►
From YouTube: IPFS GUI and Browsers Biweekly 2020-06-02
Description
Meeting notes: https://github.com/ipfs/team-mgmt/pull/1161
About IPFS GUI and Browsers Biweekly: https://github.com/ipfs/team-mgmt/issues/790
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
B
B
All
right
can
y'all
see
that
yes,
okay,
cool,
it's
tiny,
let's
make
it
bigger,
so
they
sort
of
sit
in
two
different
sitting.
Two
different
flows,
one
is
within
the
settings
screen
the
existing
settings
screen
where
we've
added
a
box
at
the
top
for
folks
to
integrate
the
pending
services
of
their
choice.
The
default
is
just
local
pinning
if
you've
got
absolutely
nothing
awesome.
It's
bringing
me
coffee,
hooray!
Thank
you.
B
So
the
default
you,
if
you
don't
have
anything
installed,
it's
just
gonna,
be
local
pinning,
but
you
add
a
service
and
it
walks
you
through
this.
This
is
sort
of
an
intermediate
step.
While
we've
still
got
a
small
number
of
providers,
we
want
to
showcase
them
and
may
have
them
have
big
splashy
icons
and
to
encourage
more
people
to
do
so.
If
one
of
these
is
not
depending
service,
you
want,
you
can
add
your
own
custom.
B
One
flow
is
pretty
simple:
you
give
it
a
nickname
in
case
you
have
to
in
Fira
accounts
you
put
in
your
secret
key
and
then
auto
upload.
There's
a
couple
of
settings
for
that.
I'll
go
into
it.
I'll!
Go
into
that
in
a
sec.
You
can
visit
the
pending
services
website.
Presumably
they
will
have
some
nice
instructions,
for
you
revolve
involving
getting
started
on
web
UI
or
desktop
at
that
link
goes
to
something
external.
B
If
it's
a
custom,
fitting
service,
the
exact
same
box
except
you,
don't
have
a
pretty
icon
and
you
have
to
find
the
API
endpoint.
There's
also
a
link
here
that
says:
if
you
want
to
make
your
custom
pinning
service
available
to
other
people,
you
can
do
that.
That
link
would
go
to
somewhere
on
our
Docs
that
talks
about
how
to
integrate
with
our
pinning
API,
which
does
not
exist
yet
so
say,
you've
got
a
bunch
of
them
because
you're
a
super,
interesting
person.
B
It
says
the
number
of
files
you've
got
on
them,
the
bandwidth
used,
which
is
still
a
little
bit
hand-wavy.
So
these
are
your
auto
upload
settings
either
nothing's
gonna
get
auto
uploaded
or
everything
you
pin
to
your
local
node
will
get
automatically
uploaded
to
that
third
party
or
all
your
services
will
get
uploaded
some
more
details
for
how
you
edit
it
or
remove
it,
and
some
little
things
like
say:
you
were
Auto
uploading
everything
or
you
were
taught
oh
uploading
anything,
and
then
you
say
I
want
to
upload
on
my
pins.
B
You
can
say:
oh,
do
you
want
to
upload
it
first
linking
all
of
the
pins
that
you've
currently
got
so
there's
a
bunch
of
and
I
no
need
to
go
into
detail
about
all
of
these,
but
there's
a
bunch
of
dialogues
and
such
for
exceptions.
So
that's
how
that
lives
in
the
settings
screen
I
can
move
on
to
the
file
screen.
If
y'all
want
to
see
ur
even
have
any
questions.
B
Very
much
you
thank
you
all
right,
so
this
guy's
got
two
flows
because
we're
not
completely
certain
yet
whether
we
want
to
allow
manual
pinning
of
individual
files
but
say
here
we
are
sitting
in
the
file
screen.
It
looks
fairly
familiar,
but
you've
got
this
somewhat
amended
pin
status
column.
This
is
up
to
a
little
bit
of
discussion
right
now.
If
it's
just
locally
pin,
you
just
have
a
pin
if
it's
only
cloud
and
cloud
pinned
if
it's
only
pin
with
a
third
party
service
which
is
sort
of
a
rare
situation.
