►
A
B
Any
see
my
screen
all
right
so
welcome
to
another
session
of
dynamic
data
and
capabilities.
We
can
do
by
the
way.
I
will.
Are
you
recording
already
there
all
right
so
for
this
session,
I'm
gonna
I'm,
going
to
coordinate
it,
I'll
be
taking
the
lead
and
I
will
need
a
note.
Keyboard.
Note
taker.
Sorry,
no
takers,
anyone
into
their
own
on
to
volunteer
for
note-taking
no.
B
B
Thank
you
very
much
that
so,
if
you
haven't
already
put
your
name
in
the
list
of
items,
please
do
it.
We're
gonna
follow
the
order
of
attendees
for
these
are
plates
and
because
we
already,
we
only
have
30
minutes
for
these
meeting.
We
just
gonna
do
the
first
part,
which
is
the
updates
session
and
you're
gonna
skip
the
second
part
of
the
demos
and
discussions
and
questions
so
I
think
that
as
I
think
the
people
missing
here
wanna
wait
a
bit,
but
if
you're,
at
the
event,
already
added
to
your
notes
to
that
gene.
B
B
I'm
Andrew
Cruz
I'm
working
for
the
working
group
that
I'm
in
beta
and
I
mainly
focus
on
identity
and
in
terms
of
my
updates,
I've
concluded
to
draft
papers
for
the
reboot
of
was
moved
by
West
Conference
that
will
be
held
in
Barcelona
in
the
in
upcoming
two
weeks.
I
think
so.
Those
draft
document
documents
were
submitted,
but
our
I
will
be
obtaining
them
and
improving
them
over
over
the
next
few
days
and
also
I've
kick-started
the
IDM
project
in
terms
of
the
management.
B
Basically,
there's
the
the
PM
IBM
pepper
on
ipfs
shipyard
I've
had
a
team
there,
I've
also
added,
but
basically
a
lot
of
issues
following
the
scrum
methodology.
So
basically
we
have
user
stories
and
then
we
have
tasks
and
we
organize
those
user
stories
and
tasks
and
screams
the
springs
have
two
weeks.
So
we
have
the
planning
by
that
by
the
start
of
the
first
week,
and
we
have
the
review
and
retrospective
by
the
end
of
the
second
week.
B
If
you
want
to
follow
the
progress,
you
may
go
into
that
link
that
I
put
there
in
the
documents,
so
you
can
follow
the
you
know
the
basic
of
the
columns
of
the
backlog
and
progress
and
then
and
review
step
we're
going
to
put
put
everything
there.
I
mean
the
Whitney
is
still
empty,
but
one
of
the
tasks
that
I
have
complete
with
me
make
everything
open
like
the
team
and
how
the
milestones
that
we
have,
and
all
of
that
I've
also
made
some
clicks
regarding
the
the
information
architecture.
B
Diagram
about
IBM
I've
been
giving
these
within
Rosa
and
I.
Think
the
one
that
we
have
now
is
is
more
final
in
terms
of
what
I'm
doing
now,
as
I
said,
I'm
working
on
the
concept
paper
and
also
The
Spectator,
it's
a
work
in
progress,
I'm
I'll
be
working
on
them
on
for
the
next
couple
of
days.
Next
steps
I
will
be
working,
as
I
said,
on
the
with
me
of
the
pmid
I'm
reco.
B
B
B
D
B
B
F
F
So,
okay,
for
this
week,
I've
completed
a
set
of
tasks
regarding
IBM.
Today,
I've
made
two
commits
one
with
the
initial
style
guide
portion
with
the
specs,
which
is
the
HTML
file
that
you
can
use
to
start
implementing.
The
this
draft
version
also
I've
made
another
commit
with
the
fonts
WebKit,
already
prepared
all
the
format's
needed
for
for
implementation.
F
Have
some
Adria
already
told,
but
we
did
some
updates
on
the
information
architecture
diagram.
There
are
still
some
naming
to
be
to
be
revised.
Regarding
some
some
details
on
flows
and
pages,
we
need
to
sync
and
just
confirm
that
everything
goes
well.
Another
thing
that
was
included
was
the
presentation
presentation.
What
is
IDM.
We
we
finished
on
last
Friday
and
presented
like
yesterday
morning
to
the
moxa
team.
Another
another
thing
that
Andrey
already
mentioned
was
the
user
stories
tasks
and
the
scrum
board
for
prepared
for
the
IDM
sprints.
F
The
next
next
steps
for
for
ADM
will
be
the
for
now
I'm,
starting
with
a
setup
block
or
user
journey.
It's
the
the
first
one
since
will
will
have
details
and
components
on
the
page
that
could
appear
could
be
translated
for
other
journeys
and
pages.
