►
From YouTube: 🖧 IPLD Every-two-weeks Sync 🙌🏽 2019-06-25
Description
A meeting every two weeks to sync up on all IPLD related topics. It's open for everyone and recorded. https://github.com/ipld/team-mgmt
B
B
And
what
do
we
have
so?
Normally
we
start
with
a
quick
run
of
updates,
and
then
we
talk
about
whatever
and
issues
you
want
to
talk
about
and
I
start
first,
because
I've
written
down
my
notes
already
so
I,
oh
I,
forgot
to
put
there
I
played
around
a
bit.
Also
I
I
will
also
put
it
there
later
and
I
put
a
bit
of
a
time
also
on
webassembly,
because
so
the
background
story
is
the
IPS
team
was.
Is
someone
from
the
IDF
s?
B
Team
has
supported
the
ribbon
Shanker
from
JavaScript
through
a
simple
script
to
compiler
to
web
assembly,
and
we
had
a
native
C
implementation,
and
the
web
assembly
stuff
is
way
slower
than
the
C
stuff,
and
it
was
just
out
of
interest
because
I
found
a
blog
post
about
compiling
C
directly
to
webassembly
without
em
scripting,
so
I
just
also
wanted
to
play
around
with
it.
So
I
just
wanted
to
port
the
original
version
and
to
see
if
it's
any
faster
and
so
the
result
was
it's
not
any
faster.
B
B
It's
dead,
I
wrote
again
on
my
dream:
stuff,
the
stack
stuff,
so
in
previous
series,
I
got
working
that
I
can
load,
it
didn't
write
on
D,
and
now
there
is
an
in
save
an
interface
called
stack.
Browser
browser
and
I
today
manage
that
I
actually
get
the
salami.
They
use
JSON
the
degradation
object
from
the
IP
LD
store
and
then
just
get
the
object
directly
into
their
web
application,
and
it
could
throw
the
data
and
that's
pretty
exciting,
because
now
I
need
to
figure
out
how
to
follow
things
and
so
on.
B
B
A
I
have
so
much
stuff,
so
this
was
a
really
hectic
and
exciting
week
or
two
partially,
just
because
I
posted
tons
and
tons
of
stuff
that
I
had
them
sitting
on
it
like
a
98
percent
completion
and
now
they're,
just
all
wasted
and
two
guests
and
PRS
in
various
places,
will
need
to
go
back
and
draw
our
link
suit
or
cook
pad
for
this.
So
the
selectors
draft
is
wasted
and
that's
not
on
a
PRN
to
the
specs
repo.
A
Yet
because
there's
a
bunch
of
examples
that
need
to
be
updated,
but
the
schema
itself
is
pushed
so
much
I
think
those
include
a
completely
reasonable
recursion
system
and
there
are
still
some
two
dudes
around
a
conditional
language,
but
the
recursion
parts
themselves
seem
to
be
correctly
shaped
out.
So
that's
cool
I
got
to
sit
down
with
some
people
from
the
falcone
project,
who
are
who
have
had
these
desires
for
selectors
for
ages
and
get
like
some
very
concrete
examples
of
these
cases.
Out
of
them.
A
A
A
A
This
is
close
and
so
he's
been
doing
a
bunch
of
other
spike
stuff
with
a
jeaious
implementation
of
some
of
the
ideas
for
UNIX
FS
me
tuition
s
and
he
has
started
to
accumulate
some
interesting
logic
about
like
how
to
split
up
files
and
so
we're
looking
at
that
and
we're
looking
at
the
schema
drafts
and
we're
kind
of
thinking.
We
can
meet
in
the
middle
here
by
putting
any
of
the
interesting
chunking
logic,
basically
anything
that's
an
if
or
above
in
power
level.
A
It's
a
nice
place
to
the
limit,
the
abstraction
so
we'll
be
able
to
change
your
mind
about
that
in
constructive
ways.
Later
and
it'll
just
be
it'll,
be
good,
so
with
all
that
stuff,
I'm
thinking
about
shifting
sometime
back
to
ski
work
and
ideally
trying
to
get
a
skeleton
of
that
that
we
could
think
about
beginning
to
power
coach
and
go
with
and
somebody
else
from
another
team.
Kook's
who's
apparently
done
some
code
general
before
and
was
saying
he
might
like
to
jump
in
and
do
some
stuff
with
that.
