►
From YouTube: NFT Community Call #4
Description
The fourth NFT Subgraph Community Call took place March 25th.
Join the community and learn more on Discord:
The Graph: https://thegraph.com/discord
The Graph NFTs: https://discord.com/invite/EdZE6EmVKC
A
Hi
everybody
welcome
to
the
fourth
nft
subgraph
call.
We
took
a
a
week
off
last
week
and
we're
back
and
we've
got
a
few
new
people
on
the
call
today.
If
anyone
wants
to
just
give
a
shout
out
and
share
either
about
their
company
or
what
they're
working
on,
we
can
do
a
round
of
intros.
If
there's
there's
anybody
new.
B
Hi
everyone,
I'm
neta,
I'm
helping
out
with
the
foundation
team
nice
to
meet
you
all.
A
Hey
nice
to
meet
you
and
I
guess,
did
you
you
contributed
mostly
to
the
to
that
google
doc
that
was
sent
out.
A
B
A
And
you
used
to
be
an
enzyme.
B
A
All
right
well
yeah,
let's,
let's
jump
in
before,
we
start
just
a
little
house
cleaning.
So
you
know
we've
got
our
housekeeping.
We've
got
this
nft
air
table
set
up.
If
you
don't
have
access
just
you
know,
shoot
me
a
message
and
here
we're
tracking
all
the
different
nfts
that
we
want
to
have
subgroups.
For
so
I
see
that
people
added
a
few
more
and
and
and
if
anyone
has
a
chance
to
go
through
and
just
kind
of
update
the
status
on
any
of
these.
A
A
Big
thing
that
we've
been
focusing
on
in
the
last
few
calls
has
been
standardization,
and
we've
talked
about
a
few
of
these
different
areas
that
we
can
standardize
on.
And
you
know
one
of
the
the
things
that
we
we
discussed
was
how
you
know
for
for
many
of
these
there's
a
difference
between
kind
of
the
erc
standard
for
the
on
chain
data
data
and
what
we're
now
calling
the
grc's
or
which
is
basically
the
the
graph
equivalent
of
standards
at
the
the
sub
graph
level.
A
And
so
you
know
it's
something
that
we
can
continue
to
look
at,
where
it
might
be
that
it's
worth
creating
some
new
erc's
for
for
the
parts
that
are
on
chain,
and
then
you
know
looking
separately
at
what
the
standard
should
be
at
the
api
layer
with
with
the
subgraph
schemas.
A
So
we
had
some
great
discussions
on
the
last
call
and
I
think
we
can
just
jump
straight
into
it
to
go
over
some
of
the
proposed
standards.
For
you
know
the
the
contracts
data,
the
nft
data-
and
you
know
this
ever-growing
metadata
field
to
see.
If
you
know
we
can
get
some
alignment
on
on
some
of
these
fields
so
and
see,
I
know
the
the
foundation
team
did
some
some
great
work
on
this
since
the
last
call,
and
would
you
guys
want
to
kick
that
off.
D
Yeah,
so
I
think
somebody
compiled
a
bunch
of
the
notes
from
the
last
from
the
last
call,
and
maybe
I
can
just
share
this
my
screen
and
walk
through
that
for
people
that
might
not
have
the
context.
D
So
everybody
see
this
yeah
so
yeah.
I
think
this
was
this.
D
Is
the
dock,
where
we're
trying
to
consolidate,
like
just
everything
that
has
been
discussed,
there's
quick
links
to
the
telegram
group
and
the
air
table,
I'm
not
going
to
go
through
so
last
time
we
talked
about
metadata
creator,
field
and
fees,
I'm
not
going
to
go
through
those
in
too
much
detail
and
I'm
actually
going
to
jump
down
here
to
the
general
subgraph
schema
which
which
ryan
from
webchat
proposed-
and
I
think
this
is
actually
a
a
great
starting
point
and
we've
actually
hashed
out
a
bunch
of
things.
D
So
I
think
here
we
we
have
a
subgraph
entity
for
what
a
contract
should
look
like
or
when
an
nft
should
look
like
and
we've
delineated.
You
know
just
what
lives
on
the
contract
itself,
so
the
name
and
symbol.
Those
are
not
necessarily
nft
specific
things,
but
like
are
more
attributes
of
the
contract
and
then
for
the
nft,
we're
saying:
okay,
the
creator
should
be
an
account.
There
should
be
a
creator
fee,
that's
self-reported,
and
that
could
be
a
fee
entity.
D
We
can
hash
out
as
well,
and
the
metadata
is
is
like
another
thing
that
we
can.
We
can
hash
out,
but
I'd
say,
let's
start
here,
because
I
I
think
we
kind
of
ended
the
last
call
with
general
alignment
that
this
is
the
right
direction,
but
I
would
say:
let's
make
sure
that
this
is
what
we
want,
and
then
we
can
drill
into
each
of
the
entities
that
we
have
not
fleshed
out.
E
No
I'm
on
board.
I
think
this
is
a
good
place
to
start.
I
have,
I
can
drop
in
the
sale
event,
transfer
event
and
account
entities
that
are
kind
of
off
the
grid.
Right
now,.
D
Cool
yeah,
I
think
getting,
I
think,
getting
those
on
paper
as
strongman
and
using
that
to
to
drive
discussion
will
be
helpful.
Sure.
