►
From YouTube: 2018-11-05 libp2p meeting
Description
Team meeting recording
A
Computer,
okay
and
reporting
is
has
started.
Okay,
welcome
to
the
lid
p2p
community
meeting
for
November,
5
2018
and
the
agenda
is
posted
in
the
chat.
It's
the
same
as
usual.
We're
just
going
to
go
through
and
start
by
talk
everyone
talking
about
what
they're
working
on
and
then
I
or
role
can
give
an
update
about
web
3
and
DEFCON
4.
A
There's
a
lot
of
little
p2p
interest
like
way
more
than
I
expected,
and
then,
if
you
have
a
topic
like
feel
free
to
add
it
to
the
agenda,
something
you
want
a
demo
question
anything
yeah.
Okay,
let's
do
it
and
if
anyone
wants
to
take
notes,
are
the
awesome,
okay,
yeah?
So
let's
go
Pierre.
What's
give
us
a
quick
update?
What's
going
on?
Okay,.
B
So
I've
been
to
web
freestyle
meets
and
that
I'm
just
like
you,
so,
yes,
you
were
going
to
give
an
update
to
me
what
happened
there,
although
my
thoughts
on
the
rest
side,
just
like
things,
are
moving
as
usual:
I
don't
have
any
precise,
any
specific
thing
to
say:
I
implemented
mdns.
So
so,
like
that's
the
thing
that
comes
to
my
mind,
yeah,
but
Syd
I
was
much
too
busy
with
the
conferences
I
just.
A
Have
one
quick
question
for
you:
a
couple
of
people
were
asking
me
about
the
new
API
stuff.
Like
basically
sorry
I,
don't
I,
don't
recall
if
top
my
head,
who
who
this
was,
but
it
was
like
they
wanted
to
contribute
something
to
Ross,
but
it
was
blocked
because
you
know
this
API
changes
stuff.
What's
the
status
of
that.
B
So,
but
also
people
who
came
up
to
me
to
ask
the
same
questions
so
it's
probably
the
same
people
and
yeah
I,
just
so
I
explained
to
them,
but
so
first
of
all,
I
didn't
know
that's
people
over
we're
using
bursty
p2p.
So
we
were
not
careful
in
breaking
things
and
I
explained
it.
I
explained
what
it
was
really
necessary
and
I
believe
it
was
necessary
for
us
to
break
everything
and
in
your
next
couple
of
weeks
it
should
be
more
stable.
B
A
C
A
B
A
E
C
Me
just
let
me
just
interject
there
yeah,
that
I
think
that's
gonna
be
a
growing
growing
priority
for
us.
There
are
a
few
earth
2.0
implementers
that
we
found
in
the
conference
that
are
gonna,
be
using
likely
Jas,
let
p2p
so
by
q1.
People
are
aiming
to
have
at
least
a
little
hint
of
a
testbed
running,
which
means
that
there
might
be
pressure
from
two
different
clients
using
goal
and
B
to
be
rusted
B
to
be
even
jenia.
A
G
Yeah,
although
I
have
been
focused
mainly
19s
over
purpose
of,
inter
up
in
the
DHT
in
the
Ibanez
over
pub/sub
pr's,
already
in
jail
and
and
then
I
was
working
in
the
interrupts
and
I
found
a
problem
with
the
topic.
I've
been
discussing
this
with
the
Steven
and
more
recently
visual.
So
it
is
ok,
nyan
about
it.
Basically,
we
left
to
change
the
topic
as
it
was
in
going
way
that
we
couldn't
talk
in
GSI.
G
So
we
need
to
decide
basically
how
we
should
go
and
make
the
changes
both
in
go
and
the
Jes
to
make
it
possible.
Then,
in
the
DHT
side,
I
got
the
HD
interrupt
test
screen
and
I'm
working
in
some
stress
tests,
which
are
will
be
like
an
initial
step
for
me
for
the
okiya
of
the
testbed
in
the
DHT
as
well,
and
so
this
week,
I
will
continue.
G
A
H
Things
are
going
well,
I've
recently
been
working
on
the
both
JavaScript
and
go
implementations
of
disjoint
paths
with
up
in
the
DHT.
That's
mostly
done
as
far
as
I
can
tell
it
works
I'm.
