►
A
A
We
want
to
put
all
the
younger
in
packages
on
my
PFS
alex
plus
cds
has
been
working
on
that
for
some
time
in
the
background,
victor
is
working
on
related
work
to
make
an
open
registry,
so
we're
going
to
kick
things
off
with
Alex
who's.
Just
gonna
give
us
a
quick
heads-up
on
where
he's
at,
like
what
his
focus
is
today
and
where
he
kind
of
sees
it
going.
Original
travel.
A
I
know
that,
like
some
new
ideas
have
occurred
to
him
in
the
last
few
weeks,
with
github
announcing
their
registry
most
interesting
for
the
people
on
this
call
is
we
are
hoping
to
release
ipfs
NPM
with
the
idea
this
desktop
app
and
the
port
there's
a
PR
from
Henry
that
does
the
boilerplate
of
making
the
command
available.
But
what
I
wanted
to
spend
some
time
on
the
school
doing
was
just
like
figuring
out
what
the
distance
between
should.
We
release
it
on
the
world
today.
A
My
gut
feeling
is
no
but
I
kind
of
want
to
like
poll
the
people
here
present
on
what
their
feelings
are
and
what
just
to
try
and
pull
together
a
short
list
of
things
that
need
to
happen
before
we're
ready
to
call
ipfs
NPM
a
thing,
so
I've
put
a
bunch
of
notes
in
the
document,
but
people
should
feel
free
to
add
more.
Let
me
share
my
screen
with
those
notes
on
it,
so
that
anyone
not
familiar
can
talk
about
it.
Can
someone
volunteer
to
be
a
note-taker,
please
I.
A
B
Sure
so
NPM
one
ipfs
two
components
to
it.
First
is
a
mirror
of
the
NPM
public
registry
that
adds
see
IDs
to
fool
the
various
packages
that
are
available.
So
each
each
package
has
a
file
called
apartament
that
describes
the
inversions
that
are
available
and
the
toggles
that
make
up
those
versions.
So
we
just
put
a
cid
next
to
each
Tarble
field.
B
There
is
the
idea
they
would
used
to
resolve
the
same
Tarble
reviews
HTTP
for
normally
so
there's
a
process
that
listens
to
changes
from
NPM
database
for
new
modules
that
are
coming
in
and
every
time
it
sees
a
new
module.
It
goes
off
downloads,
the
toggle
adder
to
ipfs
and
then
broadcast
the
cid
to
a
bunch
of
registry
mirrors
and
each
mirror
has
its
own
copy
in
the
registry
and
then
updates
them
using
these
co
DS
it
receives.
B
So
that's
a
server
side
component,
there's
a
client
side
component
that
when
you
do
an
NPM
install
using
it.
So
you
can
invoke
this
client
side
component
and
what
it
does.
Is
it
spins
up
a
web
server
and
create
an
IP
FS
node?
It
configures
your
local
copy
of
either
NPM
or
yarn
to
use
the
web
server
that
it
just
spun
up
instead
of
hitting
the
public
registry.
B
So
when
you
start
installing
things,
it
will
make
a
request
to
this
web
server
that
you're
running
locally
with
a
path
that
includes
a
package
name
and
when
it
receives
the
request
for
the
package
name,
it
looks
up
the
Packman
in
its
local
repo
using
the
IP
FS
Noda
its
funnel.
If
it
doesn't
have
it,
then
it
will
go
to
public
registry
of
copy
and
registries
kamerad
afternoon.
Add
it
to
his
local
node
and
then
server
back
to
the
client
and
then
the
next
time.
B
The
client
will
then
feed
that
back
20
Abel
says
instead
of
server
will
feed
that
back
to
the
MPM
coin
itself,
which
will
use
the
pac-man
to
work
out
which
table
at
once
to
download
in
order
to
satisfy
the
similar
version
that
the
user
has
like
is
chasing
know
then
make
a
request
to
that
terrible
to
our
server.
That's
running
locally.
B
The
online
component
will
then
take
that
terrible
user
to
look
up
in
its
local
repo
in
its
ffs,
and
if
it
has
a
CID
for
that
table
or
not,
if
it
doesn't
have
it,
then
again
it
will
hit
the
registry
to
work
out.
That's
the
ID,
which
is
odd.
If
that
happens,
because
you've
just
downloaded
the
baculum
has
all
the
celebrities
anyway,
it's
possible
that
it
happens
in
between
events
or
whatever,
and
so
anyway.
So
if
it
doesn't
have
that
that's
the
idea,
we
don't
get
it.
B
If
it
does
have
the
CID,
then
it
will
use
ipfs
to
resolve
the
tarball
and
stream
that
back
to
the
client
in
this
case
being
NPN,
which
then
goes.
Oh
I
got
a
table
from
somewhere
great
unbox.
It
puts
on
the
disk
and
it
will
just
repeat
that
process
for
every
single
dependency
that
is
needs
to
install
and
the
question
so
far.
