►
From YouTube: Core Dev Call #32
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
B
Yeah,
alright,
hey
screaming
we're
live
hello,
YouTube
so
good
morning,
everyone
morning
or
afternoon.
Whatever
time
you
are
welcome
to
the
30
seconds
status,
core
def
call
as
usual,
we're
streaming
this.
So
keep
that
in
mind.
We
are
the
denotes,
have
been
shared
on
hangouts
and
will
be
shared
publicly
later
on.
B
C
B
A
A
The
role
that
Joan
is
filling
is
called
senior
product
design,
architects,
which
is
a
role
that
we
we
really
didn't
know
how
to
how
to
call
it
so
yeah
I,
just
you
just
jump
in
with
questions
or
King
John
and
like
what
it
is
exactly
that
you
can
expect
from
him,
because,
obviously
we
all
have
different
ideas
when
we
read
it.
I
know
like
that.
A
That
requires
in
in-depth
technical
understanding,
especially
of
what
can
be
done
in
aetherium
and
then
inside
or
with
the
gossip
network,
and
this
is
what
what
John
will
be
will
be.
Focusing
on
kind
of
supporting
that
process
and
very
much
bridging
between
between
design
and
death
generally
want
to
add
anything
findings
of
introducing
yourself,
yeah.
D
Hi
yeah
great
to
you,
where
it
won't
very
happy,
could
be
here
yes,
senior
product
design,
architect,
so
one
way,
I
kind
of
explained
this
is
it
user
facing
functional
design,
so
the
design
work
happens
very
close
to
the
kind
of
initiation
at
the
beginning
of
projects
and
yeah
doing
lots
of
you
know,
use
cases,
user
journeys,
activity
diagrams
and
trying
to
work
out
the
the
user
flows
and
the
user
facing
functionality
that
some
provides
a
great
experience
and
is
achievable
with
all
the
constraints
a
little
bit
about
me
before
this
I
was
previously
at
Microsoft
for
two
and
a
half
years
and
I
was
leading
the
design
of
Visual
Studio
Visual
Studio
code
and
V
STS
that
we
renamed
was
your
DevOps
before
going
over
to
the
dark
side,
which
isn't
that
dark
these
days
actually
I
was
working
at
canonical
on
the
bun
two
for
seven
years
and
Ubuntu
unity.
D
D
B
B
The
second
thing
we
want
to
touch
on
real
quick
is
a
privacy
policy.
We've
had
a
quick
meeting
with
a
legal
firm
this
morning
with
which
was
which
was
about
giving
them
context
on
how
status
works.
The
the
idea
behind
that
is
that
we
want
to
have
a
privacy
policy
that
is
human,
readable
and
not
filled
with
legalese,
but
we
also
want
to
have
something
that
isn't
generated
from
a
website.
B
F
B
Okay,
so
this
is
gonna,
be
on
Jacob
too
much
like
everything
else:
okay,
cool,
so
you're
working
on
it
keep
supposed
to
bury
a
tissue
and
PR
for
indexing
proposal.
I.
E
G
G
G
You
know
because
I
it's
something
that
comes
up
periodically
but
I
I
didn't
know
in
terms
of
priority
number
one.
What
you
know
what's
the
priority
for
implementing
something
like
this,
it
seems
like
maybe
something
that
in
general
and
in
the
larger
scope
of
things
for
this
quarter,
the
next
water
could
be
a
low
prayer
issue,
but
I
mean
for
me
personally.
It's
something
that
comes
up
with
gaps
and
I
think
it
might
be
related
to
something
Samuel
hope
to
what
Samuel
is
addressing
I.
Think
there's
some
overlap
because
yeah
yeah.
D
H
So
just
a
question
on
that:
in
the
discuss
bread,
there
were
some
questions
around
sort
of
our
encourages,
our
usage
and
and
they
mittens
on
and
for
my
memory.
We
are
currently
at
1.5
million
requests
with
our
current
active
users
yeah,
and
we
were
rated
at
5
million
yeah,
which
basically
it's
a
question
of
what
happens
when
we
get
10x
more
users.