Just
writing
some
tests
right
now,
so
I
should
have
to
pull
requests
when
J
s
and
when
and
go
probably
tomorrow
or
so
for.
People
who
look
at
I'm
also
Henry
sent
me
a
really
interesting
looking
paper
on
Byzantine
fault-tolerant
GHD,
so
I'm
going
to
be
reading
that
as
well.
Oh.
H
I
C
A
C
Yeah,
so
so
yeah!
Okay,
so
let
me
let
me
just
focus
on
dev
I
didn't
get
too
much
given
or
the
hustle
and
bustle
with
what
Def
Con.
But
this
week,
I'm
gonna
be
looking
at
gonna,
be
continuing
with
the
diner
thanks
a
lot
for
all
your
comments
on
the
data
version.
That
was
super
useful
and
helps
me
continue,
focusing
on
the
right
approach.
There
I'm
gonna
make
another
iteration
this
week.
That's
gonna,
be
the
core
and
then
of
my
work
this
week
and
I've
got
other
stuff.
C
Yeah
so
that
I
think
oh,
it's
gonna
be,
and
then
yeah
I've
got
a
bunch
of
blog
posts
that
I
need
to
write
in
a
bunch
of
diagrams
that
people
have
asked
for
regarding
architectural
overview,
sublet
p2p
and
how
the
different
contents
have
started.
There
was
also
a
lot
of
talk
regarding
gossip
table,
and
you
know
the
lack
of
understanding
of
many
aspects,
of
course,
its
hub.
We
need
to
get
better
at
explaining
those
and
there's
a
number
of
moving
parts.
C
So
there
was
a
that
was
a
request
for
an
architectural
diagram
of
that
as
well,
which
I
think
it's
going
to
facilitate
people
pick
up
the
concepts
there,
so
I'm
gonna
be
working
on
that
along
alongside
be
so
hopefully
already.
We've
already
been
chatting
on
an
issue
and
and
yeah
basically
catching
up
on
my
backlog
after
DEFCON.
J
Playing
with
it
right
anyway,
one
thing
have
been
working
on
did
work
on
was
visit
caring
for
an
extra
month.
It
supports
the
demon,
small
issues
and
now
we're
gonna,
merge,
total
relay
so
I
wanted
to
use
knowing
pointed
out
here
the
meeting,
because
it's
been
the
hacking,
a
to
repeat
of
things,
has
been
content
to
so
anyway.
We've
been
kept
in
conversation
on
how
to
proceed
with
the
whole
service.
J
Techs
or
I've
been
thinking
a
lot
about
it,
but
I
don't
have
anything
actionable
get
on
that,
so
maybe
you
should
speaking
of
which
I
think
it's
one
of
those
issues
that
we
might
wanna
schedule
is
small
jump
me
tap
to
discuss
it.
Think
of
its
thing.
There's
I
think
you
know
like
you,
can
look
like
one
a
real
consensus
about
it.
I
don't
know
how
people
feel
about
it.
We
can
do
it.
Oh
and
I
receipt
or
or
on
the
issue,
but
who
might
wanna
weeks
has
put
same?
J
Thank
you
we're
gonna
do
some
singer
is
time,
so,
basically
that's
about
it
I'm
going
to
murder
the
relay.
That's
that's
one
thing:
I
want
to
do
waiting
from
the
meetings
so
that
we
will
know
their
times
raise
objections.
Then
I'm
going
to
work
on
the
you
know.
Ik
meriting
is
into
ipfs
to
use
it
for
the
next
release.
So
I've
been
waiting
on
that
because
goes
oh
for
18
release
and
we
don't
even
wanna
sneak
in
other
Elaine.
So
that's
why
you'll
be
waiting
with
that?
Okay,.
A
J
Really
I
mean
right
now
the
way
it's
written,
it's
the
only
way
it
could
be
written
with
a
really
lacking,
knowing
the
concept
on
every
so
I
hate
do
some.
You
know
like
there
are
there.
Gremlins
are
hiding
behind
the
public
interface
they're,
really
so
the
internal
interface,
so
I
didn't
feel
too
bad
about
it,
but
no
like
it's
not
so
bad
in
grades,
so
you
see
spidey
senses
think
now
these
two
pi
K
I
won't
be
able
to
maintain
it.
