►
From YouTube: 🖧 IPLD weekly Sync 🙌🏽 2020-05-25
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
A
Think
we've
discussed
this
a
lot
in
previous
meetings
as
well
that
it's
very
important
that
we
can
pass
on
your
custom,
coding,
cable
and
we
don't
bundle
em
huge
tables
with
implementations
and
I
was
stuck
with
a
small
point
which
is
getting
an
H
like
codec
implementation
based
on
such
a
code.
So
let's
say
you
want
to
have
takes
Ybor
and
you
only
have
the
integer
value
getting
in
the
actual
codec,
which
seems
pretty
straightforward
and
if
you
think
about
other
implementations,
but
in
rest
with
the
current
implementation.
A
It's
not
unlikely
today,
at
the
end
of
the
day,
I
found
a
solution,
I
think
which
is
yeah
exactly
you
I
think
what
I
would
need
for
the
blog
API.
It
is
not
as
general
as
it
could
be,
but
that
general
case
really
doesn't
work
the
way
things
are
implemented,
but
the
way
things
are
implemented,
I
think
makes
sense.
B
Shoot
I
did
a
bad,
I
was
still
typing.
I
did
a
bunch
of
things
this
week,
most
of
which
I
probably
forgot
and
little
think
of
tonight.
Your
poem
but
go
Cogan
got
some
major
progress
this
week,
lists
and
maps
now
both
work
fully
recursively
and
okay,
that's
not
news,
but
what
is
is
that
includes
recursion
and
when
they
are
in
representation
book.
So
previously
this
was
a
kind
of
sizable
to
do.
B
So,
like
you
could
be
partway
through
a
structure
and
like
maybe
the
struct
inside
of
this
map
is
supposed
to
be
string,
join
representation
of
a
struct
or
something-
and
you
would
Traverse
into
this-
and
you
would
forget
that
you're
in
our
presentation
mode
and
so
the
node
interface,
you
get
back,
continue
acting
like
a
map
and
what
you
actually
want
is
long
acts
like
strings
that
structure
right.
So
they
do
the
wrong
thing
now
long
story
short:
it
does
the
right
thing.
B
So
if
you
can
pass
for
some
node
value
and
you
get
the
representation
on
it,
that's
a
good
of
some
big
tree
of
data.
If
you
now
pass
that
node
to
like
any
of
the
regular
traversal
things
or,
for
example,
regular
codec,
marshaling
it'll
just
transparently
do
the
right
thing,
because
all
of
the
recursion
will
stay
where
precipitation
Mogens
so
connects.
Now
has
parently
understand,
code
generated
type,
serialization,
cool.
B
A
C
C
The
main
thing
is
speed
in
the
IPL
d,
org
JSI,
PLD,
Bitcoin
pool
request.
Number
63
is
where
the
bulk
of
my
last
couple
of
weeks
and
we're
about
pulling
that
out
into
a
second
supposedly
because
it's
it's
basically
a
complete
rewrite
of
the
JSA
PLD
Bitcoin
code.
So
and
it
doesn't
it's
not
compatible
with
the
old
one.
C
Than
that,
it's
it's
pretty
much
done
in
my
kid.
How
jeaious
Bitcoin
block
does
the
encoding
and
decoding
work
for
the
Bitcoin
format?
That's
had
to
that's
had
to
have
little
bits,
adjusted
over
the
last
few
weeks
as
I've
encountered,
these
unicorn
Deluxe's
and
calling
them
in
the
in
the
blockchain
just
blocks
and
unexpected
features-
and
you
know
I
think
we're
talking
last
week
about
how
there's
a
there's
a
financial
incentive,
block
change
for
novelty
and
and
that
that
makes
it
really
interesting
case
study
in
these
unicorn
blocks.
C
Think
that's
still
run.
Anything
I've
got
that
running
on
my
machine.
Yes,
it's
just
doing
a
verification.
It's
it's
seems
to
be
fine,
there's
no
problem
there
and,
lastly,
there's
a
pull
request:
one
seven:
six
in
the
multi
Couric
in
the
multi-format,
slash
modacrylic
repo
for
the
new
multi
codecs
table
to
add
one
more
multi
codec
for
Bitcoin
called
Bitcoin
with
this
commitment,
which
is
code
b2,
and
that's
for
this
one
little
data
structure
that
is
new
to
pull
this
all
off.
So
we've
got
three
codecs
now
for
Bitcoin
and
I've
had
enough,
but
I'm.
C
The
plan
is
to
move
on
moving
on
to
other
block
chains
and
I.
Think
at
the
moment
this
is
not
for
them,
but
I
I
have
started
looking
at
Bitcoin
cash
and
I
might
do
between
SB
as
well,
because
these
are
hard
focus
of
the
blockchain,
which
means
that
they,
their
code
is,
is
very
close
to
the
original
Bitcoin
code,
and
today
I
can
start
processing
from
the
four
point
onwards.
So
I
don't
have
the
whole
doctrine.
C
C
But
this
this
work
is
getting
the
T
is
so
I
think
I
have
to
sprinkle
any
myths.
Mother,
some
other
work,
including
getting
back
to
my
JavaScript
Siebel
possum,
sorry,
C,
ball,
encoder
or
decoder,
which
was
very
close
to
being
everything.
The
basic
functionality,
gun
and
I,
wouldn't
mind
getting
back
to
that.
But
I
do
have
a
bunch
of
other
things
on
my
plate,
my
neatest
scaler
in
there,
but
just
to
keep
myself
saying
I
think
I
need
to
focus
on
this:
a
hundred
percent
because
it's
yeah,
it's
it's
quite
consuming
I'm
finding
anyway.
D
D
D
Yes,
part
deck
for
peer
IDs
as
part
of
the
subdomain
and
the
way
we
are
encoding
currently
Lippe
to
be
keys
with
all
the
overhead
around
them
once
encoded
in
based
rt2
is
65
bytes
long,
which
is
2
bytes
longer
than
a
DNS
sub
part,
and
well.
We
don't
care
about
DNS
that
much
everything
out
there
when
it
parses
even-even
slack
when
it
parses
URLs
in
in
text,
it
will
cut
things
out
at
63
bytes.
D
So
whenever
you
do
something,
that's
URI
looking,
you
have
to
have
every
part
of
the
secured
portion
of
the
URI
be
at
most
60
bytes,
and
this
is
why
we
are
introducing
support
for
base
3
6
in
ecology
edges,
because
then
everything
fits
within
the
limits
that
we
work
with.
There
is
a
discussion
whether
we
would
want
to
switch
everything
to
base
36
or
just
use
base
36
in
basically
in
the
in
the
URL,
but
yeah
working
practice
on.