►
From YouTube: 🙌 IPFS Weekly Call 📡 💫 2019-09-02
Description
Topper Bowers from Quorum Control tells us about Tupelo, a platform for building Distributed Ledgers on top of IPLD and libp2p - https://www.quorumcontrol.com
Also: Help create a next-generation navigation for docs.ipfs.io!
Join a discussion on the existing navigation structure of the IPFS documentation site, plus a collaborative "card sort" exercise to explore new possibilities for getting around the IPFS docs.
6 September 2019
1700-1830 UTC
http://protocol.zoom.us/my/docsforce
IPFS Newsletter: https://tinyletter.com/ipfsnewsletter
A
A
A
Then
we
have
a
super
interesting
talk
from
top
of
hours
of
quorum
control,
very
much
looking
forward
to
that.
So
the
one
announcement
is
a
cool
for
help,
a
cry
for
help
to
create
a
next-generation
navigation
system
for
docs
to
ipfs
die.
Oh,
that
would
be
good.
If
you
are
interested
in
that
get
involved,
we
need
your
help.
Those
can
be
a
video
call.
The
URL
is
in
the
docs,
so
I'm
not
going
to
read
it
out.
It's
on
this.
It's
September
from
5:00
p.m.
to
6:30,
UTC,
cool.
B
Hey
thanks,
hi
everyone
looking
for
a
small
group
here,
so
one
feel
her
to
interrupt
me
and
to
have
a
little
bit
of
a
cold.
So
if
I
said
something
you
don't
understand
just
also
interrupt
me
and
say
what
did
you
say
and
I'll
repeat
it
I
think
what
I'm
gonna
do
is
give
a
little
background.
I
have
a
couple
of
like
short,
slides
I'm,
not
just
what
Tupelo
is
and
then
I'm
gonna
try
and
concentrate
most
of
it
on.
B
Actually,
you
know
against
all
the
demo
God's
doing
a
live
demo
here
against
our
test
net
and
because
I
think
that
will
help
illustrate
some
of
the
more
particle
labs,
specific
stuff,
the
IP
on
the
and
let
p2p
and
molecular
stuff
I
think
more
than
more
than
the
slides.
Well,
just
seeing
things
change
as
always
makes
that
kind
of
thing.
Click
for
me.
So
with
that
I
will
start
there
and
I
need
to
share
my
screen,
or
else
we
won't
be
able
to
see
it.
B
I
read
and
you're
hovering
over
my
present,
but
everyone
can
see
introducing
a
to
below
here:
yeah,
okay,
so
I'm
top
of
hours
on
the
CEO
of
quorum
control,
which
is
a
for-profit
company,
building
an
open
and
permissionless
tupelo
network,
which
is
a
new
layer.
One
DLT
platform
that
is
purpose-built
for
real
world
objects
and
applications,
as
opposed
to
finance
and
decentralized
finance
and
I,
think
to
illustrate
that
it
helps
to
talk
a
little
bit
about
existing
tech
here
and
I'm.
Gonna
go
kind
of
quickly
through
this
thing
stuff
so
feel
free
of
those
know.
B
If
there's
questions
and
I
will
back
up
and
answer
those,
but
I
think
it's
interesting
to
talk
about
to
blow
in
comparison
to
an
existing
tech
and
the
way
almost
every
current
blockchain
works
is
that
you
know
there's
a
chain
of
blocks
right
and
inside
of
that
block.
Every
transaction
is
put
into
an
order
right,
so
there's
transaction,
a
transaction
B
transaction
C
and
those
that
put
into
order
by
Bitcoin.
That
would
be
miners
and
some
other
thing.
B
You
need
to
see
which
one
of
those
transactions
goes
first
to
see
whether
Bob,
whether
Alice
paid
Bob
or
whether
Alice
paid
Carol
right,
and
so
they
they
spent
a
lot
of
work
just
to
make
sure
that
those
transactions
happen
in
the
correct
order,
so
that
Alice
can't
ever
spend
more
than
one.
