►
From YouTube: IPFS Weekly Call 2019-01-21 🙌🏽📞
Description
IPFS Newsletter: https://tinyletter.com/ipfsnewsletter
A
A
A
But
before
we
begin
shorten,
can
you
unmute
yourself?
Please,
yes,
excellent?
Okay,
so
would
you
like
to
jump
in
yeah.
B
A
B
Okay,
I'm,
actually
gonna
jump
out
of
here
and
do
a
little
demo
here
for
a
second
but
so
telling
lab.
We
describe
it
as
a
data
diary
app
which
a
lot
of
people
are
like
what
is
that,
and
basically
it
is
a
way
to
collect
and
analyze
data,
which
sounds
really
open-ended,
because
it
is
a
lot
of
people
when
they
start
using
it.
Their
immediate
thought
is
to
use
it
to
track
health
things
like
symptoms
or
medications,
they're
taking
supplements
things
like
that,
but
you
can
really
track
anything.
B
B
We
have
a
lot
of
tools
in
the
app
that
allow
you
to
then
say:
okay,
you
know,
I've
been
tracking.
My
chores
and
I've
been
tracking
bird
sightings
and
it
turns
out
that
every
time
I
sweep
my
kitchen
I
see
a
blue
jay
in
my
backyard
cool,
like
things
like
that.
Obviously,
that
is
a
ridiculous
example,
but
I'm
guessing
wouldn't
actually
have
it.
But
who
knows
you
can
use
Sally
lamb
to
find
out,
maybe
there's
like
secret
magic
in
the
world,
but
we
don't
know
about
okay.
B
B
B
So
the
example
I'm
going
to
use
here
is
a
common
one
that,
where
you're
trying
to
figure
out,
if
you
have
a
food
allergy,
so
the
hypothesis
here
is
that
chocolate
gives
you
headaches,
which
chocolate
gives
me
headaches,
so
not
I
have
to
eat
a
lot
of
it.
Forgive
me
headache.
Don't
be
sad.
I
only
discovered
this
like
15
years
a
long
time
ago,
when
I
always
actually
working
on
the
legs
chocolate
website
and
they
sent
us
just
like
so
much
chocolate
love
you
guys.
B
A
C
B
Okay,
so
I'm
tracking
my
chocolate
and
then
under
symptoms
I
enjoy
yeah
thanks,
shocker
and
similarly
I
can
just
add
headache
here.
I
could
also
have
set
up
the
headache
tally
to
be
a
timer
if
I
wanted
to
know
how
long
my
headaches
lasted,
and
in
that
case
this
plus
sign
would
be
able
to
play
button
when
you
hit
play
and
start
the
timer
than
them.
B
B
B
Okay,
so
that's
just
a
basic
sense
of
how
the
app
works.
There's
a
lot
of
other
detail
in
here,
but
you
see,
does
mark
join
him
very
buddy.
Okay,
great,
why
don't
I
go
back
to
the
slides
and
then
mark?
You
can
tell
man
the
slides
and
you
can
talk.
Okay,.
D
D
Okay,
so
we
knew
if
we
wanted
to
use
IP
FS
for
this
project
because
of
the
decentralized
nature
of
it,
because
for
better
or
worse,
it
was
just
something
that
was
adjacent
to,
but
not
directly,
related
to
blockchain
database.
D
If
the
OS
just
decides
to
randomly
wipe
your
data,
which
has
happened
a
couple
of
times
with
iOS
particular
and
just
recently
with
Android,
you
know
the
phone's
hard
drive
space
is
getting
low,
I,
guess
the
local
storage,
not
just
regular
local
storage,
not
just
like
the
key
value
store
in
the
browser,
but
the
index
DB.
All
of
those
seem
like
they're
just
candidates
for
removal
without
asking
the
user,
or
anything
like
that.
D
D
D
So
it's
all
encrypted
using
the
Lib
sodium
libraries
and
for
those
of
you
who
are
the
nine-oh
they're
different
keys,
so
we
can
restore
them
using
a
32
byte
seed,
so
you'll
always
have
the
same
key
pair
you'll
always
be
able
to
encrypt
and
decrypt
the
same
data
set
and
so
on
and
so
forth.
