►
From YouTube: 🖧 IPLD weekly Sync 🙌🏽 2020-05-11
Description
A weekly meeting to sync up on all IPLD (https://ipld.io) related topics. It's open for everyone and recorded. https://github.com/ipld/team-mgmt
A
Welcome
everyone
to
this
week's
IPL
desync,
meaning
it's
May,
11th
2003,
and
it's
every
week
we
go
over
the
stuff
that
we've
done
in
the
past
week
and
plan
to
do
next
week
and
discuss
any
agenda
items
you
might
have
or
answer
questions
from
anyone.
So
I
start
with
myself.
So
in
the
past
week,
I
finally
did
a
bit
of
rust
work
again.
A
Some
word
on
the
multi-format
stuff,
so
there
was
a
release
of
rust,
multi,
hash
and
rust
CID.
The
other
major
change
was
that
now
the
code
for
having
your
custom,
correct,
ableism
and
so
now,
yeah
you
can
have
your
own
custom
product
table
and
the
other
big
change
recycled
ii
was
david
contributor
to
eiffel.
A
The
recycle
DM
did
a
huge
change
on
the
kodick
trade,
which
is
kind
of
like
more
less
the
central
piece
of
one
piece
of
APD,
which
is
about
the
Codex
and
how
you
encode
it:
okay,
decode
staffed
and
there
the
trade
was
perfected
and
it's
better
than
it
used
to
be,
and
also
with
this
PR.
Some
things
moved
out
of
what
we
call
core
IP
le
into
other
parts
of
IL
d
and,
for
example,
the
storage
stuff.
So
because
I
said,
like
storage
itself,
shouldn't
be
long
early
to
the
core.
A
But
it's
just
like
a
separate
thing
and
we
couldn't
agree
on
a
block
API
yet
but
I'm
yeah.
So
my
proposals
up
but
I
now
poured
the
my
proposal
to
the
new
correct
trade
probably
this
week,
and
then
we
go
from
there
with
the
discussions
and
then
I
also
worked
a
bit
on
that
tech
choice
on
spec
the
stuff
there's
one
more
PR,
which
is
just
renaming
the
file
and
I
think
then
I
considered
egg
chase
and
probably
specified
and
done
yeah.
That's
all
I
did
so
next
on
my
list.
It's
rod.
B
Okay,
throw
me
that
my
week
has
been
swallowed
by
Bitcoin,
really
deep
in
this
stuff
in
it,
but
I
made
really
good
progress,
just
just
slower
than
I
imagined
so
I
can
now
produce
a
full
graph
from
a
block.
I
can
decompose
a
block
into
all
the
component
IP
LD
blocks,
including
the
two
merkel
graphs
and
the
witness
commitment
block.
B
I
haven't
verified
that
yet,
though,
I'm
still
stuck
and
that's
rep
stuck
on
weekend
when
I
haven't
quite
got
the
reassembly,
you
know,
so
I
can't
quite
do
that
there
and
back
again
and
test
that
it's
all
working,
that's
that
that
will
be
the
critical
pay
piece
that
I
can
take.
A
single
Bitcoin
block
and
sort
of
shatter
it
into
all
these
IP
LD
blocks
and
then
take
the
reverse.
I'm
gonna
get
bite
for
bite.
The
original,
that's
the
ultimate
goal.
Yes,
I'm
blocked
on
incompatibility
with
car
files.
B
Whole
point
of
the
whole
point
of
the
car
file
work.
I
did
in
the
chat
the
other
one
Zipcar
files,
the
reason
I
started
working
on
them
in
the
first
place
before
they
became
important
to
file
calling
and
others
was
because
I
wanted
to
use
in
protest.
Roosters
portable
text
for
test
fixtures
on
my
PLD
work
and
so
I
was
you
couldn't
try
to
plug
this
in
here
initially
for
testing,
so
that
I
can
take
a
very
large
graph
store
it
somewhere
and
then
bring
it
back
again,
and
that
was
supposed
to
be
perfect.
B
But
cuz
I'm
now
toying
with
Michael's
new
multi
formats
and
the
caliph
al
start
works
with
the
old
days
block
and
CID
model.
It's
incompatible
in
really
nasty
ways,
and
so
it
should
just
be
a
matter
of
just
you
know,
make
a
file
and
throw
blocks
out
like
it
should
be
really
simple.