A
Some,
but
more
like
ok,
that
sounds
interesting,
but
the
key
piece
of
the
puzzle
are:
we
are
running
a
service
that
is
lazy,
mirroring
and
PM's
registry.
That's
piece
one,
so
a
centralized
piece
of
the
puzzle
has
to
do
is
is
two
parts
of
the
thing
that
the
developer
runs
on
their
machine,
the
CLI
tool.
One
is
this
local
HTTP
server,
so
you
run
a
registry
proxy
on
a
local
host
on
homeland
and
the
CLI.
A
B
Yeah,
so
if,
if
the
client,
if
yawn
or
the
npm
climb,
they
were
using,
may
request
stuff
aren't
for
Parkman's
or
photographers,
then
I
just
forwards
them
straight
on
so
the
public
NPM
registry,
it's
reduced
online
login.
You
will
change
your
password
or
publish
something
whatever
all
that
traffic
go
straight
to
again,
we
don't
event
cemeteries
or
to
do
anything
that
we
already
interested.
A
Does
anyone
else?
Does
it
make
sense
to
people
here
present
some
sense
enough
enough
to
be
doing
this?
Okay?
So
so,
then,
the
next
piece
of
it
is
we're
worried,
where's
your
head
up
with
it,
the
your.
As
far
as
I'm
aware,
you're
focused
on
performance
improvements
and
refactoring
in
case
our
pivots.
B
Yeah,
so
there
is
a
problem
with
with
the
NPM
ipfs
client,
but
why
you
so
the
whole?
The
whole
of
NPM
is
represented
as
a
massive
Hampshire,
so
you
can
copy
that
into
your
local
repo,
but
was
you
know
on
a
copy,
the
whole
thing
because
it's
enormous,
but
what
you
really
need
to
feel
this
copy,
the
root
mode
of
the
show,
which
is
one
diagnose,
and
should
we
see
the
quick
and
I
turned
out?
There
was
a
bug
in
the
UNIX.
B
The
first
exported
I
think
so
long
ago
now
well,
I
can
see
is
basically
functions
and
say
it
was.
There
was
a
bug
and
I
was
trying
to
cool
down
the
entire
shower
choice:
key
yo
dude.
What
we
need
to
do
is
pull
down
one.
No,
because
when
you,
when
you
trying
to
resolve
a
patent,
it
will
say
right.
Well,
you
want
this
module
with
this
name,
so
it
will
look
in
the
show.
Ok
right!
Well,
that
is
gonna,
be
on
this
branch,
of
course,
not
that
bad.
B
B
So
there
was
like
the
exporter
is
clever
enough
to
not
load
the
whole
shot.
I
think
there
was
a
plug
in
there.
My
first
it
wasn't.
There
was
shooting
the
wrong
behavior
for
that
anyway.
So
it's
trying
to
fix
the
whole
thing.
I
fixed
it
I
grew
very
tired
of
looking
at
pool
streams
because
they
just
need
to
make
everything
really
convoluted
and
hard
to
follow,
and
we
have
this
lace
in
Kuwait
reflector
coming
out,
I
was
like
I'm
just
like
refract.
This
is
wasting
away
cool
yeah.
B
That
was
a
scope
creep
anyway,
so
I
fixed
the
bug.
You've
got
really
fast,
but
I
obviously
need
to
do
the
rest
of
the
work
to
actually
ship
the
fix-
and
that's
been
me
for
the
last,
like
three
weeks
just
to
async/await
refactoring
of
all
the
files
related
stuff
in
the
interplanetary
file
system,
which
is
like
your
tedious
max.
B
A
B
I
think
inators
is
getting
faster
like
it's
much
faster
now
than
it
was
I
was
referring
the
trickle
tag
and
yeah.
So
the
the
previous
implementation
in
order
I,
don't
interrupt
test
for
a
big
file
and
the
previous
implementation
was
taking
about
78
seconds
to
return
the
wrong
CID,
whereas
the
current,
like
the
async/await
one,
was
taking
about
three
and
a
half
seconds
to
return.
The
wrong
scale,
which
is
cool
I,
mean
it
will
just
shoot
me,
return
the
right
slowly
and.
A
Sounds
even
better
yeah,
it's
gonna
be
good
man.
So
then
the
question
becomes
so
that
you're,
the
only
factor
and
understanding
biggies
for
work,
a
you're,
imagining
that
your
head
will
be
out
of
that
next
week,
yep
and
then
with
that
in
place.
The
are
you
too
deep
to
say
what
you
would
be
doing
next,
like
do
you
know
what
your
next
few
steps
are
for.
B
Something
that's
interesting:
the
wind
his
service
on
our
server
that
receives
a
new
NPM
module
and
says
here's
a
tarball,
breaker,
hesitant
broadcaster,
see
IDs
to
all
the
nerves,
so
it
doesn't
quite
popped
up
which
is
funny
except
for
pub/sub,
doesn't
sign
messages.
So
we
have
these
registry
mirrors
that
are
like
yeah
cool.
