►
From YouTube: 🖧 IPLD Bi-Weekly Sync 🙌🏽 2019-04-29
Description
A bi-weekly meeting to sync up on all IPLD related topics. It's open for everyone and recorded. https://github.com/ipld/team-mgmt
A
All
right
so
welcome
everyone
to
this,
be
weekly
IPL,
d/deaf
meeting,
it's
April,
the
29th
11
p.m.
my
time
and
welcome
everyone.
So
as
every
time
we
just
go
over
what
we
did
and
then
discuss
any
open
issues
that
you
might
have
as
I'm.
The
gatekeeper
of
the
trip
head.
I
am
the
first
one
who
had
edit
his
items
so
I
basically
just
worked
on
the
IPR.
The
format
stuff
still
I
got
good
reviews
from
rod
in
the
past
week.
A
No,
oh
I,
don't
actually
add
animation
anyway,
it's
about
geo
data
and
about
spatial
temporal
data
and
it's
just
a
catalog
and
it
is
subscribe
normally
with
aesthetic
JSON
files
and
it's
really
build
upon
em
HTTP
at
the
web,
and
hopefully
I
can
convince
the
people
to
kind
of
it.
At
least
a
few
places
change
the
specs
or
the
also
works
with
content,
addressed
systems
and
good
news.
Is
they
don't
use
em
json-ld?
So
they
don't
use
the
full
thing
we're
just
like
having
their
own
linking
stuff
and
they
are
using
URLs.
A
A
That
I
can
influence
the
spec
and
the
plan
is
to
actually
get
a
prototype
working
because
it's
an
upcoming
conference
in
August,
where
there's
a
challenge
where
I
submitted
that
I
want
to
work
on
this
stuff
and
I
think
is
accepted
or
even
if
it's
not
accepted,
I
will
still
work
on
the
stuff
who's
in
my
class
and
I.
Think
it's
pretty
exciting
too,
because
it
would
be
the
first
step
into
the
Geo
world
and
they
don't
need
to
get
the
I.
A
Don't
like
full
idea
of
how
condom
addressing
works,
and
so
much
is
their
small
step
like
a
kettle,
obviously
a
small
step
into
this
space
and
they
hopefully
convince
them
the
people
to
use
it
and
then
a
CEO.
You
know
just
so
my
data
in
it
and
so
on
and
go
from
there
and
but
I
plea
excited
about
it
because
I
think
it's
a
nice
small
size
project
to
get
things
started.
A
B
B
I
can
mint
us
some
tokens
and
we
can
use
it
for
just
a
place
to
store
blocks
if
we
want
to
build
some
stuff
and
just
do
really
really
simple
demos,
I'm
still
fiddling
with
the
Amazon
stuff
to
try
to
get
clouds
plants
on
a
subdomain
to
work,
but
there's
just
a
different
real
structure
that
we
can
use
for
now,
other
than
that
the
meetup
is
next
week,
so
we
have
a
full
set
of
people
to
give
ritual
talks.
I
think
at
this
point.
B
If
anybody
else
wants
to
give
a
talk
there
still
room,
though
we
have,
we
were
really
big
slot
like
an
hour
and
a
half.
So
anybody
who
wants
to
say
something:
please
do
it
also
the
people
that
are
signed
up.
Remember
that
you
like
you,
need
to
write
a
talk.
So
do
that
and
then
yeah
and
then
get
ready
for
the
summit
in
a
few
weeks.
It'll
be
real
fun.
Two.
C
C
So
we
had
a
big
long
meeting
about
that
and
the
graphs
Inc
group
of
folks
and
there's
a
bunch
of
links
in
the
crypt
pad
here.
If
anyone
wants
to
check
that
out,
I've
discussed
it
with
some
people
in
one
all
the
ones
already,
but
I
think
it's
okay,
but
we'll
see
so
there's
some
spec
pr's
and
the
interface
in
the
go
code
has
been
sort
of
slowly
evolving
and
it
got
some
interesting.
C
New
features
like
selectors
are
now
going
to
have
an
interface
in
code
for
describing
whether
they
are
high,
cardinality
or
a
low
cardinality
in
there,
in
what
they're
going
to
explore.
So
we
can
decide
whether
or
not
we're
gonna
iterate
through
the
whole
node
or
if
we're
gonna
like
plug
out
individual
fields,
which
might
be
an
important
performance
optimizations
for
like
complex
data
structures,
the
camps
and
stuff.