A
B
C
So
I've
been
in
interesting
land
of
functional
programming
community
that
talks
a
lot
about
persistent
data
structures
which
are
really
just
immutable
data
structures
and
they
talk
about
them
in
the
same
way
that
we
talk
about
data
structures
in
IP
LD
and
they
think
about
it
in
you
know
almost
identical
terms
which
is
really
interesting.
You
don't
like
they.
C
Don't
they
don't
they
don't
do
it
for
the
same
reasons
and
they've
got
a
lot
of
other
concerns
like
optimizing
for
cache
locality
and
all
that
sort
of
stuff,
but
otherwise
there's
a
lot
to
mind
from
that
community.
So
that's
been
a
really
interesting
journey
and
currently
working
on
implementing
a
vector.
You
know
like
a
resizable
array.
He
basically
something
that
you
can
push
to
and
push
some
elements
to,
and
it's
really
efficient
to
to
address,
not
very
efficient.
C
Not
very
efficient
when
you
take
things
off
the
front
of
it,
but
very
efficient
to
put
things
on
the
end
of
it
and
it's
that's
working
out
pretty
well,
it's
good.
The
algorithms
were
a
bunch
of
names,
but
it's
essentially
it's
like
a
it's
like
a
hampt
without
their
hash
and
it
mapped
down
to
something
that's
very
similar
to
what
Jeremy
did
with
char
a
he's.
Sh
a
our
our
day,
Y
repo
and
his
github
I.
C
Think
if
I
pursue
this
thing
further
and
get
it
to
respect
that,
it
might
be
something
that
could
map
to
what
he's
done
in
a
certain
form.
So
with
certain
defaults
it
it
could
produce
what
he's
done
and
I
think
I,
don't
quite
understand
the
motivation
for
it.
For
his
thing,
it's,
but
it's
so
it's
it's
it's
to
do
with
file
coin
light.
Clients
really
understand
that,
but
I
believe
Lua
is
involved
and
so
I
think
it's
meant
to
be
like
very
small
and
and
efficient
or
easy.
C
But
yes,
I'll
have
to
figure
out
more
about
what
the
motivation
there
is
and
how
to
align
it
to
get.
Maybe
to
get
that
as
well
as
all
the
other
benefits
you
can
get
from
having
this.
Basically,
a
very
large
array
of
IP
LD
ordered
by
index,
or
maybe
maybe
it's
not
quite
close
enough.
I'll
see
I,
don't
quite
understand
the
the
go
code
in
there,
but
like
enough
to
see
that
it's
a
it's
very
similar
anyway.
C
B
Guess
the
the
only
thing
that
comes
to
my
mind
is
the
as
I
know,
spend
time
on
actually
using
IPD
in
the
browser
and
I
see
that
we
really
need
to
add
or
like
we
should
probably
stick
to
the
M
JSON
at
the
deck
chase
when
we
have
for
something
similar.
So
it's
really
like
you,
you
would
in
the
browser
you
really
want
to
have
some
chase,
a
representation
that
it's
just
like
easily
readable
and
possible
and
because
I
convey
this.
This
pick
out
from
I
feel
about
a
more
advanced
version.
Basically,
but
I.
B
Think
it's
not
like
it's
I,
think
it's
a
cool
thing,
but
I
think
it's
not
like
in
many
movie
called
egg
chase,
and
so
I
don't
care
about
the
names
but
I
think
it's
just
something
different
I
think
the
current
page
Jason
is
pretty
easy
to
work
with
in
it.
I
think
we
should
keep
something
like
the
step
around,
because
here
it's
just
easy.
A
Similar
I
thought
that
proposals
for
putting
some
more
stuff
and
I
was
interesting
and
like
I,
feel
like
maybe
someday.
We
should
explore
one
of
those
but
I'd,
rather
just
put
it
in
a
new
codec
and
I.
Think
it's
a
lot
easier
to
do
that
than
to
expand
the
for
better
for
worse,
well,
battle-tested,
dag
JSON.
B
Named
links
at
all
or
Ruby
just
remove
them
because
they
are
a
huge
performance
it.
So
he
did
some
benchmarking.
It's
like
8%
of
the
times
just
using
those
creating
those
named
links.
And
now
my
question
comes
to
Eric,
because
I
guess
the
go.
I
thought
this
thing
has
something
similar,
I
guess,
but
I'm
not
sure
how
that
complementation
is
doing
it.
So
is
it
within
the
ILD
Highbury
or
is
it
just
like
outside
and
they
also
do
it
some
other
freely
instead
of
a
puppy
I.