You
know
more
than
her
balance
of
tokens
and
that
both
by
only
one
of
Bob
or
Carol
get
paid,
and
that's
now
the
reason
why
they
started
out
doing
this
whole
tour
global
total
ordering.
B
But
then
we
started
to
do
with
the
LT
is
to
put
all
sorts
of
other
crazy,
crazy
stuff
on
it.
This
is
a
picture
of
a
weather
station
right
and
so
what
we
you
know,
some
people
have
started
to
weather
data
into
blockchains
and
we,
the
thing
is
that
the
underlying
system
goes
through
all
the
same
process,
even
with
this
weather
data
right.
So
not
only
is
it
doing
the
order
of
the
weather
data,
but
it's
also
putting
Alice
and
Bob's
transactions
in
order
with
the
weather
data
right.
So
this
is
like
totally.
B
Work
for
the
system
to
do.
No
one
cares
whether
the
weather
that
came
before
Alice's
payment
or
after
Alice's
payment
and
additionally
lot
teams
are
awesome
at
keeping
data
in
perpetuity,
but
they
also
have
to
keep
data
in
perpetuity.
So
all
this
weather
data
that
were
or
whatever
kind
of
data
that's
in
the
blockchain
is
kept
around
forever,
which
I
think
at
the
time
of
creation,
you
know
maybe
made
sense
for
an
auditable
blockchain
for
specifically
for
money.
But
when
you
start
talking
about
applications,
it
becomes
a
different
kind
of
concern.
B
B
This
is
back
in
2016,
so
you
know
the
choices
were
more
limited,
but
there
was
still
a
lot.
There's
the
big
coin
and
the
theory
mm,
and
what's
now
called
Nano,
which
is
the
title,
is
called
brave,
lakhs
and
there's
all
these
other
esoteric
high
speed
ones,
and
what
I
realize
is
that
almost
all
of
them
weren't
gonna
work.
None
of
them
were
gonna
work
for
the
business
for
a
variety
of
reasons,
one
all
the
transactions
were
costly
and
costly
and,
like
the
worst
possible
way
to
write
like
variable
cost.
B
Sometimes
they
cost
a
penny
and
sometimes
they
cost
$40
or
$100.
It's
when
you're
trying
to
bake
that
into
the
cost
of
running
an
application
that
becomes
really
hard,
also
inflexible,
I'm
in
some
cases
like
Bitcoin,
it's
very
inflexible,
but
in
other
cases
too
you're
always
writing
smart
contracts
and
the
problem
with
smart
contracts
are
they're
great
in
some
certain
cases
and
are
in
fact
necessary
in
some
certain
cases.
But
in
the
90%
case,
people
are
rewriting
the
same
thing
over
and
over
again
here's
this
object.
There's
some
data
about
that
object.
B
Here's
who
owns
the
object
and
that's
about
it,
and
so
when
you're
talking
about
any
kind
of
real-world
application
like
that,
whether
data
like
a
supply
chain
like
any
kind
of
track
and
trace
system
even
stuff
like
crypto
kitties.
All
of
this
stuff
is
really
just
a
crud
application
that
people
are
writing
complex,
secure,
smart
contracts
for
and
then
the
unscalable
is
the
one
we
all
know
and
love
right.
B
It's
like
seven
transactions,
a
second
and
the
electricity
consumption
of
I,
don't
even
know
what
it's
up
to
the
Ireland
or
something
like
that
nowadays
and
so
Tupelo
is
it's
pretty
different.
It
models
each
individual
object
in
the
state
of
structure
which
I'll
get
into
more
detail
in
a
second
called
a
chain
tree,
and
those
objects
are
completely
separate
from
all
the
other
objects
in
the
system
they
the
chain
tree,
is
a
combination
of
a
block
chain
and
a
tree
of
current
state.
B
All
summed
up
into
a
single
IP
LD
object,
so
localized
in
sort
of
single
tips,
and
what
that
gives
you
is
that
you
have
the
tree
of
current
state
and
the
plus
the
history
of
all
changes
to
that.
