►
From YouTube: 20181127 Dynamic Data Capabilities WG Sync
A
And
it's
recording
hi
everyone
welcome
to
the
dynamic
data
and
capabilities
working
group
thanks
for
coming
so
I'm
going
to
share
my
screen
so
that
you
can
see
the
notes
all
right
so
start,
recording
and
I
will
ask
everyone
to
put
their
names
in
their
name
in
the
list
of
attendees.
Please
and
I'll
ask
her.
Is
there
a
volunteer
to
take
notes
today.
A
And
please
everyone
ask
add
atoms
to
the
agenda
or
things
that
we'd
like
to
discuss
our
things
that
you
want
to
like
start
a
discussion
and
get
it
get
a
bunch
of
a
group
of
people
that
will
be
later
be
participating
in
another
session,
as
we
did
recently
with
the
identity
special
interests
all
right.
So,
let's
start
with
a
round
of
intros
and
updates
according
to
the
list
of
attendees
on
the
top,
all
right,
I
think
it's
that's
me.
First,
all
right!
A
So
in
that,
in
that
sense
we
had
the
first
meeting
we
created
the
IRC
channel
like
if
I
said
entity,
and
we
started
project
management,
data
repo
that
will
handle
the
DMing
of
the
IDM,
the
identity
manager,
proof
of
concept
and
spec,
and
on
on
the
code,
released
a
more
recent
version
of
the
OCI
disease
with
basically
two
well
two
small
improvements
on
the
RGA
type,
which
is
used
by
by
pier
pads
and
no
other
apps.
So,
first
one
which
is
a
better
collision
free,
UUID
generator.
A
It
was
something
that
was
only
happening
in
tests
because
it
is
theoretically
safe
across
cross
processes,
but
since
the
some
tests
were
run
the
same
process,
there
were
some
collisions
and
so
I
fixed
that
and
I
also
added
a
bunch
of
RGA
tests,
I
toughened
the
tests
and
but
but
no
required
changes.
So
the
tests
came
out
green
also
in
the
pier
star
at
realm.
I
did
a
small
initiative
of
speeding
up
the
tests,
which
was
making
some
tests
not
be
time-outs,
biased
and
be
event
based.
A
So
now,
instead
of
waiting
in
some
case
instead
of
waiting
for
some
fixed
period
of
time
and
see
if
there
is
convergence,
for
instance,
we
just
listen
to
the
events
and
test
convergence
in
that
case,
so
it's
in
the
store
at
the
time
that
the
test
run
is
it
is
optimal.
I
released
also
just
releasing
a
few
minutes
ago,
pierced
our
rap
version
0.10.0.
A
This
is
like
a
major,
it's
not
a
measure
in
terms
of
semver,
but
it's
a
major
release
with
many
fixes
and
improvements
on
the
replication
part
of
it.
So
I
was
analyzing
whether
the
deltas
were
were
optimal,
so
Delta
between
replicas
were
Ultimo,
and
they
were
in
the
case
of
two
collaborations
or
two
to
two
nodes.
Doing
a
concurrent
changes.
When
there
was
more
concurrency,
then
the
the
protocol
was
not
not
what
not
a
protocol,
but
the
internals
were
not
behaving
that
well
and,
and
so
it
will
last
resort.
A
It
sends
the
whole
state
so
sending
the
whole
state
for
large
documents
means
that's
the
encryption/decryption
happening
on
what
a
large
chunk
of
a
large
buffer
will
bog
down
any
any
app,
and
that
was
what
I
suspect
what's
happening
to
beer
bad,
and
so
hopefully
this
will
fix
some
of
the
some
some
of
the
issues
that
people
was
was
having
with
having
the
event
loop.
So
the
main
thread
occupied
a
lot
with
crypto
stuff.
That's,
let's
hope
so.
A
So
some
more
tests
coming
up
with
with
that
and
yes,
so
the
the
downside
of
that
is
that
pure
syrup
is
backwards,
incompatible
in
the
sense
of
the
replication
protocol.
But
since
we're
not
have
not
reached
yet
version
1.0,
it
first
is
hard
to
to
replicate
that
in
to
represent
that
in
cember
and
also
you
would
see.
I
don't
think,
that's
that's
a
major
major
issue
because
we
never
haven't
reach
one
zero.