So
that's
that
our
steps
here
so
we
use
IP
and
s
currently
to
do
the
backups.
D
D
We
stir
we
use
that
as
the
name
of
the
key
which
gives
us
the
APNs
name
because
they're
just
it's,
it's
bundled
there.
Together,
we
put
all
the
tallies
and
all
the
collections
up
in
a
nice
package,
unpin
the
old
one
if
it
exists
at
the
new
one,
pin
the
new
one
and
then
publish
it.
So
all
that
is
well
and
good,
except
that,
as
everybody
knows,
that
IPS
can
be
slow,
sometimes
so
Jordan.
You
want
to
talk
through
that
workflow
there
or.
B
Is
everything
involved
in
this,
which
is
maybe
not
that
interesting
to
everyone
as
well?
But
it
involves
sort
of
like
the
entire
process
like
across
page
loads
of
how
we
create
a
backup,
basically
and
so
to
what
Mark
was
saying.
B
I
feel
like
maybe,
if
I'm
gonna
go
to
the
next
one,
because
I
feel
like
it's
kind
of
important,
but
we
ended
up
making
a
couple
of
end
points
that
we
maintain
on
our
own
that
allow
some
of
these
processes
to
continue,
even
if
even
across
sort
of
like
page
sessions
or
review
sessions
for
the
user,
because
they
were
taking
long
enough
that
we
couldn't
just
like
make
the
user
sit
on.
The
page
like
here,
hang
out
for
five
minutes.
Well,
we
won't
be
make
this
all.
B
So
we
created
a
core
sort
of
like
intermediate
layer
on
the
server
there's
sort
of
traditional
API
endpoints.
That
basically
say
they
start
this
process
and
then,
as
the
user
is
continuing
to
work
around
the
app
and
do
so,
we
send
we're
listening
for
the
answers
these
calls
and
as
soon
as
they
come
back,
we
can
proceed
through
the
process.
So
that
kind
of
looks
like
this
there's
there's
a
mixture
of
things
we
created
and
I
PMS.
D
Yeah
I'll
take
it
from
here,
so
the
first
one
on
the
top
of
there
derive
a
penis
name
that
just
checks
to
see
if
the
key
had
been
created
already
and
if
not
just
return,
if
not
create
the
key
and
return
the
name
of
the
key,
which
would
give
you
the
IP
NS
name
and
if
it
does
exist,
just
return
the
name
of
the
key.
So
we're
not
like
we're
not
doing
anything
to
to
generate
the
IP
NS
name
over
again
or
to
try
to
resolve
it.
D
Somehow
we
just
sort
of
cache
it
and
store
it.
So
that
gives
us
a
pre,
immediate
thing
and
then
the
IP
n
has
published.
We
just
run
a
node
J
s
process,
an
HTTP
endpoint
that
just
backgrounds
that
task
and
talie
lab
will
just
check
in
with
it.
Every
few
you
know
every
so
often,
and
then,
when
it
finally
publishes
the
a
penis
name
it
will
it
will.
Let
us
know
the
other
trick
that
we
did
was
on
resolving
the
IP
NS
name.
D
We
reduced
the
amount
of
peers
needed
to
resolve
it
from
the
default
of
16
down
to
3,
which
works
for
our
purposes.
Once
we
scale
up
a
little
more,
we
might
have
to
increase
that,
but
we're
gonna
tune
that
as
we
go
and
that's
that's
about
it
there.
So
that's
what
we're
using
IP
best
for
now
and
I
PMS.
If
you
want
to
go
to
the
next
slide
and
what
we're
in
the
process
of
right
now
is
migrating
everything
from
our
custom,
IP
custom
index,
DB
store
and
IP
NS
to
orbit
DB.
D
So
what
that
looks
like
orbit,
DB
has
different
types
of
database.
You
know,
there's
a
key
value
database.
There's
a
document
there's
a
log
feed
and
what
its
gonna
look
like
is.
Each
tallies.
Gonna
have
its
own
orbit
DB,
which
is
a
key
value
store.
