►
From YouTube: *.ipfs.localhost in Companion, Q1 updates and Q2 plans - IPFS GUI and Browsers Weekly 2020-03-25
Description
Meeting notes: https://github.com/ipfs/team-mgmt/pull/1130
About IPFS GUI and Browsers Weekly: https://github.com/ipfs/team-mgmt/issues/790
IPFS Mirror: https://ipfs.io/ipfs/bafybeidjhqh675iqoulhhz7yowhaxes4iwabv4qo6hmpqnf4nfiwc2xqlu/
For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS
Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5
A
A
The
first
topic
is
the
preview
of
subdomain
gateway
support
in
a
TFS
companion,
so
this
week,
I
released
better
better
with
a
preview
of
support
for
those
subdomain
gateways.
I
probably
talked
what
subdomain
Gateway
is
multiple
times,
so
I
will
just
point
at
our
new
dogs
website
and
there
is
address
and
2012
article
with
H
difficult
way
section
with
the
description
of
what
our
path
gateway
is
what
our
subdomain
gateways
and
innocent
gateways.
A
A
This
is
a
person
from
last
Friday.
It
will
ship
eventually.
So
if
you
want
to
test
this,
you
need
to
run
this
test
version
and
if
you
run
that
version
of
why
DFS
and
you
have
this
better
version
of
activist
companion
and
try
to
load
CID
from
a
path-
let's
open
this
one.
So
we
can
see
instead
of
opening
that's
on
1/2.
It
landed
in
a
subdomain
here
which
is
pretty
cool,
because
it's
one
content,
identifier.
First,
then,
you
know
it's
a
TFS,
then
you
know
it's
some
local
host.
A
What
else
do
you
need?
So?
What
happened
here
is
this
is
CID
v-0
and
when
I
open,
that
it
automatically
got
redirected
to
a
sub
domain
and
the
CID
got
upgraded
to
say
DV
one.
So
it
means
the
subdomain
gateway
on
a
local
host
is
a
dropped
drop
in
replacement.
It
will
handle
all
the
requests
that
would
go
to
a
regular
path
gateway.
It
will
just
upgrade
everything
and
redirect
to
the
separate
orig
entire
content
route.
A
This
resource
will
be
opened
from
a
local
gateway
and
you
can
see
that
the
slash
wiki
is
at
the
very
root.
So
it's.
Finally,
the
content
route,
which
is
indicated
here
is
the
say,
the
content
route
of
ipfs
resource
is
the
same
as
the
root
of
the
URL
interesting
side.
Effect
of
this
is
if
we
have
a
DNS
link
website.
Historically,
we
were
not
able
like
we
are
not.
We
were
able,
and
we
did
that,
but
it
was
breaking
original
selection
of
a
website.
A
A
A
Pretty
critical
and
if
you
want
to
like
sure
you
don't
like,
of
course
you
can
like
remove
that
part
and
now
copy
and
share.
However,
here
in
the
address
under
browser
action,
I've
been
told,
people
don't
know
this,
so
I
will
need
like
I
will
explicitly
show
this
if
you
are
on
a
local
host
website
with
this
URL
here,
and
you
want
to
copy
the
URL
for
sharing
with
someone
you
go
here
and
you
copy
public
gateway,
URL
and
you
can
see
the
URL
copied
to
the
clipboard
is
just
a
regular.
A
It
will
be
either
original
URL
for
DNS
link
web
site
or
it
will
be
URL
to
ipfs
resource
at
the
public
gateway.
So
you
can
share
it
with
people
who
don't
have
activist
companion,
don't
have
I,
give
us
a
poll
and
it
will
still
look
from
as
public
gateway.
That's
just
a
tangent.
The
key
announcement
here
is
this
will
happen
soon.
I
just
need
to
question
box.
So
if
you
want
to
test,
you
can
install
it
in
Firefox
today.
A
If
you
want
to
test
that
in
chromium
based
web
browser,
you
unfortunately
need
to
manually
download,
zip
and
load
it
because
Chrome
Web
Store
for
even
for
a
better
channel
it
takes
a
week
for
to
publish
which
is
ridiculous
for
a
better
channel.
