►
From YouTube: 🖧 IPLD Every-two-weeks Sync 🙌🏽 2019-07-22
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
A
A
Don't
have
that
much
to
say
because
I
wasn't
occasions
for
tweaks
I
had
one
week
of
catching
up
with
things
and
I
clearly
spend
my
time
on
wrapping
up
BJ
s,
IPL
D,
deck
PB
work,
because
we
want
to
make
things
simpler
and
kind
of
remissed
the
plan
to
finally
getting
all
those
changes
in
and
then
I,
hopefully,
won't
need
to
touch
the
say.
Yes,
I
killed
the
things
that
we
currently
have
that
much
in
the
future.
A
C
Have
had
tons
of
exciting
time
to
do
cogent
and
typed
no
stuff,
so
that's
been
fun,
there's
a
whole
bunch
of
stuff
there
that
had
been
kind
of
like
getting
no
time
and
attention
for
a
while.
So
in
the
last
couple
of
weeks,
I
finally
got
to
start
doing
serious
prototypes
of
code
generation
and
just
immediately
found
all
sorts
of
stuff
from
the
first
draft,
noting
your
faces
in
the
NGO
libraries
that
were
just
like
wrong,
like
the
first
draft
place,
will
lose
that
and
had
to
eat
flesh
behind
them.
C
C
So
there's
now
a
distinction
between
like
type
kinds
versus
the
data
model
kinds
so
like
struct,
is
a
kind
of
map
schema
system,
but
it's
not
in
the
raw
data
model,
of
course,
and
so
now
I
have
a
type
2
node
interface
coming
along
and
it's
got
the
ability
to
do
some
inspection
on
the
type
that
describes
it
and
it
still
acts
like
a
regular
note
in
every
other
way.
The
same
your
place
is
the
data
model
stuff
does,
and
then
they
have
a
second
kind
of
note.
C
They
can
be
treated
as
which
does
the
representation
stuff
so
like.
This
is
how
the
whole
system
will
work
out
from
the
libraries
perspective
when
you
have
like
a
struct
type
with
a
tuple
representation
will
actually
have
two
nodes
that
refer
to
the
same
data,
one
of
them
black,
like
I'm
out
of
the
list.
C
The
amount
of
work
that's
ahead
in
Cochin
is
absolutely
massive
and
I'm
trying
to
sort
of
skitter
around
and
do
the
minimum
things
to
figure
out
where
all
the
hard
problems
are
supposed
to
be
so
at
the
moment,
I'm
working
on
structs,
because
those
are
by
far
one
of
the
most
complicated
things.
I
did
some
of
the
simple
scalar
stuff
like
a
type
string,
and
that
was
just
easy,
so
I'm
skipping
over
the
rest
of
those
until
later
and
I'm
kind
of
scared
about
some
of
the
complexity,
around
unions
in
the
future.
C
B
C
D
Right
so
I've
been
primarily
playing
with
a
practical
implementation
of
some
of
our
work.
So
NPM
has
a
data
set.
One
of
the
reasons
I'm
interested
in
NPM
is
I
actually
have
a
use
case
for
it.
I
have
a
website.
Node
I
drop,
see.
Oh
no!
No!
No
Tyco,
IP
years
ago,
I
made
this
service
for
putting
NPM
badges
onto
your
github.
Readme
became
really
popular,
and
you
know
a
large
portion
of
NPM
packages.
D
Their
commercial
concerns
have
stopped
them
and
making
easy
access
to
certain
bits
of
data,
and
so
you
need
to
do
you
know
interesting
stuff.
You
need
a
local
copy
of
it
to
be
able
to
have
your
own
database
to
do
queries,
so
my
thought
a
while
back
was
well
one
and
I
do
something
with
IPL
day
and
actually
use
IPL
T
as
a
local
NPM
database
and
then
use
that
as
a
way
to
extend
some
of
these
ideas.
