►
From YouTube: Progress on TCP/mDNS in Brave, Content-Type at Gateways - IPFS GUI and Browsers Weekly, 2019-08-28
Description
IPFS Mirror: https://ipfs.io/ipfs/bafybeigxei24uhaodjoex3bo4ekuycmi366cae6fqevkv6kt3f7v5lisqq/
A
Welcome
to
GUI
and
in
web
browsers
would
recall
about
stuff
related
to
those
things
and
other
things
around
ipfs
today
is
4th
of
September,
I,
think
2018
and
let's
quickly
jump
to
agenda.
If
anyone
has
any
last
moments,
topics
drop
them
to
agenda
I
just
started
costs
like
the
first
item
is
mine,
so
that
would
be
smooth
transition.
I
wanted
to
give
a
quick
progress
report
on
work
around
embedded,
J's
ipfs
in
brief,
so
there's
a
long
list
of
things
and
one
of
those
things
is
local
discovery
and
UDP
and
TCP
transports.
A
You
can
see
those
are
not
checked
here
yet,
however,
I
should
be
able
to
give
a
quick
demo
see
where
we
are
with
that
today.
So
I
got
right
nightly
and
I've
built
IP,
first
compile
from
like
developer
branch
locally
to
make
it
faster
to
demo,
but
it
will
be
soon
submitted
as
a
PR,
so
I
started
ipfs
companion
and
you
can
see
it's
offline
because
I
don't
have
go
IP
of
us
running
locally,
it
just
browser
extension
and
it's
like
in
offline
mode.
A
But
now
we
should
be
able
to
go
well.
Let
me
move
this
all
right.
We
should
be
able
to
go
to
settings
and
switch
to
embedded
node
and
in
brave
embedded.
Node
has
access
to
Chrome
sockets
api's.
So
now,
when
we
switch
to
that
various
stuff
will
happen
on
the
right
side
of
the
screen,
so
I
will
scroll
up
here.
A
A
However,
on
the
right
side,
you
can
see
the
startup
process
of
the
embedded
note.
The
nightly
has
combined
them
inside
of
brave
yeah.
So
on
the
left
side,
where
do
I
love
that
you
can
see
it
simple,
just
ipfs
and
let's
go
to
the
pure
screen
and
we
can
see,
we
got
11
peers
and
those
peers
are
connected
with
TCP.
A
A
A
Let
me
like
yep,
so
the
bootstrap
section
here
is
a
bit
big.
However,
it's
big
for
a
reason,
because
I've
put
all
the
bootstrap
addresses
from
node
setup,
which
are
just
TCP
addresses.
There
are
also
WebSocket
addresses
and
the
preload
nodes
are
listed
as
both
WebSocket
and
TCP
addresses.
So
what's
interesting
is
that
the
order
matters?
So
if
the
first
address
is
TCP
one
and
not
WebSocket,
it
will
be
used.
A
A
At
this
point,
I'm
I'm
mostly
experimenting
with
bootstraps
and
it
seems
to
work
fine,
it's
picking
up
the
TCP
ones
and
I
confirm
the
connections
are
listed.
The
connections
established
by
J's,
ipfs,
ranking
brave,
are
TCP,
even
if
we
got
both
specified
in
the
bootstrap
nodes.
So
that's
kind
of
interesting.
So,
what's
what's
next
I
stopped
sharing,
so
I
can
see
you
and
I
see.
There
are
questions,
however.
What's
next,
for
this
is
too
and
sure
like
we
can
discover
other
notes.
A
The
next
step
is
to
ensure
other
nodes
can
discover
our
embedded
node,
and
then
we
would
have
to
break
browsers
able
to
discover
each
other
right
now.
Break
browser
is
able
to
discover
your
local
go
aqui
fast
note,
but
it
does
not
like
announce
itself
and
the
separate
topic
is
we
don't
have
the
HD
yet
I
tried
enabling
the
HD,
but
it
was
not
like
it
was
using
all
the
CPU.
So
there
are
some
performance
issues
somewhere
and
we
could
use
delegated
rooting
for
until
the
DHT
is
ready.
A
However,
that
opens
a
lot
of
questions
around
performance
of
the
preload
nodes
and
delegate
roads
notes.