So
if
you
find
a
bug,
please
feel
an
issue.
It
will
be
super
useful.
It's
a
pretty
like
it
looks
like
a
small
change,
but
it's
a
pretty
big
change.
Given
that
and
companion
companion,
it
is
pretty
old
software.
A
It
started
around
like
2016
and
it's
been
like
rebooting
fully,
at
least
once
and
during
this
entire
history.
We
operated
on
those
paths,
so
there
may
be
a
lot
of
like
code
paths
where
we
only
recognize
paths
and
we
ignore
subdomains.
So
if
anything
looks
weird
just
it's
expected,
let
me
know
I
try
to
fix
it.
I
think
that's
it
for,
for
this
any.
B
A
A
The
way
we
look
at
this
is
you,
you
use
subdomain
gateway
only
when
you
load
aquifers
content
in
a
web
browser.
A
lot
of
people
may
use
ipfs
in
context
which
are
in
HTTP
context
which
are
not
brought
for
in
consumed
by
browser.
For
example,
if
you
use
W
get
or
curl
that
will
continue
to
work,
and
it
will
continue
to
work
even
with
subdomain
gateway.
We've
made
it
so
the
request,
even
the
given,
though
Karl,
does
not
follow
redirects
to
subdomains.
We
are.
A
A
They
are
super
useful
and
like
even
if
you
are
running
like
DNS
linked
website
with
this
work,
I
didn't
go
ago,
idea
fest,
it's
possible
to
like
mount
path
gateway
under
your
under
the
DNS
think
origin
of
your
website.
So
if
you
are
the
owner
of
this
specific
origin,
you
want
to
consume
I,
give
a
spot
gateway
so
that
there
are
no
spaces.
We
the
documentation.
A
From
that
another
question
existing
non
appeal
of
the
gateways:
do
they
need
to
do
anything
to
upgrade
besides
upgrading
to
0.5,
yes,
and
no
so
when
they
upgrade
to
0.5,
the
only
change
happens
on
localhost.
So
the
only
thing
that
we
will
be
set
up
out
of
the
box
is
a
subdomain
gateway
on
the
local
host
host
name.
If
you
use
the
IP
127
or
one
that
will
continue
to
be
the
path
gateway.
A
However,
if
use
the
local
host
a
human,
readable
name
that
will
start
hosting
the
subdomains,
if
you
open
it
from
the
web
browser,
if
you
load
it
from
some
tools,
it
will
continue
to
working
as
a
cupcake
way.
Do
they
need
to
do
anything
to
upgrade
if
they
are
running
on
the
s
of
the
only
path
gateway
right
now,
they
they
will
be
able
to
enable
subdomain
gateway
for
a
specific
host
name,
which
of
of
their
choosing
with
one
line
and
in
go
aquafers.
Documentation
of
the
conflict
we
already
I
believe
provides
some
recives.
A
Well,
hard
coding,
specific
user
agent
strings,
it's
tricky
because
you
always
forget
about
one.
If
you
hard
called
browser
names,
you
ignore
like
niche
browsers.
If
you
hard
called
kernel
or
double
get,
you
will
ignore
other
users
and,
for
example,
BSD
does
not
have
her
or
W
get
they
use
fetch,
which,
which
is
common
line
tool,
and
it
does
not
set
a
user
agent
at
all.
So.
B
I
wonder
what?
If,
what?
If
we
I'm,
just
like
looking
down
the
three
to
five
year
on
this
and
be
like
we,
the
majority
of
usage,
is
probably
going
to
be
browsers
from
a
Gateway
perspective,
and
maybe
we
do
something
like
we
actually
have
a
header
required
to
be
set
by
non
browser
tools
in
order
to
get
to
enable
pathways.
Because
my
concern
is
that
people,
if
people
are
opening
up
public
gateways
and
they're
not
reading
the
documentation,
the
default
experience
is
pretty
unsafe
and
will
always
default
to
path
based
gateway
for
browsers
and
browsers.
B
This
most
likely
always
going
to
be
the
Lions
here,
like
the
massive
massive
massively
overwhelmed
any
type
of
tool
based
request.
So
I
feel
like
that.
