
into
the
cycle
where
you're
constantly
using
the
software
and
that's
when
we
can
really
get
people
to
understand
file.
Coin
right.
B
Like
oh,
I
can
do
things
with
this
with
this.
Maybe
this
retrieval
thing
and
the
storage
thing,
but
until
then
you
have
to
keep
people
in
this
really
happy
cycle.
Bring
your
friends
there,
where
you're
sharing
data
between
each
other
also
kind
of
keeps
people
engaged.
You
know,
planning,
sdk
integration,
so
you
know
allowing
people
to
bring
in
their
software
into
the
experience
too
could
be
really
cool,
and
then
we
have
like
the
three
boxes
for
api
and
key
management.
B
B
I
think
this
is
kind
of
groundbreaking,
like
this
allows
people
to
kind
of
share
openly
in
the
whole
world,
on
this
open
network,
where
anyone
can
make
a
deal
as
long
as
the
miner
accepts
it
and
and
the
hub
is
so
cool
it
helps
you
find
based
on
like
a
theoretical
reputation
index
of,
like
you
know,
the
best
miners
to
make
deals
with,
and
so
you
know
you
have
some
more
guarantees,
and
so
it
just
seemed
like
all
the
pieces
were
coming
together
here
and
and
using
using
the
technology
at
a
at
a
programmatic
level.
B
A
B
Think
people
you
know,
there's
a
baseline
for
like
development,
especially
even
with
you
know,
systems
languages
like
go
or
rust
or
c,
plus,
plus
or
c,
or
haskell,
or
erlang
or
whatever.
But
you
know
when,
when
you're
working
with
javascript
there's
still
some
complexity,
you
have
to
understand
the
library,
but
I
really
like
how
they
made
their
api.
Just
super
simple,
like
you
authenticate
you
pass
through
the
secrets
that
you
have
for
textiles
only
two
and
then
you
can
basically
do
operations
with
buckets
and
then
get
a
list
of
them
and
you're
already
there.
B
So
if
like
three
lines
of
code
or
four
lines
of
code,
you're
able
to
kind
of
use
the
buckets
api,
I
think
that's
that
that
should
be
celebrated
more
like.
I
think
they
made
something
that
is
actually
fundamentally
really
really
complicated.
If
you
looked
at
previous
diagrams
in
just
a
few
like
library
methods-
and
here
you,
I
can
break
into
one
of
the
methods
that
I've
kind
of
made
more
robust,
and
so
you
know
it
at
times.
You
know
we'll
you'll
have
to
do
some.
B
B
I
think
it
just
makes
it
really
easy
for
people
to
understand
and
get
get
started
and
so
to
handle
all
this
complexity
slave
formed
a
core
team,
and
so
I
I
keep
close
to
my
heart,
the
first
days
that
everyone
contributed
code
to
slate
and
so
jason
someone
I
met
at
the
start
of
you
know
this
journey
committed
his
first
code
june,
23rd
2020..
B
B
Martina
committed
her
first
commit
on
june
28th
and
honestly
started
killing
it
from
that
point
on
tara
and
will
tara
commit
her
first
design
system
contribution
on
june
28th
will
july
1st
and
horace
my
battle
buddy
on
slate
finally
joined
july
28th.
To
kind
of
like
really
help
us
move.
B
Everything
forward
and
narrative
did
amazing
brand
work,
and
so
I
could
show
a
ton
of
slides
everywhere,
but
it's
kind
of
everywhere,
like
the
the
logo
mark,
is
really
beautiful
and
I
never
had
a
chance
to
you
know
like
in
this
whole
process
get
a
chance
to
have
another
team.
I
trusted
so
much
to
help
with
design
in
the
way
that
they
did
in
a
long
time.
B
You
know,
usually
I
I'm
kind
of
on
the
hook
for
a
lot
of
things,
and-
and
so
you
know
in
prior
work-
and
so
here
you
know,
pro
collabs
made
it
possible
to
work
with
slate,
and
I
mean
work
with
narrative.
A
B
Yeah
huge
thank
you
to
tom
mack
and
thiago
august
1st
2020.
I
put
all
the
pdfs
and
videos
on
slate
that
was
really
cool.
I
think
started
showing
that
slate
was
a
little
bit
more
multi-dimensional.
It
wasn't
just
for
like
pretty
pictures
or
whatever
we
can
store
everything,
and
I
think
we
since
then,
we've
been
expanding
daily.
B
You
know
the
code
is
really
simple
like
you
just
go
in
and
you
look
at
this
react
component
and
you
can
see
like.