A
Would
you
be
able
to
paste
it
right
in
maybe
at
the
bottom
of
the
document
on.
A
All
right-
and
I
I
saw
a
comment
in
that
document-
where
I
think
adam
was
asking
you
know
if
there
were
some
fields
that
were
added
to
that
top
level
just
for
convenience,
because
I
guess
that
that
data
was
also
included
somewhere
else,
and
you
know,
I
think
you
know
general
philosophy,
I
think,
for
kind
of
subgraph
schema
design
is
that
it's
actually
okay
to
have
some
duplication.
A
You
know,
I
think
it's
different
than
like
database
tables
or
something
where
you
want
to
have
the
data
normalized.
I
think,
at
the
api
layer.
It
actually
makes
sense
to
have
different
fields
that
are
either
like
computed
data,
or
you
know
things
like
that.
That
you
know
are
just
kind
of
for
convenience
because
that's
that's
part
of
what
you're
able
to
do
at
that
layer
is
kind
of
explode.
The
data
out
and
just
make
sure
that
it's
available
in
the
most
convenient
format
for
for
the
application.
D
Cool
yeah
ryan:
I
see
you
added
count,
sell
event
and
transfer
event.
A
Yeah
and-
and
I
just
there
was
a
message
from
hadrian
in
the
chat
saying
you
know:
if
everybody
does
that,
then
all
subgroups
will
have
a
contract
entity
and
it
all
have
different
meanings,
and
so
that's
that's
also
a
fair
point.
So
yeah
it's
it's
a
balance.
You
know,
I
think,
with
any
kind
of
design.
A
You
know
you
want
to
do
the
the
proper
kind
of
data
modeling,
but
yeah.
It's
a
trade-off
there
that
I
think
api
designers
can
kind
of
work
through.
D
So
I
think
I
think
we
give
people
time
to
people
can
absorb
the
context
for
the
earlier
parts
of
this
stock,
but
maybe
are
there
other
reactions
to
the
contract
and
nft
entity
before
we
drill
deeper
into
the
other
entities
that
are
kind
of
mentioned
here.
A
Yeah,
this
is
an
area
where
I
think
you
know
we're
interested
in
adding
some
some
new
tools
to
to
help
people
with
how
they
generate
ieds.
A
A
Do
people
assume
in
their
applications
that
these
ids
are
like
stable
and
don't
change
or
you
know?
Is
it
something
where
you
know
if,
if
in
your
subgraph,
you
just
changed
one
day
how
the
ids
were
computed?
Would
that
break
anything
in
your.
B
A
Okay,
so
so
in
this
case,
there's
actually
a
reason
why
contract
id
token
id
is
is
an
especially
good
idea.
F
I
would
say
another
reason
why
events
data
concatenated
as
an
id
is
very
important
is
because
of
the
constraint
that
you
can
only
load
by
id.
So
if
you
want
to
update
a
given
entity
in
the
middle
of
a
handler,
you
need
to
be
able
to
fetch
that
id
based
off
of
the
event
information,
and
so,
if
you
know
the
token
id
for
example,
then
you
can
grab
that
nft
entity
and
make
the
necessary
updates
that
you
want.
F
So
that
was
the
primary
modification
for
why,
in
the
zoro
subgraph,
we
we
do
that
kind
of
stuff,
at
least
well.
It
makes
sense.
F
Also,
hey
everyone,
sorry
for
missing
the
last
week,
or
so
I
feel
a
little
out
of
the
loop.
But
I'm
here.
A
Yeah
thanks
for
making
it,
I
think
it
it.
It
helps
to
get
everybody.
You
know
in
the
same
virtual
room,
yeah
get
get
to
consensus,
so
cool
like
what
do
we
think
are?
Maybe
the
more
controversial
parts
of
you
know
the
schema
here
and,
and
I
think
people
should
feel
totally
fine
bike
shedding.
You
know,
I
think
you
know
when
you're,
when
you're
trying
to
get
a
feature
out
the
door
bad
time
to
bike
shed
when
when
you've
got
everybody,
you
know
working
towards
standards.
F
I
have
a
couple
of
high
level
questions
and
maybe
a
couple
minutes.
The
first
question
is
for
the
metadata
contract.
Maybe
this
was
talked
about
last
week
the
week
before
about
super
familiar
with
contract
metadata.
E
Yeah
they
have
for
every
crypto
punk.
They
have
like
the
the
sum
of
all
of
the
punks
is
one
large
image
and
then
the
hash
of
that
image
to
verify
that
it
is
the
right
contract.
I
guess,
but
it's
one
thing
where
it's:
it's
definitely
contract
metadata,
yeah.
B
D
I
mean
I'm
imagining.
Let
me
know
this
off
like
a
lot
of
the
stuff
like,
I
think
we
should
drill
into
what
what's
useful
and
then
I
think,
there's
probably
also
a
conversation
like
what's
required
and
what's
optional
right,
so
I
think
for
most
of
us
probably
the
contract
metadata
is
not
that
useful,
but
then
the
nft
metadata
it
is
useful
and
so
at
the
contract
contract
level
like
this
is
probably
optional,
but
then
this
is
probably
required
down
here.
A
And
that
that's
what
the
bangs
for
right!
So
so,
if
it's,
if
it's
optional,
I
just
maybe
remove
the
bang.