Looking
at
the
longer
term
security
story
for
a
gif
s,
we
are
erring
on
the
side
of
a
very
small
number
of
tools
over
a
very
large
number.
The
security
of
a
very
large
number
of
end-users
and
I'm
not
sure
that
that's
the
right
calculus
to
make.
A
Just
enabling
people
to
start
using
subdomains
on
localhost.
Basically,
we
want
people
who
are
users
of
IP.
First
desktop
and
I
fix
companion
to
get
that
original
selection
in
the
future
we
want
to
do.
We
want
to
do
at
least
one
thing:
disable
cookies
and
local
storage
on
path
gateways
and
make
it
an
explicit
flag.
A
If
someone
wants
to
like
restore
that,
that's
a
great
mitigation
and
another
that
that's
one
thing
and
another
thing
will
be
more
like
more
strict
version
would
be
doing
what
you
just
described
like
making
that
like
opt-in
for
a
pass,
but
we
are
not
there
yet.
We
would
need
to
like
make
analysis
of
how
people
you
know
on
the
back
end,
but
definitely
we
we
want
to
move
into
that
direction
of
being
secure
by
default
on
the
web,
and
that's
also
like
why
we
like
ask
people
to
enable
subdomains
for
a
specific
hostname.
A
A
The
configuration
looks
the
way
it
looks,
but
yeah
there's
about
doing
that,
and
also
like
adding
some
access
controls
to
our
API
are
the
biggest
struggle
for
people
who
expose
lingo
ipfs
on
the
web,
but
going
back
to
question
to
upgrade
yeah
so,
for
example,
CloudFlare
they
already
provide
subdomain
gateway,
but
they
have
a
custom
convict
in
the
nginx,
so
they
will
what
all
they
need
to
do.
It
will
be
upgrade
and
then
yeah
and
then
enable
this
like
one-liner
for
their
hostname
and
that's
it.
A
So
it
should
be
pretty
painless
and
we
expect
more
and
more
people
having
that
one
like
one-line
experience
to
run
subdomain
gateways,
and
the
only
caveat
here
is
to
use
a
different
host
name
if
you
are
already
using
the
domain
or
upgrade
that
domain
to
being
a
subdomain
one
anything
needed
for
a
chief
SEO
after
upgrade.
Do
we
do
weird
stuff
there?
Yes,
we
do.
We
we
do
everything
there.
We
have
a
path
gateway.
We
have
a
bunch
of
web
DNS
linked
websites
and
ipfs
io
itself,
like
the
root
domain.
A
You
are
able
to
use
the
NS
link
for
the
route
and
mount
either
path,
gateway
or
subdomain
gateway
on
top
of
it
and
content
addressed
offs
always
like
take
a
priority
on
top
of
DNS
link,
so
probably
more
details
here,
but
the
existing
configuration
is
so
that
we
can
continue
doing
what
we
are
doing
on
a
DFS
I/o
and
we
don't
expect
any
change
there.
Activist
IO
will
continue
to
be
a
park,
a
path
database
gateway
for
now
the
D
web
link
will
be
the
subdomain
one
we
are
still
thinking.
A
Should
we
for
like
force
the
redirect
in
a
web
browser
from
I
give
sio2
the
web
link
that
that's
separate
conversation,
probably
for
we
need
to
check
in
with
PI
for
steam
first
to
understand
the
full
ramifications
of
that
change.
For
now,
we
just
upgrade
the
web
link
to
use
this
native
support.
Next
question
are
more
local
host
issues
bugs
in
browsers
affecting
us.
Here
there
is
one
pretty
big
thing
that
remains
to
be
a
problematic.
Each
subdomain
support
in
Firefox
and
the
way
subdomains
in
Firefox
on
localhost
are
not
secure
context.
A
Historically,
only
the
local
host
IP
was
secured
context,
but
over
time,
web
browser
vendors
started
interpreting
local
host
host
name
as
a
secure
context,
because
they
hard-coded
that
to
point
at
the
local
IP
which
made
it
actually
secure.
So
no
one
can
like
overwrite
the
destination
IP
at
the
DNS
level.