A
Don't
know
hey
I
just
saw
your
message
about
that
moments
before
this
call
I
think
that
stuff
is
done
in
the
old
generation
of
I,
peeled
the
libraries
didn't
go,
I
think
it's
also
I
should
look
before
saying,
but
I
I'm,
pretty
sure
I
remember
looking
at
something
like
that
before
me
like
this,
does
not
look
efficient,
so
yeah
I'm
not
super
sure
what
we
will
want
to
do
about
that
in
the
future.
I,
don't
think.
A
It
seems
like
every
time
we've
talked
about
what
we
want
to
do
for
the
future
of
relationship
between
the
existing
dag
PB
stuff
and
where
we
want
to
go
in
the
future.
Its
involved
technology
and
there's
been
a
lot
of
you
need
details
in
the
existing
tag,
PB
stuff,
and
maybe
we
should
like
fess
up
to
the
fact
that
basically
UNIX
S
has
to
be
1
in
all
users
and
maybe
just
start
doing
things
based
on
that
admission
and
then,
if
anything,
spectrum
you
cut
above
general
purpose
usage
every
later
date.
So.
A
Yeah
wrapped
with
caveats,
if
I
don't
know
and
I'm
going
off
the
top
of
my
head
here,
I
feel
like
we
should,
it
seems
to
me
like
we
could
be
pretty
flexible
about
what
to
do
with
these
things,
because
it
doesn't
seem
like
there
are
many
consumers
of
this
API
and
it
seems
like
the
ones
that
are
consuming.
It
are
generally
like
giving
it
a
funny
look
already
anyway.
So
if
we
want
to
improve
it,
we
should
feel
free,
I,
think.
B
Yeah
so,
and
what
we
also
started
on
the
JavaScript
side.
Only
so
there's
a
PR
open,
truly
changed
the
API
from
the
protocol
buffers
into
just
those
things
that
you're
giving
safest
needs.
So
it's
like,
like
some
of
the
other
things
that
we're
a
synchronous
because
of
like
how
I
got
involved
and
so
on
and
but
now
like
provide
purpose,
doesn't
need
any
of
this
extra
stuff.
So
we
just
removed
it
and
so
I
guess
we.
B
But
you
want
to
make
it
as
small
as
possible
because
well
less
code,
less
bags,
obviously
and
I
think
the
named
links
is
also
something
that
is
just
I
mean
it's
certainly
a
cool
thing
like
yeah,
it's
cool
thing,
but
if
no
one's
using
it
yeah
there
isn't
there's
no
really
point
of
keeping
it
if
things
still
still
work
afterwards
and
so
I
would
lean
towards
removing
them
and
I
mean
even
if
it.
If
it
turns
out
it's
a
problem,
you
can
just
add
them
back.
I
mean
if
we
sussed
to
tony
find
out.
B
A
A
B
But
basically
case
so
so
I
I
will
I
would
take
care
of
the
stuff,
so
basically
I
will
so
from
the
IPL
D
team.
Currently
the
impression
is,
we
can
just
be
a
good
to
go
to
remove
it,
but
I
will
do
I.
Ask
for
the
people
who
clear
we
have
two
involved
previously
to
see
if,
if
it
is
actually
a
problem
or
not,
but
from
the
iPod
D
side
of
things,
there's
we
don't
see
any
issues
in
removing
it.
B
It's
the
script,
okay,
cool
I'm,
even
thinking
about,
perhaps
we
can
like
in
JavaScript
it
will
be
pretty
simple
to
just
put
another:
let's
build
it
on
top,
basically,
as
doing
surface
does
it
so
if
we
want
to
like
have
some
demos,
working
or
whatever
movie
to
the
tag
PB
named
linked
library
or
whatever
it
is,
but
this
would
be
something
that
basically
I
would
saw
that
we
could
build
if
we
need
it
and
we
hopefully
never
build
it.
That's
would
be
my
plan
yeah
one.
A
Other
application,
I
can
think
of
anywhere
near
this
area
is
vector
in
working
on
some
clusters.
Stuff
said
to
me
that
he
definitely
wants
to
maintain
this
ability
to
get
in
a
numeration
of
links
all
of
the
links
in
a
block,
basically,
even
if
they're,
deeper
object
and
I
told
them
yeah,
okay,
we're
just
gonna.