B
So
I,
don't
trust
that
mirror
anymore
because
it's
entirely
possible
to
poison
it
with
spoofed
messages
so
that
the
client
is
clever
enough
to
it'll
like
standard
crime,
so
MP
understand
behavior,
the
pikemen
has
caches
of
all
the
toggles
and
it
should
in
theory
validate
that
toggle.
We
just
downloaded
against
our
hash,
but
because
they
got
the
hash
from
the
packet
which
we
got
from
us.
So
we
don't
change
the
hashes
in
the
document.
So
it's
not.
That's
not
one
rule
to
the
pub
sub
problem.
B
B
There's
also
a
bug
in
my
first
choice:
either
ghosts
around
ham,
shots
that
mean
that
is
possible
to
generate
an
invalid
shot,
which
is
fixed
in
75,
so
registries
only
regenerating
anyone,
but
this
might
all
be
moot.
So
Ollie
alluded,
sir
kind
of
things
I've
been
thinking
about
in
the
interim,
always
get
hard
to
recent
announcement
so
like
so
sneaking
into
that
like
if
the
bright
future
of
like
javascript
registries
is
some
kind
of
you
know,
federated
set
of
walled
gardens
then
hola,
all
our
merit
does
is
narrow
npm.
B
So
tomorrow
somebody
publishes
their
hot
new
module
to
github
you'll,
never
see
it
by
our
registry.
You
know
which
is
a
bit
of
a
shame
and
consider
all
about
the
distributed
web.
To
kind
of
have
this
centralized
service
the
richest
a
mirror
of
another
centralized
service.
It
doesn't
really,
you
know
we're
not
really
walking
the
walk,
which
is
a
shame
so
as
part
of
like
china
make
it
so
that
we
wouldn't
actually
need
to
have
this
separate
client
component
for
NPM
and
ipfs.
B
I
submitted
a
poor
requester
to
pocket
a
which
is
I'm,
probably
saying
Iran.
It's
it's
a
module
of
NPM
users
to
basically
choose
like
to
to
download
modules
from
from
from
someplace.
So
when
someone
from
NPM
talks
about
NPM
supporting
multiple
transports,
it
happens
in
in
potato,
so
I
added,
so
I
create
a
pull
request
to
placate
a
that
adds,
support
for
ipfs,
hashes
and
IP
NS
names,
and
using
that
in
your
package,
Jason
to
resolve
the
app
acumen
or
a
table,
which
is
you
know,
kind
of
fun
right
here.
B
B
B
But
what
I
would
like
to
do
is
kind
of
go
back
to
that
PR
and
kind
of
refactor
it
so
that,
instead
of
having
all
the
code
for
like
the
ipfs
and
the
IP
NS,
had
my
protocol
handlers
being
protected
so
having
it
just
be
like
an
external
module
that
you
would
install.
So
you
would
do
something
like
you
know:
npm
install',
G,
question.
C
B
We
change
our
mind
since
anything
anyway,
yeah
I
say
having
an
external
module
CVT,
something
like
NPM,
install
G,
NPM,
transport,
ipfs
and
then
now
would
you
know
create
that
would
put
a
modular
system
that
the
Katie
will
be
able
to
resolve
to
say
alright
cool
you've
said
my
PFS
:
/
I'm
just
going
to
hand
him
off
to
this
transformer
transformer
to
work
he'll
will
that
would
be
cool
and
then
we
cuz
actually
gives
us
control
of
publishing
updates
to
the
price
Colangelo
rather
than
having
to
get
PR.
Suppose
it's
appreciated
anyway.
A
B
D
B
Exactly
so
that
point
we
have
this
registry
mirror.
Well,
we
have
this
URL
at
moment
this
registry,
don't
chairs
I
profess
to
hire
at
the
moment.
Is
this
mirror
of
NPM
with
added
CID
crunchies,
but
if,
instead,
it
was
some
kind
of
portal
search
engine
thing
to
be
able
to
find
modules
that
were
published
using
IP
NS
names?
B
A
So
Hugo
was
demoing
his
upgrades
to
our
Guinness
using
DNS
and
other
miscellaneous
tricks,
so
I
think
we
can
I
mean
there
is
a
future
when
IPS
is
quick
assuming
an
online
assuming
you
have
connections
to
centralized
services
and
DNS
resolvers
right.
Well,
you
know
in
my
case,
but
who
is
it?
Is
there
an
issue
where
you've
written
up
your
thoughts
on
this
year?
I
said
something
we
can.
It's
just
I
mean
I
know
as
soon
as
I
finish.
This
async/await
refer
to
okay.
A
Well,
don't
forget,
while
you're
in
on
the
refactor
tractor
I
guess
so.
I
do
want
to
spend
some
time
talking
about
just
the
basic
view,
X
of
the
DFS
NPM
as
it
is
today
and
like
defining
what
we
think
needs
to
happen.
But
given
that
you're
talking
about
lots
of
registry
stuff,
it
might
be
interesting
to
hear
from
Victor,
because
Victor's
got
some
cool
project
that
seems
relevant
to
this
sure.