H
B
I
mean
there's
a
couple
things
to
consider
here.
The
first
one
is
that
in
fear
is
really
not
rate.
Limiting
anyone
on
numbers
here
and
this
I
got
straight
from
the
CEO
so
effectively.
If
we
do
cross
five
million,
unless
I
mean
unless
it
gets
ridiculous,
I
I
don't
think
we're.
Gonna
have
a
problem,
especially
because
there's
so
much
goodwill
now
and
we're
already
massively
overpaying
for
for
what
we
need.
B
Although
I
mean
to
be
fair,
the
differences
between
1
million
plan
and
the
five
million
plan
and
we're
better
yeah
better
off
be
safe
than
sorry
but
I.
If
we
do
cross
the
chasm
I'm,
you
know
I'm
very,
very
confident
that
we
can
work
with
them
to
to
either
make
sure
that
they
they
won't
cut
us
off
or
we
can
negotiate
more
requests.
D
Andrew
I,
just
it's
just
going
to
say
the
so
that
so
that
the
figure
that
1.3
million
estimation
is
I
like
an
average
over
the
last
seven
days
and
was
also
like
the
the
usage
of
the
last
but
my
last
particular
day.
But
there
is
a
big
question
mark
over
whether
a
bulk
of
that
is
not
from
our
own
internal
end-to-end,
testing
and
death
builds.
D
So
again,
I
think
that
this
is
something
we
will
probably
implement.
Hopefully,
soonish
I
know
that
there
is
a
priority
again
on
user
acquisition.
It's
not.
This
is
not
difficult
thing
to
implement,
but
that
should
hopefully
let
us
know
what
our
actual
user
consumption
is,
and
that
would
give
us
a
better
idea,
for.
Why
is
10x
of
our
current
usage
right
now?
Okay,.
B
All
right
so
will
come
will
hold
off
on
the
indexing
for
now,
and
we
can
we
can
talk
about.
We
can
talk
about
creating
a
new
project
just
for
an
e
to
e
and
see
I
want
to
see
two
things.
I
want
to
see
whether
a
new
slash
free
project,
actually
a
free
project,
wouldn't
work
because
we
need
the
historical
requests.
D
Just
just
just
quickly
a
I
mean
based
so
based
on
my
understanding
of
the
Fuhrer
and
I've
used
it
quite
a
bit
and
we
can
set
up
with
the
plan
that
we've
got.
We
can
set
up
as
many
project
endpoints
as
we
want,
so
statistics
can
be
based
on
a
project
basis.
So
what
so
and
the-
and
this
is
literally
just
creating
an
endpoint-
is
not
even
five
minutes.
A
G
Hi
yeah
I
think
there
were
a
couple
of
a
couple
of
comments
from
earlier
reviews.
My
patient's
changes
or
games,
and
things
like
that.
So
all
those
have
been
fixed
and
yes,
it
doesn't
lead
to
the
PR
has
emerged
and
my
discussion
with
oscar,
I
guess,
was
not.
There
can
be
other
enhancements
down
to
the
spec,
because
right
now,
I
think
the
protocol
interactions
between
the
status
and
status
react
native
part
of
the
app.
A
I
A
A
I
Said
before
we
were
releasing
on
a
number
of
messages,
this
is
so
basically
I
think
it
is
like
say
five
minutes
have
that
you
know.
That's
fine.
You
know,
like
the
issue,
is
that,
for
example,
to
help
you
with
what
happens
in
a
beach
at,
for
example,
you
don't
know
any
messages,
so
that
is
more
messages
will
be
blocked,
but
it
doesn't
help
you
advantage.
So
you
know
especially
someone
who
sends
like
every
second
five
megabytes
of
data
of
messages
from
one
megabyte.
I
Each
will
not
be
raised
with
it,
and
you
know:
that's
gonna
burn
your
data
very
quickly,
and
so
what
we
did.