J
Everything's
been
discussing
about
your
sexual,
it's
an
open
issue
about
introducing
the
concept
of
services
deep
inside
their
Libby
to
be
cost
because
we
need
it
and
every
design
for
relate
pretty
much
ready
ones.
Who
start
will
be
the
first
test
of
the
new
architecture?
The
refactoring
needed
to
do
you
know
way
for
the
relay
is
really
nice
aesthetically.
I
It
you
know
someone's
going
because,
like
the
constructors
are
all
looking
people
to
use
them,
but
as
long
as
we
like
have
a
resolution
to
fix
it
ASAP
and
we
actually
do
it
instead
of
putting
in
the
bathrooms
like
well
most
big
city,
but
one
more
feature
first
and
one
more
feature.
First,
let
me
end
up
with
the
monolith,
which
is
what
we
currently
have
I.
Think
because.
J
We're
going
to
fish
it
next,
so
you
know
like
I'm
putting
it
on
my
plate.
That's
the
next
thing
to
work
on
West
I
feel
kind
of
things
we
might
wanna
have
a
thing
with
because
there
seems
to
be
okay
were
discussing
about
dependency
injection
framework,
and
we
haven't
figured
out
exactly.
However,
when
I
use
it
first
I
think
when.
J
J
A
J
J
A
J
Disagree,
there's
agreement
good,
save
and
we're
both
like
yeah,
we
said
words
didn't
fix
it
later.
I
mean
the
only
the
only
thing
I
do
when
I
mess
it
first,
because
then
we
have
actual
code
to
refactor
are
supposed
to
writing
and
thinking
in
the
abstract
and
building
the
interface.
A
very
much
rather
give
life
codes.
In
fact,
it's
going
to
be
a
test
of
3d
texture,
and
you
know
it's
kind
of
weird,
because
if
we,
if
we
can't
implement
it
real
a
properly,
then
why
would
in
the
refactor
that's
going
to
travel
yeah.
K
Yeah
I
guess
mine
will
kind
of
dovetail
into
my
agenda
item
next
I've
just
been
like
poking
away
at
Damons
stuff
and
doing
a
lot
of
code
review
for
all
the
stuff
that
these
has
been
working
on
as
well.
I
need
to
do
a
more
in-depth
code
review
of
the
dialer
role
and,
if
there's
still
time,
I
can
do
that
today.
So
if
not
I'm
happy
to
wait
for
the
next
version
and
then
yeah
kicked
off
stuffed
with
Jacob
for
the
test.
Bed
I
actually
got
some
really
good
feedback
on
the
go
core
dev
sink.
A
Okay,
cool,
okay,
so
the
next
that
so
here's
what
we
got
on
the
agenda
web
three
and
DEFCON
four
update
the
testbed
stuff-
maybe
maybe
you've
already
covered
everything
you
want
to
say,
call
it
I,
don't
know,
and
then
Aiden
put
in
a
topic.
Is
this:
how
I
should
be
looking
for
lives
to
make
it
easier
to
utilize
I
feel
D,
compatible
graph
data
structures?
Add
new
command
services
to
the
demon,
also
as
stats
of
grassing?
A
A
So
Rana
are
gonna,
provide
like
a
more
detailed
update
as
they
get
a
zigg
like
a
github
issue
or
PR
somewhere
and
I'll
share
it
on
the
mailing
list
afterwards.
But
I
would
summarize
the
event
like
at
a
really
high
level,
I
would
say
between
web
3
and
DEF
CON.
There's
a
tremendous
amount
of
interest
in
lit
p2p
from
like
the
blockchain
world.
A
lot
of
these.
A
Basically,
what
I'm,
seeing
over
and
over
again
is
that
block
chains
want
a
peer-to-peer
foundational
layer
and
they
don't
want
to
feel
like
they
want
it
to
sort
of
just
work.
They
don't
want
to
have
to
understand
all
the
intricacies
and
subtleties
of
you
know
to
dhts,
and
how
do
you
prevent
attacks
like?
How
do
you
prevent
civil
attacks?
Should
we
have
a
proof
of
work
to
create
identities
like
they
just
want
us
to
figure
all
that
out
and
have
it
work?
A
And
so
those
were
the
kind
of
conversations
that
that
we
had
yeah
the
biggest
single
biggest
user
issue
that
I
see
coming
up
at
both
of
those
events
is
lack
of
documentation
like
it's
just
really
confusing
for
people
to
get
started.