So
yeah
I'll
start
with
a
setup,
blocker
and
then
I'll
continue
with
the
lock
screen
is
returning.
This
is
but
basically
for
me,
this
will
be
for
this
week
and
maybe
the
next
one
as
well,
and
yet
that's
it.
That's
almost.
It.
G
Hi
everyone
so
this
week
and
I
did
finish
the
PR
that
I
was
opening
last
week,
and
so
the
animations
under
specified
are
almost
finished
and
we
are
missing
one
or
two
animations
related
to
the
pub,
but
now
that
we
are
focusing
on
IBM
and
it
will
be
on
standby,
but
then
most
of
the
animations
are
finished
right
now,
so
the
PR
was
was
open.
It
and
we
are
I,
am
waiting
for
and
rest
review
and
for
now
I
am
setting
up
the
the
repo
for
IBM
UIKit.
G
D
B
All
right,
thank
you.
Thank
you
for
your
plates
and
by
the
way,
I'm
not
asking
questions
we
get
is
went
through.
I,
don't
have
really
much
time.
So
we
don't
have
question
if
you
have,
if
you
have
any
question,
please
follow
up
on
our
C
so
that
we
can
follow
up
on
the
questions
and
answers
all
right.
C
C
B
C
Guys
so
last
week,
I
was
working
on
trying
to
improve
the
establishment
of
membership
for
collaborations.
So
currently
we
use
kind
of
gossip
protocol
to
establish
a
membership
CR
the
membership
CITT,
and
it
turns
out
that
sends
a
lot
of
messages.
So
I
implemented
a
solution
which
basically
uses
our
existing
collaboration,
gossip
or
a
collaboration
protocol.
C
So
there's
not
quite
so
many
messages
passed
and
it
works.
But
header
pointed
out
that
it's
not
very
scalable
so
now
I'm
looking
at
doing
away
completely
with
that
membership,
CR
DT
and
instead
using
a
kind
of
hybrid
of
the
so
the
DF
set,
is
topologically
quite
similar
to
how
chord
DHT
works
so
I'm,
looking
at
using
chord
and
possibly
Kadima
ADHD
to
setup
discovery
and
kind
of
instead
of
having
to
have
knowledge
of
the
entire
network
just
having
to
have
knowledge
of
a
few
specific
nodes
that
are
close
to
us
in
our
neighborhood.
B
C
It's
the
collaboration
membership.
So
basically,
what
we
do
at
the
moment
is
we
set
up
a
gossip
network,
that's
app
wide,
and
then
we
share
all
of
the
collaboration
memberships
for
each
particular
collaboration
over
that
gossip
network.
So
I
don't
think
we
actually
need
the
gossip
network
or
the
collaboration
membership,
and
so
it
should
decrease
a
lot
of
our
set-up
time
and
a
lot
of
complexity.
If
we
can
find
an
alternative
method
and
it'll
also
make
it
more
scalable.
B
H
The
difference
is
that
this
2.1
version
it
has
chained
elites
because
script,
articles
optimizations
and
because
everything
and
change
delete
so
a
kind
of
new
thing
it
wasn't
used
in
any
previous
version
of
RJ
or
C
key,
so
I
had
to
make
and
because
this
particular
algorithm
is
kind
of
performance
critical,
it
is
like
the
performance
critical
loops,
so
I
had
to
solve
it
with
very
primitive
data
structure.
Side
didn't
have
the
option
of
using
any
sophisticated,
latest
lectures
and
III.
Think
I
did
it,
but
I'm
still
be
bugging
it.
So,
like
a
small
disaster.
E
Just
I'm
sorry
continuing
on,
we
thought
I
started
last
week
and
Pedro
had
done
some
fixes
to
the
changes
to
the
RDA
to
fix
a
previous
bug.
I
was
finding
and
I
wanted,
just
to
see
if
I
could
reproduce.
The
the
old
bug
and
confirmed
that
it
had
been
fixed
by
the
problem
is
I
only
have
the
bug
happened,
sort
of
randomly
after,
like
doing
really
large
huge
numbers,
appears,
collaborating
together
and
random
sequences
of
events
is
really
finding
a
needle
in
a
haystack
to
try
to
figure
out.
E
You
know
what
that
particular
scenario
was
I
had
some
captures
traced
and
I
spent
a
few
days
trying
to
dig
through
it
and
try
to
reduce
it
down
to
smaller
test
case
I
basically
had
to
give
up,
because
I
could
get
it
to
like
99%
do
what
it
did
previously.