B
If
you
want
to
learn
more
and
then
there's
also,
this
is
probably
gonna
be
a
little
bit
tough
to
do,
but
in
the
idea
is
floated
that
you
have
the
standard
loading
animation
while
a
file
is
being
pinned,
and
that
may
not
happen,
but
that's
the
affordance.
If
it
does
say
we
have
the
option
for
manual
pinning
you
get
this
the
three
dots
menu.
B
This
will
tell
you
what
services
you've
already
got
and
you
could
manually
override
some
of
them
so
and
then
also
if
you
need
to
go
to
your
settings,
it'll
take
you
back
to
settings.
If
something
is
strange,
if
the
individual
override
of
the
global
pinning
rules
turns
out
to
be
too
much
to
deal
with,
then
this
just
turns
into
the
pin
and
unpin
menu
as
we
know
it,
and
if
you,
if
we've
we've
talked
about
this,
isn't
set
in
stone.
B
But
if
you
were
to
just
click
on
this
pin
status,
and
this
would
only
be
in
a
world
in
which
you
don't
have
the
override
enabled
because
there'd
be
no
other
way
to
find
this
information.
If
you
clicked
on
the
icon
and
pen
status,
it
would
tell
you
which
services
it's
pinned
with
just
with
checkmarks.
You
can't
do
anything
and
then
there's
just
a
couple
of
different
states
in
here
some
error
states.
B
If
we
can't
query
whether
something
is
successfully
pinned
we'll
add
an
overlay
icon
and
a
piece,
it's
snack
bar
that
says,
go
check
your
settings,
and
that
is
it
so
far.
I
will
stop
sharing
my
screen,
but
there
is
and
fix
the
link.
There
is
also
an
epoch
that
we've
set
up
and
github.
You
can
click
on
that
to
see
all
the
constituent
parts
questions
Rackley.
B
A
C
A
So
that's
in
EM
FS
is
implicitly
pin,
and
by
that
we
mean
it
won't
be
garbage
collected.
So
it's
like
protected
from
garbage
collection,
but
it's
not
explicitly
pinned.
So
if
you
do
API
call
for
like
pinellas,
it
won't
show
up
there
and
by
which
we
are
the
remote
pinning
services,
and
you
are
able
to
set
up
that
remote
appearing
service
to
automatically
pin
everything
in
your
m
FS.
A
So
you
got
a
copy
of
everything,
but
maybe
you
don't
want
to
automatically
backup
everything,
and
you
just
want
to
manually
mark
things
here
is
where
we
like
make
those
local
pins
more
useful
cause
in
m
FS.
If
you
pin
something
it
effectively,
does
nothing.
It
just
like
shows
us
an
explicit
pin,
but
it's
protected
either
way,
but
here
this
act
of
pinning
and
that
icon,
which
was
kind
of
useless
until
now,
has
a
meaning.
A
If
you
set
up
it's
like
a
trigger
for
triggering
remote
pin
as
well,
you
can
set
up
like
pinata
to
automatically
pin
stuff
you
lock
all
the
pins,
and
you
made
that
way.
It's
like
a
bit
clearer,
visual
language
of
controlling,
what's
pin
and
what's
not
without
going
too
deep
and
the
two
flaws
that
Jessica
shown
on
the
file
screen.
The
bottom
one
is
like
the
super,
simple,
the
simplest
one
where
you
actually
are
not
able
to
override
any
pinning
strategy.
A
Apart
from
tweaking
that
on
the
settings,
you
can
either
like
disable,
automatically
pin
local
when
it's
locally
pinned
or
in
everything.
So
it's
just
repurposing
local
pins
as
a
trigger
which
kind
of
like
bridges.
The
onboarding,
like
Lex
moves
down
boarding
for
users
who
are
familiar
with
the
idea
of
pinning
but
may
not
fully
understand.
What's
the
cloud
pin,
I
I.
C
B
And
a
lot
of
that
I
think
is
going
to
rely
on
very
good
guidance
around
the
setup
flow
for
any
of
those
new
third-party
services.