So
you
can
see
what
kind
of
an
approximation
of
what
the
JSON
looks
like
down
on
the
bottom
and
then,
if
you
look
at
the
counts,
the
counts
is
its
own
sort
of
nested
orbit
DB
inside
of
the
key
value
store
which
uses
the
document
type
index
by
timestamps.
D
So
it's
sort
of
an
ad-hoc
time
series
database
kind
of
embedded
within
the
key
value
store
and
we
only
store
the
address
there
and
that
allows
us
to
do
things
like
cache.
It
compress
them
back
it
up.
Things
like
that
next
slide,
yeah
and
the
cool
thing
about
orbit
is
that
the
address
is
the
orbit
DB
addresses.
You
know
you've
seen
like
slash
orbit,
DB
/,
qm,
blah
blah.
D
Those
are
relatively
consistent
because
they're
generated
by
their
a
multi
hash
of
the
name
of
the
database,
the
type
of
the
database
and
the
access
control
list
of
the
database.
So
if
two
users
in
two
completely
different
parts
of
the
world
generate
those
same
values
and
create
a
database
they're
working
from
the
same
database
and
if
you've
ever
used,
the
orbit
chat.
D
Client,
you've
seen
people
freaked
out
about
this,
because
they're
just
well
downloaded
the
running,
localhost
and
then
they'll
just
start
chatting
with
people
everywhere
and
they're
just
confused
and
they
say
like
I'm,
just
running
this
in
localhost.
How
is
this
happening
and
that's
how
so
it's
kind
of
cool,
so
we're
looking
at
that
as
an
alternative
theist
names.
As
far
as
addressing
the
data
in
town
next
slide,
you
know
there's
obviously
there's
caveats
store
with
DB.
The
dynamic
details
are
tough
because
you
basically
have
to
clone
the
entire
database.
D
If
you
want
to
change
the
ACLs
since
they're
so
so
baked
in,
and
you
know,
disclosure
I
work
with
hiya
in
the
orbit,
DB
team
and
smart
contract
based
ACLs
are
coming,
though
we're
not
sure
if
we're
gonna
use
them.
We're
not
sure.
If
that's
the
way,
we
want
to
go
with
tally
lab
directly.
We
sort
of
need
to
figure
out
a
migration
path
with
our
own
key
pair
system
and
that
and
how
the
ACS
might
work.
D
So
we're
gonna
explore
that,
but
we're
just
not
sure
at
the
moment
next
slide
and
that's
that's
basically
it
so
I
guess
we'll
just
open
it
up
to
questions
now,
and
this
is
leave
this
up
so
just
know
how
to
get
in
touch
with
us
and
there's
a
mailing
list.
If
you're
curious
about
following
our
work,
you
can
do
that.
A
D
D
E
D
E
D
I'll
definitely
check
that
out.
Yeah
are
you
know,
I
think
the
overall
philosophy
is
sort
of
the
high
tides
rise,
all
ships
philosophy.
So
we
want
you
know
if
orbit
improves
if
ipfs
improves
if
Qri
improves,
then
everybody
gets
to
use
it
and
benefit
from
all
the
tools.
So
it's
definitely
something
that
we
look
for
sure.
D
D
D
F
Thanks
dude
and
are
those
per
device
you
have
like
a
device
key
because
unless
you're
actually
like,
if
you're
bridging
it
over
to
each
device
or
each
person,
has
the
same
thing
same
same
Drive,
diffie-hellman
key,
then
it's
always
the
potential
of
a
security
breach
that
so
you
really
want
to
sort
of
actually
have
it.
The
per
device
here,
the
paired
key
with
that
user
just
take.
D
B
D
D
But
what
we
do
we
say
if
you
want
to
use
the
backup
system,
you
have
to
go
through
this
flow.
Where
about
you
basically
answer
knowledge-based
security
questions
that
only
you
would
know,
and
we
thought
really
hard
about
which
questions
to
ask
and
we
might
be
able
to
pull
it
up
if
people
are
curious.
But
what
what
it
means
is
that
you
answer
questions
until
you
have
32
bytes
of
deterministic
information
and
that's
used
as
the
seed
to
generate
those
keys,
we're
clear
that
you
know
don't
write
these
down.
