►
From YouTube: IPFS All Hands 🙌🏽📞 Jul 23, 2018
Description
ESIP meetups, Peertunnel publishing, Nodetrust certificates, Hubot as a plugin, and native protocol handling in Firefox; all this week on the IPFS all hands call. 😎
A
B
Give
me
a
thumbs
up.
If
you
can
hear
me
I
had
phones
are
all
aware
it
awesome.
Thank
you
just
a
quick
report
out
last
week
for
most
of
the
week,
I
was
at
the
esup
summer
meeting
ISA
this
earth
science
information
partners
I
believe
they
pull
together
a
bunch
of
earth
science.
Data
managers
like
people
from
NOAA
people
from
NASA
great
folks,
I
worked
with
them
before
when
I
used
to
work
with
NOAA
I
proposed
an
unconference
session
there
and
distributed
web
stuff.
It
was
packed.
People
were
really
interested.
B
What
they
had
heard
about
distributed
web
stuff
was
a
bit
of
ipfs.
A
bit
of
DAT
in
general.
Debt
seem
more
approachable
to
folks
I'll,
be
writing
up
notes
about
all
this
and
I
will
post
it
in
a
public
place.
May
yell
no
but
yeah
some
interesting
things
there
and
lots
of
interest.
Also
now,
on
the
ease
of
slack
there's
a
distributed
web
channel
where
people
are
just
kind
of
what
is
going
on
here.
There
is
a
future
here.
We
don't
know
what
it
is,
we're
very
excited.
B
B
A
C
Wait
a
second
so
I
have
this
web
server
running
your
localhost
and
now
I
just
run
it
katana
tool,
which
then
connects
to
this
lip
PDP
server
and
then
opens
a
tunnel.
When
I
now
go
to
this
page,
it
simply
dies
to
peel
back
and
opens
a
tiny
relay,
and
now
this
page
is
published
yeah
as
I
said,
and
everybody
can
read
it
from
internet.
So
anybody
can
visit
this.
Your
end
right
now
and
also
it
looks
access
so.
C
B
C
C
So
the
second
demo
is
I
have
written
this
room
called
EPP,
no
trust
some
time
ago
that
issues
certificates
to
any
note
based
on
its
IP
address
and
I've
recently,
rewritten
the
whole
thing
in
micro-services,
so
it
can
be
managed
more
easily
and
wanted
to
demo
how
that
works,
and
so
this
is
the
demo
page.
This
basically
connects
to
the
issuing
sir,
but
it
also
runs
pops
up
where
all
no
trust
notes
announce
there
are
no
trust
addresses
and
connects
to
them.
C
So
here
are
four
tires.
The
top
left
is
the
DNS
heart.
Top
right
is
the
proof
server.
The
bottom
left
is
the
issuance
ever
and
the
bottom
right
is
the
server
or
the
client
run.
So
if
I
run,
the
client
first
connects
with
issued
server
then
connects
two
different
proof
servers,
for
example,
for
ipv4,
and
then
that
should
get
long
here.
That
would
require
two
proof.
C
C
E
C
I
would
also
like
to
get
this
deployed
and
in
the
protocol.
Let's
infrastructure
silent,
have
to
manage
this
and
maybe
get
it
into
J's
ipfs
because
would
be
really
useful
for
the
browser,
as
browsers
would
be
able
to
connect
to
any
here
instead
of
having
to
use
PDP
circuit
because
browsers
have
limited
TCP
connectivity
and
that
everyone
wants
to
run
in
jazz
at
the
festival,
web,
RTC
and
yeah.
C
G
The
actual
question
is
like
do
we?
Do
we
actually
need
it
like
we?
Thank
you
so
much
for
demoing
not
dressed
again
matching.
It
is
right
now,
just
I
can
clear,
because
we
haven't
done
the
actual.
Like
thinking
of
all
these
fits
together.
We
bleep
utopia,
collection
modules
and
like
what
users
expect
is.
Actually,
we
think
are
like
really
creates
beauty
issues
in
the
future,
and
so
there
are
all
these
questions
like.
G
It's
a
it's
a
different
step,
and
so
I
would
like
to
see
away
with
more
people
getting
involved
with
no
trust
reviewing
it.