A
The
the
problem
is
like
in
Chrome,
all
subdomains
are
secured
context
under
local
host
in
Firefox.
Local
host
itself
is
secure
context,
but
not
subdomains
under
that
are
not.
But
there
is
an
open
issue
in
Bugzilla
about
like
finishing
that,
migration
to
hard-coding
local
host
and
when
they
finish
that
I
have
seen
like
movement
last
week,
I
believe
the
subdomains
will
start
being
interpreted
as
a
secure
context.
I,
just
did
not
feel
that
we
should
like
wait
with
Gateway
support
until
Firefox
ship
start,
because
that
may
be
like
in
next
release.
A
For
now,
the
only
problem
is,
you
may
see
that
here.
Actually
here
is
this
padlock
and
it
says
connection
not
secure.
So
if
I
open
the
same
resource
from
local
hosting
in
chromium,
it
will
not
have
this
crossed
padlock,
it
will
say
yeah,
it's
fine,
so
that's
the
only
like
an
open
issue.
I
see
right
now,
we'll
oh
yeah,
so,
like
other
bars
and
browsers
affecting
us
here,
nothing
comes
to
mind.
A
There's
there
was
like
one
one
problem
with
there's:
this
header
called
clear
side
data,
it's
a
header
which
server
consent
to
a
web
browser
and
tell
it
to
purge
all
cookies.
All
like
cache
everything
for
this
specific
origin
and
I
had
this
discussion
with
you
go
hook
before
that.
There's
interesting
fact
that
if
you
go
to
like
the
support,
it's
fine,
it's
fine,
but
then
what
what
this
exact
execution
context
so
execution
context
may
forces
all
the
tabs
to
reload
after
you
receive
that
header,
so
that
does
not
work
everywhere
and
also
like
in
chromium.
A
A
If
we
made
a
mistake
and
someone
load
that
like
JavaScript
or
something
from
the
path,
somehow
still
the
browser's
will
respect
this
header
and
remove
all
cookies
and
all
storage
and
under
storage
is
important.
It's
not
just
storage.
It's
all.
The
new
API
is
like
access
to
next
be
serviceworker
registration,
so
it's
like
pretty
pretty
powerful
stuff
and
we
like
will
send
that
header
to
just
to
be
sure.
A
Either
there
is-
or
it's
in
my
notes-
I'm
not
certified
here
this.
They
will
for
sure
we
will
have
an
update
in
the
docs,
but
there
will
be
a
dedicated
blog
post
about
subdomains
I
want
to
give
some.
It
will
include
like
user
stories,
for
people
who
run
a
gateway
right
now
and
want
to
operate
for
people
who
run
kora
like
give
us
companion
users
and
want
to
understand
it
better
explain,
what's
changed
in
the
URLs
and
things
like
that,
so
yeah
for
sure
I'll
stop
talking
now.
C
C
D
So
your
story
I
talked
about
this
before,
where
I
can
give
a
fast
inside.
Basically,
it
was
planned
to
start
this
quarter
queuing
and
I
started
it,
but
basically
the
scope
of
it
got
bigger
than
expected
before
and
we
needed
to
go
into
q2
and
this
my
mind
go
now.
Basically,
the
big
goal
for
this
in
the
browser
context,
will
be
the
connectivity
with
nodes
on
a
second
start
to
for
nobuhle
be
way
faster,
because
we
will
already
have
two
peers
that
you
know
we
know
and
we
don't
need
to
discover
them.
D
This
will
eventually
open
other
opportunities
outside
the
browser
scope.
For
example,
if
we
have
peer
that
we
already
know,
we
might
not
need
to
go
and
connect
to
the
bootstrap
nodes,
which
will
be
cool
as
well,
but
for
another
context,
and
essentially
in
the
second
part,
there
is
I,
have
a
p0
which
is
from
refactoring
our
current
code
to
then
persist
and
add
the
key
book
as
well
to
the
second
part,
which
will
eventually
just
respect
for
the
q2.
D
But
the
goal
with
that
will
be
to
also
improve
the
performance
of
dials
through
McKellar
confidence
in
R
square.
Then
I
have
other
two
less
priority
goals
for
this
quarter.