E
Before
we
move
on
in
general,
with
the
multi
registry
stuff
that
we're
talking
about
and
together
with
the
new
package
registry
I
think
together,
we
can
come
up
with
a
way
to
refer
to
the
different
paths
that
the
registries
have.
Npm
has
a
global
scope
now,
and
they
also
have
the
namespaces
for
themselves.
But
now
we
have
github
as
well,
where
you
from
username
on
the
github
project
and
so
on.
In
the
future,
there
will
be
more
registries,
but
there
is
still
no
unified
way
of
referring
to
these
attributes
voice.
E
So
maybe
so
multi-format
needs
to
be
invented
where
you
can
refer
to
packages
in
the
in
the
different
projects
anyways,
so
about
open
registry.
The
the
main
idea
is
one
I
guess
a
better
sort
is
to
focus
on
the
on
the
fear
point.
So
the
fear
is
that
there
is
a
few
entities.
Only
the
infrastructure
that
we
open-source
developers
or
relying
on
open
registry
is
part
of
a
bigger,
bigger
effort
to
create
open
infrastructure,
something
that
I
call
open
services.
E
The
main
point
is
to
have
it
funded
by
the
community
developed
by
the
community
and
owned
by
the
community
itself.
So,
for
example,
the
NPM
registry
currently
is
owned
by
a
company
called
NPM
Inc
once
they
run
out
of
funding
or
they
don't
find
a
way
of
announcing
themselves.
We
don't
know
what
happens
with
registry.
They
don't
seem
to
have
a
plan
will
happen
in
case
of
a
shutdown
and
moving
on
from
this.
E
There
is
always
the
fear
that
the
registry
will
eventually
disappear,
so
open
services,
kind
of
paints
up
like
a
framework
for
what
happens
when
it
shutdowns.
How
can
the
user
export
their
data?
How
will
the
service
finance
itself
and
all
of
this
data
should
be,
should
be
public
so
compared
to
NPM?
One
idea
felt
basically
the
same
thing
except
it's
developing
the
community
rather
than
being
financed
by
a
for-profit
company
in
this
case
being
protocol
up
so
I
think
today
and
beyond
all
my
TFS.
Unless
the
license
have
changed
it's
owned
by
David
Diaz.
E
Eventually,
it
will
probably
move
to
be
owned
by
protocol
app
or
maybe
a
ipfs
foundation
in
the
future.
Right
now
is
a
private
company,
the
belonging
developing,
the
the
tools
for
doing
this.
My
fear
is
that,
once
the
funding,
if
the
funding
runs
out,
these
the
developer
will
stop,
even
though
people
are
interested
in
contributing
to
it.
E
Think
they're
called
Open
juniors
nowadays
or
something
like
that.
The
plan
is
eventually
to
start
contacting
either
the
Linux
Foundation,
which
I
understand
open
tears
to
be
a
portal,
but
right
now,
they're
only
had
the
conversation
with
like
ether
and
other
people
about
it.
I
haven't
made
any
concrete
steps
towards
the
I.
Think
open
registry
right
now
is
an
experiment
to
see
if
we
can
have
self
sustainable
infrastructure
project.
E
So
the
first,
the
first
part
of
of
our
registry,
is
the
like
the
API
back-end,
which
runs
in
a
on
a
dedicated
server.
Every
time
our
requests
consume.
It
download
the
package
package
and
Absa
to
IDF
s.
So
it's
a
it's
a
lazy
registry
but
then
similar
to
npm
on
idea
of
us.
The
result
second
CLI
utility
called
polliver,
which
basically
is
not
the
same
as
the
MPM
on
ipfs.
Just
written
in
go
line
is
but
of
note,
and
it
connects
directly
to
the
open
registry
IPA
first
node
and
it
pulls
down
packages.
A
Nice,
it
sounds
like
there's
a
lot
of
overlap
between
the
two
efforts
and
and
we've
spoken
about
this
offline,
but
so,
like
short
term,
we
have
a
goal
to
release
the
I,
give
smpm
command-line
tool
as
an
opt-in
experiment.
So
you
install
I,
give
us
desktop
and
there's
a
new
piece
of
UI.
That's
like
this
is,
if
you
use
NPM,
this
is
an
experiment
in
cotton
to
have
it
monkey
around
with
your
path
so
that
we
deployed
that
the
command
line
is
now
available
to
you.
A
Eli's
convertible
and
the
feeling
right
now
is
that,
given
some
of
the
so
it's
it's
like
an
order
of
magnitude
slower
than
NPM
today
and
so
I
think
that
at
this
moment
we
have
to
be
fairly
realistic
that
it's
it's
kind
of
an
experiment
that
we're
encouraging
ipfs
like
users,
people
who
are
keen
on
decentralization.
We
want
them
to
try
this
out
to
get
feedback
and
to
help
us
focus
performance,
improvement,
working
out
the
events,
but
it
doesn't
seem.
E
I
have
the
same
experience
with
the
NPM
on
ipfs
but
I'm,
pretty
sure.