We
also
get
limited
by
data,
so
that
eunuch
and
you
can
do
it-
might
not
be
able
to
send
much
more
than
one
megabyte
per
second,
let
me
say
that
there
are
search
those
we
should
use
directly,
getting
very
naive,
consisting
essentially
what
we
do
obviously
know
how
the
worst
you
know
like
it's,
a
constant
there's,
no
educational
service
agency.
I
A
B
A
D
D
H
Topic
of
is
alternative
is
a
reason
that
can't
be
a
topic
on
discuss
because
I
Madden
lead
a
lot
of
people
in
I
used
to
base
might
have
experienced
from
other
open
source
IRS
products
that
I've
had
the
similar
problems
in
past.
So
that
would
be
a
great
hive.
Mind
thing,
I
think
unless
there's
some
specific,
very
specific
action
or
Cori
was
supposed
to
do
no.
B
F
B
E
C
E
E
B
F
E
I
Yes,
there
is
a
bit
up.
There
is
there's
a
cost
involved.
You
know
busy
parts
intervention
there
until
new
way,
so
initially,
if
status
curve
will
part
and
if
between
dastar
would
act,
0
XP
k
that's
fairly
efficient
because
you
can
stop
boxing
as
soon
as
you
understand
it.
This
is
not
mentioned
and
mentioned
ever
very
well,
but
you
don't
have
to
go
back
alright,
so
you
cross
the
string.
Keep
saying:
is
this
a
sweetness
change
this?
You
don't
have
to
go
back.
So
that's
the
key
things.
Then
there
is
the
cost
of
at
runtime.
I
You
know
that,
especially
because
I
can
tell
you
they
can
extract
it.
Pka
just
go,
but
the
name
is
dynamic.
It
changes
as
the
ENS
name
changes.
Any
message
changes
that
we
use
subscription.
You
know
that
we
don't
have
very
much
a
lot
of
data.
How
efficient
it
is
is
definitely
not
terribly
inefficient
because
it
works
fine,
but
there
are
some
questions
about
a
little
subscription,
at
least
this
one
and
so
forth.
F
H
H
Add
between
anything
right
like
a
assume
that
would
be
uncomplicated,
but
I'm,
just
wondering
if
that's
I
mean
if
there
any
unintended
consequences
from
at
the
Sun
I
think
it
out.
Maybe
it's
hard
to
tell
but
like
what
happens,
if
I
mention
someone
with
it,
if
I
don't
know
like,
could
you
get
into
weird
situations.
H
I
C
I
C
Megan,
as
might
remind
when
you
said,
caused
was
just
thinking.
Okay,
like
I,
would
like
to
go
to
a
public
channel
which
is
very
busy
and
mentioned
everyone
by
I,
don't
know
like
either
being
very
bored
or
having
a
scripted
checks
moose
on
the
channel
and
yeah.
It's
a
form
of
caching.
So
I
can't
do
this
all
over
again
by
six
year
having
every
client
what.
I
Sorry
I
mean
a
bit.
We
are
not
actually
looking
up
through
a
bad
engine
mentions
and
do
not
cause
a
lookup
in
3
and
s
because
men
should
not
only
contain
it
publicly.
So
you
don't
have
any
reverse.
With
that.
There's
a
different
system
that
we
use
for
ENS
resolving
a
message.
As
the
user,
you
send
a
message
with
we
have
its
informational
edge.
E
E
Basically,
the
only
cost
it
will
have
if
you,
if
you
do
like
these
packets
of
mansions
in
in
messages,
is
that
it
will
check
in
your
contact
list.
So
the
DNS
names
you've
already
resolved.
If
there
is
PK,
is
matching
there
and
then
for
the
rest
it
will
use,
it
will
generate
random
names.
So
so
there
is
no
remote
calls
made
for
mentions.
C
F
E
Yeah,
but
that's
for
the
sender,
so
yes,
the
sender
might
well
anyway.
What
you're
talking
about
is
the
auto
completion
and
is
almost
done
and
the
way
it
works.