C
C
On
Thursday
night,
at
the
one
of
the
dot
project
developers
in
Osaka
organized
another
sort
of
meetup
think
this
is
small
good.
The
people
went
to
that
one
and
the
talk
event
really
well
I
thought
so
I
covered
for
different
distributed
web
things.
I
originally
wanted
to
do,
but
two
visualizations
for
it,
but
I
sort
of
ran
out
of
time,
but
there's
mostly
slides,
but
it
was
pretty
well
received.
I.
Think
people
in
Japan
don't
really
know
a
lot
about
this
any
of
these
projects.
C
C
C
So
I'm
investigating
and
cleaning
up
warnings
I
want
to
take
some
of
the
test
cases
and
sort
of
establish
performance
budget.
So
that's
really
to
the
end,
end
load,
taste,
tests
and
I'm
gonna.
Stabilize
that
a
lot
further
I
think
it's.
We
should
be
able
to
I'll
see
how
Pedro's
new
changes
he
just
released
affect
that,
but
I
like
to
see
much
much
higher
numbers
in
terms
of
my
peers.
C
We
can
run
that
once
on
the
load
test
on
me,
the
next
box
and
I
saw
had
rows,
been
putting
some
work
into
the
new
interface
for
persistence.
So
I
want
to
really
look
at
that
and
how
that's
gonna
be
used
by
the
Pinner,
and
we've
been
talking
about
how
what
how
to
get
this
latest
stuff
to
play
to
the
production
peer
pad
so
discussing
whether
to
chop
out
some
of
the
UI
that's
there
and
just
make
it
a
little
bit
more,
simpler
cuz.
C
We
just
want
to
deploy
to
internal
users
and,
of
course,
through
the
production
deploy,
there's
currently
no
printer
deployed.
So
I
gotta
figure
out
where
to
put
that
next
week,
I've
committed
to
talking
to
Google
in
Tokyo
they're
doing
the
signed,
HTTP
exchanges
work
here.
So
I'm
gonna
talk
to
Lydell
about
that
and,
of
course,
I'll
write
up
my
trip
notes
and
then
go
back
to
Canada.
So
that's
it
for
me.
Thank.
A
Have
a
suggestion,
so
if
you
approve
all
the
latest
changes
and
upgrade
pure
pad
to
the
lightest
peer
stars
and
I
suggest,
and
we
deploy
deploy
Pinner
that
we
try
it
out
as
your
feet
as
the
official
not
taking
to
next
on
the
next
meeting.
Okay,
you
could
have
a
backup
but
I
think
it's,
let's
that's
a
good,
a
good
goal
to
have,
because
I'm
I've
now
wrapped
up
the
beer
star
tests
on
that
specific
use
case
for
for
peer
pods
and
in
very
high-frequency
concurrent
changes
and
I.
C
D
C
D
D
Yeah
just
wanted
to
point
out
that,
while
using
crimp
bands
today
I
noticed
a
few
times
that
I've
been
getting
glitched
outs
and
lost
my
changes
so
I
guess
that
Lake
Jim
that
lost
your
name
as
well
so
I
feel
like
our
baseline
off
performance.
That
we
have
to
like
match
is
not
that
high.
So
we
should
be
okay,
yeah
exactly.
E
A
E
I
left
everything
on
the
repo
which
has
the
older
user
journeys
and
the
style
guide
of
theta
and
then
for
Pierre
part
I
completed
the
mobile
version,
which
is
right
now
read-only,
because
we
had
that
discussion
that
the
edits
side
of
the
application
shouldn't
be
on
on
mobile
devices
in
progress
I
have
he
started
the
discuss
if
Iran
may
know
something
that
I
stopped
before
a
bit?
First
yeah,
it's
at
least
it's
starts
it's
at
some
point.
E
The
current
status
for
the
painting
for
the
painting
for
Pierre
path,
I,
have
predicted
some
examples
and
have
this
list
on
the
denotes
and
the
next
steps
for
this
will
be
discussed
with
the
team
and
maybe
provide
a
table
with
all
the
scenario
and
even
the
bootstrap
or
for
the
pages.
What
are
the
scenarios
that
that
will
imply
to
change
something
on
the
UI
level?
And
yes,
for
now
there?
That's
it.