That's
like
implementation,
specific
issues
regarding
the
performance,
whatever
the
proxy
I
mean,
writing
is
always
faster
than
than
be
downloading,
be
HTTP,
so
I'm
pretty
sure
you
will
find
one
one
or
two
issues
where
you
can
brace
the
performance,
because
something
silly
is
happening.
It's
not
something
that
I've
seen
in
the
Golan
ecosystem
of
slippage,
be
that
same
performance
issue.
A
The
other
purpose
of
this
call
is
to
kind
of
chance
people
who
have
tried
it
out
and
see
like
what
they
think
is
the
bare
minimum
like.
What
can
we
tog
as
a
release
and
say
this
is
worth
experiment
like
throwing
out
to
the
users
and
experimenting
with
what
kind
of
feedback
are
we
looking
for
like?
What's
what
is
the
purpose
of
this
endeavor
I
guess
was
still
at
like
level
zero,
so
I'm,
looking
at
Jessica
Schilling
smiling
on
that,
because
this
is
this.
Is
where
she's,
like
guys?
Have
you
thought
about?
A
Why
you're
doing
this
like?
Well,
that's
why
I'm
reppin
is
cool,
so
it's
a
but
also
detect,
run
like
it's
really
interesting
to
I
kind
of
knew
that
there
were
similarities
between
IPs
and
Gaiman,
open
registry,
but
I
didn't
realize
how
similar
they
were.
So
it's
if
it
feels
like
a
longer
term
we
should
be
looking
to
like
is
there's
there's
a
huge
amount
of
value
in
us.
A
Just
stress
testing,
IP
investment,
making
most
performance
improvements,
but
it
seems
like
for
this
project
to
have
more
long-term
benefit
to
the
community
and
have
a
life
beyond
that.
It
seems
like
we
should
be
looking
at
how
it
merges
in
with
your
plans
for
open
registry,
or
at
least
that
would
be
a
very
kind
of
useful
like
stuff
that
we
produce
should
be
really
useful
for
you,
if
not
directly,
working
together.
E
E
Yeah,
no
it's.
The
first
focus
obviously
is
to
like
regain
the
control
of
the
private
registries,
so
Larvik
docker
and
packages
and
NPM,
essentially
like
registries,
run
by
private
companies.
The
first
one
since
I
have
the
most
experience
with
NPM
was
NPM
and
that
website
was
made
to
reflect
the
current
site,
not
the
future
type
of
project.
That.
A
Is
once
okay
well
I?
Think
now
it's
a
good
point
to
move
on
to
the
more
interactive
segment
which
is.
Do
you
I've
put
together
this
list
of
things
that
I
can
see
that
we
need
to
tackle
before
before
we
offer
this
up
to
users
as
something
that
we
say
is
good
and
useful,
so
I
can
just
whip
through
that
real
quickly.
A
Like
the
the
overarching
thing
is,
we've
got
a
goal
for
this
quarter,
which
is
to
get
this
into
people's
hands
and
start
getting
feedback,
but
I
I
still
don't
have
a
clear
idea
of
whether
it's
whether
we
intend
to
support
NPM
on
ipfs
long
term
as
a
service
or
whether
this
is
actually
you
know
a
a
useful
experiment
for
stress
testing,
idea,
fest
but
I
don't
know
if
we
we
assembled,
have
the
answers
from
it,
but
it's
worth
reminding
ourselves
that
we
don't
know
that.
A
Yet
if
we
do
know
that
someone
should
tell
me
so
that's
the
big
one
is
speed
and
I
think
I'll
access
at
this
one
hundred
times
it's
going
to
be
challenging.
But
if
victory
of
saying
is
true,
then
clearly
there
is
some
hope
there.
Also.
We
need
to
do
a
lot
of
work.
Around
testing
I.
Think
one
of
the
like.
What
are
the
benefits?
What
was
it
the
situation
of
using
this
one
of
them
is.
A
They
shouldn't
have
to
go
outside
the
network,
so
I
think
we
need
to
do
some
testing
to
make
sure
that
that
is
the
case
that,
like
I,
just
haven't,
seen
it
and
seen
in
action,
and
there
were
some
reports
on
the
the
Cote
pull
request
that
suggesting
that
the
speed
of
like
ipfs
already
has
a
tarball
and
it
wasn't
much
faster
than
if
any.
There
was
some
question
about
whether
local
LAN,
IP
FS
connection
fetching
a
terrible.
A
A
B
Fine
I,
just
I
just
wanted
to
be
stable
and
deployable
and
repeatable,
and
all
that
and
it's
not
it's
just
running
on
a
Big
Easy
sob
easy
to
books
on
them,
which
is
not
no
idea
at
all,
but
I
think
yeah,
they've
kind
of
realized
that
they're
not
gonna
have
times
doing
themselves
so
they're.
Just
can
you
give
me
the
keys
to
do
so,
which
is
great
that.
A
A
So
then,
then,
it's
back
to
like
what
are
what's
the
benefits
of
using
this
over
NPM.
For
me,
it's
always
like
the
thing
that
gets
me
most
excited.