D
D
It's
going
it's
going
well
and
it's
making
me
explore
a
lot
of
it
interesting
areas
that
I
think
are
really
useful
for
me
to
have
my
head
around
so
far,
I'm
I'm,
avoiding
the
DHT
I,
don't
do
anything
with
live
p2p
at
the
moment.
It's
because
just
could
just
blocks,
and
you
know
fetching
block,
storing
blocks
the
really
simple
operations.
D
D
The
there's
this
challenge
as
I
keep
on
thinking
about
in
this
domain.
One
of
them
is
how
do
you
update
I
know
the
package
managers
folks
have
to
deal
with
this
stuff
and
I
saw.
There
was
a
discussion
in
there
about
IP
and
IP
n
s.
I
didn't
was
starting
to
sync
with
them
about
their
needs
on
IP,
NS,
and
so
hopefully,
I
PMS
becomes
a
a
useful
thing
one
day,
but
for
now
it's
like
how
do
you
so
you've
got
a
you've
got
a
route
block
for
through
this
complex
array
of
data.
D
How
do
you
first
of
all
trigger
an
update
and
then
how
do
you
perform
an
update
on
that?
So
been
thinking
a
lot
about
the
idea
of
mutable
links?
You
know
these
these.
These
blocks
that
are,
you
know,
have
some
way.
Has
some
mechanism
of
saying:
there's
an
X
there's
some
sort
of
expiry.
Where,
if
you
beyond
this
time,
then
maybe
you
should
explore
if
you
can
getting
a
new
version
of
me
and
here's
a
signed
is
a
way
of
verifying
that
the
replacement
of
me
that
you
fetched
from
this
place
is
actually
valid.
D
It's
just
thinking
about
that
that
sort
of
problems,
and
then
also
on
the
client
side.
So
when
you're
querying
this
thing
and
you're
fetching
blocks
and
you're
building
your
own
local
version
of
what
you
need
of
it,
how
are
you
storing
it?
I've
been
looking
at
the
I
think
it's
the
block
service
API
that
everyone's
using
for
ipfs
looking
at
performance,
concerns
there
and
see
if
there's
any
any
interesting
optimizations
for
this
particular
use
case.
D
But
you
know
it's
just
it's
just
an
interesting
problem
set
because
it's
there's
a
whole
bundle
of
concerns
that
it's
forcing
me
to
think
about.
It
I
think
it's
helping
me
to
get
into
similar
headspace
to
everyone
else
that
has
been
working
on
this
stuff
for
longer
than
I.
So
that's
pretty
much.
What's
been
absorbing
my
time,
cool
things.
A
No
okay,
it's
fine
I've,
just
posted
two
new
gen
items.
That
came
to
my
mind
because,
like
first
of
all,
there
was
a
lot
of
stuff
happening
while
I
was
locations
so
kind
of
the
whole
person
thing
kind
of
blew
up
or
like
become
big
and
to
attempt
to
catch
up
and
when
I
was
finally
caught
up.
Michael
was
on
holidays.
So
we
could
really
think
about
that.
A
Much
about
this
part
that
much
and
yeah
I
just
want
to
discuss
with
the
team,
because
yeah
I've
also
talked
with
some
of
you
already
that
I
have
the
impression
that
Mike
was
really
excited
about.
This
wasn't
stuff
was
cool,
I'm,
also
kind
of
excited
but
I,
think
or
differently.
So
I,
always
sleep
less,
like
researchy
thing,
and
not
really
so
it
should
really
block
anything
because
I
think
in
some
document
there
and
notes
about
what
the
relation
between
was
them
and
they
Phoenix
FSB
to
work
and
I.
B
Then
I
just
give
some
comments
like
yes,
like
we
started
going
down
to
wasn't
rabbit-hole
because,
like
us
all
these
things-
and
you
know
as
waited
this
is
going
to
be
really
complicated.
We
need
something
now
so
Deb
únicos
to
and
then
we're
gonna
eunuch
service,
3,
eventually
and
eunuchs
best
3
will
be
like
basically
we'll
be
able
to
interpret
eunuchs
mess
to
using
eunuchs
best
3.