E
F
Hey
guys,
what's
going
on
not
much
to
report
after
thanksgiving
week,
but
I've
just
been
writing
tasks
for
social
proof
and
thinking
more
about
how
it
might
be
deconstructed
and
used
elsewhere
as
modules.
So
that's
that's
really
all
that's
going
on
couple
bugs
with
Firefox
2
also,
hopefully,
maybe
there's
a
few
minutes
afterwards.
If,
if
anyone
wants
to
just
spend
two
minutes
talking
real
quick
about
software
security,
slash,
you
know
what
would
you
call
it
and
I
just
I,
don't
know
if
all
the
crypto
has
been
vetted
properly,
so
just
a
real
quick.
F
D
Okay,
okay,
so
not
a
not
a
ton
of
updates
so
successfully
completed
dental
surgery
and
Thanksgiving.
So
that's
that's
great
aside
from
that.
I
did
kind
of
like
a
mini
products
session,
which
focused
mostly
on
immediate
kind
of
needs,
as
opposed
to
the
little
bit
further
outs
like
it's
simmering
to
polish
stuff
that,
like
whatever
has
been
looking
at
so
basically
the
idea
there
was.
How
do
we
make
this
the
best
most
focused
tool
for
internal
company
note-taking
today,
without
thinking
necessarily
about
how
it
would
look
to
someone
from
outside
peel
and
so
on?
D
So
there's
an
issue
that
basically
focuses
on
kind
of
like
ruthlessly
tearing
things
out.
Look
everyone
is
froze,
am
I
was
a
glitching
out
just
now:
you're,
okay,
okay,
all
right
so
yeah.
So
there's
an
issue
linked
in
my
updates.
Please
take
a
look
at
that.
If
you
have
moments,
let
me
know
your
thoughts
basically
focuses
just
taking
things
out.
We
can
get
a
more
streamlined
thing
that
we
can
focus
on
getting
the
core
functionality
really
grates
before
we
kind
of
build
it
up
again.
D
E
C
D
I
think
I
think
that
it's
been
kind
of
pushing
for
like
just
more
openness
across
various
things,
which
I
mean
I,
think
it
could
be
a
little
annoying
to
have
a
kind
of
like
people.
Chiming
in
was
like
minor
design
input
when
it's
not
really
the
best
medium.
For
that,
but
that's
I
mean
that's
a
fine
cost,
so
I'm,
okay
with
making
public.
If.
G
Hi,
okay,
so
the
basic
sort
of
I
have
a
couple
of
graphs:
I
want
them
to
be
synchronized
in
the
same
state.
It's
working,
fine,
I'm,
pushing
off
the
more
sophisticated
approaches
to
this
then
just
sort
of
the
brute
force
send
nodes,
merge
graphs,
both
because
I
think
defining
the
interfaces
is
more
important
right
now
and
more
pressing
than
making
the
implementation
better,
but
also
because
the
I
talked
with
the
IP
LD
folks
at
their
bi-weekly
and
then
after
for
a
while
yesterday
and
they're
interested
in
this.
G
G
G
Once
this
ties
a
little
bit
into
the
diverging
stuff
that
I'm
Jeb
put
out
there,
because
it's
all
sort
of
the
same
problem
of
how
do
we
access
and
talk
about
virgin
documents
and
change
them
I'm,
putting
up
some
new
working
on
putting
up
more
specs
and.
G
A
G
G
Moving
around
graphs
that
just
contains
C
IDs
is
not
so
expensive
and
synchronizing
them,
even
with
the
you
know,
very
basic
flooding
sync
I
have
now
and
probably
will
replace
it
with
other
more
sophisticated
things
later.
Is
it's
not
too
bad
and
you
can
use
this
instead
of
having
either
a
centralized
either
needing
consensus
and
sort
of
a
centralizing
agreement
protocol
or
making
it
sort
of
you
know
single
writer
like
IP
NS
does
all
right.
So
this
is
this
is
sort
of
an
attempt
at
that.
We
will
see
how
it
turns
out.
G
My
guess
is
that
it's
going
to
be
okay,
because
it
may,
you
may
need
to
tweak
your
workflow,
for
instance,
I'm,
not
sure
how
well
this
would
work
if
you
decided
to
use
it
with.