So
that's
the
car.
Does
they
like
it
just
an
object
where
we
are
so.
The
next
step
is
to
figure
out
how
to
announce
ourselves
to
the
network
using
local
for
local
DNS
discovery
to
work
and
then
probably
will
invest
some
time
to
get
DHT
to
the
point
where
we
don't
need
any
third
party
and
notes
any
questions.
Sorry
for
the
plonk
update,
I
think
Alan
was
first
and
then
ditching.
B
This
is
super
cool,
I'm,
very
excited.
That's
great
I
have
a
question
regarding
like
the
discovery
and
the
transport
like.
How
are
they,
how
are
they
built?
Are
they
a
separate
module,
or
are
you
just
using
different
like?
Is
it
a
fork
of
the
existing
ones
who
have
different
requires
or
like
how?
How
is
it
working
and
can
you
like
link
to
them
so
that
I
can
just
have
a
look
that
was
I'm
really
interested
so.
C
A
The
problem
is
I'm,
not
sure
I'd
need
to
look
if
we
can
make
it
discoverable
itself,
because,
right
now
the
transport
is
mostly
in
the
client
mode.
I'm
able
to
like
connect
to
multi
others,
TCP
multi
others,
but
I'm,
not
sure
if
I'm
others
are
able
to
connect
to
me.
I
need
to
see
how
it's
like
the
TCP
transport
itself
is
wired
in
Jessica
fest,
so
it
updates
its
like
ads
it
own
TCP
address
to
its
list
to
the
list
of
its
own
addresses.
Right
now.
A
B
A
Yeah
yeah
yeah,
so
yeah
we
use
the
same
polyfills
just
a
matter
of
configuration.
We
probably
will
to
have
support
for
things
like
websocket
star,
listen,
WebSockets
are
multi.
We
would
need
to
create
a
custom
lipid
to
be
bundler
function
because
right
now,
I'm
just
like
passing
the
config,
but
for
more
advanced
setup,
we'll
need
to
create
a
custom
bundle.
E
Of
had
a
same
similar
question
as
Alan
I
think
will
be
really
useful
to
share
kind
of
the
overall
architecture
approach
to
ease.
So
if
there's
either
some
documentation
about
how
you
went
through
that
process
of
building
the
polyfills,
where
the
where
they
live
so,
like
you
know,
your
poly
filling
node
API
is
that
live
in
the
web
extension
context
of
a
certain
browser
that
are
regular
genius,
I
confess
code.
Can
it
thinks
it's
in
a
friendly
environment,
the
sports
all
the
stuff
that
it
needs,
but
really
it's
just
a
fake.
E
Maybe
this
is
a
short
blog
post
or
some
documentation,
or
something
like
that
describes
that
overall
topology
of
what
lives
we're
so
that
I
think
broadcasting
up
would
also
really
help
other
browser
vendors
whenever
thinking
about
how
they
can
like
what
because
I
think
some
brow,
you
know
most
browsers
are
looking
at
this
and
they
see
I
give
SS
this
really
big
standalone
project
with
this
entire
stack
and
I
have
to
buy
it
to
all
this
stuff.
E
But
what
you
have
shown
here
is
that
you
don't
you
actually
just
need
a
few
of
these
core
network,
primitive
api's
and
an
operating
vironment
that
looks
like
node
and
call
it
good
everything.
Everything
just
works,
so
I'd
love
for
you
to
build
this,
for
you
to
share
the
simplicity
of
this
approach
so
that
others
can
have
an
idea
of
the
nuts
and
bolts
of
how
it
happened.
Yep.
A
E
So
as
soon
as
we
started
listening
on
the
network
or
broadcasting
that
we
were
accessible
as
a
resource,
one
network
from
mdns,
but
also
the
TCP
server
code
inside
a
giant
ball,
HTTP
client
think
things
we
found
places
we
found
we
found
challenges,
so
it
might
be,
it
might
be,
it
might
be
really
different,
and
this
is
a
first
time
frame
is
probably
ever
well.
Maybe
not
that
break
is
used
to
those
chrome.
Sockets
api
is
so.