Like
sharing,
sharing
their
thoughts,
probably
like
get
no
trust
to
be
like
to
have
a
proper
speken
RFC,
so
other
people
can
really
understand
what's
happening
underneath
and
I
quote
are
the
shortcomings
that
they
try
to
solve.
Otherwise,
we
we
might
just
like
keep
adding
something
that
we
don't
clearly
understand
the
implications
you.
G
F
G
G
Yes,
basically
like
we
should
have
this
discussion
like
any
people
are
super
interested
like
they
should,
or
at
least
we
can
self
appoint
themselves
right
now
or
later,
but
like
weren't
here
to
review
and
work
on
this
with
magic
in
math.
You
can
like
follow
up
on
the
issue
on
fleek.
It
appear
about
no
trust
and
defeating
the
group
here
to
see
if
everyone.
A
H
And
then,
if
you
want
to
see,
you
know,
works
exactly
like
the
way
ones
did
where
you
know.
I
can
ask
my
help.
I
can
ask
FS
help
and
then
I'm
gonna
grab
a
hand,
you
could
just
save
hash
and
weight
and
it
pins
your
hashes
for
you
and
that's
it.
But
that
means
that
if
we
make
updates
to
plugins
centrally,
you
know
it's
as
easy
as
a
little
NPM
update
inside
of
your
own
boo
box
to
get
the
latest
version,
which
just
makes
an
easier
thing
to
easier
to
maintain
I
think
that's
most
option.
I
Let's
try
to
lower
my
screen.
Okay,
let's
read.
I
Can
you
see
my
screen
yeah
cool,
alright,
so
I
created
a
Lib
D
web
branch
in
aquifers
companion,
our
companion
browser
extension
and
you
have
to
use
scripts
to
build
it.
The
sleepy
web
experiment,
you
just
run
lib,
do
a
build
and
then
you
need
to
download
Firefox
nightly,
make
sure
it's
on
your
path
and
then,
if
you
run
in
this
context,
you
just
start
Firefox
nightly
with
our
little
experiment,
and
hopefully
yeah
looks
like
it.
A
I
I
think
I've,
muted
myself,
when
I
was
like
dragging
this
sorry
about
that.
So
we
don't
have
many.
We
don't
have
many
hashes
to
test
so
I.
There's
this
issue
with
this
experiment
and
I
put
some
hashes.
So
if
you
copy
a
hash
and
put
it
in
the
address
bar,
it
will
load
natively
and
it
stays
there.
So
what's
happening
behind
the
scenes.
Is
that
there's
no
HTTP
anymore,
at
least
in
the
browser
context?
We
are
stringing
corresponds
from
ipfs
node
right
now.
I
It's
like
my
local
node,
which
I
connect
over
HTTP,
but
you
can
switch
to
ember
that
one
that
will
work
at
some
point.
What's
cool
about
this?
Is
that
not
only
images
work
but
also
like
regular?
Okay?
Maybe
if
I
don't
make
a
typo
yeah,
if
you
copy
without
a
typo,
you
will
see
there
is
like
a
regular
website,
loading
from
ipfs
protocol
and
not
only
that
we
also
have
directory
listening
back
again.
I
I
I
That's
the
current
state
of
things.
We
also
have
like
basic
Rohan
link.
So
if
you
paste
invalid
CID,
you
will
get
an
information
that
you
should
upgrade
to
base
42.
It
will
make
this
like
user
experience
better,
but
for
now
that
starts-
and
there
are
like
a
list
of
open
questions
and
known
issues
which
will
be
talking
over
time
with
friends
from
Mozilla
watts.
That's
the
current
status
updates
of
a
native
protocol
handler
and
I'll
stop
sharing
right
now,
so
I
can
see
if
there
are
any
quick
cuts.
D
I
A
good
question,
I
suspect
the
problem
is
that
the
way
streaming
of
video
sources
work?
This
may
be
like
internal
hundred
of
our
content
range
header
in
HTTP.
The
problem
is
that
when
you
move
away
from
HTTP,
you
don't
have
this
abstraction,
so
it
may
be
that
we
may
need
additional
like
metadata
layer
or
like
way
of
signaling
ranges
for
those
whose
cases
that's
like
an
open
question.
It
might
look
like
it's
possible
that
it
started
working
because
I've
opened
it
directly
before
so
it
was
like
in
Firefox
memory.