Oh,
you
know
we're
just
checking
the
type
and
we're
rendering
the
right
html5
component
to
it
right.
So
you
know
video
audio
image
object.
B
These
are
all
available,
and
I
think
this
is
the
genesis
for
expanding
on
this
even
further
like
what,
if
we
combine
them
together,
what
if
we
did
other
things-
and
I
think
in
a
later
class
someday
harass-
will
probably
cover
some
of
the
future
vision
we
have
for
site,
because
you
know
this
is
just
this
is
so
naive
at
this
point,
but
we
could
go
so
much
further
august
17
2020.
I
did
a
class
on
the
developer
api
for
slate.
B
This
was
kind
of
a
fun
class
because
at
the
time
slate's
api
was
a
little
bit
buggy.
So
I
had
a
course
issue
which
prevented
anyone
from
actually
following
the
tutorial
at
the
time
of
the
presentation.
But
you
know
there's
a
lot
of
really.
B
As
well-
and
one
thing
to
highlight,
is
you
know
at
this
point
the
slate
team
felt
like
just
the
coolest
group
of
people
to
work
with,
so
we
had
just
contributors
from
everywhere,
and
this
is
the
developer
api
screen
as
it
stands
today
and
we're
seeing
on
twitter.
You
know
shout
out
to
adrian,
but,
like
you
know,
they're
using
slate
now
to
take
all
their
data
from
products
like
instagram
and
and
use
it
as
a
back-end
replacement,
and
I
think
that's
awesome,
and
so
we
provide
like
this
really
simple
developer
api.
B
You
know
that
lets
you
do
this
and
watching
people
do.
That
is
really
cool,
and
so
the
code
is
really
simple.
You
know
you
just
have
your
authorization
key
header
in
your
fetch
request
and
fetch
is
a
common
library
amongst
javascript
developers,
and
if
you
have
any
confusion,
I
can
always
walk
you
through
it
later.
But
you
know
you,
you
you're
a
user
and
you
make
a
slight
api
request
and
ends
up
in
a
bucket
simple,
as
that
and
I
think
like
when
we
were
able
to
provide
this
experience.
B
You
know,
and
if
you
saw
in
the
other
diagram,
the
the
bucket
is
the
gateway
to
file
coin
right.
So
you
have
all
your
data
in
maybe
a
directory
or
a
single
file,
and
then
you're
ready
to
make
take
that
root
c
id
and
make
a
deal
with
the
file
coin
network.
It
unlocks
a
ton
of
you
know,
possibilities
for
engaging
with
file
coin.
B
So
getting
this
flow
and
having
seen
people
use
it,
I
feel
like
we
got
people
through
the
first
step
and
now
we
can
get
them
to
other
steps
and
so
august
23rd.
This
is
a
special
day
for
you
guys.
I
love
monet.
I
think
his
art's
legendary
not
much
else
to
be
said
there
and
it
was
awesome
to
kind
of
like
take
some
of
the
the
commons
work
and
put
it
onto
slate
and
be
like
hey.
B
So
I
have
a
slate
and
you
can
see
the
preview
of
everything
and
you
know
it's
on
a
data
bucket
but
check
this
out.
So
let's
say
you're
two
different
cases:
you're
a
person
who
wants
to
preview
a
file
coin
deal
and
or
you're
just
a
person
who
wants
to
use
slate
regardless
now
using
the
file
coin
network.
So
if
auto
automatic
fog,
when
are
archiving
for
your
account,
you
know
if
you
want
to.
If
you
want
the
data
set
to
be
open.
B
Let's
say
you
collected
all
the
art
from
like
museums
that
were
willing
to
share
it
at
high
resolution.
Now,
every
week
or
so
we
know
we're
putting
on
the
falcon
network
or
people
can
retrieve
and
keep
local
if
they
want
or
or
help.
You
know
support
in
some
way
and
so
once
that,
once
that
user
relationship
is
happening
and
people
can
bear
witness,
you
know,
then
we
can
start
designing
more
experiences
around
that.
So
you
know
in
the
code.
For
that
you
know,
that's
all
textile.
It's
super
easy!
It's
just
bucket
style
archive.
B
You
know
we.
We
do
some
of
the
heavy
lifting
to
get
you
that
cid
and
organize
the
file
in
the
data
bucket
right.
But
then,
once
the
data
is
there,
you
know
calling
bucket.archive
just
works,
and
so
that
was
kind
of
you
know
really
awesome
to
have
during
implementation
time,
because
you
know
we
didn't
have
to
deal
with
all
of
the
what's
running
on
my
computer.