E
E
B
Is
that
a
four
it
would
it?
What
would
be
cool
is
if,
if
we
had
some
some
module
or
polyfill
that
would
make
the
chrome
API
is
look
like
Lib
D
web,
because
then
we
already
have
live
like
in
Firefox.
We
can
use
live
d
web
and-
and
we
have
literally
TCP
or
the
D
Web
transport
and
discovery,
and
we
could
use
the
same
ones
in
chrome
like
browsers.
If
X
bar
is
a
little
Web
API.
A
Sort
of,
but
but
the
other
way
around,
so
that
the
problem
with
do
it
is
that
then
our
integration
from
2018
was
based
on
the
protocol
handler
API.
So
we
did
not
like
spam.
Local
HTTP
gateway.
We
just
like
injected
payload
to
the
handler,
and
that's
not
something
we
can
polish
you.
However,
we
had
discussion
around
the
develop
the
web
project
that
similar
to
what
we
have
are
in
brave.
On
top
of
life,
chrome
sockets.
We
got
polyfills
for
net
diagram
modules,
I
believe
we
have
diagram
polyfills
on
top
of
Liberty
web
I.
A
Don't
think
we
have
the
net
one,
so
there's
a
leap
D
web
or
org
on
github
when
we
started
collecting
those
polyfills
for
lib,
the
web,
yeah
I
believe
we
will
end
up
having
net
and
D
Graham,
probably
maybe
DNS
as
well,
for
on
top
of
Lib
the
web,
and
we
could
do
the
same
thing
there
that
we
are
doing
here
for
brave.
The
only
difference
is
this
like
protocol
handler
and
which
changes
the
dynamic
in
the
location
bar,
but
also
how
the
browser
handles
stuff.
So
it's
no
longer
regular
HTTP.
It's
like
something
else.
A
I
had
a
brief
discussion
with
brave,
rising
in
me
like
that,
there's
this
API
in
the
web
and
in
in,
if
they
plan
at
some
point
in
the
future
of
adding
they
more
low-level
integration.
That
would
be
a
template.
So
we
don't
invent
this.
Multiple
ideas,
just
reuse,
something
that
works
and,
what's
proven,
but
yeah,
exciting.
B
There
and
my
understanding
of
the
D
Web
was
more
kind
of
wheat.
It
was
created,
as
a
kind
of
this
is
what
we
could
expose
and
and
like
that,
so
what
I'm
getting
at
is
in
no
Jess.
If
we,
if
we
just
say
that
we
need
to
have
like
node
like
AP
eyes,
it's
kind
of
all
or
nothing
and
there's
no
like
well.
You
know
you
can
have
d
gram,
so
you
need
all
of
the
bits
of
that
for
it
to
eat,
for
people
to
be
able
to
use
it.
B
E
A
A
At
the
same
time,
we
may
not
need
like
all
the
API
is
we
I,
don't
think
we
actually
like
the
polishing.
We
use
right
now.
Does
I,
don't
I'm,
not
sure
if
it
implements
entire
net
package,
it
just
implements
enough
to
have
like
HTTP
server,
to
run
fine
to
open
sockets
by
our
transport
without
any
need
for
changing
any
code.
A
A
Dropping
config
to
overwrite
things
like
contact
type
at
HP
gateway,
so
let's
make
it
less
green,
happy
topic
yep.
So
we
have
HTTP
gateways
and
some
people
liked
would
like
to
use
ipfs
for
hosting
websites,
and
there
are
some
edge
cases
around
content
type
when
a
file
is
returned
by
our
HCP
Kate
way,
with
invalid
content
type,
which
either
break
some
images
like
SVG
images,
sometimes
get
interpreted
as
XML
or
plain
text,
and
then
browser
refuses
to
render
it
even
though
it
loved
it.
A
So
there
are
multiple
use
cases
where
a
content
type
returned
by
HTTP
gateway,
causes
problems
or
makes
things
harder
for
people,
and
we've
been
discussing
potential
potential
solutions
of
how
to
solve
this,
how
to
enable
people
to
specify
or
overwrite
content
type,
because,
right
now,
when
you
request
a
file
from
the
Gateway,
what
happens
behind
the
scenes
is:
go
80,
FS
or
jessep
your
files.
