►
From YouTube: libp2p Call - Q4 Roadmap
Description
The libp2p call from October 10th, discussing the Q4 Roadmaps. See https://github.com/ipfs/pm/issues/217.
A
B
You
I
just
killed
my
brother,
so
I
need
to
open
my
time.
Okay,
all
right.
So
there
is
everyone
as
the
other
path
open
all
right,
so
only
peer
to
peer.
So
we
peer
to
peer
he's
always
in
this
call
to
talk
about
gold
peer
to
peer
and
Jesse
peer-to-peer,
mostly
the
energy
that
has
been
put
in,
we
p2p
as
being
like
to
fulfill
the
needs
of
a
PFS.
So
like
everything
that
we
like
everything,
is
recessed
on
the
roadmap
for
which
it
appears.
B
I
was
like
what
GPFS
or
ago
I
see
past
needs,
however,
especially
with
recent
events
from
the
last
couple
of
months.
Maybe
more
one
more
people
are
interested
in
using
repeaters
here
and
right
now,
it's
pretty
exclusive,
like
it
is
really
hard
to
get
into,
because
there
is
no
entry
point.
There
is
no
documentation
like.
Basically,
there
is
no
work
done
in
developer
UX.
Of
course,
there
is
a
lot
of
features
that
need
to
be
built
and
like
things
that,
like
pub/sub,
are
on.
B
We
feel
free
of
scope,
and
we
already
talked
about
my
dad's
routing
the
DHT
and
also
know
peteyboo
scope,
but
go
real
talk
about
that.
There
is
things
like,
including
connectivity,
which
we'll
cover
in
a
second
but
yeah,
like
one
of
the
things
that
we
need
to
step
up
our
game
for
this
quarter
so
that
we
get
like
we
get.
B
We
need
to
write
better
tutorials
right
now,
so
if
it
appears
back
it
does
something
that
hasn't
been
done
anywhere
else.
But
sometimes
writing
aspect
can
like
hoping
at
this.
Back
is
the
glue
that
you
points
for
people
to
learn
is
not
a
good
idea.
What
we
really
need
is
like
use
cases
or
even
a
little
tools
will
appear
appear
that
people
can
use
after
they
can
seek.
I
can
check
out
and
like
they
can
see
the
power
I'm.
B
A
strong
believer
of
screencasts
I,
just
like
two
minute:
videos
explaining
hey
I,
just
connected
to
know
the
four
behind
Nets
and
now
my
kids
are
all
dedicated
like
a
lot
more
strings
like
it
works
and
that's
very
powerful
to
communicate
out.
These
things
are
useful.
One
of
the
things
that
we
don't
text
so
far
is
the
interface
with
peer-to-peer,
which
makes
more
sense
than
ever
in
in
JJ's
world.
We
already
have
a
pseudo
interface
defined
because
we
asked
so
like.
B
We
have
two
implementations
of
bleep,
it
appear
javascript
and
go,
but
we
actually
have
3d
builds
of
whippet
appear.
Don't
wanna
choose
by
ipfs
enjoy
prefers
the
one
that
choose
by
professing
I
pee
pads
for
no
jails
and
the
one
that's
used
by
ffs
in
applicants
for
the
browser.
So
there
are
three
builds,
and
this
idea
for
the
browser
note
follow
that
same
interface,
so
that
we
can
change
between
node
and
the
browser
without
changing
the
code
base.
The
discovery
mechanism
primitives
are
the
same.
The
dialing
primitives
are
the
same.
B
Routing
primitives
are
saying
all
things,
I
would
say,
but
it
is
just
like
made
using
rule
of
thumb.
Like
just
trying
it
out
and
say,
oh,
this
looks
the
same,
doesn't
break
it
works.
We
don't
have
a
battery
of
tests
to
prove
that
B
is
a
p2p
organization,
actually
follow
the
same
interface
as
we
do
for
all
of
the
other
modules
so
yeah.
B
We
need
to
get
an
interface
which
will
appear
something
that
people
can
check,
how
it
works
enough
to
use
it
and
actually
can
have
more
builds
to
fulfill
what
the
picture
specific
foods
cool
so
yeah.
A
lot
of
these
will
be
discussed
this
week
and
like
again,
following
the
same
jets,
IP
first
proposal.
We
need
to
come
up
with
like
subtasks,
for
these
endeavors
and
like
be
very
comprehensive,
so
that
we
understand
the
amount
of
effort
and
energy
that
has
to
be
put
onto
these
things.
B
See
any
hands:
okay
next
up
website,
so
richard
has
been
working
a
lot
on
the
way
peer-to-peer
website,
and
this
is
going
to
be
very
important
for
developer
experience
as
well.