Not
they
all
go
and
start
reading
the
code,
but
we
don't
provide
them
any
high-level
guidance
of
like
what.
A
What
are
the
abstractions
like
diagrams,
so
I
think
that's
where
we
should
work
work
on
some
creating
some
content
and
I
think
rolls
gonna
try
to
create
like
a
series
of
maybe
blog
posts
or
other
documents.
He
did
an
awesome
job
of
this
kind
of
like
life
life
of
a
peer.
It's
similar.
It's
is
kind
of
inspired
by
the
talk
you
gave
Jacob
once
about
like
the
life
cycle
of
a
Lippe
to
be
connections,
so
role
did
kind
of
like
a
walkthrough
of
like
a
like,
hey
I'm
up
here.
A
I
have
pure
ID
now
I
want
to
discover
other
peers.
Okay,
now
I've
discovered
other
peers.
Now
you
know
I
want
to
publish
on
hopes
up
and
you
just
walking
through
the
different
abstractions
and
I.
Think
turning
that
into
text
and
diagrams
is
not
going
to
be
too
hard
and
I.
Think
it'll
be
super
high
leverage
for
the
project,
so
that
was
more
than
2
minutes.
That's
my
summary.
C
Yeah
I
think
I,
just
wanna,
that
was
great
Mike
thanks
I
just
tried
to
add
a
couple
more
things
here.
One
of
the
focus
was
to
rope
people
into
the
community
so
rather
than
have
have
us
build
stuff
for
blockchain
communities
and
in
general,
decentralized
infrastructure
and
so
on.
We
really
want
to
enroll
people
in
so
we
are
onboard.
The
investor
team
d
thean
foundation
is
in
the
process
of
finally
committing
to
select
a
Libby
to
be
grantee.
C
They
open
up
call
for
grants
for
proposals
and
a
team
that
we
interacted
with
a
lot
during
the
during
the
conference
made
a
proposal,
and
they
are
a
group
of
students.
They're
super
talented
they've
already
made
progress.
Are
we
going
to
be
bringing
this
red
plane
to
the
Lippett
of
the
organization
as
well?
There's
a
Java
Libby
to
be
implementation
likely
starting
as
well.
At
some
point,
there
are
two
actors
that
are
interested
in
contributing
to
that.
C
So
so
that's
that's
good
and
yeah
definitely
documentation
from
the
from
the
point
of
view
of
like
yeah
people
can
read
code.
That's
great
but
like
have
them
get
this
mental
picture
of
the
entire
system.
That's
the
challenging
part.
So
that's
what
personally
I'm
gonna
be
working
on
in
the
next
weeks
and
yeah
I
think
we
need
to.
We
definitely
uncover
the
fact
that
we
need
to
start
thinking
about
community
governance
as
well.
C
So
it's
great
that
you
here
Michael
because
as
we
bring
in
or
I
don't
know,
if
you
still
here
yeah
as
we
bring
in
new
people
into
the
into
the
ecosystem,
you
know
new
discussions
will
start
popping
up
and
like
arriving
to
consensus
about
stuff
and
keeping
specially
like
so
many
implementations
can
sink
and
lockstep
with
one
another
and
interoperating.
That's
that's
gonna
be
challenging
right,
so
we
definitely
need
to
start
start
thinking
about
about
these
aspects
and
and
yeah.
C
I
C
I
At
the
end
of
the
day,
we'll
probably
end
up
moving
more
towards
the
daemon
as
default
for
everything
you
can
go
because
simply
because
it's
named
ESP,
you
don't
want
to
have
a
million
connections
open
for
each
implementation.
So
I
guess
we're
mostly
AB
standpoint.
You
really
want
one
daemon,
but
also
from
a
like
a
integration
standpoint.
I
really
think
we
can.
I
G
C
Yeah
yeah
in
general,
in
general,
I
think
I,
think
I,
think
that
makes
sense,
I
think
as
we
as
we
move
towards
the
multi-tenancy
approaches
well
for
the
demon
and
will
start
making
more
sense.
Where
you
have.
You
know
a
single
demon
running
for
as
many
applications
as
you
might
have
in
them
in
that
particular
machine,
but
in
general
the
the
community
is
tending
towards
building
native
implementations
in
each
in
each
language,
with
their
native
constructs
and
so
on.