If
I
just
can
get
it
converge
to
the
exact
error
situation
on
the
old
source
code
that
it
did
so
so
that
just
sort
of
told
me
that
you
know
my
sort
of
hybrid
approach
of
doing
manual
trick.
E
Do
you
a
lot
of
manual
work
with
the
automated
tracing?
It
just
needs
to
be
fully
automated
all
the
way
through
in
terms
of
generating
these
test
cases.
So
I
ended
up
moving
on
to
the
next
step,
which
is
how
to
replace
all
the
little
console
logs
I'm
doing
with
a
real
distributed
tracing
framework
and
there's
a
few
things
out
there.
One's
called
open
tracing
one
is
called
open
census.
Open
they're
used
a
lot
at
places
like
Google
and
uber,
and
I
experimented
a
bit
with
that
and
I
have
a
demo.
D
All
right:
do
you
have
any
sort
of
test?
That's
sort
of
at
a
lower
level
choose
what
order
the
connections
come
in.
Basically,
what
I'm
thinking
is,
if
you
had
the
right,
the
the
places
where
you
have
variability
is
both
in
the
order
of
like
the
characters
that
get
typed,
but
also
in
the
way
that
the
data
flows
between
the
nodes
are
you
able
to
have
lower
control
over
how
that
goes,
so
that
you
can
then
say,
like
okay,
I'm,
randomly
gonna
try
all
these
different
orderings
we'll
see
which
one's
working,
which
ones
fail.
Yeah.
E
D
E
So
it's
it's
a
bit
like
it's
basically
fuzzing
like
we
have
a
random
there's
already
a
test
and
there's
this
random
collaboration.
So
to
me
like,
if
you're
doing
a
random
test,
that's
fuzzing,
so
we
are
sort
of
fuzzing
it
already.
It's
just
we're
sort
of
changing
that
we're
keeping
the
static
parameters
and
we're
just
letting
it
do
its
thing,
and
it's
it
just
there's
a
little
random
which
keyboard
the
delay
between
the
keys
being
pressed
and
that
besa
fuzzing
test,
so
that's
definitely
exercising
it
and
it
is
even
with
all
the
new
changes.
E
It's
it's
really
really
stable.
Now,
it's
like
maybe
works
nine
out
of
ten
times,
but
one
out
of
ten
times
with
I'm
talking
about
like
huge
numbers,
appears
and
larger
collaborations.
It'll
still
get
stuck
sometimes
so.
This
is
probably
like
little
race
condition.
Type
things
going
on
that
we
haven't
really
will
really
thought
so
at
this
one
I'll
be
able
to
capture
the
trace
capture.
All
the
spots
then
we'll
have
we'll
know
if
this
serious
events
happened.
E
So
so
we
can
say,
if
you
put
this
in
here
and
put
this
here
this
here,
they're
defector
clocks.
We
have
that
we
have
the
CR
DT,
so
we
can
reproduce
that
and
it'll
be
like.
Oh,
it's
usually
something
like
the
vector
clocks
are
being
handled
wrong.
Something
and
then
I
think
would
be
very
interesting-
would
be
making
changing
the
the
the
sink,
the
sinking
machinery.
So
it's
a
little
bit
more
testable
and
that
we
can.
We
can
sort
of
factor
out
the
randomness
that
you
get
from
the
networking.
E
D
I
D
D
We
couldn't
quite
figure
out
what
was
going
wrong,
but
we
figured
out
where
it
was
happening.
So
I
rewrote
the
integration
test
to
use
a
local
like
a
test,
specific
bootstrapping
node
that
you
would
sort
of
instantiate
and
have
all
of
your
internal
nodes
connect
to
instead
of
using
your
local
daemon-
and
that
was
that's-
that's
working
nicely
and
quickly
now
and
then
I've
been
putting
some
thought
and
chatting
with
with
Durkin
with
Cole
a
little
bit
about
how
I
want
to
do.
Basically.
Also
this
membership
problem,
I
was
originally
gonna.
D
Just
try
and
use
like
pub/sub
has
the
same
issue
right
when
you
try
and
get
into
it
when
you
try
and
get
into
a
channel
it's.
How
do
you
figure
out
how
to
propagate
the
messages
amongst
the
members
of
the
channel
and
seeing
if
there's
like
a
common
interface,
that
I
can
extract
from
mesh
or
gossip
sub
and
go
ipfs
and
the
stuff
that
we
have
going
on
in
peer
base
and
make
something
similar
and
then
also
trying
to
sort
of
document
a
little
bit
of
my
thoughts
while
this
is
going
along?
B
D
B
B
For
instance,
you
could
module
operation
based
er
duties
on
on
the
X
and
so
on.