Basically,
one
of
them
is
related
with
the
user
experience
in
just
a
bit.
We
are
having
one
of
the
biggest
pain
points
of
users,
it's
always
around
configuration,
and
so
we
are
going
to
move
in
a
completely
different
approach
in
configuration.
D
There
is
also
already
one
issue
if
any
of
you
want
to
discuss
with
us,
so
check
have
already
made
a
proposal
and
we
will
probably
be
iterating
on
that
and,
finally,
the
other
one
will
be
regarding
Liberty
node
flows,
because
these
are
things
that
we
discussed
before
we
title
as
well.
We
need
to
have
only
of
all
the
possible
flows
that
a
note
has
so
that
we
understand
what
should
be
the
priorities
that
we
have
to
improve
because
we
sometimes
might
be
working
on
stuff.
D
D
So
this
is
because
we
have
no
Liberty
noise,
which
was
implemented
by
not
factory
under
grant,
but
they
are
using
a
basically
a
crypto
library,
which
is
a
really
pain
for
our
been
delayed
as
it
will
grow,
like
maybe
twice
the
size
it's
already,
and
we
don't
just
want
to
do
it
so
before
we
go
into
enabling
noise
by
default
in
Jessup,
HP
and
consequently
GSM
FS.
We
want
to
basically
move
our
elliptical
a
bit
big
crypto
in
a
modular
way.
D
Where
you
can,
you
can
choose
what
pieces
that
you
want
to
have
and
not
like
bring
all
the
crypto
just
for
a
minor
set
of
things
that
you
might
need
to
use.
This
will
be
eventually
just
for
more
power
users
until
we
get
to
like
really
easy
documents
which
new
users
can
like
easily.
Yet
then
the
other
one
for
Jacob
is
multi
other
his
offers.
We
have
this
issue
for
a
while.
Basically,
we
need
to
be
able
from
a
multi
adder
to
just
resolve
it.
D
It's
Dainius
content
and
we
currently
don't
have
that,
and
it
will
be
extremely
important
for,
for
example,
we
have
to
know
an
issue
on
WebRTC
deploy
because
of
that
and
the
remaining
one.
It's
something
that
already
came
from
q1,
which
is
the
WebRTC
signalling
and
the
connect
connection
operate
specs,
and
that,
since
any
questions.
B
Yeah
but
the
pierced
or
improvements
who's,
the
best
person
like
it
should
we
engage
our
collaborators
to
be
able
to
just
test
that
early
I
know
with
go
I'd
be
fast.
We
have
like
that
early
testing
program.
What
would
it
be
worth
talking
to
folks
like
three
box
or
other
people,
that
we
know
that
are
shipping?
D
D
At
least
I
wouldn't
say
that
for
the
initial
for
milestones,
it
would
be
useful
because
it
will
be
just
a
question
of
the
second
start
of
the
node,
but
when
we
start
talking
about
the
multi
other
confidence
in
peer
scoring,
that's
where
I
think
it
would
be
cool
to
have
input
from
people,
because
by
the
time
we
will
probably
need
to
tweak
some
configurations
of
the
peer
scoring
functions
and
it
might
be
good
to
have
feedback.
So
I
would
say
just
for
the
seconds.
B
Lippe
config
improvements.
This
is
something
where
I
don't
know
how
what
the
you
know
t-shirt
sizing
for
this
is,
but
the
feedback
that
we've
gotten
from
our
collab
interview
so
far
is
that
this
is
a
as
you
said,
something
people
really
really
comes
up
in
conversation,
so
I'd
love
to
be
able
to
see
if
we
can
make
this
and
sooner
rather
than
later.
So
if
it
is,
if
there's
a
priority
discussion
here,
we
have
direct
input
from
collabs
around
the
importance
of
any
config
improvements.
B
I
I
wonder
to
is:
are
these:
it
would
be
good
to
have
more
detail
around
what
type
of
configured
permits,
because
we've
kind
of
had
we've
had
conversations
multiple
times
around,
saying
faults
for
for
web
versus
non
web
use
cases,
and,
and
things
like
that,
and
how
we
don't
have
those
really
easily
set
up.