We
want
to
make
sure
that
the
documentation
we
provide
both
they
really
good,
concise
definition
of
pinning
on
its
own,
but
then
also
explains
really
what
Lyle
just
said
in
terms
of
why
maybe
introducing
the
metaphor
of
using
a
local
pin
as
a
means
for
doing
more
useful,
global
or
other
people
accessible
ends,
for
example,
so
yeah
it's
gonna,
it's
gonna
hit
a
none
really
going
on
mourning
documentation.
B
So
we
put
out
that
survey,
I
guess
it
was
about
two
weeks
ago.
That
said,
how
do
you
use
IP
FS
GUI
tools?
We
had
56
responses,
which
is
cool,
that's
enough
to
really
kind
of
trust,
the
numbers,
but
then
also
still
be
able
to
get
some
usefulness
out
of
the
individual
verbatim
responses.
I
went
over
those
to
start
with
yesterday
and
haven't
dug
too
deep,
but
a
couple
of
interesting
points
is
kind
of
everyone
uses
companion,
that's
cool
desktop
is
a
little
bit
more
of
an
all-or-nothing
sort
of
scenario.
B
People
either
said
they
used
it
often
or
never,
which
was
interesting
web
UI,
peek
sort
of
in
the
middle
of
that
range
and
then
CLI
is
fairly
universally
used.
This
is
something
interesting
because
this
aligns
with
what
we
have
seen.
You
know
what
we're
the
the
audience
assumption
we're
working
on
right
now
is
that
most
respondents,
they're
saying
there's
about
seventy-two
percent,
say
they're
using
IP
FS
for
learning
how
to
use
IP,
FS
and
hacking
on
ipfs.
B
But
this
it's
interesting,
because
that
motivation
is
nearly
a
first-place
tie
with
people
who
are
saying
I'm
using
IP
FS
to
publish
websites
or
share
things
share
or
publish
articles.
The
share
files
was
a
close
third
place.
This
is
something
interesting
that
I'll
go
through
in
a
second
will
note
that
the
best-loved
features
are
just
like
real
super
gooey
stuff,
visual
indicators
for
network
traffic
for
bandwidths.
B
The
number
of
nodes
connected
in
your
bar
in
your
browser
right
click,
actions,
drag
and
drop
stuff
like
that,
so
this
I
think
at
a
very
high
level
is
telling
us
we're
not
gonna,
be
able
to
align
on
a
very
well-defined
funnel
if
you
want
to
use
sort
of
marketing
language
around
this,
to
get
people
to
learn
more
and
engage
more
with
ipfs,
and
that
may
be
what
we
need
to
be
thinking
about.
You
know:
we've
been
so
focused
on
the
single
hacker
audience
really.
B
So
so
this
is
gonna
get
interesting.
You
tentatively
I
might
think
that
that
splits
out
along
the
hackers,
the
the
very
well
informed
business
users
and
then
the
people
who
just
wanted
to
build
a
website
on
ipfs.
Now
the
the
people
who
were
early
adopters
but
who
may
not
necessarily
care
passionately
about
the
web,
for
example,
is
like.
Oh,
this
is
a
cool
new
thing.
B
Let
me
play
with
it
so
watch
this
space
I
will
share
more
stuff
as
I
get
more
stuff,
but
but
it's
interesting
just
to
see
that
there
really
was
not
a
clear
winner
in
any
of
this.
Any
questions
on
that
I.
B
Alright
I
mean
they're,
pretty
close,
they
are
pretty
close,
but
it's
very
it's
very
definitely
publishing
websites
sharing
files
and
hacking
and
then
there's
like
everything
else,
and
that
includes
people
who
responded
that
they're
using
IP
FS
as
part
of
its
you
know
critical
to
their
open
source
project
or
it's
critical
to
their
to
their
business
model,
or
something
like
that
which
you
know
I
expected
to
see
those
low,
but
maybe
not
that
low.
Oh,
that
was
interesting,
is.
C
B
B
Summon
some
so
it
was
a.
It
was
a
good
set
of
focus,
questions
to
to
get
you
some
an
initial
answer
to
that
question.
The
answer
to
that
questions
fairly
murky,
even
just
based
on
the
distribution
of
who
uses
what
there's
some
some
nuance
and
some
of
the
other
questions
where
people
are
like
oh
yeah,
I
was
I
was
very
surprised
to
see
a
lot
less
confusion
around
like
accidentally,
trying
to
open
multiple
nodes,
for
example,
then
we
were
suspecting
and
then
the
distribution
of
the
well.