You
know
peer
pods,
sending
every
single
character
right,
but
if
you
want
it,
if
you
batch
them
off
into
slightly
larger
groups,
then
maybe
maybe
things
get
a
lot
easier.
G
Delta
C
Rd
T's
are
good,
but
they
only
work
in
this
one
particular
use
case
in
which
you
can
optimize.
For
so
the
question
is
in
using
a
more
generic
structure
like
this.
Is
there
a
way
we
can
accommodate
more
without
compromising
too
much
by
just
saying
you
can
do
anything
and
we
will
see
what
happens.
G
A
Nice
I
suggest,
in
terms
of
of
the
pub
sub
and
the
topology
gossip
sub
I,
think
there's
been
a
recent
creation
or
gossip
well
implementation
of
gossip
suck
in
ipfs,
so
you
may
want
to
either
use
that
or
be
inspired
by
by
it
it's
so.
The
use
case
is
that
it's
optimized,
it's
not
optimized,
so
it's
not
optimized
for
single
source.
A
Whether
the
epidemic
broadcast
tree
protocol
is
typically
optimized
for
the
single
source,
so
it
like
it
works
really
well,
if
you
have
like,
as
once,
one
node
is
producing
all
the
data,
but
it
kind
of
how
do
you
say
it?
It
flaps
when
when
there
is
multiple
sources
of
data,
so
that
may
be
interesting
in
in
pursuing
the
gossips
protocol,
which
is
something
that
pier
star
I
think
should
also
at
a
given
time
use
once
it's
implemented
and
JSF
EFS
or
even
drive
the
implementation
of
that
in
joyous
at
the
offense.
A
G
A
Yeah
and
and
you'll
think
just
no
one
to
along
too
much
but
I
think
you'll
probably
reach
the
the
same
stage
as
at
pier
star
we're
right
for
rapid,
saying
you.
You
create
a
specific
protocol
for
for
seeking
me
in
our
case,
in
this
case
sorry
Delta's
and
then
synchronizing
Delta's
over
that
that
that
are
have
affected
clocks.
It's
a
very
specific.
A
So
it's
a
very
specific
protocol
because,
because
doctors
have
had
this,
this
property
of
being
joinable
and
but
not
so
much
with
operation
bias
charities.
If
you
want
to
make
sure
it's
generic
dag
like
operation
and
make
it
that
pliable
to
all
sorts
of
different
use,
cases
I
think
that's
a
good
generic
approach.
I
agree.
H
Hi
guys
first
I'm
talking
today
so
check
my
wrists
so
coming
coming
back
from
from
my
vacation,
so
first
I
needed
to
sync
with
design
working
group.
Few
things
affected
our
working
group
as
well,
so
right
now,
I
need
to
update
our
design
file
structure
based
on
the
discussions
that
we've
had
with
the
design
working
group,
and
this
will
be
very
useful
for
for
anyone
who
is
doing
work
that
has
any
any
type
of
design
work
embedded
in
it.
So
pure
your
pad
would
be
a
good
example
of
this.
H
Then
one
thing
we
still
need
to
decide
for
now
for
to
specify
we're
still
using
github
to
store
all
the
design
assets.
That's
something
that
the
design
working
group
is
still
deciding
on
where
we
will
store
all
that
information,
because
one
thing
is
storing
the
information.
The
other
thing
is:
how
do
you
actually
get
people
to
collaborate
on
those
files?
Some
of
these
all
of
these
projects
are
open
source.
H
H
Another
thing
that
is
in
discussion
there
is
how
the
DDC
and
other
working
groups
should
interact
and
request
the
request
stuff
from
the
design
working
group.
That's
we're
talking
about
the
lab
OS
how
this
should
actually
work.
It's
still
working
progress
as
well,
so
moving
on,
we
are
right
now
creating
a
table
of
possible
statuses,
as
andre
Sosa
mentioned,
especially
intervene,
ian's
status
and
messaging
for
each
of
those
intersections.
H
Once
we
have
a
solution,
we
will
present
you
to
gather
feedback,
then,
in
our
view,
this
this
this
is
a
problem
that
is
typical
to
all
peer
star,
apps
and,
ideally
on
peer.
We
should
provide
essentially
a
table
of
recommendations
of
user
experience
for
anyone
building
on
top
of
pure
star.
