►
Description
Welcome standup discussion and track intro - Data and IPFS: Models
----------------------------------------
Missed last year's IPFS þhing? Register to attend IPFS þhing 2023 in Brussels, Belgium here: https://2023.ipfs-thing.io/
A
So
we've
got
20
minutes
of
sort
of
a
track.
Intro
and
you
know
I
do
have
a
couple
of
slides
but
which
I'll
go
through,
but
I
kind
of
want
to
hear
a
little
bit
from
like
this
is
this
is
going
to
be
a
more
collaborative
session
will
be
a
few
talks
but
like
I
want
to
hear
from
folks
like
kind
of
what
what
they
are
thinking.
What
their
perspective
is
on
this
chunk
of
of
the
space.
A
What
you
want
to
see
from
working
with
data
in
ipfs,
which
is
sort
of
the
the
prompt
on
the
intro
session
thing
for
for
those
of
you
who
filled
it
out.
So
maybe
I
guess
we'll
actually
we'll
start
with
that
kind
of
go
around
the
room,
say
your
name
and
maybe
a
little
bit
about
what
you
kind
of
wish
was
doable
with
the
ipfs
and
data
things
that
isn't
today
and
like
a
couple
of
sentences.
A
So
I
will
start
my.
My
name
is
adin
work
on
ipfs
things
like
kubo.
The
thing
that
I
really
want
to
see
from
working
with
data
is
like
at
the
very
very
least.
I
would
like.
All
of
the
hash
linked
descriptions
of
files
and
directories
to
all
be
traversable
and
usable
with
with
a
set
of
tools
like,
for
instance,
a
gateway
api.
So
if
you
are
git
or
you
are
bittorrent
or
you
are
a
checksum
of
a
large
file
or
you
are
unix
fs
or
you
are,
are
we
or
you're?
A
Wnfs,
all
of
these
things
should
should
work.
If
you
haven't
done
that,
then
we've
made
our
lives
really
hard
because
we
make
things
you
know
you
want
to
make
these
things
change
over
time.
We
want
to
make
better
structures
going
forward
that
was
kind
of
long,
but
we'll
move
along.
Yes
hannah.
You
want
to
go
next,
oh.
C
Yeah,
hey
everyone,
I'm
hannah,
so
I've
worked
a
lot
with
ipld
because
I
work
on
a
network.
We
use
something
called.
C
I
would
most
like
to
figure
out,
like
I'm
interested
in
ipld
from
an
application
perspective,
as
well
as
from
a
thing
we
built
units
on
top
of
perspective,
and
I
think
they
said
yesterday
in
the
lotus.
Talk
that,
like
you
know,
lotus
is
essentially
an
ipld
application,
because
the
entire
vm
is
idlp
data,
and
I
think
we
should,
I
think,
that's
a
really
interesting
fit
because
there's
actually
extensive
use
of
operating
assets,
essentially
an
application
storage
there
and
what
I
would
like
to
get
out
of
this
topic.
C
I
I
think
I
don't
have
a
strong
opinion
about
ipld,
but
I
would
love.
I
would
really
love
it
if
we
could
run
selectors
at
the
level
of
like
a
database
where
you
know
it's
all
executed
based
upon
these
indexes
instead
of
computation,
and
that
would
probably
require
rewriting.
B
D
I
So
I
work
on
yeah
right
now
supporting
the
actors
in
the
poplin
from.
I
And
they
use
it
as
their
data
structure,
so
they're
different
applications.
They
use
ipld
as
their
persistent
memory.
So
my
whole
ability
really
is
to
get
basically
network
transparent,
automatically
serialized
memory.
It
just
works.
I
don't
even
want
to
think
about
the
fact
that
it's
started
multiple
data
structures
or
where
it
lives,
or
anything
like
that.
B
G
A
All
right,
thanks
all
of
you,
for
for
sharing
that
one
of
the
reasons
I
I
ask
for
this
or
push
on
this
is-
I
find
at
least
for
for
me
personally.
A
Oh
yeah,
we
have
people
on
zoom,
hello,
jonathan
ollie,
reba
you're,
asking
folks.
I
think
you
can.
I
don't
know
if
you
heard
us
at
the
beginning,
but
a
little
bit
about
what
you
want
to
see
from
working
on
this
kind
of
tooling
working
with
ipfs
and
and
data,
and
what
you
want
to
see
out
of
this
workshop.
J
A
J
And
it
works
cool,
well,
not
so
cool
because
I
don't
have
much
to
say.
I
am
not
entirely
sure
what
to
expect.
So
you
know
not
being
there
I'll
I'll,
mostly
just
listen
but
hoping
to
get.
I
guess
a
little
more
delineation
between
the
different
things
that
build
ibfs
as
a
concept,
because
we
seem
to
be.
J
From
my
perspective,
one
of
the
main
problems
is
that
we
work
on
several
competition
related
technologies.
We
call
them
the
same
thing
and
everybody
kind
of
holds
a
different
part
of
that
and
that's
what
brings
us
to
the
place
where
we
are
today
and
that's
what
I
want
to
get
more
clarity
on
what
you
know.
Others
think
kind
of
thing.
A
Thanks
jonathan,
you
want
to
go
next.