It's
basically
doing
mind
sniffing.
It
looks
at
the
header.
It
looks
at
a
few
bytes
from
the
header
of
the
file
and
based
on
some
magic
byte
codes.
It
recognizes.
A
Oh,
this
is
this
type
of
image,
always
with
those
bikes,
so
this
is
probably
a
JPEG
or
something
so
that's
a
problem
because
it's
raw
it
relies
on
third-party
library
to
always
do
content
sniffing
in
a
deterministic
way
and
those
libraries
get
updated.
Also,
we
look
at
the
file
extension.
However,
new
the
file
extension
may
change
or
may
not
always
be
present.
A
So
there
are
some
examples
here
on
the
issue
I
linked.
However,
the
potential
solutions
are
one.
First,
we
discussed
is
to
just
embed
content
type
with
the
data
itself.
So
when
you
add
a
file
to
ipfs,
you
would
pass
a
flag
and
specify
this
is
an
XML
file
or
this
is
a
zip
file
and
that
content
type
would
be
added
to
metadata
and
would
travel
with
the
data
itself.
The
problem
is
that
requires
like
changes
to
low-level
plumbing
and
also
it
changes
the
dug
itself
cause
when
you
die
add
something
to
the
dog.
A
It
changes
the.
If
you
add
something
to
metadata,
it
will
change
the
final
CID.
So
if
you
already
have
stuff
on
ipfs-
and
you
want
to
add
content
type
to
it,
this
way
the
CID
would
change.
So.
The
second
idea
is
something
we
discussed
with
Steven
on
Monday
and
also
eric
mentioned
made
a
good
case
along
those
the
same
lines
in
a
related
discussion
about
metadata
for
unique
service
v1.
A
So,
basically,
the
idea
is
to
instead
of
embedding
content
type
in
the
dag
itself
or
in
the
blocks
of
the
data,
we
would
come
up
with
some
special
directory
or
a
special
config
file
that
people
could
just
drop
without
having
them
to
no
low-level
plumbing
staff
of
ipfs.
So
the
idea
is
maybe
a
good
abstraction
is
how
historically
we
had
that
htaccess
files-
or
maybe
that's
not
a
good
example.
A
Maybe
a
better
example
is
how
git
attributes
a
file
called
git
attributes
could
be
dropped
in
any
directory
of
your
project,
and
you
could
define
you
could
override
the
type
of
file
you
could
define
that
this
file
is
a
binary
and
it
should
be
not
have
like
underlined,
characters
translated
and
stuff
like
that.
We
could
come
up
with
something
similar
for
content
types,
so
just
a
file
you
drop
in
a
directory
saying
this
file.
Is
this
content
type
and
then,
when
someone
loads,
the
fat
data
from
HTTP
gateway,
the
HTTP
gateway
itself
would
check.
A
Initially,
that
could
be
just
the
content
type,
but
also
we
could
customize
error
pages
that
way.
So
it's
just
like
an
idea
separate
from
embedding
content
type
in
the
unit's
FS
metadata,
and
it's
probably
something
we
should
discuss
with
the
Gateway
team,
but
I
think
it's
also
related
to
web
browsers,
so
wanted
to
pick
your
brains
if
anyone
has
any
questions
or
yeah,
so
the
question
from
Alain
is
who
would
create
this
file
so
the
way
I
see
it
is
that
those
config
files
would
be
created
by
website
owner.
A
B
A
B
I
get
it
on
the
kind
of
reading
end,
but
I'm
more
like
when
it's
first
written,
oh
that's
created
like
is
it
like,
if
I
add
stuff
to
I
BFS,
there's
ipfs
create
that
file
for
me
so
that
when
the
Gateway
reads
it
can
it's
there
like
this,
there's
a
kind
of
keeping
it
in
sync
issue,
but
also
creating
it
in
the
first
instance.
Yes,
yeah
I
have
questions
yeah,
like
just
question,
marks
a
better
idea.
Yeah.
A
Yeah
so
I
think
the
good
way
to
think
about
this
is
it's
an
override,
so
by
default
nothing
changes.
The
Gateway
will
continue
doing
content
sniffing.