So
that
was
interesting
and
yeah.
Some
of
that
stuff
seems
to
actually
be
shaping
up.
Really
nicely
now
it's
back
to
a
recursive
format.
C
C
But
that's
like
a
very
partially
big
thought
right
now.
So
if
anybody
has
more
time
the
next
week
or
so
and
wants
to
talk
about
that,
I
really
like
some
other
brains
and
in
other
news
we
had
Jeremy
and
I
had
talked
about
like
what
will
be
needed
for
Falcone
to
start
using
go
IPL
D
prime
ntldr
they're,
probably
going
to
need
to
bind
two
node
implementations
and
I
have
no
idea
how
to
prioritize
that.
Yet
so
there's
some
ongoing
discussion.
D
Guess
I
had
a
question:
if
we
have
time
Michael,
you
can
decide
we.
This
is
for
another
time
or
not,
but
I
I
had
been
bugging
Mike
a
little
about
about
the
what
I
perceives
like
features
and
I
guess.
D
Maybe
it's
maybe
it's
a
little
slow
or
maybe
it's
a
pain
for
people
to
run
cluster
if
they
don't
know
how
to
run
cluster,
so
you
want
them
to
be
able
to
use
like
hosting
on
Amazon,
instead
err
or
something
like
using
s3
instead
of
running
nodes
that
have
to
process
all
the
traffic
I
think
those
are
like
different
problems.
I
guess
I
just
want
to.
If
people
have
time
and
are
curious,
I
gonna
be
interesting.
Well,.
B
I
mean
that's
like
that.
That
question
is
the
answer
to
that
question
is
different,
depending
on
the
use
case
right
like
for
some
use
cases,
it's
gonna
work,
fine
and
for
some
that's
not
and
I
think
that
I'm
not
entirely
clear
on
what's
the
top
line,
use
cases
are
for
ipfs
other
than
just
package
manager
and
if
there's
a
particular
package
manager
and
it's
data
that
we
know
that
we
need
to
ask
for,
and
we
know
how
many
records.
B
That
is,
that
we
can
come
up
with
a
pretty
clear
answer
on
like
whether
or
not
that's
going
to
work.
But
if
it's
just
an
open-ended
question
that
I
don't
really
know
how
to
solve
it.
Right,
like
I,
know
that,
when,
like
some
of
the
work
that
our
katie
has
done
to
try
to
talk
to
partners
or
potential
partners
and
their
use
cases,
some
of
those
have
been.
You
know
in
the
millions
of
record
and
it's
just
been
kind
of
a
no-go,
but
I
don't
know.
D
B
Like
crypto
kitties
is
one
of
them
so
like
how
many
kitties
are
there
I,
don't
know,
there's
a
lot.
There's
gonna
be
more
in
the
future.
It
could
be
the
record.
B
I
mean
I
think
the
issue
with
like
with
something
like
crypto,
kitties
and
and
and
a
lot
of
these
cases.
There's
this
big
long
tail
of
like
stuff.
That's
really
high
access
that
you
that
everybody's
gonna
be
asking
for
and
then
like
90
plus
percent
of
its
there,
like
nobody
ever
asked
for
or
the
or,
if
they
do
ask
for
that's
worth
like.
B
Incredibly
rarely-
and
you
know
if
we
put
it
in
a
ham,
it's
just
going
to
be
like
really
random,
which
directories
are
going
to
need
or
that,
and
then
it's
not
necessarily
going
to
fix
that
for
us
like.
We
don't
have
a
great
way
to
know
ahead
of
time,
which
of
those
like
need,
that
kind
of
action
pattern
of
which
don't
the.
E
The
vault
look
at
the
concern
I'm
trying
dress.
There
is
actually
confident
writing
well,
I
can
be
put
into
my
camp.
Then
at
least
we
won't
announce
one
route.
The
problem.
There,
though,
is
it's
like
if
someone
outside,
so
you
know,
I
wanted
item
directly
instead
of
this
beautiful
I
penis,
then
that
just
doesn't
work
because
I
won't
be
an
opportunity.
Each
and
every
single
individual
record.
B
B
B
E
B
Mean
like
like
these,
these
are
all
like.
We
have.
We
have
a
bunch
of
different
answers
that
we
can
apply
to
different
use
cases
right,
I.