That
would
be
really
great
to
have
salts
from
our
site
so
that
you
can
language
and
use
that
to
build
more
awesome
things
on
top.
This
is
related
to
the
things
that
we've
discussed
in
ibid.
If
we
opted
for
ipad2
to
be
like
operation
based,
for
instance,
where
you
had
devices
with
no
devices
and
so
on,
we
can,
you
know
basically
weed
out
the
deck
and
converse
to
a
state.
B
D
C
B
If
you'd
get
this
out
just
informants
and
we
we
start
the
so
in
terms
of
the
second
part
we
have
at
least
one
demo
from
which
is
which
is
named,
you
see
with
the
tracing
of
P
waves
with
open
sea
and
Susan
Jagger
I,
don't
know
who
was
was
the
person
that
put
here
me?
It
was
no.
Has
you
have
the
your
time
in
the
demo?
You.
E
E
Okay,
the
screen
okay,
so
it's
not
much.
Basically,
I
can
run
the
random
collaboration.
Random
tested
from
pure
pad
previously
before
might
modified
it.
So
it
logged
a
bunch
of
stuff
out
to
the
screen
I'm
not
doing
that
anymore
I'm,
sending
all
the
trace
data
to
a
server
and
I'm
using
a
library
from
Google
called
open
census.
E
There's
a
competing
library
called
open
tracing,
so
this
few
different
ones
to
try
and
I'm
using
a
web
interface
called
the
server
that
I'm
sending
to
is
called
jeager,
which
is
from
uber,
so
so
I
ran
the
trace,
and
the
only
thing
different
here
is
at
the
end.
I've
got
the
captured
facing
I
click
on
here,
and
this
is
the
jaggery
UI
and
you
can
see,
though
this
Tracy
that
was
captured,
I
realize
I'm
doing
something
a
little
bit
wrong
here.
E
E
And
so
so
you
can
see
all
these
letters
got
typed
and
then
eventually
the
this
is
all
happening
pretty
quickly.
Then
it
decided
to
just
they
compared
the
vector
clocks
decided
to
push
a
batch,
and
you
can.
You
can
see
it
put
a
whole
bunch
of
letters
together
and
made
a
batch,
so
eat
efg,
hijk
and
then
it
is
sent
it
to
the
from
curiam
st-pierre
sh
seats.
So
you
can
see
down
here.
This
is
where
it
applied
it.
So.
E
E
E
Really
really,
as
well
I'm
going
to
do
some
more
work
on
this
one
I'm
going
to
try
to
be
able
to
take
take
this
raw
stuff.
You
can
get
it
out
of
the
system
as
JSON.
You
can.
Click
on
here,
increase
JSON
and
that's
all
the
data.
So
I
can
like
slurp
that
in
and
build
sort
of
determinate
reproduce
the
search
of
the
events
that
happen
in
a
more
deterministic
way,
and
if
we
actually
find
bugs,
we
can
turn
them
in
unit
tests.
E
Think
the
big
thing
is
we've
driven
most
of
the
small
bugs
out
of
the
system
with
manual
techniques
that
was
happening
is
the
bugs
that
are
still.
There
are
really
really
really
hard
to
reproduce
and
find,
and
these
sort
of
fuzzing
techniques
and
spitting
out
huge
number
of
peers
and
lots
lots
of
collaboration.
It
does
tickle
it
in
lots
of
different
ways
that
things
happen,
but
then
you
can't
figure
out
what
happened,
because
it's
a
needle
in
a
haystack
sort
of
situation.
E
So
this
will
help
us
really
take
it
to
make
it
like
a
hundred
percent,
solid
and
I
think
that'll
be
really
good
for
pure
pad
and
any
other
apps
that
we
build
on
it
based
on
peer
base
yeah
and
this
type
that
these
tools
are
actually
useful.
Like
there's
a
bit
experimentation
on
the
file
coin
team
and
ipfs
cluster
team
with
distributed
tracing
team,
so
I'm
talking
with
those
people
to
compare
notes.
C
I
E
Yeah
I
think
there's
a
lot
of
things
we
can
do
with
the
code
after
having
this
get
into
it,
I
think
like
just
adding
on
additional
layers
of
testing
forces
you
to
clean
up
the
code,
because
you
can't
instrumented
in
the
right
places
unless
it's
structured
in
a
certain
way
and
I.
Just
just
doing
this
type
of
work
forces
forces
the
code
to
get
better.
So
you
know
I,
think
you're.
E
Having
some
more
events,
I,
don't
know,
I
think
I
think
it's
this
getting
bad
with
it
like
there's
a
whole
there's
a
lot
of
things
that
I
understand
with
these
tracing
tools.