It
would
be
great
if
we
had
a
better
view
into
what
these
configure
boots
are.
F,
maybe
can
link
to
an
issue
or
something
like
that.
Yeah.
D
I
will
link
to
the
issue,
but
yes,
one
of
the
things
that
we
want
to
do
is
precise,
with
it
to
get
browsers
better
way
of
doing
things
instead
of
just
as
we
currently
do
mostly
go
for
notes,
defaults
and
that's
it.
We
want
to
definitely
go
on
a
way
where
we
put
the
good
defaults
for
each
runtime
that
we
have
and
I
will
point
the
link,
because
we
already
discussed
that
English.
D
A
Square
yeah,
it's
working
progress
was
overall
a
I,
see
it
like,
like
a
maintenance
quota,
so
on
the
subdomain
part,
I
did
not
finish.
The
JJ's
implementation
just
also
provides
a
ste
gateway,
so
part
of
this
could
be
like
extending
this
to
adding
interrupt
tests
for
a
gateway.
I,
don't
think
we
have
it
right
now.
A
Those
in
another
thing
is
like
the
think
of
technical
depth
in
like
a
companion,
it's
seducing,
the
old
api's,
need
to
switch
to
the
new
ones
clean
up
addressing
stuff,
that's
both
internal
and
external.
On
the
internal
side,
we
we
got
like
addressing
specs
in
multiple
places.
We
need
to
coalesce
them
close
all
the
issues
which
are
years
old
and
something
I
wrote
down.
A
They
may
be
using
WebRTC
or
Stardust
transport
and
signaling
and
they
use
the
default
one,
because
it's
there
and
it's
not
easy
clear
or
easy
to
do
that.
I
know.
Jacob
and
park
are
doing.
It's
super
super
much
easier
by
creating
docker
images
and
stuff
like
that,
but
we
have
a
problem
of
not
having
tht
in
a
web
browser
so
that
Jessica
first
may
not
be
able
to
do
everything,
for
example,
resolve
IP
and
s.
A
So
that's
why
I
probably
merge
those
everything
under
like
more
clear
roadmap,
but
for
that
we
may
leverage
delegated
routing
modules
and
people
should
be
able
to
host
their
own,
create
nodes,
delegated
routing
modules.
So
that
would
include
like
fixing
but
what's
missing
in
go
IP
affairs
and
then
going
through
that
ensuring
everything
works,
not
just
it's
written
down
that
actually
doing
their
thing
and
that
putting
heat
and
then
writing
down
sort
of
like
a
cookbook
for
someone
who
would
like
to
use
it
and
like
self
cost
as
much
as
possible.
A
B
Cool
thanks
thanks
for
sharing
it
I'd,
really
appreciate
both
you
and
Vasco
included
things
about
documenting
how
somebody's
connection
types
work
and
how
some
of
the
user
flows
and
access
pathways
work,
and
this
is
really
really
going
to
be
important
as
we
as
we
start
doing.
Improving
the
documentation
around
that
and
you
know,
Jessica
did
some
work
on
the
audit.
B
So
thanks
for
including
that,
even
even
though
it's
the
lower
priority,
this
also
might
be
the
kind
of
thing
where
for
for
really
pulling
back
a
little
bit.
This
quarter
on
some
of
these
bigger
harder
projects,
leavings
time
for
for
flexible
things
or
creative
things
are
time
to
write
things
down.
These
might
be
some
places
where
it'd
be
fun
to
maybe
even
organize.
Some
group
documentation
hacking
sessions
where
we
try
and
write
out
some
of
these
scenarios.
B
It
was
really
fun
to
be
able
to,
like
I,
took
that
one
spreadsheet
I
just
tried
to
think
of
all
the
different
ways
that
every
permutation
of
how
somebody
might
get
on
or
off
or
connected
to
I
think
that's,
not
work,
and
very
few
were
documented.
So
far,
there
isn't
really
nothing
to
point
to
for
a
lot
of
this,
so
it's
cool
that
this
is
showing
up
in
the
quarterly
goals,
any
other
questions
that
folks
have
for
Vasco
Lydell
and
the
fasco
proxy
to
take
him
since
he
couldn't
make
it
today.