You
should
approach
it
this
way
and
communicate
it
this
way
for,
in
order
to
be
simple
for
the
user,
this
will
likely
be
adopted
by
Pierre
pad,
which
is
something
is
already
working
on.
H
H
What's
next,
okay,
so
one
of
the
things
now
that
we're
reaching
a
certain
stage-
and
we
want
to
make
it
easier
to
launch,
updates
and
gather
feedback
from
from
people
who
are
testing
right
now
you
have
to
install
it
manually,
so
we
want
to
put
it
more,
a
bit
more
public
on
on
the
Chrome,
Web
Store
and
advanced
for
Firefox,
again,
not
a
real
real
launch.
So
the
idea
is
not
to
have
it's
fully
functioning,
really
functional.
H
The
idea
is
to
make
it
easier
for
better
testers
to
come
on
board,
to
gather
feedback
from
those
people
and
be
easier
to
release
the
plates.
So,
where
Chrome,
Web,
Store
and
other
add-ons
for
Firefox
a
few
things
need
to
improve
it's
not
blocking,
we
will
likely
improve
the
banner
and
there's
an
issue
there,
things
that
will
likely
improve
and
there's
the
the
pier
and
pinning
connectivity
and
because
this
is
these,
the
consistency
status,
which
needs
to
be
reviewed
with
which
salsa
and
Android
groups
so
blocking.
In
order
to
launch
this,
we
need
to
finish.
H
First
began
integrating
into
discuss
safai,
that's
Andre
was
doing
there's
the
integration,
and
this
directly
relates
to
the
following
point,
which
is
that
in
order
to
so
that
we
don't
have
permanent
data
in
the
peers,
because
we
know
this
is
still
data.
We
want
to
discard
information,
so
we
will
recommend
to
implement
time
to
live
on
that
data,
so
that
we
can
keep
moving,
releasing
new
versions
and
anything
that
was
discussed
until
then.
It's
just
Victor.
H
We
will
combine
that
with
versioning
just
to
make
sure
that
new
versions
of
the
extension
don't
use
old,
stale
data.
Another
thing
that
needs
to
be
decided
is
who
owns
the
key
to
publish
extensions.
The
Chrome
Web
Store
uses
private
public
keys
for
Firefox
add-ons.
It's
something
that
Android
will
need
to
do
is
there
is
a
slightly
different
manifesto.
That
means
we
will
have
likely
a
template
that
is
used
to
generate
both
the
Chrome
and
Firefox
manifests.
H
H
Providing
feedback
on
their
own
or
just
capturing
metrics
from
from
usage
errors,
and
that
sort
of
thing
just
to
make
sure
that
we
know
what's
what's
happening.
We
will
also
probably-
and
this
is
something
we
need
to
discuss-
I
need
discuss
with
Pedro.
We
should
probably
find
influencers
and
and
the
channels
or
mediums
where
we
could
announce
this
effort
to
gather
interest
around
the
project.
This
could
be
helpful
for
the
whole
peer
pad,
sorry,
pure
app
ecosystem
and
finally,
hopefully,
this
next
cycle,
we
will
have
it
on
the
stores.
B
Doesn't
say
something
that
I
I
remember
Marco
regarding
the
versioning
and
the
time
to
leave.
We
can
actually
just
in
this,
and
we
were
that
easy.
This
is
just
a
simple
solution
so
to
to
actually
like
leak,
the
old
data
we
can
just
use
the
version.
The
major
version
and
prefix
are
prepared
it
to
the
collaboration
names.
So
that's
you
know
each
time
a
major
version
or
a
minor.
Let's
say
I'd
say
like
that,
is
bumped.
There's
a
new
collaboration
going
on
yeah.
B
Just
for
you
to
be
aware
like
we
are,
we
are
we
have
like
we
are.
We
are
being
careful
regarding
the
data
and
migrations,
so
this
is
something
that
needs
to
be
careful.
We
thought
out
and
for
now
we
were
just
looking
into
a
simple
solution
where,
whenever
the
data
changes,
the
data
is
lost
completely.
B
So
this
is,
of
course,
something
temporary,
and
this
is
something
only
for
the
alphabet
a
better
stage
until
we
we
plant
the
migrations
parts
carefully
so
yeah
we
we
can
just
test
with
two
things
that
we
can
do.