These
disbeliever
do
this
for
our
clinic.
A
A
D
D
How
can
we
target
a
future
where
those
in
practicalities
might
be
able
to
come
and
that
this
thing
is
insurgent?
So
that's
why
you
know
whenever
I
talk
to
Michael
about
this
stuff,
it's
like
you
know,
he's
taking
some
of
some
of
the
decisions
that
might
otherwise
look
strange.
I've
got
this
was
a
future
in
mind,
and
so
that's
always
you
know
some
stuff
around
composites.
D
B
I
think
it's
inevitable,
because,
like
every
time
we
look
at
this
problem.
It
always
leads
back
to
shipping
code
to
people
right,
but
then
I
always
treasure
to
this
I.
Well,
it's
like
a
lot
of
copy
of
poems.
You
just
saw
first,
we
deal
with
like
interoperability,
upgrade
ability
to
people
writing
or
writing
different
code
and,
like
actually
I
mean
Allah,
sees
work
together.
It's
yeah,
it's
versatile
security
concerns.
So
that's
why
a
unity
was
kept.
Lee
block
on
it,
but
right,
given
that
Rossum
was
I,
could
be
shipping
arbitrary
code.
B
D
It's
and
so
I
the
way
I've
been
going
with
my
code
and
even
the
algorithms
I've
been
running
for
the
specs
is
trying
to
trying
to
boil
them
down
into
single
operations.
So,
instead
of
saying
this
thing
is
a
a
set
of
you
know:
it's
a
data
structure
with
these
algorithms
associated
with
it.
Now
here's
a
data
structure
and
here's
an
operation,
and
this
is
how
you
perform
that
operation
and
then
even
Michael's
code.
It's
it's
like
you
want
to
traverse.
D
Okay,
we
want
to
get
from
here
to
here
and
you
wanna,
you
wanna,
do
a
get
with
a
key:
that's
an
operation
that
should
be
self-contained
and
there
should
be
the
interfaces,
it's
the
kind
of
thing
that
wasn't
might
have,
and
it's
not
necessarily
constrained
by
that,
but
its
interest.
It's
an
interesting
constraint
to
force.
D
You
know
really
clean
algorithms,
where
you're
traversing
block
by
block-
and
you
know
where
you're
gonna
get
and
you've
got
these
various
things
where
you
might
need
to
even
need
to
store
continuation
information
and
those
those
little
operations
are
turning
into
the
kind
of
things
that
would
convert
really
well
into.
Wasn't
we
just
what
I
can
take
Geely.
B
A
Yes,
I
think
the
like
would
put
truly
big
discussion
is
about
exactly
this
being
is
side-effect
free
and
so
basically,
if
your
side-effect
free,
it's
easy
to,
do
it
wasn't
or
easier
to
do
it
wasn't
and
so
I
think
yeah.
So
there's
benefits
to
think
about
in
those
terms,
but
actually
what
I
think
what
in
terms
we
were
thinking
is
more
like
functional
programming,
basically
doing
it
it,
for
my
Erlang
background
is
like
we're
just
doing
freshman
programming
and
if
Michael
has,
though,
wasn't
everything
in
mind.
A
A
All
right
cool-
and
this
was
basically
Vietnam
I-
want
to
talk
about
what
there
wasn't
stuff
and
the
other
one
is
I.
Think
Eric
brought
it
up
if
we
should
think
about
making
this
a
weekly
meeting,
because
there's
plenty
of
stuff
now
going
on
and
it
might
make
sense
to
catch
up
more
frequently
but
I'm
like
I,
don't
never
spoken
about
it.
So
on
it's
open
for
discussion.
C
D
A
B
D
We
could
do
it's
just
that,
then
it
becomes
my
late-night
and
these
guys
early
morning.
It's
it's
just.
It
sucks
either
way.
I
I
think
I.