Where
they
talk
about
they
can
you
can
Artie
I'm
not
using
it,
but
you
can
actually
like
match
packs.
So
you
can
have
things
running
on
multiple
servers
and
you
know
messages
are
being
sent
back
and
forth
either.
You
know
RPC
or
HTTP
load
balancers,
that
sort
of
thing
and
these
tools
all
have
ways
to
match
the
packets
together
and
propagate
context.
E
B
B
E
Yeah
I
mean
especially
in
JavaScript
land.
You
can
do
things
like
you
can
use
proxy
JavaScript
objects.
You
can
actually
create
folks
to
like
map
into
things
that
you
don't
want
to
modify.
You
can
do
things
like
that.
There's
a
whole
there's
a
whole
other
world
of
apart
from
this
distributed.
Tracing
there's
like
Chrome
has
a
whole
tracing
framework
built
into
it,
which
is
separate
for
the
things
in
this.
So
you
can
tracing
at
that
level,
but
it
it
would
be
up
to
you
totally
get
the
data
from
one
chrome
instance
and
others.
E
I
know
I
know
all
those
guys
knew
I
was
gone
over
the
Linux
world
and
their
greasy
reproducing
all
the
work
we
did
in
Solaris,
and
this
Linux
world
is
like
Netflix
and
there.
So
you
can
I
think
my
dream
is
to
get
like
the
the
Chrome's
world,
the
chrome
tracing
stuff,
working
together
with
the
slow
level
Linux
system
tracing
stuff
together
with
this
sort
of
like
distributed
instrumenting
swarms
of
you
know,
hundreds
thousands
of
peers
talking
together
and
capturing
a
full
system
trace
of
that
I
think
that's.
That
would
be
really
interesting.
B
B
Okay,
can
you
see
my
screen
yeah?
Yes,
all
right.
So,
regarding
the
questions,
I
have
I've,
put
one
question
for
for
has
to
discuss
so
a
question
arise
from
Pedro
about
how
can
we
report
progress
on
IBM
more
in
terms
of
the
media
meetings,
progress
reports
just
like
the
ones
you
we
have
today,
we
had
today
like
reporting
progress
in
terms
of
what
we're
doing
what
we
done
and
what
what's
next
and
it's
more
like
face-to-face
meetings
and
so
on.
B
Of
course,
we
have
the
progress
being
reported
on
the
management
repository
when
and
so
on,
with
issues
and
tasks
and
user
journeys,
but
perhaps
we
should
also
report
progress
in
a
weekly
minutes,
just
just
like
the
one
we
have
today.
So
the
question
is:
should
me
the
people
that
are
working
on
actively
on
the
IDM
be
reporting
the
progress
as
part
of
this
Meetup?
B
B
There's
two
disabilities
I
mean
we
can
do
what
what
we
have
done
today,
like
Mike
and
Reza
and
I
and
and
Pedro
gave
support
about
stuff
there.
They
are
doing
they
were
doing,
or
we
can,
for
instance,
in
the
next
hour,
like
like,
say
to
from
5
to
6
p.m.
GMT
plus
0.
We
could
have
meeting
dedicated
to
that
to
the
onion.
B
My
feelings
is
that
if
we
report
progress
of
IDM
inside
these
Mira
will
over
extend
it
probably
every
time
because
we
are
already
3
and
Paulo
Paulo
is
already
joining
joining
the
team,
so
perhaps
we
we
will
always
be
over
schedule
and
doing
the
record
session.
So
I
don't
know
if
you.
What
are
you
sure
about
this,
and
what
are
your
thoughts
about
this.
E
D
Yeah,
so
it's
two
things:
aye
aye,
there's
probably
like
a
little
bit
of
an
intermediary
you're
right.
It
probably
should
be
a
separate.
There
should
be
a
separate
thing
for
a
lot
of
like
the
details
of
what's
going
on
inside,
but
it
may
be
worth
having
at
least
some
part
or
some
interface
of
like
what
it
is.
D
That's,
like
the
dynamic
data
thing
and
what
is
the
separate
project
thing
and
since
we
I
think
it's
hard
to
see,
cuz
I'm,
a
mobile
I
think
we
thought
we
have
Molly
in
the
call
and
I
feel
like
she's,
been
doing
a
lot
of
this
recently
with
how
different
groups
interact
with
each
other
I'd
be
interested.
What
she
has
to
say.
I
As
long
as
you
can
also
get
new
people
excited
about
stuff
and
they're
aware
enough
about
what
you're
doing
that
they
can
come
in
and
help
out
and
suggest
things
to
you.