B
We
just
flat-out
asked
them
if
we
worked
out
desktop.
What
would
that
make
your
life
better
worse
or
not
care
that
was
sort
of
an
interesting
one
too?
So
so
we
may.
We
may
actually
want
to
approach
this
more
as
just
sort
of
getting
an
idea
of
audience
segmentation
at
this
point,
especially
now
that
we're
sort
of
distracted
with
the
pinnings
there
is
stuff,
but
but
it
does
give
us
it
does
give
us
some
initial
guidance
on
both
of
those
and
answer
your
question.
Yes,
thank
you.
A
A
Assume
you
can
see
something
so
yes,
subdomain
support
for
long
LC
IDs.
There
are
two
fixes
I'm,
not
sure.
If
I
mentioned
that
on
this
call,
specifically
I'll
mention
it
on
the
other
calls.
So
long
story
short
is
the
longer
see
IDs
do
not
fit
within
63
character,
limit
of
a
single
label
for
the
end
as
compliant
the
main
name.
A
It's
not
a
problem
right
now,
but
in
the
future,
when
we
decide
to
switch
to
a
longer
hash
by
default
or
when
people
start
using
edy
205
519
keys,
which
would
probably
at
some
point
switch
go
FS
as
a
new
default
that
will
go
over
when
encoded
to
base
42.
So
the
ways
to
fix
it,
involve
either
splitting
the
label
or
changing
the
encoding,
and
each
one
has
pros
and
cons.
A
Splitting
is
easy.
Does
you
can
simply
take
the
CID
and
add
a
dot
after
63
characters?
The
problem
is
that
causes
problems
at
subdomain
gateways.
When
people
want
to
provide
TLS
certificate,
there
are
no
multi-level
certificates.
You
can
only
create
a
wild-card
for
a
single
level,
so
the
problem
is,
if
you
split,
you
can
only
get
a
third
for
single-level,
so
we
looked
at
ways
to
fix
it.
A
A
A
A
Problems
will
arise
anyway
from
base
36.
We
need
to
add
support
to
dots
across
the
ecosystem.
It
will
land
in
the
next
version
of
go
IP
affairs
and
Jessica
fest,
but
you'll
take
at
least
like
one
or
two
release
cycles
until
our
ecosystem
is
capable
of
recognizing
this
new
encoding,
which
is
not
that
popular
like
base,
52
was
much
more
popular
and
it
still
was
pretty
hard
to
get
supported
everywhere,
and
there
are
still
like
some
libraries
like
the
Java
HDPE
library,
which
I
looked
at
today,
which
does
not
support
base
52.
A
So
that's
this
problem,
space
and
sort
of
where
we
are
right
now,
I
think
the
action
item
here
if
anyone
is
interested
is
to
comment
on
the
base
46
by
default
in
all
contexts
and
I
just
want
to
highlight
that
like
paste,
42
are
those
see
IDs
that
start
with
Buffy
like
that
they
start
with
B,
but
the
rest
is
usually
the
same,
and
the
base
46
gives
us
CA
DS,
which
start
with
k2,
which
is
also
cool
yeah.
So
that's
it
mechanic
hope.
Hopefully
it
was
useful
to
a
bit
of
information.
Any
questions.
A
It's
actually
not
it's.
It's
a
hard
question
because
depends
what
you
mean
by
compatibility.
So
that's
the
CID
is
the
content.
Identifier
and
the
identifiers
used
in
IP
FS
have
text
representation
and
they
have
like
the
byte
representation
on
the
wire.
So
here
we
don't
talk
about
changing
the
bytes
representation,
so
all
like
the
networking
lit
b2b
stuff
everything
on
the
wire
and
programmatic
interfaces
that
will
continue
to
work
and
will
remain
backwards
compatible.
The
problem
is
the
tech
representation,
because
when
we
create
a
text
representation,
we
add
base
prefix.
A
So
when
you
convert
from
a
text
representation
to
the
binary
representation,
you
withdraw,
you
drop
the
prefix
and
the
problem
disappears.