To
that
data
structure
in
containing
the
single
data
structure,
and
it
doesn't
have
to
interact
with
any
other
chain.
B
Trees
have
its
own
state,
it'd,
be
verifiable,
and
because
it
we
can
define
tuple,
it
defines
some
certain
transactions
you
can
take
on
those
chain,
trees
which
I'll
get
to
in
a
second
again,
but
basically
they
provides
the
ability
to
set
data
set.
Who
owns
this
chain
tree
and
then
do
some
some
token
work
on
that
chain
tree
which
again
I'll
get
into
in
a
second?
But
the
big
deal
with
having
individual
objects,
modeled
separately,
is
that
the
consensus
algorithm
can
be
much
simpler
and
much
faster.
Also,
transactions
can
be
happen
asynchronously.
B
So
when
a
chit,
when
a
transaction
happens
on
one
chain
tree,
we
don't
have
to
wait
for
a
block.
We
don't
have
to
wait
for
transactions
and
three
other
chain
trees
to
happen.
We
can
deal
with
that
single
transaction
which
again
speeds
up
the
whole
process
and
doesn't
have
unnecessary
work
and
so
yeah.
We
have
a
new
chain
tree.
A
new
data
structure
called
a
chain
tree.
We
have
our
own
consensus
algorithm,
which
is
I
mean
it's
a
it's
not
entirely
our
own.
It's
based
on
other
people's
good
work.
B
We
have
this
other
cool
thing,
which
is
that
the
signers,
which
is
like
our
version
of
validators,
don't
actually
need
to
keep
the
history
of
the
blockchain
around
what
they
do
is
they
keep
a
a
basically
a
hash
table
of
a
ID
of
the
chain
tree
and
its
current
tip,
and
so
what
that
means,
though,
is
that
when
the
next
transaction
comes
in
the
chain
tree
can
pass
back
in
any
relevant
state
that
is
needed
for
that
transaction.
So
that
can
be
a
little
complicated.
B
B
You
know
Minna
Scrabble
dive
in
a
little
more
teeth
on
that.
What
you're
actually
doing
is
sending
a
message
to
the
peer
to
peer
like
pub/sub
network,
but
from
a
programming
language.
It
feels
like
you're
like
calling
an
api
and
getting
that
tip
back
like
the
signed
tip
of
your
chain
tree,
which
I'll
show
you
in
the
in
actual
code
in
a
second.
B
This
is
what
a
kind
of
a
conceptual
view
of
a
chain
tree.
There's
the
route
which
points
to
a
chain
and
a
tree
this
one
on
the
Left
would
probably
be
better
off
horizontal.
But
it's
like
you
know,
it's
abnormal
block
chain,
there's
the
last
lock
and
it
goes
out
to
the
first
block
and
the
tree
points
to
a
user
set
of
all
data
section.
Whatever
thing
user
can
do,
whatever
it
wants,
and
this
under
to
underscore
Tupelo
section
where
changes
are.
The
whole
thing
is
a
finite
state
machine.
B
B
Yeah,
so
train
trees
are
finite,
state
machines
and
they
transition
from
one
state
to
another,
so
from
an
IPL
DSP
for
one
tip
to
another
by
using
a
transaction
and
will
actually
by
appending
a
block
to
the
chain
section,
which
is
a
list
of
transactions
you
play,
those
transactions
go
the
chain
tree,
you
get
a
new
tip
and
then
the
signers
on
the
network
sign
that
tip.
And
then
the
world
can
know
what
the
new
state
of
their
chain
tree
is
so
I
already
went
through
these
and
I
can
demo
now,
but
I'll
pause
there.
B
There
any
questions
that
people
have
right
now
look
so
we
we
have
a
new
SDK,
which
is
our
to
blossom
SDK,
which
is
it's
kind
of
cool.
It's
it's
actually
assisting
a
uses,
I
think
like
all
of
go
and
all
of
JavaScript
of
protocol
labs,
the
p2p
and
and
IP
LT
so
be
rather
than
having
to
redo
how
the
transactions
happen.