One
is,
you
know,
as
I
said,
prefix
or
suffix
the
collaboration
names
another
one
is
to
whenever
I
receive
an
a
plate
on
the
App
Store
and
it's
a
breaking
change.
I
can
just
you
know,
deleter
the
FS
repository
and
it's
gone,
and
also
the
Pinner
should
reflect
that
as
well.
So
yeah.
H
Want
one
suggestion
considering
we're
working
forward
for
the
permanent
well,
one
could
argue
that,
even
even
if
there
is
the
old
data
there
is,
that
is
no
longer
valid.
Considering
we
will
be
well,
we
have
repositories,
so
anyone
who
wants
to
actually
build
an
old
version
can
build
an
old
version
if
they
need
to
so
that
that
would
be
an
argument
for
keeping
information
in
the
penis.
Just
in
case.
A
All
right-
let's,
let's,
let's
address
this
in
a
negative
issue-
I
think
just
just
general
comments
in
that
that
subjects
that
I
well
quickly
things
before
we
settle
on
a
like
some
sort
of
versions:
dag
like
the
version
that
we'll
be
discussing
and
things
that
Adina
has
been
discussing
before.
We
gather
all
the
use
cases
and
try
and
settle
on
on
one
data
model.
I
think
that
data
loss
is
going
to
be
an
issue
I
think,
and
then
we
have
to
factor
in
things
like
like
migrating
from
one
collaboration
to
the
other.
A
Well,
that
you're
part
of
the
migration
or
not
all
of
the
strategy
or
not,
and
how
we
do
need
consensus
on
that
or
not,
and
so
it's
I
don't
think
we
do.
We
do
need,
but
are
some
drawbacks
on
that's
okay,
we
can.
We
can
perhaps
still
like
a
longer
longer,
along
a
conversation
that
we
can
have
offline
on
on
github,
for
instance,
because
there
is
like
a
lot
of
ramifications
and
on
on
any
solution.
A
B
Suggestion
Marco
is
for
you
to
create,
like
an
issue,
release,
release,
classifying
in
chrome,
store
and
Firefox
add-ons
and
just
you
know,
create
a
list
of
todos
there,
and
we
include
the
points
mentioning
this
part
of
the
data
and
the
loss
of
data
and
how
to
approach
that,
and-
and
we
know
it's
just
linked
to
that
issue-
and
we
can
discuss
that
separately.
Okay,
so
it
allows
you
allows,
has
to
know
what's
the
status
of
the
launch
and
what's
missing.
B
Okay,
so
concluded
I
finished
the
replies
in
disgust
if
I
and
also
the
lazy
load
of
the
comments,
including
the
replies,
so
I
will
actually
make
demo
if
possible
by
the
end
of
the
like
tea
session,
so
that
we
can
see
that
working
so
keep
the
creep
body
saying
that
is
their
plates
available.
So
I
can
see
my
notes.
Let
me
refresh
the
browser.
B
You
better
deal
like
okay,
so
yeah
so
concluded.
That's
it.
The
replies
regarding
the
in
progress
basically
is
is
I
I
need
to
finish
the
histories
and
the
history
and
the
versioning
of
comments
so
that
anyone
can
see
the
history
of
a
comment
either
because
it
was
elated
or
because
it
was
plated,
and
these
resulted
in
a
discussion
of
creating
a
module
called
actually
two
modules.
One
called
versa
dog,
which
is
you
know,
I
profess,
a
non
related
I
profess
module
that
supports
versioning
via
doc.
B
Issues
I
will
place
it
there
afterwards,
but
essentially
it
will
allow
it
will
allow
discuss
if
I
and
other
apps
to
add,
add
versioning
to
anything
to
any
document
or
any
comment
or
any
other
data
structure
that
you
might
want
is
really
very
abstract,
and
the
goal
is
delighted
to
develop
also
I
profess
mercy
that
which
is
just
a
wrapper
around
versa,
doc
that
simply
configures
the
right
node
and
there
with
no
node
functions
to
use
ipfs
so
that
it's
easier
to
use
so
I.
My
status
on
that
I
have
a
working
version
of
lucida.
B
I
will
do
with
just
a
few
tests.
Working
I
need
to
finish
finish.
The
test,
and
you
know,
write
to
read
me
and,
and
so
on.