B
Sure,
apologies
if
I'm
repeating
what
was
being
said
in
the
room,
because
it
was
a
little
bit
difficult
to
hear
what
other
people
were
saying.
Borrowing
yourself
as
the
speaker,
which
is
okay.
It's
fine
yeah.
I
think
the
the
an
idea
of
how
the
ecosystem
may
move
forward
in
terms
of
driving
the
adoptions
of
schemers
interchangeability
between
the
ability
to
to
to
use
those
more
intelligently.
B
I
think
you
know,
as
other
people
are
sort
of
heading
towards
it,
the
the
there
is
a
distinction
between
ipfs
and
l
ipld,
which
you
know
is
not
always
obvious
from
the
tooling
and
I'm
again
very
interested
in
in
ipld
as
a
sort
of
a
technology
for
for
application
development
and
not
simply
for
file
storage,
and,
I
think,
there's
a
huge
potential
there
and
some
of
the
parts.
B
K
Hello,
I'm
mostly
lurking
with
extreme
fomo,
but
I'm
here
for
the
chat
around
making
unix
fs
less
special
gateways
how
to
present
ipld
any
ipd
data
over
the
http
gateway.
That's
my
interest.
A
It
helps
when
dealing
with
some
of
these,
like
large
systems,
to
understand
where
people
are
coming
from
in
the
ipfs
context,
like
one
that
I
I
frequently
use
as
I
sort
of
ask
people
what
about
ipfs,
they
find
most
valuable
and
some
folks
will
say
things
like
you
know:
immutable
data
content
addressing
so
mutable
data
is
pretty
cool.
My
cache
numbers
go
to
infinity
like
I
don't
have
to
have
like
an
eviction
policy
other
than
when
I
don't
feel
like
having
you
anymore
and
other
people
will
say
what's
really
cool
about.
A
I
understand
both
sides,
and
so
I
can
I
can.
It
makes
it
easier
for
me
to
understand
where
someone
else
is
coming
from,
which
means
that
we
get
to
talk
into
some
of
the
discussions
later
today.
A
A
A
You
know.
Multi-Formats
has
sort
of
like
the
official
slogan
like
not
going
to
change,
considered
harmful
right.
Multi-Hash
sha-1
is
the
best
and
then
someone's
like.
Oh,
how
about
shot
two
and
then
someone's
like?
Have
you
considered
shot?
One
is
kind
of
broken
right
and
if
you're
stuck
in
sha-1
get
in
bittorrent,
then
then,
like
life
is
hard
for
you
right.
There
are
different
base.
Encodings
everyone
has
their
favorite
base
encoding.
There
are
different.
You
know
there
are
different,
addressing
schemes
for
networks
right,
lib
p2p
uses
this
an
abstraction
layer.
A
You
know
just
give
me
a
stream
right
there.
Many
ways
to
do
this
with
packets
and
request
response
too,
but
just
like
give
me
a
stream,
I
don't
care
how
you
do
it.
That's
like
what
I
want
to
build
on
top
of
and
ipld
is
trying
to
do
this
in
some
way
for
hash
link
data
structures.
Some
of
these
are
harder
than
others
right.
I
think,
like
using
streams
for
networking,
seems
like
fairly
reasonable,
given
the
prevalence
of
things
like
tcp
hatchling
data
structures,
we're
kind
of
pushing
the
boundaries
there.
A
We
have
to
find
the
abstraction
layers
that
work
for
us.
We
got
a
bunch
of
them
already.
There's
the
data
model,
there's
codex,
adl
schema
selectors,
even
some
of
like
the
pathing
stuff
that
exists
in
like
ips
gateways.
Are
there
enough?
Do
we
need
more?
Do
we
need
to
change
some
of
these
right?
These
are
things
that,
like
we're
building
abstractions,
because
we
feel
we
need
them
not
because
we
think
they're
cool
so
like
yeah.
How
do
we?
What
what's
going
to
make
it
work
for
us?
A
We'll
talk
about
ipld
on
gateways
with
urls,
which
is
basically
another
abstraction
layer,
that's
being
added
to
help
reason
about
or
work
with,
ipld
data
make
it
more
friendly.
A
We'll
have
a
discussion
about
abstractions
on
hash
link
data
and
basically
look
at
this.
What
do
we
need
thing?
A
little
bit
I'll
talk
a
little
bit
about
some
knowledge
I've
gained
from
working
on,
and
I
feel
the
implementation
that
does
webassembly
things,
yeah
and
sort
of
what
what
I've
learned
and
give
a
little
demo
and
see
how
that
goes.
And
then
people
seem
to
want
to
talk
about
unix
ffs.
A
So
let's
talk
about
uxfs
and
what
comes
next
and
what
we
need
in
order
to
make
what
whatever
we
want
to
come
next
happen,
and
this
is
part
of
a
sort
of
a
joint
there's.
This
there's
data
transfer
tomorrow
and
then
we're
gonna
have
some
uncomps
later
tomorrow,
so
figure
we
should
get
at
least
15
minutes
to
like
figure
out.
If
there's
anything
that
has
sprung
up
immediately
is
like,
we
should
schedule
those
sort
of
uncomp
things
and,
of
course,
more
can
spring
up
tonight
or
tomorrow
and
we'll
do
that.