But
the
problem
here
is:
if
someone
takes
that
text
and
input
it's
in
an
application
which
talks,
Li,
p2p
or
talks
ipfs,
but
does
not
support
that
specific
base
encoding.
That's
the
compatibility
problem
we
need
to
like
address
before
so.
D
C
A
A
C
Know
I
get
it
so
I
guess.
Let
me
first
I
think
switching
from
footing,
CID
after
the
slash
and
now
moving
it
to
the
sub-region
is
mostly
to
get
a
separate
content.
Separation
is
that
okay?
Is
there
a
reason
why
this
could
not
be
hidden
away
from
the
end
user
thing
like
what
I'm
imagining
gateways
could
do
the
internal
ready,
Rex,
but
all
the
credit
bureaus
might
still
reflect
older
way
of
doing
it.
C
A
I
totally
agree
what
our
subdomain
gateways
do
right
now,
if
you
use
a
subdomain
gateways
such
as
D
web
link,
and
you
put
ipfs
paths
or
D
web
link,
/
IP,
FS
and
C
ID,
and
that's
the
ID,
no
matter
in
which
encoding
you
put
it
there.
Our
subdomain
gateway
will
automatically
normalize
that
CID
subdomain
safe
version
and
redirect
you
to
a
subdomain
and
if
you
use
ipfs
companion,
companion,
takes
advantage
of
that
and
redirect
you
to
that
root
r
which
automatically
converts.
A
But
the
problem
is
user
ends
up
on
some
sub
domain
and
it
that
sub
domain
there
is
that
CID
and
user
may
either
copy
that
link
and
open
it
in
companion
or
some
other
tool
which
understands
ipfs,
but
does
not
understand
that
see
ad
encoding
or
they
may
copy
just
the
cid
because
they
know.
Oh,
this
is
a
Content
identifier
and
I
can
like
to
simply
take
it
and
paste
it
in
ipfs,
Cape,
capable
application,
and
it
will
just
fetch
the
content.
So
does
that
then
the
gap
yeah.
C
So
this
is
kind
of
what
I
was
wondering.
I
did
something
similar
in
the
experimentation
that
I've
been
doing
before
and
I
think
one
thing
that
worked
well,
there
are
some
issues
that
I'm
overlooking
is
to
have
the
top
level
navigation.
We
give
the
URL
just
for
the
end
user
browser
whenever
it's
all
to
navigation.
C
You
can
just
navigate
to
the
iframe
that
internally
do
the
thing
which
pretty
much
takes
a
whole
UI
space
and
you
wind
up
with
the
result
where
it
looks
like
say
URLs
that
you
typed
with
slash
I
could
have
slash
cid,
but
internally
it
will
be
still
origin
separated,
because
the
thing
everything
will
actually
be
loaded
under
I,
I'm
I,
don't
want
to
take
up
too
much
time.
Maybe
this
is
something
worth
talking
about
in
the
future
somewhere
else.
I.
C
A
Yeah
yeah,
actually
we
do
something
similar
right
now
in
subdomains.
So
if
you
ask
subdomain
for
a
path
using
non
web
browser
user
agent
which
does
not
follow
redirects,
you
still
get
the
payload
response,
but
we
add
like
clear
cache
that
to
avoid
situation
when
there's
like
origin
leakage,
yeah,
definitely
it's
it's
yet
another
option.
We
may
consider.
F
So
yeah
in
libel
to
be
side
of
things,
so
first
Jacob
push
the
button
was
trick
and
we
released
0.28
release
candidate
and
we
hope
that
we
get
the
final
release
this
week.
So
it
is
basically
we
last
week
we
think
about
coming
work,
that
we
expect,
whether
in
Lippitt
appeal
for
the
upcoming
weeks
and
I
will
probably
share
my
screen.
F
Okay,
are
you
saying
my
chrome,
yep,
yep?
Okay,
so
first
in
the
0.29
release,
we
we
are
working
on
shipping,
three
main
things,
which
are
the
signal
signal
ensign,
pure
hacker
trendabl
protocol
in
the
gossip
Selvi
1.1.
Basically,
the
sign
peer
records,
it's
needed
posts
for
gossip
servant
for
Geneva
Protocol
and
specifically
to
the
browser
context.