I
will,
like
finish
finish,
the
module
in
one
to
two
days,
worst
case
scenario
and
then
I
need
to
integrate
the
model
into
I,
mean
to
discuss
if
I,
to
have
the
histories
and
the
version
so
blocks
in
terms
of
the
persistency
of
the
data,
in
this
case
the
Pinner
and
the
sink
while
language.
B
So
the
Pinner,
as
far
as
I
know,
is
being
integrated
into
into
the
infrastructure
and
the
CI
workflow
I
think
so
I
mean
I,
don't
know
the
status
of
that
I.
Think
Victor
is
not
here,
but
I
will
try
to
reach
him
out
or
read
the
issue
about
about
that
in
the
infrastructure.
I've
been
subscribing
to
that
issue.
B
So
they
are,
they
are
coming
with
a
solution
for
peer
pod
and
also
writing
the
table
of
all
the
combinations
of
statuses,
so
I
will
wait
for
for
what
they
come
up
with
and
then
I
will
get
back
to
it
once
we
have
a
deaf
definite
solution
for
that
next,
one
yeah,
it's
integrates
to
integrate
ipfs
for
seed
are
into
the
specify
so
that
we
have
URI.
This
will
require
a
smaller
factor,
because
the
CR
DT
at
the
moment
has
a
few
fields
that
can
be
removed
in
favor
of
diversity.
B
Lack
more
most
specifically
I,
have
the
light
step
latest
timestamp,
and
also
the
CID
of
the
comment
that
now
leave
in
the
in
diversity'
University
dog
nodes,
essentially
the
hands.
The
heads
of
diversity
dog
note
contained
all
that
information
for
me,
so
I'll
I
will
just
get
rid
of
that
and
use
diversity
dog
data
first,
for
that
by
the
way
versus
I
exist,
the
name
then
I
came
up
with
I
I.
Think
it's
like
like
a
good
night,
but
I
I,
don't
know
if
you
guys
have
any
suggestions
for
it.
B
So
next
on
next
on
I
have
a
point
here,
which
is
either
I
dedicate
the
next,
the
next
one
to
two
weeks
to
launch,
specified
or
either
I
kick-starts
identity,
slash
IDM.
So,
as
Mark
who
said,
we
really
need
to
you
know,
make
easier
for
users
to
experiment,
specify
and
user
specified
via
the
chrome
store
and
the
Firefox
add-ons
start.
B
But
to
do
that
we
need
to
do
some
stuff.
More
specifically,
we
need
to
prepare
the
manifest
of
the
accession
to
be
compatible
with
Firefox.
There
there's,
basically
a
few
fields,
some
fuels
that
are
exclusive
to
chrome
and
I
need
to
exclude
those
fields
for
for
Firefox.
So
this
is
more
like
a
build
step
phase
where
I
create
a
manifestation
for
Chrome
and
a
manifestation
for
Firefox.
B
We
need
to
come
up
with
a
solution
for
the
versioning
of
the
data
model,
so
that
you
know
we
don't
deal
with
migrations
for
now.
Also,
we
need
to
implement
a
new
banner,
slash
header,
which
is
smaller
and
also
pre
prettier,
and
also
we
need
to.
We
need
to
publish
on
the
star
by
having
your
logo,
decide
a
copy
of
the
description
of
the
app
and
also
the
the
public
key.
B
We
need
to
either
generate
a
public
key
and
past
the
private,
slash
public
key
to
the
infrastructure
team,
I,
don't
know
or
or
they
supply
a
public
key
for
us
or
they
publish
the
the
thing
we
don't
know.
I
think
we
need
to
reach
out
the
PFS
extension
or
the
team
that
did
the
extension
so
that
we
probably
follow
what
they
were.
What
they
did
in
relation
to
the
public
key
and
the
private
key
to
publish,
but
as
as
you
know
and
I've,
been
talking
about
a
few
a
few
points.
B
B
We
have
n
drivers
here,
which
is
a
Mozilla
developer
that
has
been
working
on
leave
and
also
have
Alex
from
from
protocol
as
well
to
talk
about
probably
IP,
offense
I.
Don't
know,
and
I
will
be
talking
about
this
stuff
as
well.