Think
that
what
we
really
need
to
understand.
It's
like
what
is
the
use
case
said:
ITSs
is
going
to
end
up
max
like
optimizing
for
that
we
definitely
need
to
be
able
to
fix,
and
then
we
can
say
like
okay.
Well,
it's
this!
It's
potentially
this
research
project.
B
D
D
B
B
E
Would
be
painful,
but
any
network
change
is
painful
basically,
so
like
well
I
think
there
may
be
ways
of
doing
that.
We
saw
less
painful,
basically,
no
matter
what
we
do.
We'll
have
a
transition
period,
we're
like
a
big
server
starts,
a
server
start
supporting
it
and
then
I
have
deep-sea
clients
start
using
it
effectively
like
yeah.
B
E
B
You
mean
that
there
will
just
be
a
time,
I
mean
so
I'm
trying
to
get
clarity
like
by
that.
Do
you
mean
there's
a
period
in
time
where
there
are
records
in
the
DHT,
using
this
new
thing
that
a
bunch
of
peers
just
don't
know
what
to
do
it
or
literally,
like
those
peers,
will
blow
up
when
they
see
that
data,
because.
E
It's
only
that
the
servers
will
reject
it,
probably
depending
how
we
implement
something.
That's
rejecting
these
records
entirely.
So,
in
fact
like
to
make
this
useful
without
like
adding
a
bunch
of
round
trips
like
try
something
and
then,
if
that
doesn't
work,
try
something
else
like
we
could
either
up
the
beach
team
protocol.
E
So
if
you
learn
ahead
of
time,
oh
you
speak
this
person
protocol,
therefore
I
know
it
used
to
be
that
you
can
store
these
kinds
of
records
or
we
just
like
say
well,
like
we
add,
support
on
the
server
side,
for
this
tip
record,
wait
a
couple
releases
get
most
the
network
upgrade.
We
can
look
at
the
temerity
HD
crawlers
like
see
like
one
most
network
is
acquitted
and
then
once
this
happened,
then
we
upgrade
all
the
clients
and
say
like.
Okay,
now
the
clients
can
start
to
fittings.
E
E
E
D
E
E
E
Like
like
address
HTTP
endpoint,
not
massaged,
to
be
file,
it's
like
there's
everything
like
HTM
point.
That's
used
like
an
RPC
library
or
something
like
that.
Like
a
piece
of
content
on
some
server,
it's
like
most
things,
just
it's
not
used
for
content
so
like
the
way
we'd
use
it
to
say
like
okay.
Here
is
like
this
directory
in
the
server
and
then
under
this,
or
it's
really
like
here's.
This
API
endpoint
somewhere
query
this
API
endpoint.
E
For
this
thing,
the
problem
it
just
like
I
think
we
have
three
fields
won't
be
like
here
is
the
endpoint.
You
talked
to
here's
the
protocol.
You
should
speak
with
this
endpoint
and
then
well.
Yeah
really
started
to
things
yeah,
but
here's.
The
endpoint
here
is
the
protocol.
We
should
speak
over
this
like
the
employ
described
by
people
together.
We
could
potentially
put
them
all
theaters
like
this.
Space
is
not
very
well
defined.
E
A
F
Yeah
so
Maya
place.
Look
I,
don't
have
anything
major
that
the
to
report
I
haven't
had
the
most
productive
couple
of
weeks
because
of
some
health
problems,
but
I
spent
over
a
bunch
of
time
on
schemas
I
have
been
building
at
a
little
test
suite
for
a
schema
quaza,
hopefully
trying
to
get
this
right
and
also
bumping
up
against
some
of
the
issues
that
I've
been
going
back
and
forth
with
Eric
about
its
regard
to
Union
representations.
F
So
that's
been
kind
of
interesting
and
also
the
other
thing
was
the
pull
requests
that
the
volkl
put
in
to
all
the
IPL
DJ's
libraries.
That
was
a
really
good
chance
to
dive
into
the
details
of
them
all
like
I
hadn't,
really
looked
in
the
protobuf
I
peeled
a
library
and
I
spent
some
quality
time,
understanding
that
so
that
was
that
was
all
very
helpful
for
understand
the
stage
of
things.
F
So
those
were
the
two
big
highlight
items
for
me
and
going
ahead.