Do
that
with
a
selector
that
happens
to
do
that
and
just
gonna
make
that
a
expose,
like
that'll,
be
a
constant
selector
in
the
library.
B
B
B
A
C
Maybe
we
should
note
that
there
has
been
a
lot
of
discussion
about
advanced
layouts
and
this
has
been
Oh
generics
or
whatever
it
is.
This
has
been
occupying
a
lot
of
our
time
in
discussion
time
anyway.
Over
the
past
couple
of
a
couple
of
weeks,
maybe
I'll
put
some
links
in
the
in
the
doc
to
that,
but
there's
been
just
discussion
about
where,
where
it
fits,
how
it
relates
to
schemas
how
did
do
signaling
and
all
those
pieces.
C
A
B
A
A
A
We
should
have
a
powwow
about
the
naming
things
I,
don't
know
if
a
github
issues
like
the
greatest
way
to
reach
a
final
resolution
on
that
github
issues
have
this
tendency
to
become
in
forever
and
you
can
like,
there's
no
way
to
vote
or
signal
consensus
and
a
github
issue.
It's
just
not
a
conversational.
No,
that
you
can
do
in
that
technology,
she's
a
continuing
irritant.
A
A
C
C
Think
a
lot
of
a
lot
of
the
meaty
stuff
comes
out
of
Berlin.
We
didn't
put
in
person
stuff
yeah,
there's.
Actually
those
note
there's
something
I
should
mention
is
in
the
IPL
d,
a--
/,
IP
LD
repo,
there's
a
pull
request,
still
waiting
for
reviews
on
with
notes
from
billing,
so
be
good.
If
you
guys
have
a
look
at
that.
A
A
The
idea
being
that,
like
some
of
these
organizations
that
were
known
for
doing
good
research
in
the
past,
like
the
appointment,
people
would
have
a
lab
notebook
write
things
in
it
and
anybody
else
could
come.
Look
at
your
lab
notes,
but
it's
basically
where
you
put
your
ideas
in
it
as
they
come
to
you
and
that's
just
right.
There's
no
review
process.
It's
just.
A
A
Don't
know
if
I
have
any
great
ways
in
minds
to
like
execute
that
strategy
and
like
make
it
fit
well
with
him
the
substrate
of
get
but
I
think
it
would
be
cool
if
we
did
more
of
that
sort
of
stuff
correspondingly
closed
a
bunch
of
issues
because
they
can
live
somewhere
else.
But
that's
no!
That's
a
half-baked
father.
C
A
C
No
you're
right,
though
I
did
I
feel
like
this
there's
too
many
repos
on
too
many
issues
and
also,
at
the
same
time,
not
enough
capturing
of
some
of
the
things
that
are
going
on
in
their
own
heads,
but
also
when
we
collectively
talk.
You
know
we
do.
These
meetings
are
a
good
place
to
get
that
stuff
public,
but
there's
still
so
much
more
context.
That's
missing
from
this
and
so
I
think
I.
Imagine
you
know.
C
I
know
that
people
from
the
outside
can
try
to
come
in
and
try
to
get
up
to
speed
with
what
we're
collectively
thinking
and
talking
about
it's
probably
going
to
be
impossible
just
from
what
we
put
out
in
public.
Even
if
you
were
to
mine
slack
for
that
stuff,
it's
just
not
going
to
be
enough
yeah,
so
we
do
need
some
way
of
capturing
it
like
I.
Even
yesterday,
I
was
stumbling
across
some
resources
with
that
we
just
I
just
thought
this
is
amazing.
C
A
A
It's
when
I
try
to
do
it
for
myself
and
like
localhost
repo
I,
pretty
much
just
use
date,
stamp
files
really
really
heavily,
because
that
de
facto
causes
like
a
sort
of
mark
and
sweep
of
garbage
collection,
which
I
think
is
an
important
factor.
If
you're
not
going
I'm
not
trying
to
edit
things,
DOS
makes
it
easier
to
emit
new
content
that
lasts
for
any
given
project
for
like
a
year
or
so
before.
A
B
Yeah
I
guess
it's
something
like
like
to
me:
it
sounds
something
like
a
tiddlywiki
which
is
actually
nice,
I
kind
of
like
yeah,
like
this
killing
things
and
so
on,
but
yeah
like
yeah
I
I
tried
a
few
times
to
really
like
tiddlywiki,
but
I.
Don't
like
it's.
It's
close
like
a
concept
sounds
so
good,
but
I,
don't
know
yeah.