But
it's
just
error
for
very
sad,
so
I
have
to
now
choose
which
react
to
shave
to
make
that
work.
C
B
B
C
B
B
B
B
C
I
want
to
ship
the
new
block,
API
and
all
the
new
codec
changes
and
all
the
new
hashing
functions
once
all
of
that
is
out,
and
we've
updated,
like
all
the
internal
IP
of
these
stuff
to
the
new
block
API.
On
top
of
all
of
this,
then
I'll
prepare
a
talk
and
documentation
for
the
other
teams
like
to
sort
of
explain
what
it
is,
how
it
all
works.
So
that
would
probably
be
the
time
when
we
want
to
fit
them
in.
B
Anyway,
this
is
not
it's
not
a
lot.
Nobody
read.
This
is
a
criticism
because
I
actually
really
enjoy
bringing
you
not
before
that
stuff.
It's
just
it's!
It's
so
nice
to
work
on
these
things
in
isolation.
Cuz
in
pass.
We've
have
to
you
know
you
just
want
to
say
you
just
want
to
write.
Seawall
blocks
and
you'll
install
that
and
suddenly
you've
got
Z
cash
in
your
dependency
tree
and
this
thing's,
just
they
just
have
so
big
and
heavy
weight.
B
This
the
new
model
that
Michaels
come
up
with
is
you're
just
dealing
with
these
isolated
pieces
that
you
want,
and
even
in
your
tests
you
can
narrow.
It
went
down
just
one
thing
and
you
can
make
stuff
up
and
I
can
certain
marks
and
I
can
create
new
codecs
and
you
know
do
it
all
in
a
very
isolated
space,
which
is
how
it
should
be.
So
here's
a
nice
API,
it's
just
it's
missing
the
rest
on
stack,
so
that's
the
company
so
anyway,
what
I
might
do
today,
I've
been
pondering
it.
B
D
So
I
have
not
a
ton
of
technical
things
reported
this
week,
because
I
ended
up
doing
a
lot
of
coordination
things,
but
there's
plenty
of
fun
days
on
that
front.
So,
first
and
foremost
there's
a
bunch
of
people
working
on
high
POV
and
selectors
and
rust,
who
are
hanging
out
in
the
IRC
channels
and
that's
really
exciting,
because
lots
of
good
questions
are
getting
raised.
D
I
think
some
of
the
discussions
that
we're
having
there
we
want
to
totally
want
to
go
back
through
the
IRC
logs
at
a
later
date
to
end
turn
them
into
more
documentation
where,
wherever
there's
gaps
or
just
if
those
things
are
documents
to
make
them
more
discoverable.
You
know,
and
some
of
the
stuff
that
we're
uncovering
there.
D
In
addition,
just
being
good
to
talk
about
the
people
working
on,
it
are
also
doing
a
bunch
of
tests,
vector
development,
and
so
we
spend
some
time
talking
about
what
will
be
a
good
way
to
layout
test
vectors
and
like
which
parts
are
not
obvious.
But
you
really
want
to
get
the
vectors
to
cover
and
I'm
really
excited
about
that.
So
there's
already
draft
test
vector
content
that
these
folks
are
pushing
out
and
super
exciting.
So
I'll,
probably
at
some
point
in
the
future,
have
more
work
that
I'll
take
on
to
make
the
go.
D
Libraries
that
already
exist
use
those
same
test
vectors
but
I.
Think
that's
going
to
be
a
awesome
use
of
time.
So
super
excited
about
that
and
this
might
barely
belong
in
our
I
peel
de
team
notes.
But
in
case
anybody
else
out.
There
wants
to
do
go
and
it's
starting
to
profile.
Things
I've
been
taking
a
bunch
of
notes
on
the
lessons
learned
and
put
anyone
else
and
do
a
guest
format
that
are
some
good
getting
started.
A
D
A
PR
comes
in
from
outer
space
on
the
internet
and
we
try
to
figure
out
if
these
things
are
a
good
idea
where
they
fit
in
with
other
ideas
and
like
what
some
balancing
factors
are
in
trying
to
find
the
virtue
of
the
thing
and
so
like
checklist.
So
if
we
have
this
feature
here,
will
it
increase
the
complexity
of
this
comic,
or
will
it
make
this
harder
to
port
to
a
new
language
or
something
like
this
things
that
we
informally
think
about
all
the
time?