I'm
mainly
focused
on
schema
stuff,
because
I
want
to
I
want
I
really
want
to
make
it
to
a
point
where
I
can
actually
start
implementing
something,
because
I
can
imagine
this
being
a
rabbit
hole
that
it
lasts
for
months.
I.
Don't
really
want
that
to
happen.
I
want
to
be
able
to
use
this
for
something
interesting
soon,
so
that
we
can
prove
its
utility.
So
that's
that's
the
aim
to
make
it
useful
soon,
cool.
That's.
C
Work
on
the
stuff
that
I'm
drafting
for
selectors,
for
example,
just
make
sure
I'm
not
doing
anything
BS
in
the
selectors.
That's
another
thing
that
would
give
me
it's
at
you
really
soon
and
that's
not
stuff,
that's
fully
landed.
It
would
go
libraries
either,
but
I
wish.
It
was
for
an
example
of
how
to
make
it
concrete.
F
F
Still
want
to
get
this
Union
stuff
right,
and
so
what
I'm
going
to
do
in
the
next
couple
of
days
is,
is
trying
I'll
play
with
the
proposal
that
you
had
for
putting
the
representation
for
the
fields
in
line
with
friends
and
see
if
that
solves
some
of
the
annoyances,
because
it's
it
really
is.
When
you
look
at
a
parser,
it's
then
unions
that
stand
out
as
the
problem,
because
all
the
reference
representations
do
something
completely
different
to
what
you're
reading,
so
it
just
doesn't
seem,
doesn't
seem
right.
F
Working
on
the
boat
on
those
yesterday
and
that's
see,
it
didn't
seem
that
that
problematic,
I
haven't
done
the
string
doing
stuff
any
of
the
string
formats.
Man,
maybe
they'll
end
up
being
weird
as
well,
but
it
just
seems
like
the
union
ones
were
fundamentally
weird
because
each
one
of
them
was
was
different,
but
yeah.
You
might
be
right.
B
So
one
in
this
case
that
I
have
that
I
need
like
I
want
to
have
in
JavaScript
is
basically
just
a
change
log
data
structure.
So
the
note
itself
would
just
have
the
current
route
like
a
link
to
the
current
route.
I,
don't
link
to
the
prior
change
log
route
so
that
you
could
just
Traverse
back
through
all
prior
versions
as
you
update
it
so
and
it
has
sort
of
a
mild
transform
in
that.
B
As
you
apply
paths
to
it,
those
paths
all
just
apply
directly
to
the
to
the
current
to
the
current
route,
but
there's
also
a
property
for
versions
to
go
back
in
the
time,
but
those
yeah
those
would
get
exposed
to
little
different
reasons.
They
actually
look
in
the
data
structure.
So
there
is
like
something
like
I
transform
to
the
reader
and
yeah
I.
Don't
know
I,
don't
know
how
close
we
are
to
having
that
be
a
thing.
F
F
Know
one
of
it
was
I'm
trying
I'm
trying
to
get
to
that
quick
quicker
by
trimming
off
some
of
these
stuff.
That
just
seems
a
little
bit
extraneous
but
necessary,
but
in
the
future,
but
not
so
useful
for
getting
to
utility
early,
so
I
mean
I.
I
need
to
I
need
I
do
need
to
draw
a
line
under
under
my
approach,
because
I
don't
really
as
I
said.
I
don't
want
to
be
spending
the
next
two
months
on
just
pausing
and
getting
it
perfect,
because
that
doesn't
seem
like
a
good
place.
Good
use
of
time.
F
C
F
F
And
I'll
try
and
get
the
code
up
to
somewhere
on
github
soon
it's
just
not
in
a
good
State
for
putting
anywhere
and
it's
better
to
make
it
a
bit
more
structured.
So
what
I
would
have?
What
I
wanted
to
do
is
share,
be
the
test
case.
Fixtures
with
you,
so
I
did
them
in
a
form
on
writing
them
in
a
format
that
is
not
specific
to
JavaScript.
F
I
haven't
got
to
that
bit
yet,
but
that's
what
else
I've
been
geeks
til
next
week,
so
I'd
like
to
share
those
with
you,
because
they
that's
where
I'm
starting
to
see
more
of
these
patterns
of
it's.
The
examples
really
will
make.
This
thing
come
alive,
at
least
for
me
anyway.
So,
there's
these
building
out
these
small
snippets
of
what
this
thing
looks
like
and
what
it
what
it
should
be
represented
as
to
be
useful.