A
B
This
next
one
is
me
to
the
minor
update
from
opera.
It's
gonna
ship
57
skinship.
Instead
of
this
week,
it's
gonna
ship
next
week,
so
we
should
expect
Monday.
The
current
plan
is
that
opera
Android
will
ship
with
I,
prefer
support
and
hopefully
have
a
blog
post
up
in
time
for
that
next
one
is
me
as
well
just
a
link
to
a
galley,
a
maze
of
comments
on
the
protocol
and
their
PRS
is
our
first
kinda
for
a
working
with
them.
B
Small
projects,
too,
will
hopefully
lead
to
bigger
projects,
but
these
are
experts
in
and
Linux
stuff
from
that
core
could
be
in
a
court,
could
bases
of
chromium,
WebKit
and
Catco.
So
it's
interesting
to
see
their
perspective
when
they're
evaluating
and
the
type
of
changes
that
we're
proposing
that
web
extensions
any
changes
to
web
extensions
is.
It
has
an
extra
layer
of
awesome
complexity,
because
it's
a
group
of
browsers
that
are
standardizing
around
the
technology,
that's
offered
by
one
browser
that
refuses
to
standardize.
So
you
have
a
very
interesting
market.
Sorry,
standardization,
dynamic.
A
A
E
So
that
exactly
ipfs
specific
but
I
think
hanging
up
mostly
I,
feel
you
guys
and
they
have
the
schema,
define
if
you
don't
know
and
I
screwed
together
a
very
simple
web
app
that
allows
you
to
edit
IP
LD
schema
and
that
old
syntax
highlight
it
even
show
errors,
and
so,
if
I
come
in
here
and
type
in
download
type
comment,
it'll
parse
it
and
show
the
error
bring
up
here.
So
this
is
a
very
simple
little
web
app
I
thought
I'd
show
it
since
I
killed.
A
That's
pretty
cool
I
wonder
because
we
got
I
peeled
the
Explorer,
which
gives
a
very
basic
ability
of
traversing
the
tree,
but
it's
on
a
back
burner.
But
always
we've
been
discussing
that
it
would
be
really
cool
to
move
that
if
we
move
that
project
further
to
give
ability
for
people
to
create
like
arbitrary
sibour
tags
or
things
like
that,
and
it's
super
useful
to
see
experimentation.
A
B
If
you
clear
private
data,
all
your
I,
professor
local
repos,
it
webpages,
are
gone,
gone
forever.
A
complete
and
total
data
loss,
but
the
model
of
ipfs
doesn't
really
talk
about
that
data
in
that
way.
So
as
somebody
who's
a
web
app
builder,
for
example,
who
instantiates
JSI
confess
on
their
web
page
and
the
user
starts
spinning
and
storing
data
there,
they
don't
have
that
expectation
that
that
data
is
temporary
or
we
will
be
blown
away.
B
If
we
want
a
web
page
to
be
a
first-class
application
platform,
that's
what
native
apps
had
to
do
and
they
have
to
request
access
to
local
storage
and
that's
part
of
their
permissions
model.
So
this
explicit
asked
upfront
I
think
in
the
long
term
will
lead
to
I,
could
just
being
treated
more
more
seriously
and
trusted
more
eventually
by
users.
B
If
we
set
up
these
expectations
that
their
data
is
actually
safe,
when
they're
used
against
a
webpage
and
then
the
people
building
on
top
of
us,
like
the
textiles
in
the
three
boxes,
that
people
who
are
shipping
things
instead
of
a
webpage,
also
how
these
types
of
this
expectations
set
up
by
default
and
there's?
No,
no
that
really
reduces
the
gap
for
misunderstanding.
The
safety
or
treatment
of
locally
stored
data.
Until
we
get
sync
premiums
where
people
have
different
expectations
around
whether
people
lives
generally.
B
A
So,
like
some
details
and
thoughts
on
that
which
I
to
note,
however,
each
this
change
proposed
by
Safari
is
probably
the
the
final
push
to
start
thinking
about
leveraging
the
fact
that
there
may
be
multiple
ipfs
nodes
in
web
browser
or
on
the
user's
machine.