B
F
Okay,
I'll
go
a
couple
of
other
comments
on
nft,
I'm
a
believer
of.
We
should
store
the
block
hash
over
the
block
number
just
like
for
forks
and
stuff,
and
I
it
looks
like
block
I'm.
Imagining
is
block
number.
F
Okay,
cool
yeah
like
so
so
we
use
like
we.
We
have
a
couple
of
fields
for
like
when
things
are
created,
so
we
have
like
a
created,
app
block
number
and,
like
I
created
that
time
stamp,
and
then
we
also
have
time
stamps
for
like
when
things
are
burned.
So
I
don't
know
how
many
of
the
contracts
that
are
standardized
are
inherit
the
erc721
burnable
standard,
but
I
think
it's
helpful
to
have
like
any
of
those
like
contract
level.
F
Events
on
a
single
nft
having
just
like
explicit,
timestamps
and
block
number
relations
and
hashes
so
that
people
can
derive
when
those
things
happened
and
make
sure
that
things
are
correct.
A
Yeah
now
one
thing
that
we're
going
to
be
working
on
is,
you
know,
I
think
we've
talked
about
both
of
these
features.
One
is
network
subgraphs,
so
like
an
ethereum
subgraph,
where
we're
indexing,
the
raw
like
blocks
transactions
accounts
all
that
kind
of
data
and
then
step
graph
composition,
so
it
it
might
be
that,
ideally,
once
we
get
those
two
features
in
what
we
have
here
is
just
like
a
reference
to
an
actual
like
block
or
transaction
entity.
A
Now
it
could
be
like
the
it
could
be,
it
could
be
both
and
then
that
way
you
could
traverse
and
get
all
of
the
additional
metadata
about
those
raw
blocks
or
transactions.
D
D
D
Which,
I
think
hadrian
I
see
a
comment
here.
I
think
the
intent
is
like
this
is
for
the
mid
right
right,
right,
yeah.
B
A
D
Yeah,
I
think
I
think
you
need
you're
asking
why
why
sales
and
transfer
not
like
mint
as
like
a
separate
thing
here,
and
I
think
it's
because
mint
and
burn
happened
once
whereas
sales
and
transfer
can
happen
multiple
times.
B
D
So
couple
couple
directions:
I
think
we
can
take
this.
I
think
meta
nft
metadata
is
a
big
hairy
one.
So
maybe
we
save
that
for
later,
but,
like
a
transfer
event,
I
think
is
pretty
straightforward.
Sale
event.
I
know
there's
been
concern
over
before
in
terms
of
just
like
how
how
standard
it
is
across
all
nfts.
Do
we
want
to
chat
about
sale
event
first
and
then
go
into
the
metadata.
D
Okay,
hot
takes
on
sale
event.
F
Yeah,
I
I
saw
a
proposal
that
I
thought
was
really
interesting
about
like
having
a
notion
of
like
general
fee
recipients.
I
I
don't
know
how
that
conversation
went
last
time,
but
I
can
speak
selfishly
in
zorah,
there's
more
than
just
a
a
buyer
and
a
seller.
There
are
actually
multiple
entities,
potentially
that
can
receive
or
be
the
recipients
of
a
successful
sale.
F
So
obviously,
in
my
in
my
book
or
from
my
perspective,
I'd
love
for
it
to
be
more
general
but
yeah.
I
honestly
have
not
thought
enough
about
it.
I
know
that
I
was
sort
of
tasked
with
doing
this,
but
things
have
gotten
a
little
crazy
in
the
last
couple
weeks.
So
I
apologize.
A
F
Yeah,
so
in
azura
nft
there
are
three
participants:
there's
an
owner,
so
the
current
owner
of
the
nft
there's
the
creator,
and
then
there
is
this
notion
of
the
previous
owner,
and
so
in
the
contracts
in
storage.
There
is
like
bid
shares
or
what
we
call
the
shares
of
the
sale
for
all
of
those
participants.
F
The
creator
and
the
owner
are
like
obligatory
and
the
previous
owner
is,
depending
on
the
like
the
providence
of
the
nft,
and
so
whenever
a
sale
is
is
finalized
before,
like
as
a
part
of
the
finalization
of
a
sale,
it
like
will
split
the
sale
up
into
the
necessary
shares
for
each
individual
entity
and
transfer
to
each
entity.
Discreetly,.
A
So
is:
is
this
a
place
where
it
would
make
sense
to
have
different
types
of
sale
events,
and
maybe
there's
like
an
interface,
and
you
know
if
it's
would
you
call
it
as
a
znft
or
a
zft?
If
it's,
if
it's,
that
type
of
nft
then
use
you
know
this
other
type
of
sale
event.
F
Ooh
and
like
an
azura
nft
is
composed
of
multiple
sale
events,
and
I
think
this
is
also
nice
for,
like
a
regular
nft
sale
where
there's
an
owner
and
a
creator
as
well,
where
it's
composed
of
two
different
sale
events
right
and
maybe
maybe
the
fields
can
be
more
generalized.
It's
not
like
buyer
and
seller,
but,
like
I
don't
know,
recipient
and.
F
And
another
word
for
buyer-
I
don't
know,
I
don't
want
to
hijack
the
conversation
as
well,
but
that's
an
interesting
thought.
I
like
the
idea
of
having
like
almost
even
like
a
sub
unit
like
sale
event,
is
a
subunit
and
you
can
have
numerous
sale
events
and,
like
an
actual.