It's
like
I
can
co-host
the
things
that
I
need
depend
on
that
seems
to
be
for
me
like
if
it's
slow
I
can
deal
with
it.
If
there's
some
value
to
it,
things
like
but
I'm
struggling
at
the
moment
to
find
a
way
of
presenting
that
back
to
users
beyond
saying,
like
hey
good
work,
you
did
the
right
thing.
You
are
covered
co-hosting.
A
Your
modules
I
can't
see
a
good
way
today
of
being
able
to
tell
the
user
how
much
good
work
they're
doing,
just
because
of
the
lack
of
an
API
for
being
able
to
give
bandwidth
per
cid
but
I'm
as
anyone
like
Alex.
If
you
have
any
thoughts
about
anyone,
Connie
and
well,
how
do
we
present
to
the
user
like
the
value
of
using
this
tool?
I
guess
is
the
question.
A
B
You
know
so
I
mean
am/pm
could
remove
a
version
from
a
document,
but
we
moving
versions
from
a
mirror,
so
always
be
there.
The
github
registry
of
Heine
and
you're
saying
that
they
can
have
some
kind
of
automated
blocking
thing
where,
if
you
get
a
vulnerability
published,
they
will
just
yank
your
package
until
it's
resolved
in
one
way
or
another.
You
know,
obviously,
if
the
price
of
your
dev
dependencies
and
all
that
kind
of
stuff
which
you
know
typically,
he
don't
really
care
that
much
about
vulnerabilities
unless
it's
like
steering
stealing
you're
like
cryptocurrency.
D
Okay,
I
think
like
there's
the
sort
of
altruistic
aspect
where
hey
I'm,
providing
some
bandwidth
I'm
helping
close
this
thing,
but
I
think
that
it's
it's
not
even
just
altruistic,
I
think
it's
like
you
know.
There's
people
I
just
went
to
the
CSV
comp
conference
in
Portland,
where
it's
all
scientists
and
they're
doing
research
reproducibility
and
it's
a
hard
requirement
for
them
that
they
need
to
be
able
to
have
their
data
published
and
that
the
software
tools
that
they
use.
D
That
goes
along
with
the
research
and
have
it
be
reproducible
by
other
scientists,
so
they
can
build
on
their
data
and
cycler
data
and
so
a
lot.
A
large
part
of
this
is
sustainability
and
Prevention
of
software
collapse,
which
is
what
happens
when
npm
goes
offline.
It
probably
won't
happened,
but
you
know,
but
also
you
know
like
this.
A
A
D
But
you
know
that
that's
not
necessary
not
necessary
for
keeping
the
data
alive
in
the
future,
but
for
very
small.
You
know
if
you're
publishing,
I
say
a
module
and
you
don't
want
to
use
NPM.
You
know,
and
it's
a
very
niche
sort
of
thing
then
you're
doing
it
in
a
decentralized
manner.
You're
gonna
need
some
other
people
to
keep
that
module
alive.
A
Is
that
that
sounds
like
something
that
we
need
to
message
in
difficult
in
the
CLI
then
certainly
I'll.
Give
us
desktop,
could
talk
about
it,
but
I,
don't
because
you're
sort
of
talking
about
modules,
individual
modules
and
preserving
then
forever,
whereas
I'm
thinking
about
the
dev
user
story
of
I've
installed,
desktop
and
now
I
have
this
new
tool.
Why?
Why
would
I
use
it.
A
A
A
A
A
B
A
A
A
It's
just
saying
it's
definitely
doing
something:
okay
and
then
there's
yeah,
there's
a
few
simple
ones
like
we
invoke
ipfs
for
commands
that
don't
need
it
and
we
don't
invoke
ipfs
for
commands
that
do
you
need
it,
so
they
seem
like
small
things
that
we
can
fix
before.
Releasing
then
there's
the
kind
of
usability
question
of
should.
A
G
Question
throwing
out
a
suggestion
that
is
probably
a
horrible
suggestion.
It
would
make
us
deeply
deeply
unpopular
but
by
having
an
alias
out
to
say
you
install
the
desktop
client
and
it
Elisa's
out
and
him
as
a
default
interesting
just
to
see
how
many
people
notice
see
how
many
people
notice
the
change
in
speed
I
mean
I'm,
not
saying.
That's
now
saying
that's
a
good
idea,
but
it
would
be
very,
very
interesting
to
see
I.
A
Mean
I
was
kind
of
of
the
mind
that
if
we
tidy
up
the
logging
output
and
you're
installing
this,
it
deliberately
and
this
flow
would
be
installed
desktop
and
it's
offered
to
you.
Would
you
like
me
to
upgrade
your
NPM
with
ipfs
skills,
like
my
feeling
was
when
I
first
heard
about
it?
The
first
thing
I
did
was
alias.
I
can
listen
again
to
the
MDM,
so
that
I
would
just
not
forget
to
use
it
at
the
time.
A
The
registry
was
even
more
unreliable,
so
I
quickly
had
to
stop
doing
that,
because
I
couldn't
actually
do
any
dev
work,
but
it
was
what
it
was.