In
JavaScript,
we
have
a
Lawson
file
generator
from
go
and
go
also
those
stuff
like
walk
the
hash
array,
Maps
tree
but
yeah
and
then
but
then
uses
JavaScript
to
do
stuff.
B
B
What?
Yes,
okay,
we
can
see
it's
pretty
easy.
So
this
we
get
a
community
which
is
a
layer
on
top
of
to
blow
they
actual
a
consensus
algorithm
so
to
blow
that
provides
some
things
for
the
pub
sub
Network
and
also
for
some
storage
for
that
public
members.
We
run
a
community
member.
Now
this
community
layer
is
decentralized
like
email
as
opposed
to
DLT
and,
like
anyone
can
run
one
of
these
nodes
where
to
blow
itself,
is
a
pretty
stake
system
working
off
the
state
coins
and
become
signers.
B
B
What
we're
gonna
do
is
we're
gonna,
create
a
key
and
we're
gonna
create
a
trading
card,
so
a
chain
tree
that
represents
a
trading
card,
we're
passing
in
a
block
service,
just
the
regular
ipfs
block
service
and
its
key
I
know
C
we're
setting
the
property
of
the
trading
card
ID
and
then
we're
gonna
make
a
block.
So
this
play
transactions
creates
a
block
on
the
trading
card
tree.
We're
gonna,
do
three
set
training
set
data
transactions.
So
we're
setting
slash
series
to
that
slash
item
two
number
48
here
and
slash
condition
to
mint
condition.
B
There
are
three
community
members
we're
just
waiting
for
those
to
update.
Then
we
get
the
ID
again
and
then
we
go
back
to
to
glow
and
get
the
new
tip
as
known
by
the
assigning
members
and
we'll
see
that
new
tip
here.
Those
are
assertions
which
won't
show
up
and
we'll
see
here
successfully
that
mustard.
So
assuming
that
everything
goes
well.
B
And
you'll
see-
and
this
is
actually
running
against
our
real
test
net
here-
not
against
the
local
of
mine,
so
you
can
see
the
speed
and
that
kind
of
thing-
and
this
is
finality-
it's
not
probabilistic.
This
happened,
I
can't
can't
be
undone
and
you'll
see
the
new
tip
carver
successfully
registered.
So
we
take
the
D
ID
and
if
we
go
back
to
this,
also
I
should
point
out
so
this
app
that
we're
gonna
use
here.
That.
C
B
Okay,
so
that's
this
one
check
yep.
Okay,
this
app
has
no
server.
I
mean
it's
running
on
github
io,
so
I,
guess
that
is
a
server,
but
I
mean
there's
no
like
back
inside.
It's
all
front-end
and
watch
them
running
in
the
browser
here,
and
so
this
is
a
Machane
straight
Explorer.
So
you
can
see
here's
the
D
ID
of
the
change
we
made
and
the
tree
we
can
click
into
it.
B
B
Just
quick
one
here
will
do
ownership
transfer.
So
with
this
one
we
do
we
get
the
community.
We
get
analysis
key.
Also,
it
gets
a
training
part
of
our
own
Bob
gets
a
key,
Bob
gets
an
address
and
then
Alice
immediately.
The
first
transaction
is
to
change
the
owner
of
the
trading
card
over
to
Bob,
and
so
we'll
do
that.
B
B
B
Okay,
let
me
go
to
the
tree
with
this
one
and
now
I
see
this
one
has
the
underscore
two
below
that
I
was
talking
about
before
authentications,
and
this
is
now
this
is
Bob's
address,
and
so
only
only
Bob,
whose
key
matches
this
address.
They
make
any
changes
to
this
chain
tree
in
the
world
and
tokens.
B
Make
sure
we
could
ID
this
time
around
okay
and
then
well
we're
gonna
each
chain
tree
convinced
it's
only
tokens
and
it
has
sort
of
human,
readable
names
or
arbitrary
names
inside
of
the
chain.