E
D
I
wonder
if
there's
something
we
could
do
with
the
id
the
id
so
there's
couple
groupings
already
built
in
right.
We
have
the
the
web,
the
block
and
we
have
a
transaction.
We
have
the
hash
transaction
hash,
and
so
maybe
like
I'm
in
favor
of
like
turning
this
into
like
an
interface
and
basically
like
most
people
would
just
use
the
vanilla
cell
event.
But
then,
if
you're,
if
you
are
building
things
that
are,
that
have
a
little
more
functionality
than
what
what
you
can
do.
D
For
example,
like
breakfast,
I
think
for
zora.
You
could
probably
just
have
the
same
transaction
hash
just
happens
to
have
multiple
events
and
you're
just
reporting
each
of
those
events
separately
and
then,
if
you're
needing
more
fields,
you
can
extend
the
sale
event.
Interface
to
add
those
fields
for
your
xor
specific
sale
events.
But
then
everybody
else
will,
like
consumers,
will
be
able
to
just
walk
through
all
cell
events
and
not
have
to
know
the
underlying
implementation.
A
D
We,
I
think
I
I
think
majority
of
nfts
will
probably
just
have
like
buyer
and
seller,
but
then
I
think
for,
for
example,
for
the
zora
case.
This
is
where
maybe
you
can
add
a
field.
That's
like
the
previous
owner
or
something
like
that,
and
then
you
leave
one
of
these
fields.
Blank
and-
and
then
I
think
at
that
point
it'd
be
a
conversation
between
like
zora
and
whoever's
consuming,
there's
a
graph
on
like
how
to
interpret
that.
A
Yeah-
and
I
think
there
could
still
be
a
conversation
about
if
it
makes
sense
to
to
just
change
the
names
a
little
bit
so
that
it
would
work
for
both
use
cases.
I
think
that
that
could
still
be
on
the
table.
F
Yeah,
I
I
tend
to
agree:
how
do
you,
how
do
you
all
envision,
creatorship
like
in
the
case
where
a
sale
leads
to
the
creator
receiving
some
funds
back?
D
That's
a
good
question.
I
mean
one
thing:
we
could
also
do
to
make
it
more
explicit.
Opening
hearing
feedback
here
is
like
there
could
be
like
a
like
a
fee
event
or
like
a
like
a
royalty
event
right
where
you're
there's
like
marketplace
fees,
that
everybody
is
collecting
and
then
there's
royalty.
D
That
goes
to
creator,
and
that
could
be
his
own
event
and
we
don't
have
to
lump
into
sale,
because
I
think
the
sale
event,
if
I'm
understanding
correctly
for
for
platforms,
say
like
for
upshot
for
whoever
you
guys
are
probably
using
this
to
figure
out
like
when
the
entity
changes
hands
right.
So
I
don't
want
to
pollute
that
with
too
much
metadata
with
too
much
data.
D
So
this
can
be
like
a
let
me
know.
People
have
other
better
names
for
for
this,
but
this
is
basically
like
a
fee.
Then.
C
Is
the
thinking
of
putting
a
royalty
in
every
event
that,
like
even
if
creator
fee,
is
global
for
the
contract
or
global
for
a
particular
creator
that
it
might
be
updatable,
so
you'll
want
the
historical
record
of
like
if
it
was
updated
for
past
events,
what
the
fee
was
at
the
time?
I
think
I
think
that
makes
sense.
If
that's
what
the
thinking
is.
F
D
F
F
F
Cinderworks
and
then
maybe
another
thing
is
like
the
currency,
I
I
don't
know
this
is
another
net,
and
this
is
another
specific
thing
for
zora,
but
we
accept
bids
in
any
currency
and
so
generally,
whenever
we
use
amounts.
We
also
like
are
explicit
in
the
currency
field
and
I
imagine,
as
erc
721
evolves.
More
contracts
will
also-
or
you
know,
market
contracts
will
accept
bids
and
you
know
transfers
and
stuff
in
arbitrary
currencies.
So
we
might
want
to
add
that
as
well.
C
How
do
we
handle
protocol
upgrades
for
currencies
like
side
to
die?
For
instance,
would
we
use
address,
or
a
short
name
or
other.
F
Yeah,
I
I
think
address
is
definitely
the
the
primary
key
but
then
dealing
with
e.
A
This
is
another
one
where,
where
I'd
love
to
have
like
a
separate
sub
graph,
you
know
with
entities
for
like
you
know,
I
don't
know
if
it'd
be
tokens
or
coins
or
assets.
I
guess
I
guess
it'd
be
assets,
and
then
you
could
just
reference
like
the
the
asset
across
subgraphs
like,
like
you
know,
we
mentioned
with
with
the
block
eventually.
D
D
It
seems,
like
said
by
saying
the
same
thing,
so
this
could
be.
What
would
this
be
address
or
contract.
F
Yeah
and
then
that
way
we
can
easily
move
into
like
an
actual
currency
entity,
that's
provided
by
the
graph
in
a
future
date.
Right,
I
know
unit
swap
does
something
similar
where
they
actually
like,
pull
in
all
the
native
fields.
So
they'll
fetch
like
pertinent
data
from
the
erc20
contracts,
things
like
the
symbol,
the
decimals
etc,
but
it
sounds
like
we
don't
need
to
do
any
of
that
stuff.