If
in
the
future,
we
have
this
metadata
in
UNIX
FS,
then
it
would
use
that.
However,
there
would
be
a
simpler
way
of
overriding
content
type
that
does
not
require
changing
in
UNIX,
FS
or
anything.
A
Yes,
it's
not
something
people
would
do
all
the
time.
It
would
be
only
for
those
rare
cases
when
someone
needs
to
fetch
data
from
HTTP
gateway,
but
it
needs
to
have
a
specific
content
type
and
it
and
that
person
wants
to
have
the
same
content
type,
no
matter
from
which
gateway
it's
fetched.
So
that
is
why
this
configuration
would
travel
with
data
itself
similar
to
like
the
get
attributes
config
file.
A
Instead
of
like
what
hugo
linked
in
Android
X,
you
can
override
content
types,
but
that's
like
specific
to
the
server
and
that's
the
problem
like
if
we
change
that
only
on
our
Gateway
other
gateways
would
continue
having
the
same
thing.
So
that's
why
the
idea
is
to
have
it
travel
with
data
itself
and
it
just
scoped
to
the
HTTP
gateway,
yeah.
E
A
A
Specifying
content
type,
when
you
publish
a
website
to
ipfs,
you
don't
have
just
one
text
file:
you
have
a
directory
and
in
that
directory
you
have
index.html.
You've
got
some
assets,
you
may
have
some
some
directories
and
that
how
you
would
overwrite
a
content
type
in
this
scenario
is,
you
would
just
add
this
dot.
Ipfs
subdirectory
with
additional
metadata.
A
E
But
I
guess
I'm
asking
it's
like
the
gateways
are
a
transition
point
to
a
world
where
we
don't
need
gateways
anymore,
right
and
and
in
like
you
know,
right
now,
where
were
open
five
or
less
and
not
even
20.5
in
these
implementations.
So
if
there's
any
time
to
think
about
what
are
the
walls
that
were
hitting
in
terms
of
protocol
design
and
I'm
in
use
cases,
this
would
be
the
like
run
out
of
time.
E
If
you
wanted
get
to
a
little
point
out
at
some
point,
and
so
I
guess
for
is
I,
guess
I'm
asking
the
bigger
the
broader
question
is
it?
Are
we
gonna
hit
this
again
later?
Are
we
gonna
just?
Is
this
pushing
kicking
the
can
down
the
road
till
I've
confessed
this
everywhere
and
we're
like
man
I
wish
we
had
content
types
built
into
the
protocol
built
into
the
metadata
and
I
feel
like
there's?
E
No,
there
will
never
be
a
better
time
to
ask
that
question
and
then
we'll
never
be
a
better
time
to
make
backwards,
incompatible
changes.
So
I
guess
like
that.
I
feel
like
the
context
for
from
making
these
types
of
decisions,
we
can
pretty
much
to
work
into
something
that's
gateway,
specific
and
website
specific.
But
if
the
protocol
has
a
broader
issue
around
this,
that
I
would
like
to
answer
that
question
before
going
down
the
solutions.
Road,
oh
yeah,.
A
Totally
and
I
sort
of
linked
that,
like
on
the
ideas
to
Explorer
on
the
issue,
I
linked
there
is
link
to
addressing
this
at
the
protocol
level.
So
I
want
to
be
a
very
specific.
This
discussion
is
only
about
content
type
at
HTTP,
HTTP
gate
is
there
anything
we
can
do
sooner
to
make
it
easier
for
people
to
publish
a
website
and
host
the
websites
on
ipfs
before
that
happens.
A
So
if
there's
proposed
change
in
UNIX
FSB
to
to
support
embedding
content
type
with
the
data
itself
and
that's
sort
of
what
you
mentioned,
that
it's
all
sort
of
solves
the
problem
at
the
protocol
level,
because
once
you,
when
you
add
this
metadata,
when
you
add
data
to
ipfs,
it
will
be
always
there
and
when
gateway.
When
you
request
that
data
from
ipfs
the
Gateway
will
see
that
oh,
this
data
already
has
content
type
and
it
will
not
know
to
do
anything.
It
will
just
use
the
content
type,
that's
shipped
with
the
data
itself.