A
No,
so
it's
demo
time
so
I
can
have
like
five
minutes
I'm
going
to
try
and
make
it
a
quick
demo
and
if
the
demo
gods
allow
me
so
pure
pads
and
to
end
load
tests,
I'm
going
to
demoed
the
new
version
of
piercer
app
running
on
on
pure
pads.
So
does
everyone
see
my
my
screen?
Alright,
okay,
so
I'm
on
pure
pad,
checked
out
npm
run
tests.
It
should
be
loads.
A
Are
using
ciphers,
no
I'm
using
puppeteer
and
so
I'm?
What
I'm
doing
is
like
well,
typically,
the
puppeteer
tests
are
not
are
not
our
headless
but
I'm,
just
just
for
demo
purposes,
I
took
the
headless
option
out
and
so
that
we're
I'm,
injecting
random,
not
random,
sequential
characters,
but
each
pier
takes
one
character
and
so
they're
there
at
the
same
time,
injecting
characters
on
the
pads
and
hopefully,
by
the
end
of
that
that
face
they
will
all
be
converging.
If
not,
there
is
an
issue.
A
A
A
Yeah
so
demo
gods,
they
didn't
have
yeah,
they
didn't
help
them
do
that
blend.
Oh
God
didn't
help
on
this.
One
I
bet
it's
related
to
yeah.
Something
is
weird.
A
More
or
less
all,
oh,
there
are
something
blocking
some
of
these
guys,
so
there's
random
changes
being
propagated
across
across
them,
and
then
there
is
this
that
then
there
is
this
process
on
our
behind
that's
falling
for
the
states
and
making
sure
that
the
state
they
reach
the
state
of
of
convergence
which
I
believe
they've
reach
and
then,
after
that,
they
make
random
changes.
Now
we
entered
the
face
of
random
changes
on
a
patch.
A
So
again,
this
is
inserting
and
removing
characters
at
random
and
then
waiting
a
bit
and
seeing
if
all
the
the
peers
are
in
sync.
This
is
how
I
found
out
that
we
were
exchanged.
A
lot
of
timing
works
in
this
scenario:
changing
a
lot
of
state
instead
of
just
the
deltas
and
now
I'm,
printing,
the
deltas
and
the
deltas
are
pretty
small
as
you
as
you
can
see
so
back
to
the
demo.
So
the
changes
have
stopped
now
and
hopefully
they're
all
on
the
same
path,
States
and
now,
after
some
time
out,
the.
A
The
timeout
our
timeouts
are
conservatives
because
we
had
bigger
sink
times
before,
hopefully
now
they're
there
they're
smaller,
as
you
can
see
there,
they're
not
changing
for,
haven't
changed
for
a
while.
So
the
the
injection
phase-
and
it's
all
good,
so
if
you
want
to
check
check
this
out
and
and
have
in
the
tests,
are
pure
pad
tests
each
wheel
loads.
A
A
A
B
B
And
few
more
okay!
So
if
I,
if
I
refresh
I
see
that
the
replies
are
still
there
and
I
see
there,
they
are
being
collapsed.
So
only
the
latest
two
replies
are
being
shown
to
the
user
and
you
can
lazy
load
them
as
you
want.
So
you
can
expand,
expand
them.
I
will
add
a
few
more
just
quickly,
so
that
you
can
see.
E
B
We
go
for
the
level
yeah
yeah.
You
can
and
also
help
me
feel
some
comments.
Ladies
faster
I
just
need
to
reach
out
ten
comments
so
that
you,
you
see
the
load
more
working
on
okay,
yeah
Marco
is
helping
to
thank
you.
Michael
yeah
I
will
try
to
recruit
I'll,
try
to
refresh
that
and
see.
If
it's
working
for
you
yeah.
We
have
49
points.
So
if
I
open
that
it
shows
like
it's
expanded,
ten
more
and
it
expand
the
last
four
ones,
so
I
will
actually
show
just
a
lazy
load.
B
Part
I
will
just
add
a
few.
Actually
a
bunch
of
comments.
This
is
for
you
to
see.
There
was
the
lazy
load
part
working,
hopefully
that's
enough,
and
if
I
refresh,
if
I
I
scroll,
hopefully
the
connection
delay
is
not
visible.
So
if
you
see
my
scroll
position
here,
it's
you
see.
You
see
it's
lazy
loading.
The
list
of
comments
I
will
replay
that,
but.