D
B
D
D
Please
make
sure
if
you're
sharing
this
doesn't
get
until
a
while
and
cause
a
bunch
of
mess,
but
you
guys
should
all
have
edit
access.
D
D
I
know
the
open
c
team
looks
like
dan
and
dan's
in
here
from
open
c
yeah,
I'm
wondering
if,
if
you
guys
have
any
any
thoughts
on
these,
because
I
know
you
guys
have
dealt
with
a
bunch
of
different
currencies
and
like
a
bunch
of
different
types
of
like
trades
and
sales,
and
I
wanna-
I
think
you
guys
probably
have
perspective
here
on
what
a
good
schema
might
look
like.
C
G
D
Have
oh
yeah,
I
mean
this:
this
stock,
you
guys
should
have
access,
so
you
guys
can
review
and
absorb
the
context,
and
then
I
think
we
can
continue
to
work
on
this.
D
I'm
I'm
loving
the
live
refactoring.
So
as
we
speak,
it's
getting
better.
F
I
think
the
same
thing
I
don't
know
if
this
comment
already
exists,
but
the
block,
hash
and
timestamp
things.
It
might
honestly
be
nice
to
standardize
it
across
all
things
that
we
save
it
on
it's
just
like
this
is
these.
Are
the
these?
Are
all
the
block
information
that
this
that
this
event
was
created
at
so
like?
F
I
know
we
use
mented
for
for
nft,
but
like
created
at
for
all
entities,
might
just
be
more
standard
and
easier
to
parse
through,
and
everyone
knows
what
it
is
versus
having
different
terminology
for
the
same
thing
on
different
entities.
D
Yeah
that
makes
sense
to
me
too.
I
mean
I
think
you
need.
What
are
you
guys
like
the
chain
native
entities
like?
Is
that,
like
a
is
that
coming
soon
like?
Should
we
should
we
design
the
entities
assuming
that
that's
going
to
come,
live
or
should
we.
A
I,
for
now
I
would,
I
would
assume
it's
not
there,
so
I
mean
we're
going
to
be
working
on
it
for
a
kind
of
q3
sort
of
time
frame.
So
I
imagine
you'll
want
to
have
this
before
then.
A
Right
yeah
just
and
I
guess
you
could
stuff
in
all
the
relevant
blocks
and
you
don't
need
to
have
every
single
block
you
know
ever
mined.
You
know
it
could
just
be
like
you
know,
a
block
cache
of
all
the
blocks
where
nft
things
have
happened.
A
F
A
From
from
the
last
year
forward
for
sure.
D
D
I
think
the
next
big
thing
that
hatch
out
is
probably
metadata.
I
know
that's
a
hairy
one.
So
are
people
ready
to
move
on
to
that.
B
I
just
wanted
to
add
that,
as
I
already
said
in
many
of
these
calls,
I
have
a
library
with
tools
that
I
linked
again
in
the
comments,
and
I
already
have
standard
interfaces
for
events
and
transaction
that
kind
of
stuff.
I
don't
think
I
have
something
for
block,
but
I
use
it
none
by
sub
graph,
and
I
invite
you
to
extend
so
here
you
can.
There
is
event
a
transaction
and
both
have
a
cs
file,
and
so
this
is
creating
it
and
you've
got
a.
D
D
Okay,
I
think
I'm
gonna
move
on
to
metadata
and
for
this
one
I'm
actually
gonna
put
sam
on
on
the
and
put
sam
on
to
share,
screen
and
walk
through,
maybe
because
sam's
put
a
lot
of
thought
into
metadata.
Just
because
he's
been
interacting
with
a
lot
of
different
nft
platforms.
So
we'd
love
for
for
him
to
walk
through
his
thinking
and
his
reasoning
and
figure
out
what
we
can
drive
to
alignment
on
as
a.
H
Group
cool
thanks
el
pizo.
Let
me
share
my
screen.
A
Cool,
maybe
just
zoom
in
a
little,
if
you
can
command
plus.
H
So
yeah,
so
this
is
kind
of
going
over
for
those
of
you
that
were
here
two
weeks
ago.
This
is
obviously
the
the
document
that
was
shared
out
from
then
and
then
shared
in
the
the
telegram.
H
Essentially,
I
think,
where
we
landed
on
at
the
end
of
that
conversation
was
following
a
similar
solution.
That
zora
is
good
actually
to
have
breck
on
the
call,
because
a
lot
of
this
was
stemming
from
their
concepts
of
how
they've
created
these
metadata
schemas,
which
I
think
is
a
really
solid
idea.
Moving
forward.
H
H
If
we
wanted
to
go
that
specific,
but
these
are
just
json
schemas,
so
they
could
essentially
be
kind
of
composed,
together
with
platform,
specific
metadata,
the
idea
being
that
we
could
incorporate
this
obviously
into
subgraphs
and
because
the
json
schemas
can
be
introspected
by
just
we're,
essentially
apply
a
kind
of
schema
name
and
similar
to
how
the
zora
the
sdk
works,
where
they
have
a
repo
that
just
collates
them
all
together
and
there's
a
function
that
can
kind
of
reach
out
to
those
pull
down.
H
The
metadata
schema
using
a
you
know,
json
schema
library
and
then
kind
of
understand
like
what
fields
are
expected
to
be
in
the
metadata.