It
is
again
it's
going
to
be
the
TLDR
of
the
spec,
the
the
things
that
you
need
to
know
the
features
the
where
you
can
get
example
where
you
can
get
the
tutorials
right
now.
I
know
rich
I'm.
B
And
with
the
website
will
also
come
the
questions
of
making
sure
that
we
have
the
right
examples:
the
right
interfaces,
the
right
language
to
explain
the
repute
appear
and
so
on.
Cool
next
type
is
peer,
routing
and
content.
Routing
has
like
we
mentioned
it's
my
profession,
recipe
fest,
so
I,
clearly
zdt
getting
a
way
to
find
content
on
the
network
job
having
to
dial
directly
to
the
notes,
I
believe
this
was
in
out
like
sufficiently
covered.
Would
someone
like
to
have
anything
on
this?
B
C
So
there
is
a
number
of
quick
implementations.
At
least
one
one
is
live,
quick,
which
is
somebody's
attempted
extracting
quick
from
chromium.
This
is
not
yet
the
supported
a
supported
version,
though
the
quick
team
has
made
some
directions
in
that
route,
but
there's
a
standalone
go
implementation
of
quick
written
by
Lucas
Clemente.
He
is
presenting
it
at
dot.
C
Go
I,
think
this
week,
so
there's
only
the
server
part
of
it
at
the
moment,
but
I
actually
reach
out
to
him
and
talk
to
him
about
writing
the
client
part,
and
so
we'll
be
working
with
him
on
that,
and
he
is
also
going
to
write
it
into
a
loopy
transport
and
then
we're
also
and
that
feeds
into
the
XTP
work
as
well.
That
I
can
talk
about
in
a
moment,
but
so
quick
is
a
very
real
possibility
within
the
next.
C
You
know
few
months
potentially
thank
you.
This
quarter
or
the
following
at
least
for
go
for
JavaScript.
Getting
access
to
quick
will
be
harder,
though
the
chrome
team
is
working
right
now
on
putting
WebSockets
on
top
of
quick.
So
there's
no
head
of
line
blocking
so
once
that
that
is
there,
then
you
can
get
WebSockets,
naturally
using
quick
in
the
browser
in
node.
It
will
probably
take
so
longer
time.
I,
don't
know
what
the
story
quality
is
there,
but
I
think
we
can
probably
see
quick
as
a
transport
route
solution.
There's.
B
C
B
C
So
although
this
transfer
work
requires-
or
we
think
requires
this
too-
all
that
I've
been
calling
xtv
for
now
XD
peace,
external
transport.
The
point
of
XP
is
to
isolate
a
transport
in
a
different
process,
and
this
is
to
get
around
a
couple
of
problems.
One
is
that
you
accidentally
ship
a
an
implementation
that
is
very
young
and
has
a
ton
of
problems,
and
then
your
network
ties.
C
This
happened
to
us
with
a
UTP
implantation
and
another
problem
is
the
implementation
may
not
be
in
a
language
or
distribution
packaging
of
your
choice,
meaning
in
my
B
and
C,
and
you
don't
want
to
use
C
go,
it
might
be
in
C
and
you
might
be
in
the
browser
like
there's
a
bunch
of
thanks
right
so
to
get
around
that
problem.
Xep
isolates
some
transport
limitation
in
the
process
and
then
allows
you
to
connect
to
that
to
some
server
and
like
control
the
transport
from
a
from
an
enterprise
perspective.
C
There
is
an
interface
being
written
about
this
now
and
then,
once
the
interfaces
is
once
I'm
pretty
happy
with
it,
then
we
can
think
about
like
actually
implementing
it.
So
this
could
happen.
This
quarter
could
happen.
Exporter
I
would
think
this
is
like
a
nice
to
have
I,
wouldn't
think
that
unless
we
decide
that
hey,
this
is
absolutely
super
important,
let's
just
ship
it
which
I'll
be
very
happy
to
to
get
to,
but
it
is
something
kind
of
on
the
fringe.
It
is
required
for
us
to
be
able
to
go
into
all
these
other
transports.
B
B
B
We
keep
trying
it
around,
but
it
really
means
like
just
making
sure
that's
like
ipfs
can
hire
in
situations
that
are
kind
of
I'm
not
used
to
deal
with,
and
this
comes
from
like
doing
stuff
like
relay
packet,
switching
better
than
traversal,
even
I,
just
making
sure
that
we
humble
connections
properly,
because
we
kind
of
like
keep
a
lot
of
connections
open
for
a
long
time
and
apparently,
as
Jeremy
says,
like
it's
impossible
to
kill
connections,
because
it
just
come
back
alive.
B
B
It
was
known
as
a
mouse
on
Oh
bleep.
It
appear,
which
was
on
that's,
really
correct,
and
we
have
to
rethink
that
and
then
comes
the
network
lab.