Trade
unique
to
that
change,
like
that's,
be
unique
within
the
chain
tree,
but
then
the
the
top
full
token
name
is
the
full
chain
treaty,
ID
:
token,
which
means
you
know
so
that
allows
you
to
do
kind
of
interesting
things
like.
If
you
have
a
piece
of
artwork
represented
by
a
chain
tree,
you
could
have
that
those
tokens
be
shares.
B
Other
things
you
can
do
with
them
is
like
usage
tokens
and
like
we
actually
have
a
scooter
company
building
and
they're
gonna
use
these
tokens
to
represent,
like
you're,
allowed
to
ride
on
a
scooter.
That
kind
of
thing
excuse
me,
so
then
you
establish
the
token
and
the
reason
why
you're
established
is
so
that
you
can
have
a
maximum
amount
ever
created.
That's
the
only
monetary
policy
we
currently
support,
but
looking
at
with
others,
and
then
you
mint.
B
B
We
do
need
to
establish
a
relationship
between
send
and
receive,
and
so
we
have
a
send
transaction
that
happens
and
you
get
a
receipt
of
that
and
then
you
give
it
to
the
next
person
and
they
call
it
they
they
will
receive
transaction,
and
so
it's
extend
that
off
chain,
some
of
the
payload
and
then
a
receive
transaction,
and
what
that
does
is
establishes
an
order
between
two
transactions
that
do
actually
need
that
order,
but
only
but
only
between
those
two.
So
let's
go
see
tokens.
B
Tokens-
and
you
can
see
that
there's
this
Alice
token
here
now
and
I-
know
we're
coming
up
quickly
on
time
here.
So
I'll
just
do
a
quick.
The
this
is
like
a
wallet
representation
of
that
same
thing,
basically
different
chain
tree
than
we're
working
with,
but
you
can
see
you
can
establish
token
at
the
MS.
Koken
will
make
it
so
that
there's
only
how
many
people
around
here
and
ever
to
be
allowed.
B
After
refresh
right
now,
UI
is
not
perfect
in
all
of
us,
but
now
there's
IPS
token
with
a
max
of
10.
We
can
mint
five
of
them
and
now
there's
a
balance
of
five
and
we
could
send
them
to
someone
with
a
D
ID
and
all
that
is
happening
in
the
same
systems
that
you
saw
so
I.
Think
that's
my
my
quick
overview
there.
B
If
that
was
ever
fixed,
we
could
probably
just
have
to
operate
directly
on
the
ipfs
network
and
the
whole
I
mean
changes
are
built
entirely
on
the
IP
LD
right
anybody
it
would
be.
It
would
be
reinventing
the
wheel
to
do
them
some
with
some
other
different
technology,
and
then
we
use
you
know
bit
swap
we
don't
actually
have
a
whole
ipfs.
B
C
Oh!
What
hey
thanks
thanks
for
this
I'm,
trying
to
wrap
my
head
around
the
way
that
you
get
around
the
total
ordering
and
and
also
like
the
implications
of
that
is
that
is
the
main
like,
like
innovation,
decoupling,
the
send
and
receive.
B
So
decoupling
the
send
and
receive
allows
for
the
tokens
to
establish
an
order
again,
but
for
likes
things
like
set
data
and
stuff.
There's
no
1s,
you
don't
have
to
have
any
any
tool
right,
think
the
transactions
happen,
but
yeah
the
decoupling
the
send
and
receive
is
for
sure
what
allows
on
the
receive
and
to
know
like
it's
like
it's
like
a
cashier's
check.
B
I,
don't
know
how
many
US
people
are
on
there
and
Europeans
tend
to
not
know
what
that
is,
but
so
it's
like,
if
the
send
token,
actually
deducts
the
balance
from
the
chain
tree
at
that
time,
and
then
you
have
this
payload
which
doesn't
need
to
be
kept
secure
because
it's
like
written
to
the
receiver
and
then
we
make
when
the
receiver
goes
to
receive.
The
network
is
already
signed
the
same
and
so
you're
establishing
this
order
between
the
two
of
them
again.
That
makes
sense
yeah.