It's
mostly
the
Hannibal
protocol,
and
so
this
Phillies
will
be
focused
on
possibly
be
hardened
and
improve.
The
browser
experience
in
the
field
of
food
protocol
I
already
started
implementing
its
so
I
have
appeared
with
an
initial.
F
For
the
0.30,
we
are
planning
on
working
on
improving
the
connection
manager,
which,
from
until
now,
it's
really
basic
implementation,
and
we
want
to
as
also
a
thing
that
we
need
to
improve
in
the
we
want
to
really
focus
when
Jill
is
on
the
connection
manager
and
yeah.
That's
its
I
know.
If
anyone
has
any
questions,
I.
A
Got
sort
of
tangential
question
if
someone
asks
me
today,
what's
the
best
we
can
do
for
connectivity
in
a
web
browser.
What
is
a
better
advice
to
tell
people
to
stay
on
that
older
version
of
the
p2p
and
Jays
ipfs
before
the
refactor,
which
still
has
WebSocket
star
or
is
one
of
existing
intermediate
options,
better
I.
F
Why
I've
been
answering
some
issues
and
I
know
that
there
are
people
having
problems
with
WebRTC
star
when
using
pipiens
and
other
kind
of
stuff,
and
so
for
that
specific
use
case,
maybe
it
would
be
better
for
people
to
wait,
but
I
would
hopefully
want
people
to
update
sooner
rather
than
later,
but
yeah
like
we
are
now
we're
starting
to
work
on
the
0.29,
so
I
think
will
not
take
much
longer
until
we
have
a
better
solution
for
our
users
in
the
browser.
Yeah.
G
G
Go
is
also
potentially
looking
at
implementing
rendezvous
for
different
reasons,
so
we
may
be
able
to
leverage
some
servers
there,
but
I
think
we
also
need
to
couple
which
couple
that,
with
some
just
generic
would
be
TP
nodes
running
as
relay
servers
and
get
those
up
to
really
replace
like
WebSocket
star,
so
that
we
also
then
have
web
RTC
star
and
then
rendezvous
with
relay,
which
is
kind
of
a
more
reliable
direction
that
we're
looking
to
go
so
I.
Think
making
sure
that
we
coordinate
all
that
for
o29
with
infra
will
be
important.
A
A
However,
that
TLD,
however
companion,
is
able
to
recover
from
DNS
errors
now.
So
if
DNS
fails
companion
checks,
if
DFS
is
able
to
resolve
DNS
link
for
ETH,
we
have
a
special
logic
which
uses
eth,
/,
dns
gateway
and
dns
link
is
detected
and
companion
is
able
to
recover
and
open
that
website
from
a
local
nodes
in
the
same
tab.
So
all
that
stuff
happens
behind
the
scenes
and
if
you
enter
IP
FS
that
that
eth
in
the
address
bar
it
will
just
open.
So
that's
an
interesting
side
effect
of
contribution.
A
We
received
to
do
the
recovery
in
the
same
tab.
It
also
improved
UX
of
ENS
websites
and
I
made
a
quick
release
of
do
I,
mostly
to
squeeze
it
in
to
go
ipfs,
but
also
to
not
fall
behind
with
all
the
other
projects.
New
files,
progress
feedback
for
buy
raffle.
So
when
you
import
files,
you
you
don't
forget
what
you
imported
some
ux
improvement
for
you
windows
users
when
you
need
to
set
up
course,
headers
in
your
local
node
and
catalan
language
added.
A
Thank
you
to
everyone
who
contributes
translations
to
ipfs
translation
project.
I
know
we
not
always
are
super
fast
in
shipping.
Those
translations,
but
don't
worry,
they
are
not
lost.
We
just.
We
try
to
bump
translations
before
every
release.
Sometimes
we
need
to
skip
one
release
for
technical
reasons,
but
eventually
those
contributed
translations
will
get
to
our
projects
I'm
at
the
end
of
my
teleprompter.
So
unless
you
have
any
questions
to
releases
or
any
ad
hoc
topics,
I
will
make
30
seconds
of
uncomfortable
silence
and
then
I
will
end.
The
meeting
so
think
twice.