C
C
All
right
so
I'm,
very
interested
on
so
first
of
all,
this
is
great.
I
would
love
to
have
these
for
myself
I
keep
trying
to
track
my
own
data,
and
it's
hard
like.
Please
help
us
solve
this
problem,
any
any
interesting
properties
on
the
data
model
or
our
easy
data
getting
stored
on
tally,
just
raw
files
for
now.
C
D
Yeah
we've
done
work
and
Jordan
has
done
a
lot
of
work
in
particular,
making
this
as
malleable
as
possible
for
people,
so
we're
not
trying
to
shoehorn
people
in
she
did
a
lot
of
research
on.
You
know
like
a
competitive
analysis
of
all
the
different
other
types
of
quantified
life
apps
out
there,
and
they
really
are
specified
to
one
thing.
You
know
it's
like
medication
and
symptoms,
app
or
fitness
app.
D
B
Add
to
that
and
say
that
so
the
basic
units
and
the
app
are
there's
a
collection
which
is
just
a
bunch
of
talent.
So
then
there's
a
talent
which
is
the
thing
you're
tracking
and
then
inside.
The
tally
is
a
single
instance
of
that
thing
happening.
We
call
account
and
each
count
has
data
that
you
can
associate
with
associate
a
numeric
value
as
I
get
out
a
note.
You
can
add
you
every
every
count
is
is
a
date.
B
It's
a
timestamp
that
you
can
also
add
an
end
Santa
which
will
make
it
a
duration
right
and,
as
we
go
obviously
like
more
use
cases,
crop
output
that
need
more
different
things,
and
so
we're
constantly
sort
of
thinking
through
how
we're
going
to
add
things
you
can
add
to
account,
and
so
like
the
first
thing
is
going
to
be
multiple
numeric
values.
So
you
know
maybe
you're
tracking
your
runs,
and
you
want
to
add
like
miles
to
expiration
aspect.
To
that
there's
a
mileage,
a
stipend.
B
B
So
it's
like
it's
like
this
balance.
You
know,
and
what
we
want
is
for
anyone
to
be
able
to
set
up
any
kind
of
data
that
they're
looking
to
capture
while
at
the
same
time
having
stock
templates,
where
you
can
just
install
how
it's
that
you
know
if
you're,
if
you're
tracking
your
runs,
you
can
assume
some
things
about
here:
distance
all
that
Sally.
It's
about
awesome.
C
B
C
B
And
then
some
really
good
feedback
I
mean
you're,
making
it
would
be
nice,
they
allow
users
to
import
their
own
famous
keys,
a
purse
specific
tally.
This
is
how
the
status
should
look.
This
is
what
I
want
you
to
present
me
when
I
add
in
the
account,
and
that
would
be
pretty
easy
for
us
to
do
so.
That's
that's
a
great
idea.
Awesome.
G
D
Yeah
I'll
answer
Molly's
question
a
little
bit:
well,
actually
yeah
Jordan
chime
in
here
too.
We
we've
just
started
working
with
behavior
therapists,
which
is
kind
of
an
interesting
one.
To
do
so,
you
know
my
my
son
is
followed
by
a
bunch
of
psychiatrists,
a
bunch
of
what
they
call
applied.
Behavior
analysis,
people,
and
you
know
they
still
keep
data
in
binders
and
you
know
draw
line,
charts
themselves
with
graph
paper
and
stuff,
so
we're
sort
of
like
hey
check
this
out.
This
might
be
useful
to
you.
D
B
Also
I
mean
it
should
be,
maybe
implied
from
how
we
describe
it
so
far,
but
it's
typical
and
its
GDP
are
compliant
a
box
like
the
way
you
constructed
it.
We
can't
actually
see
anything
that
anyone
is
tracking,
even
as
people
making
this
app.
So
it's
a
great
health
care
these
case
you
know,
because
we
don't
have
to
do
anything
special.
It's
just
already.
D
A
We
really
are
running
short
on
time.
Yeah
I
really
do
want
to
get
Molly's
question
in
please.
A
D
G
A
Thank
you
very
much
and
I
will
see
everyone
next
week,
so
have
a
great
week.
Take
care,
bye,
bye,.