So
I
guess
you
know
one
one
potential
path
for
you
guys
is
to
continue
using
this
as
kind
of
status,
update
and
then
have
a
second
identity
meeting
to
talk
about
all
of
the
hard
nitty-gritty
details
of
like
what's
blocked
this
week,
kind
of
like
the
sprint
process
that
you're
trying
to
go
through.
What
are
people
trying
to
accomplish
on
how
to
do
it?
I
I
They
have
like
a
status
update
section
where
people
try
really
hard
to
fill
out
status
updates
ahead
of
time,
and
then
they
have
a
second
half
of
the
meeting
that
people
can
drop
off
work
they're,
not
interested,
which
goes
into
kind
of
more
detail
on
things
things
you
could
kind
of
go
that
route
again:
I'm
I'm
agnostic
here,
I'm
just
happy
to
be
here.
Okay,.
B
Thank
you
Thank,
You,
Molly
I,
think
we
are
more
doing
the
third
option
that
he
said
right
now.
We
have
like
a
30
minutes,
wonderful
dates,
and
then
we
have
30
minutes
of
questions
and
and
demos,
and
perhaps
if
you
want
to
discuss
some
something
in
in
deep
and
we
can-
we
can
do
that
during
the
second
time.
B
Regarding
the
other
two
options:
I'm
more
inclined
to
do
the
first
one
with
a
catch
which,
which
is
you
know,
have
a
separate
identity,
mediate
meeting
or
call
just
like
this
one,
but
with
it
with
a
twist,
which
is
one
of
one
of
the
the
persons
doing
doing
the
identity
projects
will
be
also
part
of
this
meeting
dynamic.
Later
to
just
give
you
give
you
a
quick
overview
of
how
the
project
is
going
and
a
quick
up
later
on
it
as
well.
B
I
I
B
Mean
I
think
is
also
in
transit,
so
it's
quite
difficult
for
him
to
be
here
and
we
had.
We
had
issues
about
the
zooms
account
because
I
think
only
few
people
control
the
working
group
zoom
accounts,
so
we
can't
really
eat
or
start
the
meeting
and
records
using
the
geniculate
and
capabilities,
zoom
account
or
channel
public
zoom
zoom
channel.
So
that's
why
we
are
using
the
zoom
ins
account
three
of
these
meeting.
B
So
all
right.
So,
regarding
this
specific
question,
I
think
I
will
suggest
doing
a
second
identity,
meetup
or
meaning
and
I
will
try
to
schedule
perhaps
right
after
this
meal,
so
that
if
people
want
to
be,
you
know
interested,
they
can
just
you
know,
take
two
hours
of
their
time
to
dedicate
to
the
DBC
and
if
they
are
also
interested
in
identity,
they
they
take
the
second
hour
right
after
to
attend
to
the
identity
meeting
as
well.
C
B
C
B
J
The
challenge
also
is
at
in
10
minutes.
Is
the
w3c
credentials
working
group,
which
is
a
lot
of
the
sustained
research
actually
coming
out
of
that?
So
I
usually
attend
that
meeting
and
I
apologize
I.
You
missing
this
meeting
for
months
and
months
and
so
I
think
it's
just
tough
to
for
scheduling.
I
did
get,
can
give
it
just
a
quick
update
before
I
in
the
next
ten
minutes.
All
right
think
I
took
me.
I
couldn't
log
on
early
enough
to
actually
give
you
an
update.
J
So
if
you
guys
recall,
I
wrote
the
IP
ID
spec
for
the
dig
method,
so
basically
as
publishing
at
first
just
JSON
to
IP
NS,
but
now
actually
the
updated
version
of
it
is
IP
LD
and
there's
some
issues
right
now
with
the
the
way
the
w3c
is
actually
approaching.
This.
It's
not
deterministic
and
I've
been
pushing
them
to
use
C
bore
as
a
deterministic
serialization
of
the
the
JSON
they're
actually
coming
around
to
that
and
actually
they're
coming
around
to
using
cozy
versus
Jose
for
the
signature
suites.
J
So
that's
a
C
bore
serialization
of
signatures.
Are
the
Venge
ADA
beauties
in
the
Jose
package,
put
the
link
in
for
my
updated
paper
from
Toronto
from
the
rebooting
web
of
trust,
which
is
basically
using
IP
LD
as
a
framework
for
publishing
the
did
document,
this
actually
Kim
Duffy
who's
on
the
credentials
working
group,
the
Derby
3c,
is
on
the
co-author,
as
well
as
Kristen
mud
fits
from
you
port.
So
actually
there's
a
lot
of
traction
with
you
port
to
scale
there.
J
A
big
issue
is
in
transactions
is
that
they
have
to
pay
the
gas
cost
for
each
of
their
transactions,
so
they
have
yet
to
go
to
name
net
because
it'll
be
cost
prohibitive
to
actually
for
them
to
use
any
of
the
their
current
infrastructure.