I
A
B
Have
a
question-
and
you
might
be
able
to
redirect
me
somewhere
the
biggest
question
that
came
up,
but
he
said
well,
maybe
there
are
two
separate
questions.
One
of
them
was
what
happens
on
ipfs,
for
something
like
New,
York
Times,
or
something
like
that.
Where
there's
a
place
that
you
want
to
go
a
lot
but
the
stuff
underneath
the
changes.
Is
that
something
that
someone
has
written
up
or
does
that
make
sense
like
you
might
have
a
Content
address
to
an
article?
B
Probably
I
mean
the
general
concept
that
I
couldn't
explain,
it
cuz
don't
buy.
Neither
the
key
was
that
what
happens
then
was
something
like
the
New
York
Times,
where
it
changes,
but
you
want
to
have
you
want
to
be
able
to
put
in
it
yeah?
You
could
also
think
of
it.
I
mean
related,
but
maybe
a
separate
question
is
like
streaming,
but
that's
kind
of
a
separate
one,
but
we.
J
F
B
Think
that
was
their
their
way.
Their
example,
probably
about
the
use
case
for
big
data
stuff,
would
be
a
new
version
of
data
comes
up
and
you
want
to
be
able
to
like
say:
oh
here's
all
the
stuff
and
then
okay,
here's
the
new
thing,
there's
anything
here,
there's
new
thing
but
not
lose
the
connection,
but
this
is
the
new
version
of
the
thing.
A
lot:
okay,
yeah.
F
J
Built
on
that
with
the
web,
companion
and
ipfs
colon
slash,
slash
is
that
is
the
IP
NS
colon
slash,
slash
and
ultimately,
that's
going
to
be
a
hash
of
a
public
key
and
in
base
thirty-two
encoding
how
to
handle
the
multi
keys,
especially
for
elliptic
curve,
public
keys.
You
know
what
was
the
challenges
to
to
that
I
suppose
the
question
is
really
for
Lydell
yeah.
I
So
that's
a
very
timely
question
because,
like
right
now,
if
you
have
IP
NS-
and
you
want
to
refer
to
the
to
appear-
use
like
your
ID
right
and
the
problem
is
that
those
are
in
base
58
and
I.
Don't
think
we
have
like
a
convention
or
do
we
even
allow
different
like
at
least
multi
base
and
coding
yeah
of
tier
IDs.
So
that's
like
an
open
question
and
until
that
solved
we
won't
be
able
to
use
IP
NS
in
your
protocols
on
the
IPS.
J
Okay,
do
you
have
a
documented
issue
somewhere
so
I'm
working
on
this
right
now
with
as
I
presented
last
week?
Was
this
IP
ID
project
with
Edie
two
five,
five,
one,
nine
and
there's
emerging
standard
for
HD
wallets
using
two
five,
five,
one,
nine
and
so
I
think,
but
it
ultimately
since
using
the
peer
ID
has.
This
did
unique
identifier,
its
I'm
writing
to
this
issue
about
base
32
encoding
and
it
suddenly,
like
not
gonna,
be
possible.
Yeah.
I
F
F
Things
that,
like
their
bolsters,
bare
multi
hashes
and
can't
tell
the
difference.
There
are
some
issues
that
try
to
figure
this
out,
mostly
by
saying
well,
there's
no
real
CID
version,
0,
it's
just
contextual,
based
on
the
prefix
feel
like
you
know
it
/
IP,
FS
/,
the
ad
version.
0,
you
interpret
it
one
way.
If
you
have
our
sessions,
they're
hash
interpret
it
one
way.
If
you
have
/
p2p,
fair
hashers,
new
tradition
boy.
Unfortunately,
we
also
have
like
we've
been
using
/
ipfs
for
the
protocol
as
well
vertical
name.
F
D
A
D
I
Yeah,
like
anything
apart
from
the
native
protocol,
will
work
because,
unless
you
want
to
put
a
PID
like
the
key
hash
in
sub-domain
and
I,
think
a
part
of
using
that
in
subdomain
will
work
perfectly
like
it
works
right
now.
Subdomains
and
native
protocol
handler
are
kinda
the
same
abstraction
you
put
base
32
case-insensitive.