So
the
network
lab
is
a
name
for
a
thing
that
we
have
been
talking
about.
For
many
months-
and
it
it's
just
like
having
hardware
a
test
hardware
to
test
in
network
connectivity
like
a
bunch
of
tiny
computers,
raspberry
PI's,
5/64,
whatever
and
routers
that
we
can
configure
and
I
can
have
test
scripts
to
run
over
this
setup,
and
we
can
act
on
those
situations.
B
B
People
still
do
a
lot
of
these
tests
manually
or
just
like
setting
up
stuff
manually
and
like
trying
to
infer
what's
going
on.
If
we
get
like
this
network
lab
working
in
a
wet
weekend
like
automate
it
and
repeat
it
often
like
setup
different
coffee
questions
very
quickly,
then
we
can
use
like
leave
it
running
and
then,
like
almost
all
of
these
tests
for
us
and
for
other
people,
I
do
know
that
Jeremy
has
been
clinging
or
about
it,
and
I
was
already
a
bunch
of
routers
at
his
place.
B
That
he's
like
working
towards
getting
it.
I
also
have
a
bunch
of
brothers
in
my
place
and
a
bunch
of
like
raspberry
PI's
that
I
want
to
use
for
this,
which
means
that
we
can
have
even
like
two
natural
collabs
across
continents,
which
might
be
fun
as
all
and
again
like.
We
don't
have
a
comprehensive
list
yet
of
things
that
I
have
to
happen
for
this,
because
it's
kind
of
like
a
huge
endeavor
and
you
generate
and
they're
taking,
but
something
that's.
B
D
B
That's
a
good
point,
probably
should
be
one
of
the
items
on
that
killer.
Issue
template
on
go
ideas
if
people
are
able
to
give
their
router
description
and
model,
this
yeah
yeah
all
right.
So
ten
minutes,
okay,
good
sometime,
the
next
one
is
ECC
key-
is
in
JavaScript,
so
now
I
go
support,
CC
keys,
but
not
on
peer
IDs
and
pyramids
actually
I,
just
like
I
studies
recently
and
I
do
have
the
chance
to
reviewing
it
all.
B
E
G
B
B
C
You
should
bad
that
I
think
it
would
be
valuable
to
like
that.
Crypto
stuff
is
really
it's,
not
the
I
profess
to
crypto,
it's
the
the
p2p
crypt
I,
think
and
that
could
be
a
module
on
its
own,
like
if
people
could
like
selectively
taken.
However,
I,
don't
know
how
you
do
the
keys,
because
multi
key
will
matter,
and
so
the
kids
will
have
to
be
real
world.
So.
B
What
we,
what
I
was
just
thinking
is
it
might
be
useful
to
have
like
IP
assault,
sleep
appeal
of
crypto
exposed,
but
to
the
ipfs
distance,
so
that
people
don't
have
to
require
again
and
I
have
to
go
through,
like
the
mass
of
like
having
different
PP
crypto
versions
and
I
was
like
bundling
several
times
the
same
thing,
although
like
now,
a
little
bit
about
boy
is
getting
way
smaller.
With
all
the
changes
to
things
with
you
and
Freda
was
something
say,
go
ahead,
feel
yeah.
F
They'd
be
just
about
the
PTP
Krypton
module.
Actually,
it's
it's
getting
smaller
and
larger.
At
the
same
time,
it's
gonna
small
in
size,
but
much
larger
functionality
at
the
moment,
because
what
I
did
in
j/s
with
the
latest
changes
is
that
I
actually
pulled
all
crypto
methods
effectively
that
I
used
throughout
the
PTP
into
the
pit
in
crypto.
F
H
B
I
And
then
it'll
come
up
for
you,
Stream
events
to
the
client
and
maybe
some
basic
stats.
So
the
states
the
sensor
propagation
time
right
now,
but
give
them
a
little
bit
usage
on
it.
For
people
who
are
going
to
develop
some
strategies,
that's
gonna
kind
of
inform
the
rest
of
it
so
feel
free
to
I
got
the
repo
and
you
guys
can
wear
out
next.
B
Okay,
definitely
we'll
check
it
and
and
work
on
getting
like
the
foot
on
top
of
these
tests
net.
We
probably
can
work
a
lot
into
making
all
these
who
are
agnostic,
like
independent
of
what
shall
we
pick.
You
know
they're
just
happy
bastard,
because
after
all,
the
primitives
for
dying
are
the
same,
because
it's
the
same
thing
that
filing
and
then
we
can
use
this
after
to
spawn
alike,
just
be
personnel.
B
B
Any
I
would
would
you
like
hair
trial?
Would
you
have
anything
if
you
are
planning
to
do
now
from
now
or
I
know
that
we
are
walking
you
by
not
having
for
what
we
will
be
covering
it
with
you
and
do
the
integration,
but
is
there
anything
that
we
can
provide
right
now
terms
of
information,
communication.