What
fields
are
obviously
optional
or
mandatory.
I
think
I
don't
think
there's
been
a
huge
amount
built.
On
top
of
this.
I
tried
to
add
some
of
the
feedback
that
I
got
from
last
week
from
the
last
call
that
we
had
into
this
by
adding
like
optional
kind
of
params
into
the
into
these
kind
of
mock
interfaces
or
mock.
Schemas.
H
Sorry,
not
mock
schemas
kind
of
examples,
but
yeah
I'm,
I'm
kind
of
like
I'd,
be
interested
breck
to
kind
of
get
your
take
on
it
as
well.
Obviously
you
guys
have
the
the
zorro
one
so
I'd
be
interested
to
kind
of
see
what
what
your
lessons
learnt
from
that
are.
F
Yeah,
I
so
high
level
thoughts.
I
really
like
the
idea
of
having
a
core
schema
for
like
an
image
type
or
not
internet
shape,
but
a
a
type
of
content
right
now.
Our
schemas
are
definitely
geared
towards
like
platform
specific
metadata
so
like,
if
someone's
building
on
zora-
and
they
have
you
know
like
they're,
an
audio
platform
or
you're
like
a
script,
writing
platform
or
whatever.
You
might
have
your
own
metadata
for
your
own
platform
use
case.
H
Yeah,
I
guess
the
one
thing
that
I'm
definitely
less
familiar
with,
and
we
touched
upon
it
at
the
end
of
the
last
school-
is
how
to
how
this
could
be
incorporated
into
a
graph.
Obviously
one
way
is,
I
guess,
there's
a
few
different
options.
Right,
like
one
way
is,
there
is
a
url
leading
out
to
this,
which
I
guess
is
like
how
some
things
work.
H
Now,
where
you
just
add
in
the
metadata
field
in
the
subgraph,
it
would
just
link
directly
out
to
the
metadata,
and
you
would
have
to
do
like
passing
of
the
schema
and
look
up
and
everything
like
that.
Another
thing
could
be
to
kind
of
like
potentially
stringify
or
just
dump
the
contents
into
the
metadata
field
as
like
a
sub
object.
H
Potentially,
if
that's
even
possible,
I
don't
even
know
if
that's
even
a
thing
that
could
happen
or
stringified
and
the
other
one,
I
guess
would
be
to
kind
of
construct
a
a
kind
of
entity
similar
to
how
the
scheme.graphql,
like
building
our
own
scheme,
based
on
like
a
an
introspected
version
of
the
schema
and
kind
of
like
constructing
it,
either
dynamically
or
ahead
of
time.
But
I'd
definitely
be
open.
A
Yeah,
I
think
you
know
definitely
parsing.
The
data
you
know
is
is
the
way
to
go
so
that
you
know
you
can
actually
query.
You
know
on
specific,
you
know
fields,
it's
just
a
much
more
powerful
way
to
go.
A
I
guess
the
first
question
I
would
have
is
you
know
you
mentioned
having
these
schemas
for
the
individual
types
as
like,
json
schema
and
then
having
that
something
that's
parse
later,
what
other
clients
would
be
consumers
of
that
json
schema
specifically,
if
you
know
all
of
that
data
was
also
kind
of
you
know,
indexed
this
way.
H
Yeah,
so
I
guess,
if
it's
indexed
purely
I
mean,
I
definitely
think
that,
like
the
subgraph
would
be
one
client,
I
think
really.
The
other
use
cases
are
other
dapps
right
that
are
trying
to
kind
of
like
pull
down
specific
contracts
like.
I
would
have
been
super
helpful,
for
example,
with
like
nfte.app.
H
If
there
was
a
way
that
I
could
have
kind
of
like
gone
ahead
and
pulled
the
schema
which
related
to
some
kind
of
json
schema
remotely
and
then
kind
of
understood
what
I
would
actually
have
access
to
ahead
of
time.
So
in
my
eyes
it's
kind
of
like
any
platform,
that's
trying
to
consolidate
or
trying
to
kind
of
like
aggregate,
multiple
or
just
one
type
of
contract,
because
bear
in
mind
that
these
contract,
these
schemes
could
be
updated
over
time
right.
H
They
could
be
upgraded
to
different
types
of
schemas,
so
one
one
nft
from
a
contract
might
be
might
have
a
different
nft
schema
if
there
is
at
some
point
in
the
future,
like
a
roll
out
of
an
upgrade
and
therefore
like
more
fields
might
be
added
on,
whereas
like
previous
ones
might
not
have
had
it.
I
had
it
so
the
schema
kind
of
changing
throughout
the
lifespan
of
my
ads
nfts
get
minted
could
be
potentially
interesting.
A
Yeah,
you
know,
I
think
all
of
that
could
be
supported
just
natively,
you
know.
So
I
guess
I
guess
here's
here's
my
my
thought
process
and
it
could
go
either
way.
So
I
saw
basically
the
question:
was
you
know
if
we
had
the
stepper
jason
schema?
How
would
it
get
parsed?
Is
it
you
know,
could
could
the
the
the
sub
graph
kind
of
dynamically
generate
the
tables
for
the
fields
and
and
the
subgraph
can't
dynamically
generate
tables?