How
do
I
make
sure
everything's
well
connected?
It's
just
a
simple
api
call
and
you.
A
B
I
I
can't
say
it's
enough:
I
think,
having
all
that
stuff
there,
you
know.
Jason
has
like
crazy
perception,
so
he's
able
to
like
see.
Okay,
I
see,
what's
going
on
with
the
current
slate
product.
Just
kills
it
with
this
chrome
extension
blows
blows
all
our
minds,
it's
like
awesome
as
hell,
and
you
know
it's
open
source,
total
game
changer
and
now
you
can
right
click
on
anything
on
the
internet
and
if
you
go
back
to
the
slide
here,
you
know
in
your
user.
B
You
mix
api
requests
and
you
put
it
into
a
textile
bucket
from
here.
Now
we
can
get
those
into
the
filecoin
network
so
easily,
so
people
don't
realize
it,
but
they're
now
like
helping
grow
the
file
coin
network
just
by
using
the
internet
daily,
just
by
collecting
things
that
they
like,
and
so
I
think,
that's
just
that
that
will
help
people
understand
the
falcon
like.
Oh
you
know,
I
get
things
off
of
these
like
websites
that
their
websites
are
open.
B
The
rules
are
like
you
can
use
the
images,
but
I
get
them
off
of
this
like
web
2
website,
and
I
put
them
on
this
network
where
anyone
globally
can
retrieve
them.
It's
open
data
for
everyone
in
the
world,
and
so
with
that
work
you
know
horace
and
martina
went
to
town.
I
think,
like
you
know,
there
are
just
countless
design,
improvements
and
implementations
that
actually
happen
throughout
the
product.
So
you
can
just
see
you
know
like
we
have
a
directory
this
directory
thing.
I
always
wanted,
and
now
we
have
this
these
tiers.
B
So
we
have
people,
you
can
make
requests
to
that
be
trusted.
You
can
just
subscribe,
and
so
we
add
activity
features
you
can.
You
know,
follow
people
and
you
can
see
what
they're
storing
on
the
file
coin
network
eventually.
So
then
you
can
retrieve
that
back
into
your
own
account,
and
so
I
think
that's
going
to
be
really
fun.
You
know
you
have
a
data
view
and-
and
I
was
kind
of
stubborn,
like
I
I'm
kind
of
sometimes
one
dimensional
as
a
designer.
B
I
think
there's
like
one
key
solution,
but
my
eyes
opened
when
they
killed
it
with
like
this
multiple
views,
you
can
view
a
grid,
you
can
view
a
list
and
you
can
see
that
we
have
diversity
of
files.
Here
we
have
you
have
music,
you
have
images
this
page.
I
love
it
like
space
to
see
the
slate
previews
in
the
first
version.
B
You
can
only
see
like
you
know,
you
could
see
only
the
name
of
this,
which
is
kind
of
silly
now,
but
like
being
able
to
see
the
previews
made
a
huge
difference,
it's
easy
to
curate
what
you
have,
and
you
know,
tara
and
will
obviously
killing
with
the
marketing
page
concepts,
and
so
we
want
to
make
it
known
that
when
it
came
to,
you
know
slate
that
a
lot
of
people
are
involved,
and
so
it
was
awesome
to
kind
of
have
like
them
kind
of
dig
into
the
community
and
work
on
slates
and
make
marketing
pages
and
trying
to
figure
out
the
voice
and
we're
learning
every
day.
B
And
it's
super
imp,
I'm
I'm
learning
from
them
at
this
point.
So
it's
really
fun
and
you
know
kind
of
we
are
being
ambitious
too.
So
I
think
we're
trying
to
offer
more
free
storage.
B
Really
cool,
I
mean
like
what
I'm
learning
a
lot
from
people
who
are
younger
than
me
today
is
that
they
have
cooler
influences
than
I
did.
You
know
when
we
were
doing
web
apps
in
like
2010.
You
know
the
space
was
very
young.
I
think
like
it
was
coming
off
of
the
dot-com
crash
and
you
know
a
lot
of
cool
websites
and
it's
it's
like
early
facebook
days
too,
and
I
think
you
know
now
the
because
of
the
vast
array
of
data
on
the
internet.
B
A
lot
of
my
mentees
and
people
I
work
with
who
are
younger,
have
a
ton
of
really
insanely
cool
influences,
so
that
includes
van
over
bush
douglas
and
engelbart.