K
Know
tonight
can
I
talk
about
what
I'd
like
to
do
with
this
all
right
sure.
Please
do
so.
Hi
I'm,
Thomas
I
come
to
you
from
the
land
of
aetherium.
I
have
a
chrome
extension
called
meta
mask
and
they
injects
the
etherium
application
standard
into
your
browser,
so
it
has
about
3,000,
installs
I,
don't
know
how
many
concurrent
users
we
have
but
I'd
like
to
move
to
an
ethereal
light
client
built
on
the
p2p
sometime
in
the
near
the
medium
term.
In
the
meantime,
it
would
be
really
cool
to
take
some
of
these.
K
These
tests
I
know
this
PSTN
libo
existed,
and
this
is
fantastic.
If
we
could
take
this,
make
it
run
on
WebRTC
and
then
just
put
it
in
meta
mask
and
kind
of
hide
it.
In
the
background
we
could
do
long-running
Network
tests
and
see
how
the
network
does
on
periods
of
time
and
I
would
love
to
do
this.
So
tell
me
now,
as
you've
said,
there's
it's
it's.
A
Debu.
X
is
a
little
wonky
at
this
point.
K
J
B
So,
thank
you,
that's
exciting.
Let's
make
it
happen.
So
as
as
we
as
we
discussed
doing
this
cause
like,
we
definitely
want
to
beef
up
our
documentation,
but
I
think
we
can
even
like
start
working
on
it
and
like
like
get
something
working
and
use
that
as
an
example,
so
that
we
can
like
cooperate
in
the
meantime.
While
we
don't
have
like
this,
like
really
beautiful,
Docs
well
know
where
Mark
is
yeah.
B
I
come
on
the
web
RTC
part,
so
right
now,
I
deposit
test
that
really
uses
like
runs
everything
in
node.js
for
them
for
enabling
it
with
the
web
RTC
it
mean
to
go.
You
just
have
like
to
add
the
web.
Rtc
start
a
sport
and
on
the
nodes
in
chrome
instead
of
nodes,
I
dunno
they're
like
like
from
bottle
experience
or
scars,
live
like
spawning
3000
nodes
in
browsers.
It's
not
the
same
thing.
You
know.
One
machine
is
not
the
same
thing.
B
I
was
like
note
so
like
we
need,
we
can
actually
work,
but,
like
probably,
we
can
like
reduce
the
size
of
the
tests
for
the
folks.
I
first
met
and
then
I
just
get
that
working.
So
that
we
can
deploy
it
on
meta
mats,
can
I
see
the
notes
coming
from
several
machines
and
connecting
to
this
won't
work,
hopefully
without
being
enemies.
C
B
Doesn't
have
what
word
you
see,
but
we
can
then
we
can
spawn
one
chrome
instance,
and
you
know
laptops,
but
when
you
over
over
the
browser
with
a
lot
of
artistic
channels,
all
linking
because,
like
again
what
Jafar
will
try
to
convince
the
browser
to
try
out
all
the
other
notes,
I
creating
a
huge
mesh
and
nano
if
we
start
like
propagating
messages
over
those
channels,
it's
like
the
web
part
is
explored
all
over
again.
Just
like
kills
the
memory
of
the
browser
so
yeah.
B
C
Should
make
sure
that
the
that
is
not
really
connecting
to
all
nodes
so
before,
like
canvas
puts
it
into
mathematical,
and
we
should
make
sure
that
the
topology
forming
is
such
that,
like
you,
don't
end
up
with
a
one
or
a
huge
like
super
connected
mesh
that
it
it
like,
keeps
a
degree
small
right
like
that.
Ade
sense
so
then
connect
to
like
say
ten
nodes
each
and
then
it
may
be
like
cycles
them.
So
this
is
a.
C
People
are
running
that,
so
it
needs
to
be
like
a
very
small
resource
constrained
thing
right
only.
You
said
only
to
know
that,
like
the
thing
works
with
like
thousands
of
nodes
and
run
very
little
like
message
on
it,
because
these
are
people
like
you're
gonna,
have
these
brothers
running,
you
can't
just
like
peg
their
network
or
prints
of
you.
A
B
Yeah,
let's
make
sure
that,
like
that
is
one
of
the
items
of
this
week
and
if
we
come
up
with
the
thing
that
we
need
to
do
and
we
kept
working
away
to
this
week
or
earlier
next
week
to
make
sure
that
we
booked
all
the
exhibits
a
bit
cool.
Thank
you.
This
is
awesome,
any
other
thing,
no
I
guess
we
had
our
for
me.
It's
over
this
was
announced
meeting
so
like
yeah
think
you
have
money
for
not
getting
so.