The
approach
that
I
kind
of
instinctively
went
for
so
then
it
was
the
approach
I
suggested
that
we
go
for
desktop,
but
I
went
to
sanity
check
it
with
people.
If
there
is,
is
it
can
easily
cause
a
fence
if
you
go
and
mess
with
people's
paths,
but
I
think
if,
if
you've
told
a
big
switch
in
a
UI
that
says
please
make
my
NPM
like
get
messy
with.
F
That
work
with
other
tools
that
you,
like
I,
regularly
switch
between
own
versions
and
have
different
versions
NPM
using
NPM
and
then,
if
I
switch
between
two,
then
that
would
rewrite
that
again
and
again
or
good
pickles.
So
it's
like
if
we
have
to
monitor
that
and
constantly
switch
back
with
that
cause
more
issues.
H
Mdm
and
related
switchers
use
like
a
shim
that
will
essentially
run
a
program
to
see
what
version
of
a
thing
in
my
on
so
you're,
not
actually
changing
the
direct
path
to
NPM
every
time
that
you
switch,
so
it
it
shouldn't
cause
too
many
problems.
Obviously,
like
crazy
people
do
crazy
things
and
they're
always
going
to
be
edge
cases
I,
guess
like.
H
A
Just
as
a
quick
put
your
hand
up
vote,
who
would
prefer
it
to
become
their
default
if
you
use
NPM-
and
you
talk
to
this
in,
would
you
like
this
to
be
in
p.m.
without
get
this?
What
would
you
like?
It's
a
separate
command
so
hands
up
for
NPM,
one,
two,
three
and
four
five
and
four
separate
command
one
tea,
please
yeah!
That's
pretty
split
250
and
a
question,
and
some
apps
are
yeah.
G
So
questions
so
say
you
know
so
say
we
default
that
and
what
about
the
viability
of
you
know
you're
sitting
there
you're
sitting
there
looking
at
your
terminal
and
knowing
you're
installing
a
package
in
is
taking
forever
and
ever
and
ever
and
ever
and
ever
and
then
you
know
maybe
the
fall
back
on
that
is
to
have
a
reminder
in
the
CLI
that
just
says
hey.
If
this
is
taking
forever,
and
this
is
making
you
angry
you
can
you
can
temporarily
on
ipfs
and
goodness
your
npm.
By
doing
this,
don't.
G
A
So
this
is
something
this
is
a
recurring
problem
that,
like
the
performance
issue,
so
companioned
had
to
put
the
toggle
in
so
it's
like.
Can
you
ask
companions
to
stop
doing
what
companion
does,
because
it's
slowing
my
browsing
experience
chris
mode
would
be
good,
but
yet
that
might
be
the
remember
the
way.
I
can't
immediately
visit
what
would
look
like,
but
that
is
something
to
consider.
A
I
And
in
the
context
of
a
broader
like
lump
of
experiments,
this
is
one.
This
is
one
experiment
like
some.
Do
you.
A
I
In
and
I
thought
perhaps
it
could
be,
you
know
when
you,
when
you
launch
the
UI,
it
gives
you
a
notification,
you
know
and
it
asks
you
whether
or
not
you
want
to
know
what
we
you
know.
We've
got
an
experiment
or
it's
been
installed.
You
know.
Maybe
it
says
it's
been
installed
at
any
rate,
it's
surfacing
what's
going
on,
because
this
is
such
an
experimental.
You
know
ipfs,
it's
an
experimental
thing,
so
it
lets
them
know
hey.
We
have
this
idea,
Monica
idea.
We
have
M
p.m.
I
on
a
ipfs
running
and
then
an
option
to
you
know
click
over
to
just
see
what
that's
all
about,
but
in
terms
of
configuring
it
in
an
ongoing
basis.
I
think
maybe
it's
in
the
settings-
and
you
know
you
can
either
ask
before
installing
or
automatically
install
these
and
experience,
and
maybe
the
toggles
are
there
as
well.
G
A
A
For
sure
questions
in
the
chat,
where
does
desktop
have
a
cm
I
know
the
next
release
of
desktop
is
going
to
make
my
PFS
CLI
available
on
your
path,
but
that'll
just
mean
that
if
you
didn't
have
white
give
s
on
the
sit
on
the
command
line
before
now,
you
do,
it
doesn't
have
its
own
extension
command
line,
but
I
think
it
could
Lydell.
What
do
you
reckon
I
think
this
desktop
as
a
command
way
to
just
saying
out
loud
sounds
stupid,
but
it
could
just
work
unless.
E
A
F
A
D
B
Stocki
us
know,
which
is
great
and
after
the
let's
say
you
type
in
IPM
I
have
any
idea
that
and
after
that
anything
you
type
just
get
pumped
straight
to
the
package
manager.
It's
using.
So
we
haven't
like
re-implemented
like
installer.
Anything
crazy,
like
that.
Maybe
we'll
just
goes
to
whichever
one
you're
using
tell
yes,
you'll
support
any
and
yeah
new
hot
package
manager,
X
that
comes
after
you
on
t
so.