So
they
are
actually
are
thinking
about
adopting
IP
LD
in
their
ther
stack
and
using
a
side
tree
so
basically
creating
a
Merkle
tree
and
IP
LD
and
basically
aggregating
all
of
the
transactions
into
one
transaction.
That's
rooted
in
it,
one
smart,
smart
contract,
so
I've.
Actually,
basically
you
can
get
thousands.
J
J
Peters
stuff
I
from
Bitcoin,
so
so
anyways,
but
the
good
news
is
that
I
have
a
couple
of
issues
filed
for
PRS
for
the
w3c
work.
The
issues
with
the
w3c
is
that
it's
not
deterministic
and
it's
it's
mutable
data
structures.
So
it's
is
vulnerable
from
the
get-go.
So
this
base
based
you
know,
major
security
concerns,
they're
I,
think
they're
reluctant
to
use
IP
LD,
because
IP
LD
is
a
competitor
to
Jason.
Ld
Jason
LD
is
what
man
knows.
Boonie
created
this.
It's
got
a
lot
of
traction
as
a
pseudo
RDF
model
and
I.
J
So
we
actually
Christian,
Heather
and
I
are
actually
getting
some
traction
and
other
standards
bodies
for
identity
working
outside
of
the
w3c,
which
has
a
very
much
web
2.0
API,
old-school
technology,
sort
of
approach
to
it
that
they're
that
they're,
that
they
bring
with
their
baggage,
so
I
think
so,
I
think
identity.
A
lot
of
the
work
is
coming
out
of
w3c
but
I
think
there's
some
traction
for
the
the
IP
LD
approach
that
I
implemented
it.
An
IP
ID
might
actually
get
standardized
in
other
standards
bodies.
J
So
that
said,
I'd
love
to
join
you
guys
in
in
Barcelona
and
then
in
the
end
of
the
month,
I'm,
not
sure
if
I
can
convince
my
my
main
investor
aka,
my
wife,
that
I
need
to
fly
to
Barcelona
without
her
but
I'll.
Let
you
know
I
think
I
will
submit
the
paper.
The
the
thing
I
put
in
the
in
the
in
the
notes
is
the
paper
I
wrote
from
Toronto
and
it
still
hasn't
been
finalized,
I'm,
not
sure
what
the
the
issue
is.
J
B
Right
Thank,
You
Johnny,
but
by
the
way,
could
you
please
include
any
relevant
things,
the
notes
so
that
we
can
and
I
specifically
I
really
wanted
to
dig
further
on
all
the
material
that
you
have
regarding
what
what
what
you're
doing
with
HIV
and
a
key
ID
using
IP
LD
and
all
the
you
port
guys
more
quick,
inari
on
using
ipfs
and
ability
to
storage.
Although
the
idea
of
metadata,
if
there's
many
references,
it's.
D
J
In
Duffy
for
easy,
IP
LD
as
a
framework
for
the
did
document,
I'm
I'm
using
four
did
authentication
and
the
P.
If
this
is
just
it
has
to
do
with
standards.
So
my
approach
for
did
off
is
you
basically,
you
offended
Kate
using
the
quick
protocol
using
lib
p2p,
and
that's
that,
if
you
indicated
with
a
bi-directional
cryptographic,
self
signed
certificate,
you've
authenticated,
but
we
all
have
to
pay
the
issue.
Is
everyone
has
to
be
on
the
same
curve
in
order
to
do
that?
So
he
or
you
have
to
use
the
library
like.
J
You
basically
need
to
spin
up
an
ephemeral
node
for
the
p2p.
The
issue
right
now
has
to
do
with
key
exchange.
So
if
we
do
key
exchange,
everyone
has
to
use
the
same
key,
but
everyone
has
different
keys.
There
is
either
RSA
key,
which
is
the
default
for
lib
p2p,
there's
edie
to
55.9,
which
would
sovereign
uses.
Then
there
is
sec,
P
256,
which
is
what
Christian
you
for
uses.
Then
you
have
DLS
signatures,
which
is
what
z
cash
uses.
J
So
all
of
those
are
not
compatible
in
order
to
do
did
authentication
so
that
this
is
where
we're
stuck.
As
far
as
we
have
to
have
a
a
in
our
case,
lib
peter
peas.
Endpoint
is
basically
the
p2p
peer
that
you're
dialing
up
and
connecting,
and
then
you
you,
then
I'll
send
a
kate
using
an
ephemeral
key,
and
then
you
exchange
a
shared
secret
to
actually
to
do
it
over
some
curve
to
actually
then
connect
so
the
way
I
approached.