It's
just
yeah!
So
it's
it's!
I'm
learning
about
the
computing
ecosystem.
While
I
work
with
will
and
tara,
so
that's
awesome
and-
and
we
had
a
product
on
launch.
So
after
four
days
we
ended
700
users,
10
gigabytes
of
data
stored
and
some,
and
I
think
it's
cool,
because
it's
kind
of
like
a
crowded
product
hunt,
so
september
14th.
B
I
thought
I
told
you
I'd
explain
this
later
on,
but
it
became
official
martina
long
day
because
we
had
zero
downtime
and
the
product
looked
great
and
I
think
martina
is
responsible
for
a
lot
of
that
work.
So
thank
you
again
and
yeah.
We
we
brought
in
the
filecoin
features
so
in
our
settings.
I
think
it's
really
important
when
you're
designing
a
system
like
this
to
have,
you
know,
really
clear
settings
about
what's
happening,
so
we
let
users
opt
out
on
automatic
archiving.
They
don't
want
to
have
their
stuff
automatically
archived.
We
force
encryption.
B
So
only
you
can
retrieve
your
storage
deals
or
give
permission,
and
you
can
give
permission
through
slate
too,
so
that
would
be
cool
and
that
code,
you
know,
textiles
services
are
so
well
typed
like
you
could
just
go
into
their
documentation.
B
You
can
really
see
how
it
all
works
and
not
have
to
ask
them
any
questions,
and
so
you
know
we
at
slate
really
wanted
a
way
to
make
a
one-off
storage
deal,
and
so
we
were
able
to
kind
of
like
get
be
given
documentation
like
this,
and
it
became
very
clear
how
to
use
that,
and
so
you
know
instead
of
just
automatic,
follicle
and
archiving.
B
You
know
anything
you
want.
Is
a
storage
deal
through
site
too,
and
so
we
all
the
stuff
that
we
had
built
and
working
with
textile
got
to
this
point
where
it's
just
like.
Okay,
now
we
can
really
configure
things.
We
can
really
fine
tune
them
and-
and
I
know
right
now,
in
slingshot-
some
miners
are
concerned
that
you
know
they're
not
getting
the
deal
flow,
that
they
want,
and
so
now
that
we
have
this
config
that
landed
recently.
B
I
think
in
this
way,
designing
a
product
and
building
a
product
this
way
and
allowing
areas
to
be
iterated
on
you
know,
allow
us
to
be
flexible
and
grow
with
this.
This
new
ecosystem
and
the
changes
that
occur
with
it-
and
you
can
see
here-
you
know
it's
just
in
the
middle,
where
you
see
bug,
saw
create,
and
you
can
see
it's
like
some
of
my
code
is
really
scratchy.
B
So
it's
not
even
like
textile
code's
messy
it's
my
code,
so
you
know
I
can
determine
if
I
want
to
encrypt
a
deal-
and
this
is
how
we
actually
for
that
deal,
just
create
a
new
bucket
turn.
The
encryption
to
true
grab
the
root
cid
and
we
make
an
a
new
bucket
grab
that
key
and
then
make
an
archiving
deal
so
simple
as
that.
If
you
want
to
see,
if
you
want
me
to
explain
more
of
this
later
on,
I
totally
can
and
here's
our
storage
worker.
B
So
behind
the
scenes
you
know,
because
we
don't
hide
anything
here.
The
whole
code
base
is
open
source.
You
can
see
the
worker
that's
kind
of
going
and
looking
at
everyone's
configuration
cloning,
the
buckets
and
cloning
them
encrypted
or
open
and
making
the
storage
deal,
and
you
can
see
there's
there's
in
progress
archives
here
and
so,
when
we
know,
there's
more
successful
deals,
we'll
have
pages
where
you
can
see
all
the
deals
that
are
successful
and
I
think
that's
how
we
put
meaningful
data
on
the
file
coin
network.
B
I
think
you
start
from
user
experience
that
compels
people
to
put
stuff
that
matters
to
them.
I
think
you
look
at
the
union
between
two
people
and
you
see:
okay,
like
person
a
and
person
b
love,
this
type
of
data
sets
and
they're
growing
it
together,
and
you
start
gathering
all
this
wonderful
content
and
meaningful
stuff
like
music,
visuals,
diagrams
pdfs,
you
know
just
anything
right
and
then
boom.
You
just
make
a
storage
deal.