A
A
H
Wish
I
was
going
with
that
question
of
you
need
to
be
able
to
talk
to
NPM
registry
to
be
able
to
install
the
the
the
way
the
NPM
works
offline.
In
theory,
mostly
thinking
about
ipfs
camp
I
mean
like,
is
there
a
way
that,
if
you're,
if
you're,
surrounded
by
people
with
this
thing,
can
you
get
that
thing?
H
I
A
B
H
A
Definitely
a
for
further
future
question
all
right.
We
are
out
of
time
if
people
wants
the
ground,
I
will
keep
talking
if
people
need
to
drop
off,
feel
free
to
do
that,
but
the
the
vvv,
the
remnants
of
the
call,
are
open
mic.
So,
if
anything
had
anything
they
want
to
share
now
would
be
a
good
time
to
say
it
shall
hallow.
F
One
thing
maybe
would
be
thinking
back
to
the:
why
a
bit
more,
maybe
we're
not
gonna
hit
that
with
the
first
version,
and
perhaps
the
longer
tail
of
that
is
the
whole
story
about
the
centralized
package
ready
tree
and
which
it
feels
like
we
could
enable.
So
it's
like
what
would
be
the
first
prototype
version
of
NPM
and
IFIF
idea.
First,
we
could
build
that
would
basically
allow
you
to
publish
a
module
into
my
PFS
and
then
fetch
it
from
my
defense.
F
B
Know
the
poor
requester
poco
today
that
enables
like
DNS
names
as
ways
to
resolve
par
commences.
Those
will
do
that.
You
can
totally
do
that
either
than
that,
like
that,
the
interesting
things
were
I
see
the
speed
of
our
penis.
Resolving
we're
just
going
to
improve
the
other
thing
is
distributed,
identity
that
you
can
treat
at
the
moment.
B
There's
no
dignity,
some
way
of
saying
hi,
I'm
aching,
praying
I
approve
this
package
like
so
me
teacher
can
enjoy
its
and
interesting
conversations
with
Microsoft
about
their
own,
just
riveted
identity
platform,
which
is
kind
of
just
kicking
off
and
using
my
PFS
to
store
things
which
is
got
cool
but
like
at
the
moment
I
can.
Realistically.
If
you
wanted
to
start
typing
today
and
like
that
program,
yeah,
the
identity
is
coming.
The
whole
thing
this
whole.
A
Anybody
got
anything
they
want
a
demo,
yeah
anything
fun.
Alright,
then
the
only
thing
left
took
well.
There
were
some
other
things
on
the
dock.
You
can
look
at
them
at
your
leisure
mermaid
diagrams
changed
my
life,
so
rad
so
where's
this.
You
don't
need
me
to
tell
you
that
they
went
well.
It's
a
lot
of
notes.
Thanks
Jim
no
made
diagrams
you
just
you
just
type
out
the
flow
and
it
keeps
you
a
diagram.
It's
amazing
anyway.
Oh
my.
J
A
Your
dependency
diagram
pian
hop
from
now
on
and,
of
course,
the
big
shout
out
for
hack
nd
is
my
new
favorite
way
of
writing
mock
then,
because
it
supports
murmured
diagrams
in
line
you
just
give
it
the
type,
give
a
code
block,
type
and
OMG.
If
you
want
to
elevate.
Oh,
oh,
it's
happening
it's
happening.
A
Yes,
there
you
go
see
that
Jam
I
think
we
were
going
to
be
communicating
more
clearly
from
now
on
about
our
DAGs.
Talking
of
communicating
clearly
about
our
DAGs
check
out
Allen
jaws
I
give
a
stag
visualizer
service
I
started
off
working
on
it,
but
today
he's
just
made
it
look
beautiful,
so
you
should
check
it
out.
There's
a
link
in
there.
You
can
visualize
all
your
decks
and
change
all
athletes.
A
So
when
you
add
a
file,
there's
a
whole
bunch
of
options
for
how
the
bag
gets
constructed
and
it's
sometimes
a
little
difficult
to
understand
what
those
dag
options
are
slowly
with
benefit
of
posterity
that
could
file
look
into
the
Internet
there
he
is
come
on.
I
live
between
12
by
trunks
gives
you
some
an
eye
laughs,
that's
in
a
balance
back
wall.
If
it
was
in
a
trickle
bag,
oh
my.
G
F
A
A
A
So
my
suggestions
were
simply
preschool
is
getting
a
lot
of
work
done
on
it
at
the
moment
and
we'll
be
the
center
of
people's
also
I
professed
camp
thanks,
Andrew
I
always
think
in.
We
should
talk
about
I
practiced,
school
and
visualizers
and
camp
content.
Maybe
that
would
be
something
that's
important,
that
we
should
catch
up
about
people
demoing
of
but
I'm
open
to
suggestions,
let's,
let's
leave
it
at
preschool
and
camp
content
for
now.
But
we
shouldn't
talk
about
it
off
this
cool
all
right,
any
other
business
big!