A
A
Cause
we've
been
talking
about
UNIX
fSV
to
for
years
and
I,
think
it
stills
but
got
possible
spawned,
and
maybe
we
could
make
it
easier
for
people
sooner
than
later
and
there's
a
separate
discussion
that
there
could
be
times
when
someone
wants
to
take
an
existing
data
that
has
content
type
embedded
and
still
override
that
they
may
want
to
have
some
some
specific
format
that
has
like
versioning
in
the
content
type
and
they
wanted
to
like
artificially
bumped
a
version
or
something
that
could
be
a
solution.
So
it's
not
like
or
it's
and
you
could.
E
Yeah,
sorry
I
guess
my
concern
comes
from
so
I
totally
agree
that
you,
even
once
you
have
melody
since
the
protocol.
You
might
also
need
to
be
able
to
overwrite
at
times
totally
something
we
could,
but
we
can
cross
that
bridge
when
we
get
to
it.
We
understand
better
what
the
shortcomings
that
they
don't
any
better
do.
This
approach
would
be
and
I
guess
my
concern
comes
from
things
like
the
conversation
around
WebSocket
star
and
where
we
have
lots
of
there's
like
a
history
of
stock
gaps.
A
A
It's
just
like
a
separate
feature
of
the
Gateway
itself
and
we
it
was
like
mentioned
multiple
times
so
I,
just
like
documented
the
idea,
and
it's
it's
there,
maybe
we'll
pick
it
up
with
Gateway
team
at
some
point,
I
would
I
know
that
they
are
planning
to
support
custom
error
pages.
So
if
you
like,
drop
4
for
that
HTML
that
will
like
override
the
default
Eero
page
so
I
feel
it's
sort
of
like
a
similar.
A
Yeah
yeah,
that's
true,
but
that's
what
also
so
Allen
mentioned
that
the
overwrite
won't
work.
If
you
request
data
directly
using
it's
the
idea
I
totally
agree,
but
the
specific
use
case
for
this
is
like
hosting
websites
or
like
hosting
data
set
with
specific
content
type.
So
you
are
always
wrapping
them
in
some
directory.
A
We
cannot
do
much
with
requesting
direct
like
file
directly.
There
is
like
a
filename
parameter
that
you
can
pass
and
that
would
overwrite
both
how
the
URL
ends
and
also
would
add,
content
disposition
header,
so
the
Gateway
would
return
and
suggest
the
file
name
in
content-disposition
header,
but
yeah
I
I
think
that
this
is
out
of
scope.
We
mostly
are
like
personally
I.
Will
I
just
want
to
make
it
easier
for
people
to
house
stuff
on
ipfs
house
websites
of
an
IP
offense,
and
for
now
that
means
some
sort
of
a
gateway
access.
A
E
E
A
Yeah,
so
III
did
not
look
at
the
updates
on
on
that
issue.
However,
I
know
that
we
had
issues
with
various
static
website
generators
and
the
problem.
There
was
that
in
the
code
you
have
like,
you
wrote
in
the
template.
You
use
like
those
valid
relative
paths.
However,
the
engine
which
was
generating
static
website
took
those
paths
and
broke
them,
either
by
defaulting
to
the
root
on
the
root
or
or
replacing
them
with
absolute
paths,
but
usually
when
stop
breaks.
A
It's
when
static
website
generator
replaces
those
relative
paths
with
like
related
to
the
root
and
makes
them
starting
at
the
root.
I
know
that
we
had
those
those
problems
when
Hugo,
but
there's
now
like
an
option
to
tweak
it
off.
I'm,
not
sure,
are
you
using
any
specific
website
generator
or
is
it
like
handcraft
I?
This.
E
Was
a
handcrafted
HTML
file
with
a
handcrafted,
CSS
and
JavaScript
file,
but
I
know
exactly
what
I
published
and
uploaded
and
I
tried
it.
Mr.
play
and
indirectly,
through
I,
give
her
sad
as
well
all
right.
It
would
truly
means.
It
seems
like
that
it
that
that
the
Gateway
configurations
are
different.