B
B
B
B
So
the
reason
why
we're
open
source-
and
it's
not
the
only
reason,
but
it's
because
we
want
to
show
everyone,
our
mistakes
and
all
the
pitfalls
that
we
experience
along
the
way
so
that
everyone
can
learn
and
so
slingshot
is
about
getting
applications
onto
the
filecoin
network
and
there's
no
better
way
in
my
opinion
than
just
ourselves,
throwing
ourselves
in
there
and
trying
our
best
to
prove
that
the
stuff
works
and
so
yeah.
I
think
like
stable
enough
to
be
taken
seriously
by
a
business.
You
know
all
services
I
use
have
bugs.
B
Cool
if
there
are
no
other
questions
or
well,
I
didn't
realize
that,
like
carson,
andrew
and
aaron
are
all
in
here,
wow
glad
you
guys
got
to
tune
in
hey
horace.
B
I
think
I
did
a
good
time
too.
I
was
hoping
to
end
at
9
45
but
looks
like
I
ended
at
9
51..
Thank
you,
zachary
yeah.
I
can
answer
any
questions
or
I
love
to
get
more
into
technical
details.
B
I
know
that
it
I
I
was
supposed
to
get
into
a
lot
of
deep
technical
stuff,
but
I
can't
help
that
you
know
slate
is
so
it
appears
to
be
simpler
because
textiles,
api
is
just
really
clean,
so
I
I
actually
can't
I
wish
the
deep
technical
side
was
actually
more
intimidating
or
like
had
a
lot
of
methods
I
had
to
worry
about,
but
to
use
slate,
it's
actually.
B
Thank
you
carson.
I
I
think
like
this
is
like
a
dream.
Come
true
for
me
is
to
have
like
such
a
passionate
group
of
people
who
want
to
make
a
thing
work
and
learn
at
the
fringes
together.
You
guys.
Are
you
really
like
for
a
year?
That's
been
pretty
crazy.
You
guys
are
probably
like
one
of
the
best
parts
of
this
year.
For
me,
it's
rare
to
have
this
kind
of
dedication
to
vision,
so
I'd
love
to
help
out
any
slingshot
teams.
B
If
I
could
as
well
to
get
oh
carlos
you're
hitting
like
a
really
really
special
place
for
me,
I
I
really
want
to
support
3d
modeling
software
and
slate.
If
I
can-
and
so
you
know,
if
we
look
back
at
one
of
my
older
slides,
you
know
talk
about
like
the
object
preview
right,
so
you
can
view
a
pdf
and
view
video
if
we
get
modeling
software
into
slate
and
we
start
creating
the
sdk
for
all
of
that.
So
you
know
slave
becomes
a
viewer
for
so
many
things
gosh.
B
A
B
Other
people,
when
you're
researching
certain
things
and
you
see
the
the
the
phylogenetic
genome
for
so
one
of
my
favorite
products
out
there
artsy
the
carter
cleveland,
had
this
concept
of
the
the
art
genome
where
you
can
trace
the
lineage
of,
like
you,
know,
art
throughout
history
and
where
it
came
from,
and
you
go
to
museums,
you
kind
of
see
this
too.
I
think
for
us.
B
So
it's
just
the
opportunities
are
endless,
and
then
I
think
you
know
textiles
infrastructure,
we're
only
touching
like
I
I'm
only
really
working
with
powergain
and
buckets,
but
I
could
build
real-time
chat.
You
know.
B
We
have.
We
have
the
api
now,
thanks
to
some
help,
help
from
digital
mob,
where
we
have
the
api
for
a
block
explorer
too.
So
we
can,
you
know,
start
doing
cid
search
on
the
file
coin
network.
We
can
start
doing
block
exploring,
and
so
you
know
we're
we
kind
of
we're
kind
of
a
place
where
I
think
we're
we.
We
can
innovate
in
many
directions,
and
so
I
love
to
hear
from
the
community
and
like
what
they
think
is
interesting.
B
You
know
I'm
I'm
kind
of
at
my
wit's
end,
sometimes
because
it's
like
you
know
sometimes
it's
so
galaxy
brain
for
me,
I'm
like
wow.
There
could
be
a
ton
of
stuff
that
happens
here.
What
have
we
not
explored
because
of
the
last?
You
know,
10
years
of
tradition
in
the
software
industry,.
A
Questions
you
know,
I'm.
B
Cool
cool,
well,
it
looks
like
we're.
I
think,
we're
good
to
wrap
up.
I,
if
there's
no
letter
or
questions
or.
B
Thank
you,
carlos
means
a
lot.
I
I
feel
like
just
really
lucky
to
even
be
able
to
work
on.