J
It
was
I'm
right
now,
I'm
using
SEC,
p25
6k
one
and
they
did
document,
but
that's
published
to
the
master
secret,
which
is
actually
is
in
the
EDT
EDD
to
55.9.
So
that's
the
secret
key
that
actually
is
hidden
and
it's
never
revealed
to
the
user.
But
all
of
what
I'm
dealing
with
is
actually
is
in
Sec,
be
2
5
6,
K
1
is
ephemeral
as
a
hierarchal,
deterministic
wallet,
so
actually
I
can
I
can
use
one
at
a
time.
I
can
I
can
know
if
I
want
to
actually
make
it.
B
Yeah
I
really
have
to
dig
into
that
document
and
read:
what's
coming
up
with
and
I
wanna,
you
know,
step
tell
you
that
we
are
discussing
right
now
is
really
important
to
the
IDM
projects.
As
you
know,
we
are
one
of
the
formations
is
very
foul
crimes,
but
but,
as
you
said,
they
are
using
Jason
ditional
D,
which
has
the
contacts
problem
of
putting
two
two
links
to
two
HTTP
links.
B
That's
a
problem
because,
as
I
said,
they
are
multiple
and-
and
it
has
the
almost
sort
of
problems
on
that
on
the
other
end,
if
we
opt
out
on
using
verifiable
clients,
we
might
be
left
behind
in
terms
of
using
open
standards
and
being
interoperable
with
other
kind
of
wallets
or
or
ages
user
agent
that
use
those
kind
of
credentials
and
then
that's
where
we
need
to
come
up
with
a
with
a
solution
or
decision
about
that.
We
we
aren't
doing
that
without
credentials.
B
C
J
So
I
use
IP
and
s
just
because
it's
actually
easy,
and
so
it's
a
general-purpose
key
value,
store
cryptographic.
Key
value
store.
You
port
is
actually
thinking
about
using
IP
LD
but
publishing
the
IP
LDC
ID
to
a
spark
contract
and
at
one
approach
sovereign
approached
me
they're
interested
in
using
IP
LD
for
pairwise
signatures,
and
so
basically
you
don't
yet
you
basically
you
hash
it
and
you
don't
store
it
and
I
as
a
block
in
IP
FS.
J
So
it's
an
ephemeral,
so
you
can
recreate
it
and
it's
basically
you're
using
a
CID,
that's
not
stored
into
the
public
resolvable
nature.
So
it's
used
really
using
IP
LD
as
a
reusable
framework
for
publishing
added
document
and
that
that
CID
can
be
published
anywhere.
It
could
be
in
the
simplest
way
with
like
zero
lines
of
code.
When
I
originally
wrote,
it
was
basically
just
publishing
it
to
IP
NS.
The
is
the
there
is.
J
D
C
J
Heading
towards
being
good
enough
for
very
general
purpose
use
case.
So
it's
you
stole
timing
out
at
over
a
minute
for
16
peers
to
come
back.
I
resolved
it
to
three
and
and
it's
Altima
about
how
the
optimization
of
the
DHT
I
have
it
down
to
three
and
it's
working:
fine,
because
it
it's
waiting
for
three
and
then
it's
really
waiting
for
the
pub/sub
you're,
basically
listening
to
updates
for
four
pub/sub
tonight
and
I.
Think
caching,
those
so
that's
been
good
enough
for
my
use
case
and.
B
J
Then
back
it
in
Berlin
talked
about.
We
have
to
have
enough
nodes
that
I
using
elliptic
curve
by
default
and
Kyle
actually
spun
up
a
bunch
of
nodes
on
digital
ocean
to
improve
the
robustness
of
the
DHT
and
did
some
testing
and
it
it
was
better.
But
it's
still
after
an
hour
or
two,
it's
stalled
back
out,
and
actually
it
was
at
another
minute
again
but,
like
I
said
I'm
just
waiting
for
three
peers
and
it
should
be
resolving
over
experimental
pub/sub
now
so.
B
Just
to
penalize
Johnny
and
we're
going
to
start
as
we
discussed
a
new
identity
meeting
and
about
that
yeah
man
that
it's
project
so
I
will
really
have
to
have
you
there
for
a
few
meetings
that
it's
possible.
Of
course,
I.
Think
your
influence
on
my
PID
and
the
other
stuff
is
real
because
of
their
work.
So,
if
possible,
I
will
post
an
issue
on
camera
demo
with
all
the
details
about
the
meetings,
if
possible,
folded
up-
and
you
know
it's
possible
appearing
in
one
of
to
one
or
two
whenever
possible.
If.