D
We
will
try
and
reason
about
this
I
try
to
put
the
Middle,
East
and
so
yeah,
maybe
that'll,
be
useful,
but
it'll
probably
need
a
lot
of
room.
You
too,
because
it's
also
one
of
those
very
meta
things.
That's
very
hard
to
apply
so
yeah.
Looking
forward
to
that,
and
the
last
thing
that
I
did
in
this
week,
that
is
also
very
early
IPL
D,
but
I
want
to
mention
it
briefly
is
I
made
a
new
JSON
codec
for
absolute
giggles
this
weekend,
just
just
for
fun,
it
emits
JSON
tables
and
I.
D
Probably
shouldn't
try
too
hard
to
describe
this
out
loud
because
you
kind
of
have
to
see
a
picture
and
that's
actually
worth
a
thousand
words
in
this
case.
But
basically,
if
you
pipe
certain
structures
of
JSON
format
into
it,
this
couric
can
auto
detect
what
that
data
looks
like
and
be
like.
You
have
a
list
here
and
it
has
a
bunch
of
Maps.
Maybe
this
list
is
the
table,
and
maybe
these
map
keys
are
all
column
headings
and
then
it'll,
basically
pretty
print
this
using
the
non-semantic
whitespace
to
make
nice
human,
readable,
skimmable
stuff.
B
D
So
I
did
it
turned
out
that
this
was
really
easy
to
do
with
all
the
api's
in
modern
go
activity
pride,
so
that
was
cool
proof
of
concept
in
the
future.
I'm
thinking
this
might
be
a
fun
thing
to
carry
further,
because
we
could
use
selectors
the
configuration
language
to
let
people
say.
Oh
yes,
please
highlight
this
sub
section
in
this
way.
I
don't
know
could
could
be
a
future.
Fun
could
be
totally
irrelevant.
We'll
see
it's
out
here
and
that's
it
for
me.
A
C
Hello
yeah
so
who
okay,
sorry,
I'm
a
little
out
of
it
today,
yeah
so
last
week
finished
up
the
gist
multi
for
my
stuff.
I
took
the
buffer
removal
PR
that
vulgar
sense
and
put
another
couple
days
into.
It,
took
me
a
lot
longer
to
finish
that,
but
I
thought
it
would.
But
I
was
a
lot
happier.
I
thought
it
would
be
so
you
know
I
paid
off
after
that
gets
reviewed.
That's
basically
ready
to
go.
I
can
start
bringing
over
the
block
API
and
all
the
codec
implementations.
C
The
legacy
API
got
way
bigger
to
handle
a
lot
more
than
it
used
to
so
it
actually
does
sort
of
full
conversion
from
the
old
CIT,
so
the
new
CID
is
and
for
all
the
codecs
and
it
also
swaps
out
buffer,
for
you
meant
eight
array
on
all
the
old
codecs
stuff.
So
it's
like
quite
a
bit
more
code
than
it
used
to
be,
and
it's
definitely
gonna
slow
down
the
legacy
codecs.
But
if
you
care
about
performance,
you
should
just
get
off
of
the
legacy
stuff
and
fix
your
problem.
C
Put
some
time
in
the
dag
give
me
over
the
weekend:
I
have
functional
indexes
Maps
and
you
can
do
property
indexes
and
then
do
like
sums
and
counts
and
stuff
on
that's
pretty
cool
and
yeah
I'm
I'm
off
for
a
couple
weeks,
starting
on
Wednesday,
so
make
sure
you
get
anything
you
need
to
me
today
or
tomorrow.
Cuz
then
I've
got
a
I'm
gonna
move
during
shelter-in-place,
which
is
going
to
be
an
adventure.
That's
why
I
need
a
couple
weeks
to
finish
it
all
out
so
and
that's
it
for
me.
B
A
B
A
B
Say
how
how
strange
it
is
that
Russ
might
get
selectors
before
JavaScript
does
I
was
thinking.
I
was
thinking
how
cool
there
would
be
for
some
of
these
work.
Now,
because
I've
got
these
really
interest
in
graphs,
it
would
be
so
cool
to
have
select
in
the
way
you
could.