But
yes,
I
see
what
your
point
is
and
also
there
was.
C
There
was
some
discussion
about
whether
it
would
be
feasible
at
all
to
implement
a
canonical
version
of
Lapita
being
wasn't
as
well
at
that
composite
on
to
Wasson,
so
that
it
then
any
application
that
that
is
written
in
any
language
that
can
also
compile
down
to
once
and
will
be
able
to
use
it
I
think
in
practicality.
There
are
a
number
of
challenges
with
with
dealing
with
sockets
and
so
on
and
wasn't
especially
because
they
are
not
as
far
as
far
as
I
understand
it.
It's
not,
it
is
not.
C
A
K
I
guess
as
I
thought
about
the
daemon
stuff,
but
that
can,
but
if
I
can
happily
jump
into
the
testbed
thing,
so
one
I
would
just
like
to
invite
anybody.
Who's
interested
in
panco
I
think
you
probably
qualify
for
this
to
join
Jacob
and
I
on
a
call
tomorrow.
If
it's
convenient,
it's
going
to
be
1
p.m.
Eastern
Time,
which
is
actually
now
the
daylight
savings
time
just
happens.
K
I
have
no
idea
what
time
it
is
UTC,
but
it's
a
time
I
can
invite
anybody
to
the
event,
and
you
can
just
stick
fine
or
accept
or
whatever,
but
yeah
we're
gonna
be
talking
about
kind
of
practical.
First
steps
to
the
testbed
I
spent
Friday
poking
around
kubernetes
ipfs
I
got
a
lot
of
feedback.
That
IPTV
is
really
where
we
should
be
looking
for
this.
That
kubernetes
ipfs
is
kind
of
dead.
K
I
think
what
drew
us
to
the
kubernetes
project
was
that
it
has
like,
like
integrated
support
for
testing,
but
one
of
the
things
that
Jacob
and
I
were
talking
about
last
week
was
that
we
were
probably
gonna
have
to
rewrite
the
testing
side
of
it
anyway.
So
it
seems
like
IPTV
might
be
just
the
best
candidate
for
now
but
yeah.
If
John
I
got
your
message,
so
I'll
definitely
invite
you
to
that
afterwards,
but
yeah.
K
A
A
J
J
So
so
the
other
part
I
wanted
to
say
is
that
link
for
the
demo.
I
mean
the
the
application
can
size,
the
fastest
lucidum
only
the
bindings
for
the
language.
In
fact,
apart
mm
very
effectively,
so
the
API
we
did
in
the
go
silent
basic
human
actions.
Let
me
just
be
the
client
object,
so
I
think
we
definitely
I.
J
A
C
A
C
J
Okay,
yeah,
that's
fine,
but
all
I'm
saying
is
that
the
default
implementation
and
the
first
implementation
on
the
agreement
was
seven
here
that
the
first
implementation
here
I
should
be
used
a
demon
and
that
second
complementation
for
building
applications,
and
they
need
to
you
decide
that
demon
doesn't
give
you
the
native
performance
that
you
want
out
of
their
applications,
which
we
have
already
verified.
Then
you're
free
to
do
is
knowing
your
native
landing
yeah.
C
I
completely
agree,
and
maybe
we
can-
we
can
reinforce
thinking
about
evaluating
or
implementing
native
implementations,
to
relook
at
the
demon
and
to
you
know.
But
maybe
this
is
just
a
suspicion
on
there
and
that
you
know
they
prefer
to
have
a
native
implementation
in
whatever
language
and
the
demon
might
perform
better
for
them,
especially
especially
for
Python
right.
So,
but
at
the
end
of
the
day,
Python
is
not
really
in
a
theorem
is
not
really
a
a
client
that
is
used
in
production.
It
is
used
for
prototyping
and
especially
for
the
research
team.
C
C
A
On
just
one
second
I
want
to
respect
people's
time.
So
look
I
want
to
end
this
meeting.
I
don't
want
to
end
the
discussion,
I'm
gonna
schedule
to
follow
up
meetings,
as
you
can
see
in
the
notes
and
we'll
have
ample
time
to
keep
going
on.
This
I
mean
yeah,
but
I
think
we
should
try
to
respect
the
time
boundaries
for
the
meeting.
So,
let's,
let's
call
it
a
day.
Okay,.