A
The
the
schema
is
something
that
needs
to
be
like
part
of
the
sub
graph
definition,
but
what
you
could
do
is
have
tooling
on
top
that,
like
you
know,
generates
a
new
subgraph
for
you
based
on
like
parsing,
say
a
you
know,
schema.org
schema
and
then
it
could
kind
of
like
generate
the
corresponding
graphql
api
for
you.
A
So
you
can
build
some
tooling
around
that
and
then
the
question
is
you
know
is:
is
that
worth
it
or
is
that
just
kind
of
like
you
know
redundant
in
this
in
a
sense
because
you
know
graphql
itself
is
introspectible,
so
you
know
you
can
introspect
to
see
like
you
know
the
whole
schema.
A
You
know
subgraphs
are
upgradable,
so
you
know
you
can
have
different
versions
of
a
subgraph,
and
you
know
I
don't
know
that
you
can
do
it
today,
but
you
can
imagine
you
know
being
able
to
like
query
for,
like
the
history
of
a
subgraph,
to
get
past
versions,
see
the
past
schemas
things
like
that.
So
I
guess
it's
an
open
question.
F
Well,
I
I
think
the
reason
why
it
needs
to
be
defined
outside
of
graphql
is
like.
F
There
are
many
different
ways
in
code
that
you
may
want
to
consume
the
metadata,
and
I
think
the
subgraph
is
like
lower
down
the
stack
but
like
there
are
plenty
of
applications
that
might
want
to
consume
data
from
an
ft
that
doesn't
use
a
subgraph,
for
example,
sure
and
and
they
need
to,
in
that
context,
be
able
to
parse
and
understand
different
types
of
metadata
and
what
the
field
expected
fields
are
right,
and
so
I
think
that's
the
primary
innovation
for
having
like
a
central
source
of
truth
in
a
schema
repository,
and
then
you
know
when,
when
updates
to
the
schema
are
made,
they
can
propagate
down
to
a
subgraph
another
another
context
in
which
that
that
core
data
would
be
consumed.
F
A
Are
you
looking
at
storing
that
repository
itself
on
kind
of
a
decentralized
platform
of
sorts.
H
Yeah
I
think
like
as
like
a
bootstrap
there's.
Obviously
you
know
you
know.
Github
is
like
a
straight
away.
I
think
there
is
definitely
I
think,
yeah
there's
a
lot
of
value
in
storing
it
decentralized
to
match
the
whole
rest
of
the
ecosystem,
so
yeah.
I
would
push
for
it
to
be
stored,
decentral,
stored
decentralized
as
well.
Yeah.
A
So
that
that
could
be
something
to
to
work
towards,
and
you
know
I
I
buy
that,
there's
kind
of
you
know
value
there
and
kind
of
having
the
the
raw
kind
of
storage
schema.
If
you
will,
you
know
have
that
represented
separately,
using
something
like
json
schema,
and
so
so
then.
A
I
think
the
approach
would
be
to
build
tooling
around
kind
of
like
auto,
generating
a
subgraph
based
on
that,
where
you
know
it
can
generate
the
graphql
schema
that
you
want
based
on
the
json
schema
and
then
it
would
just
be
a
you
know
in
an
upgrade
to
the
sub
graph
whenever,
whenever
new
types
are
added,.
F
So
so,
how
does
that
work,
if,
like
many
nfts,
can
have
many
types
of
metadata
metadata?
Because
right
like
we're?
Not
it's
not
an
enforceable
thing
at
the
protocol
to
say
that,
like
the
uri
that
this
nft
is
attached
to
must
conform
to
this
schema
so
like
anyone
could
put
in
anything
at
the
uri,
and
the
challenge
that
I
keep,
I
don't
under
fully
understand
is
how
the
subgraph
would
be
able
to
parse
and
store
data
according
to
a
schema
that
it
doesn't
necessarily
know
it
will
conf
like
conform
to
yeah.
I
think
at.
D
That
point,
I
think
it's
at
that
point
it'd
be
up
to
you,
know
whoever's,
maintaining
the
subgraph
to
to
do
that
translation
and
I
think
the
the
like.
The
reason
for
doing
it
is
just
to
make
sure
consumers
of
those
sub
graphs
have
like
a
standard
guys
way
of
reading
it.
D
So
I
think
the
owners
is
on
on
the
subgraph
maintainer,
so
I
think
it'd
be
like
you
would
have
knowledge
of
how
you're
storing
things
on
chain
and
on
ipfs
and
then
you
would
index
it
in
a
way
that
you
think
it's
like
adheres
to
standards
so
that,
like
consumers,
have
have
a
predictable
path.
A
A
That's
kind
of
like
a
general
philosophical
issue
with
kind
of
structured
versus
unstructured
data
right.
If
the
data
is
unstructured,
then
you
either
have
to
have
that
knowledge
of
the
application
or
you
have
to
do
more
brittle
kind
of
like
you
know,
processing
and
that's
kind
of
the
trade-off,
and
so
in
general
you
know,
I
think,
we're
in
favor
of
structured
data.
A
You
know,
I
think,
like
it
just
makes
things
more
sane,
and
so
you
know,
wherever
you
can
do
things
like,
have
a
field
that
describes
the
type
of
the
metadata
or
something
like
that.
That
then
just
like
allows
it
to
be
a
pretty
like
dry,
deterministic,
parsing
kind
of
process.
That's
an
improvement!