You
could
do
something
like
take
a
block
and
extract
values
out
of
it,
look
at
different
components
of
it
and
make
a
really
cool.
Little
presentation
talk
on
this
stuff
that
I'm
doing,
though
we
just
don't,
have
those
tools
available
to
us.
A
Yeah
so
but
know
one
thing
is
that
so
what
and
it
might
be
surprising,
but
the
selectors
currently
don't
use
what
I
call
rust
IPL
D.
So
so
so
as
the
selectors
work
comes
from
500
when
they
currently
don't
use
the
rust
I
believe
that
I
work
on,
but
it's
all
under
control,
so
no
worries.
It's
just
like
we
had
different
timelines,
but
basically
my
so
after
I
finished
that
block
API,
probably
I,
will
then
I.
Think
rust.
I
believe
would
be
the
good
shape
to
basically
then
start
looking
into
okay.
C
A
C
A
A
This
was
actually
triggered
by
me
seeing
that
they
that
rusts
icon
people
are
doing
a
tag
PB
code
egg,
which
is
kind
of
to
use
it,
which
is
using
the
eye,
peeled
the
data
model
and
therefore
kind
of
like
also
used
what
could
be
described
as
a
schema,
so
they
have,
it
makes
sense
to
see
with
technically
we
agreed
on
how
does
take
PP
look
like
in
the
IPL
the
data
model
they
see.
This
was
triggered
by
this,
so
they
also
do
so
your
so
it's
actually
right.
A
They
also
use
tech
PB,
but
they
use
I,
think
that
could
be.
They
only
need
to
this
utilized
egg
PB,
probably,
and
they
only
encode
yeah.
They
only
encode
seaboard.
Therefore,
they
basically
can
go
so
I'm,
not
sure
what
they
could
do,
but
they
could
even
go
directly
from
rust,
negative
types
to
see
more
if
they
want
to
without
the
data
model,
but
I
think
they
use
the
data
model,
because
I
told
them
to
use
the
data
model.
A
For
this
reason,
so,
as
I
said,
it
should
like
it
should
work
out,
but
I
just
haven't
had
the
time
and
still
don't
have
the
time
but
I'm
confident
that
and
we
can
just
use
the
selectors
work
that
is
currently
done.
Also
with.
Then,
the
rust
I
believe
that
I'm
working
on,
let's
totally
the
plan
and
yeah
and
I,
will
make
sure
this
happens.
So
yeah
cool.
A
E
A
So
there's
really
no
real
update
and
I
still
do
care
with
the
5
on
4.
But
what
curly
it
looks
so
I
would
say
truly
looks
like
that.
The
the
people
that
need
it,
the
most
are
the
Russ
ficon
people,
so
it's
likely
that
they
saw
sa
moving
very
fast,
also
with
the
Spector
stuff.
So
we
see
they
do
this
select
the
stuff
now
to
do
the
grousing
stuff.
So
I
would
expect
that
they
do
the
graphs
at
least
to
a
grousing
implementation.
A
So
yes,
but
yeah
other
than
that
I
don't
know
if
I
should,
if
I
do
it
this
week
to
get
also
the
other
parties
involved.
That's
true
yeah,
so
yeah,
but
no
further
updates
on
this.
There
yeah
yeah,
but
so
they
will
be
nap
T
next
week.
So
I
promised
an
update
next
week.
So
I
won't
look
on
its
this
weekend,
then
yeah
cool.
A
C
So
I
pulled
him
off
to
do
a
go
library
this
week,
so
the
Jas
stuff
definitely
isn't
happening
really
soon
like
within
the
next
week
or
two.
So
we
won't
have
the
selector
engine
nsj
as
soon
as
I
would
have
liked,
because
we
just
had
another
priority
pop
up,
for
we
need
the
this
string
syntax
for
selectors
to
have
a
go
library
for
file
coin,
so
yeah.
That's!
They
got
delayed
a
little
bit.
E
C
I,
don't
I
mean
Tim
just
started
working
on
it
like
maybe
a
week
ago,
so
he
hasn't
been
actually
putting
in
a
ton
of
time
just
to
it
until
pretty
recently,
but
other
than
that
I.
Don't
I,
don't
know
exactly
what
the
status
is,
but
Tim
is
a
on
IRC.
So
if
you
wanna
ping
him
in
and
check
in
with
him
about
it
go
for
it.