We
may
at
some
point
those
limitations
in
your
browser
context,
without
the
browser,
extension
or
native
support
of
ifs
make
it
so
like
so
severe
to
the
point
that
Jessica
first
note
running
in
a
browser.
Page
context
is
always
like
just
an
ephemeral
cache
and
it
even
its
identity.
A
A
Creating
user
experiences
around
ipfs
notes,
discovering
each
other
and
maybe
just
like
EFS,
run
it
in
a
web
browser
asking
activists
desktop
note:
hey
I
got
this
data.
I
would
like
you
to
persist
that,
because
I'm
not
trustworthy,
definitely
that's
the
space.
We
will
have
to
explore
and
I
I
feel
with
like
announcements
like
this.
We
are
getting
like
step
by
step
closer
there
and
and
we
may-
and
we
may
explore
that
discussions
with
you
directly
about
the
notes,
the
duplicated
notes
and
also
synchronizing
data.
A
Ask
like
notes
asking
each
other
to
persist
on
data
right
now
when
I
talked
with
people
who
are
using
Jessica's
in
the
web
browser,
but
they
have
a
regular
web
app,
which
has
a
back-end.
What
people
do
is
they
effectively
already
use
JSF
FHS
ipfs
rank
in
a
web
browser.
They
use
it
in
that
ephemeral
fashion.
So
they
don't
assume
that
to
be
something
permanent.
They
don't
even
care
about
pure
ID.
A
Of
that
specific
note
running
in
a
browser
they
just
care
about,
like
I,
give
as
ad
ipfs
cat
cat
and
they
get
other
data
they
get
cid
and
then
they
immediately
are
either
connected
to
bootstrap
notes
or
preload
notes
on
the
backend
which
they
self
host
and
then
on.
The
the
persistence
happen
on
the
backend,
so
there
may
be
a
middleware
which
takes
that
CID
and
pins
it
anatta
or
some
pinning
service,
or
maybe
like
self
hosted
in
service.
But
that's
the
pattern.
A
I
observed
that
people
sort
of
already
don't
have
too
much
trust
in
that
local
storage.
Or
are
they
always
like
what
developers
always
think
about
the
storage
as
an
ephemeral
and
if
they
need
anything
to
persist
more
than
one
a
few
days,
they
usually
have
additional
like
orchestration
on
the
back
end,
so
that
will
be
interesting
both
making
it
easier
for
people
to
self
host.
That's
like
the
q2
priority
for
me
of
sorts
will
be
to
to
make
it
like
easier
for
people
to
to
enable
that
sort
of
orchestration.
A
That's
like
a
short
term
and
the
long
term.
Probably
a
keifa's
project
in
a
web
browser
could
do
much
more
in
the
regard
of
people
leveraging
a
desktop
note
or
a
native
support
that
exists
in
the
browser
on
operating
system.
At
some
point,
those
are
my
thoughts
on
this.
I
want
to
be
able
to
fight
this
and
the
direction
that
the
current
climate
around
like
privacy,
those
are
very
valid
concerns
and
that
storage
will
get
more
and
more
ephemeral,
yeah
I.
B
Well,
the
I
think
the
work
that
we're
gonna
do
into
I
unifying
the
desktop
companion,
workflow
and
experience
coming
into
this
next
couple
of
quarters
will
really
support
our
ability
to
actually
realistically
point
people
towards
that
as
a
solution
and
I
think
like
for
a
net.
For
now
that
don't
know
that's
something
that
I
would
want
everybody
who
builds
on
top.
A
A
Made
at
some
point
end
up
with
some
way
of
describing
states
of
chase
ipfs
not
running
on
a
page
that
note
will
probably
have
like
this
basic
stage
will
be
ephemeral
and
then,
if
it's
able
to
like
have
some
means
of
persistent
storage,
that
may
be
like
a
local
NGO
idea.
Fest
or
it
may
be
like
a
service
and
like
talking
with
the
service,
overly
p2p.
That
remains
to
be
seen.
But
there
will
be
probably
some
sort
of
like
estates
for
web
developers
to
it
will
make
it
easier
for
web
developers
to
reason
about
it.