If
you
don't
have
that,
then
you
know
you,
you
can
write
more
brittle
code
that
basically
introspects
it
tries
to
parse
it.
A
You
know
potentially
fails,
looks
for
specific
fields
based
on
that,
like
makes
kind
of
you
know,
heuristic
decisions
of
like
what
am
I
looking
at
that.
Could
you
you
can
do
that
in
mapping,
but
it
just
makes
things
a
little
bit
messier,
yep.
D
I
see
someone
dropped
a
schema.graphql
for
the
metadata.
Does
someone
want
to
talk
chat
through
that
sam?
If
you
scroll
down,
I
think
yeah
right
there,
someone
dropped.
D
Whoever
did
that
do
they
want
to
chat
about
what
they're,
what
they're
thinking
here
it
seems
like.
I
do
think
we
want
to
drive
towards
like
yeah
a
subgraph
scheme
right
here.
B
Yeah,
where
is
that
same
yeah?
It's
the
same,
schema
that
some
proposed,
but
just
related
to
a
graphql,
schema
having
a
particular
base
metadata
interface
and
having
a
particular
type
for
each
known
madrid.
So
probably
this
is
a
draft.
It's
not
the
final
version.
I
just
posted
to
provide
an
idea
on
that,
hopefully,
for
for
next
meeting,
I
will
share
with
you
a
small
proof's
concept
of
that.
Having
a
couple
of
shades
on
file
is
supporting
different
metadata
and
parsing
it
within
nasograph.
G
G
So
any
advice
on
structuring
that,
while
conforming
to
standards
and
another
point
that
I
thought
was
pretty
interesting,
but
we've
actually
already
received
requests
for
localization.
So
actually
you
know
supporting
multiple
languages
and
local
characters.
I
know
this
is
slightly
specific,
but
I
think
like,
if
we're
really
going
for
you
know,
scale
and
supporting
a
lot
of
users
having
that
sort
of
support
early
on
might
be
useful.
A
Yeah
I'll
I'll,
let
others
speak
to
the
first
I
can
say
I
think
localization
is,
is
definitely
something
that
we
should
tackle
kind
of
protocol
wide
and
it's
a
good
point
to
follow
up
on.
H
Language
yeah,
I
think,
to
the
first
one
where
essentially,
you
have
kind
of
like
multiple
different
outputs
like
of
the
source.
So
I
think
yeah,
there's
a
very
naive,
take
on
this
that
I
kind
of
added
in
here,
and
it
is
by
no
means
like
a
kind
of
well
thought
out
solution.
H
But
I
imagine
there
being
like
and
resolutions
again
is
like
a
terrible
name,
but
some
kind
of
key
value
store
where
there's
like
kind
of
like
key
values
of
like
common
or
like
a
group
pre-agreed
upon
different
either
like
resolutions
or
like
sprite
sheets,
or
you
know,
high
definition,
low
definition
like
a
3d
like
you
can
imagine
with
video
like
they
could
be
hdr.
There
could
be
kind
of
like
3d,
like
there's
all
types
of
like
different
core
formats.
That
need
to
be
considered
that
a
video
could
be
available
in.
H
So
I
think,
like
some
kind
of
like
key
mapping
could
go
a
long
way
there,
and
I
think
this
is
where,
like
the
ability
to
compose
and
like
build
upon,
is
kind
of
quite
core
because,
for
example,
like
your
solution
might
have
like
many
different
types
of
formats.
That
are
specific
to
you
in
the
video
realm,
even
in
like
your
video
realm,
compared
to
other
video
kind
of
like
dapps
that
are
out
or
protocols
that
are
kind
of
like
coming
out.
H
So
having
you
know
the
ability
to
kind
of
like
have
a
base
layer
and
then
compose
on
top
of
that,
I
think,
is
kind
of
pretty
crucial
for
any
think.
Moving
forward.
Yeah
yeah,
if
we
can
kind
of
like
agree
to
something
on
a
lower
level,
then
even
better,
but.
G
Sounds
good
inheritance
might
be
interesting,
but
I
think
that
would
be
difficult
with
graphql.
I'm
not
sure
another
just
like
to
share
knowledge.
We've
been
doing
a
lot
of
experiments
with
video
and
hls
seems
to
work
very
well
over
over
ipfs
better
than
mp4.
B
G
Yeah,
okay,.
A
Okay,
great
well
we're
getting
towards
the
end
of
time
here
so
yeah.
I
feel
like
the
the
weekly
cadence
feels
like
a
lot
generally.
I
also
know
that
the
nft
space
is
moving
super
fast
and
it
feels
like
we
never
run
out
of
things
to
talk
about,
so
you
know
I'm
for
for
keeping
it.
If
you
guys
are,
I
feel
like
we
can
continue
from
here
next
week,
and
you
know
we
can
try
to
work
on
a
an
agenda
over
the
telegram
ahead
of
time.
A
If
people
have
things
that
come
up
during
the
week
that
they
want
to
make
sure
that
they
have
a
chance
to
discuss
on
the
call
does
that
sound
good,
all
right,
awesome
well
super
productive
day,
and
you
know
special
shout
out
to
the
foundation
team
for
doing
such
a
great
job
capturing
all
of
those
you
know,
proposals
from
from
the
the
last
call
so
yeah
thanks
a
lot
everybody
and
see
you
at
next
week's
call
thanks.
Everyone
thanks.