E
E
A
We
we
should
like
cross-check
with
at
least
one
other
third-party
gateway
just
to
ensure
it's
not
like
a
configuration
oblem
with
our
gateway,
I
think
we
we
had
some.
No,
that
was
not
related
to
this
one
I.
Probably
we
should
look
at
this
I
think
because
I'm
not
sure
if
we
can,
unless
someone
has
any
idea
on
this
code,
what
could
go
wrong.
E
It
really
looks
like
it
might
be
a
different
like
there's
the
same
to
file
else,
work
fine
when
I
infrasound
and
they
don't,
when
I,
add
patients
to
play
in
it
and
let
this
little
through
in
Fira,
so
maybe
I'm
just
a
make
a
test
that
pushes
it
to
a
bunch
of
different
gateways.
Jackson
sees
what
the
return
hundred
years
are.
Yeah.
E
That's
lots
of
gateway,
stuff
yeah
anyway,
we
don't.
We
don't
need
to
fix
this
here.
I
just
thought.
If
anybody
had
thoughts
about
what
that
issues
might
be,
that
next
side
of
my
head
was
kind
of
about
issues
like
this.
Where
do
we
fall
like
developer
experience
stuff?
So
if
I
basic
web
developer
just
doing
I'll,
give
us
add
my
files
and
I
need
some
issues
in
their
workflow.
Do
we
have
a
place
where
we
kind
of
collect,
developer
experience
and
developer
workflow
issues
broadly
I.
E
Maybe
that
yeah-
maybe
this
is
a
section
of
the
docs
I
mean
we
still
need
to
track
the
issue
somewhere.
So
if
that,
if
there
are
developer
workflow
issues
that
we
need
to
be
able
to
document,
maybe
we
file
those
and
then
in
the
docs
repo.
We
have
a
representative
somebody
who's
been
working
with
the
Ducks
group
here
today
with
us.
Mr.
Waring,
do
you
have
any
suggestions?
How
have
you
all
been
sucking
in
that
growth
around
12
or
experience
issues
and
how
to
track
them
and
where
to
put
them.
C
E
C
Essentially,
as
things
have
been
coming
in
and
we've
got
a
bigger
bucket
for
it,
we
start
organizing
things
so
I'd
say
if
you've
already
got
a
system
for
collecting
or
Clayton
those
things,
then
I
would
continue
with
that.
That's
a
good
home
for
the
meantime,
but
we
have
been
just
opening
issues
on
the
docks,
repo
tracking,
a
lot
of
stuff
yeah,
we're
trying
I
think
Deb.
You
excited
things.
We
haven't
gone
that
deep
in
the
rabbit
hole,
yeah
it's
more
kind
of
more
in
the
content
landscape.
A
C
C
C
E
E
C
E
A
There,
but,
like
all
the
mobile,
it's
pretty
interesting
proposition
because
I'm
always
every
time
I
use
like
Firefox
for
Android,
it's
like
so
it's
like
so
good
cuz
like
the
sink.
Is
there
extensions
work
and
you
have
like
one
click
to
switch
between
like
mobile
and
desktop
rendering
mode,
not
sure
if
everyone
is
aware,
it's
really
good
and
and
compiling
works
with
embedded
nodes.
It
worked
like
it
works
for
uploading
files
to
ipfs
on
the
go
register.
E
A
The
utility
aspect
of
just
like
not
inventing
too
much
but
just
like
getting
what
we
already
have
and
packaging
it
together,
not
sure
how
it
looks
on
iPhone
on
iOS.
However,
on
Android,
you
got
all
this
scheme
of
intense,
so
you
can
like
share
something
to
other
app
and
this
app
could
be
like
receiving
text
or
data
or
pictures
and
would
just
take
care
of
adding
it
to
IP
offense,
and
that's
probably
that
number
one
use
case.
People
yeah.
E
One
of
the
experiments
that
I
wanted
to
do
I
started
poking
around
at
the
web
share
API.
Is
that
that
chrome
implements
for
a
pwace
on
Android,
so
there
you
could
actually
have
like
just
ipfs
upload
to
I
PFF
says
in
the
share
menu
know
that
probably
spot
doesn't
even
require
folded.
So
some
there's
some
fun
experiments
to
be
had
you.