Think
these
guys
are
you
guys
are
okay
with
the
evening
night
hours
on
you.
So
you
know
I
know
to
me.
You
know:
I've
got
kicking
up
in
the
morning
for
kids
anyway,
so
I,
my
my
night
out
a
bit
capabilities
are
more
limited
than
they
used
to
be
so
I
make
hay
with.
A
This
video,
okay
cool
as
so
should
be
so
it
should
be
smith.
You
say
we
start
and
next
week
with
the
first
weekly
meeting,
basically,
okay,
I
I
can
make
sure
that
I
updated
all
the
like
repository
saying
we
oh
wow,
I
I
bought
updated
for
now.
We
first
do
the
weekly
meetings
and
I
laugh
at
them.
Whatever,
so
I
will
update
the
repositories
to
see
weekly
after
they
resumed
cool
am.
Is
there
anything
else?
Anyone
wants
to
talk
about.
D
A
D
Don't
think
so,
I
think
I
think
there's
it's
just
this
shared
mental
state
that
we
have.
We
continually
have
trouble
with,
and
we
have
these
discussions
about.
Well
things
like
the
terminology
in
advance.
They
have
versus
composites.
That's
that's
a
shared
mental
state
problem
where
we
really
need
to
get
on
the
same
page.
D
C
On
that
train
of
thought
wondering
if
it
might
be
a
good
time
to
think
about,
like
checking
in
with
some
very
central
read
knees
and
like
mission
statements
of
stuff
I,
don't
know
if
that
will.
Maybe
we
get
I,
don't
know
if
that'll
be
easier
than
having
concrete
stuff
or
not.
But
there's
some
conversations
we've
had
recent
war
I
feel
like
we
could
short-circuit
to
the
endgame
much
much
faster.
C
C
A
Yes,
I
think
that
that's
a
good
idea
because,
like
when
we
like
this
was
done
I'm,
mostly
by
Michael
and
I
I,
agree,
basically
agreed
on
what
he
was
doing,
but
it
was
like
a
year
ago,
nor
nine
months
ago,
I
think
there
was
like
the
bank.
Then
the
mission
was
pretty
clear
and
was
good
to
have
it,
because
we
finally
had
an
agreement
on
the
scope
and
things.
C
A
D
B
D
C
B
C
C
C
Think
we've
been
stuck
on
three
layers
for
whatever
reason
for
longer
than
we've
even
had
schemas,
which
is
kind
of
funny.
So
now
we've
got
like
schema
layers,
roughly
layer,
three
and
that's
been
kind
of
ill-fitting
since
the
beginning-
and
we
know
it-
we
just
didn't
come
up
with
any
better
suggestion
right,
so
we
haven't
been
able
to
figure
out
where
exactly
advanced
a
collapse.
In
there
we
kind
of
hooked
a
mogila
schema
layer,
but
it
never
quite
fit
and
just
ignored
it
right
and
it
hasn't
been
a
super
clear
way.
Selectors
stuff
is
gone.
C
It
doesn't
fit
clearly
onto
any
of
those
three
layers
right.
It
fits
on
top
of
either
types
or
data
model,
or
maybe
something
like
yeah.
The
the
three
strict
layers
doesn't
work.
Super
good
I,
don't
know
what
other
models
we
can
replace
it
if
that
are
going
to
be
simple
either,
but
it
was
really
really
tricky
to
talk
about
selectors
in
the
layer
model
on
some
of
that
stuff
during
camp
when
people
do
already
stood
in
it,
though
yeah.
D
As
I
said
to
you,
yesterday,
I
mostly
don't
care
what
the
hierarchy
is
as
long
as
there's
some
kind
of
hierarchy
to
this
stuff,
because
there's
so
many
when
you
approach
this
from
the
outside
there's,
so
many
little
rabbit
holes
to
go
down.
We
need
to
be
able
to
organize
this
in
a
way
that
somebody
new
coming
up
to
it
can
see
where
the
rabbit
holes
are
and
their
importance
in
the
stack.