You
already
know
the
the
NS
name,
because
what
what
you
were
to
completing
with
is
all
the
DNS
names
of
contacts
you
already
know,
plus
the
names
of
people
in
random
names
of
people
in
the
current
chat
that
are
loaded.
So
even
in
that
case,
but
but
first
of
all,
in
that
case
the
only
person
you're
forcing
to
do
remote
calls
would
be
you.
C
Right,
let's
say
if
I,
if
I
have
some
spare
time
and
would
do
a
lot
of
E&S,
lookups
I
mean
just
like
as
a
takeaway
I'm
I
don't
want
to
like
go
and
too
deep
into
this.
No,
but
if
there
is
a
way
that
I
can
trigger
someone
else
to
perform
a
lookup,
this
would
be
like
in
yeah
some
some
cost,
so
they
could
make
another
mode
pay.
Basically,.
E
I
I
Best
strategy,
essentially
because
it's
what
happens,
we
back
off
exponentially.
So
you
know,
like
you
said
you
know.
Basically
you
send
as
a
name
we'll
have
to
resolve
it,
but
a
zombie
it
straight
like
in
30
seconds
and
then
we'll
have
only
90
seconds
if
it
fails
and
then
with
whatever
272nd
and
then
so
on
and
so
forth.
And
after
the
day
we
stop,
we
are
trying
to
any
times
in
a
day,
so
no
unlock
it
fails.
If
it's
successful
or
busy,
we
try
it
every
30
seconds.
You
know,
there's
a
separate.
I
E
I
G
How
maybe
I
missed
this?
How
how
how
long
is
sort
of
the
relationship,
the
the
username,
the
usernames
cached
for?
G
I
C
C
I
F
It's
bad,
so
the
only
I
think
what
what
status
should
do
for
updating
these
ENS
usernames
is
subscribing
to
the
event
regarding
this
ENS
node
in
the
resolver
in
in
the
ENS
root.
You
can
ping
me
to
give
more
details
on
that
later,
but
essentially
we
we
need
a
way
to
get
events
from
the
blotchy.
Then
we
can
update
that
on
demand.
H
I
H
G
A
Have
an
epoch
which
collects
issues
related
to
ENS,
so
I,
don't
know
who's
most
likely
to
create
an
issue
this,
but
please
ping
me,
so
we
can
plan
for
it,
yeah,
I'm
thinking,
so
because
it
sounds
like
you're
you're,
identifying
the
issue
here
or
buried
your
mic,
creating
an
issue
either
of
you.
I
have.
H
G
G
I'd
say
passed
away,
and
but
it's
not
so
it's
so
what
we
want
to
do
is
to
use
OneNote,
which
will
be
sending
the
notification
via
a
notification
service
and
to
send
with
Notification
Center
means
to
send
message
to
this
vaca
model.
So
currently
we
have
stables.
Where
implemented.
It's
you're
able
notification
service
without
firebase
or
any
other
pathway
services
and
also
I,
was
using
the
goal.
Library
for
triggering.
B
I
You
want
to
give
you
know,
because
I
haven't
had
time
to
you
know,
put
everything
in
writing
still
have
quite
a
few
unknowns,
and
you
know,
like
certain
things
at
work,
still
want
to
figure
out.
I.
Think
you
description
of
where
we
are
at,
and
it's
going
to
be
technical,
but
I
don't
know,
I
want
to
do
the
time.
Instead
of
supports
quick.
I
So
essentially
like
diminishes
centralization
in
your
life,
because
it's
a
notification,
we
need
an
apple
certificate
that
certificate
is
coming
from.
You
know
that
come
from
Apple
and
it's
not
supposed
to
be
shared
right,
and
if
my
understanding
is
that,
basically
what
firebase
does
is
exactly
that
you
put
the
certificate
in
fibers,
you
trusts
and
firebase
with
your
certificate,
and
it's
gonna
take
care
of
that.
So,
if
we're
not
going
to
describe
rest,
we
need
to
have
some
service
that
is
run
by
status.
I
That
cannot
be
run
by
others
by
other
people
with
disability,
unless
you
want
to
share
the
certificate
which,
in
terms
of
the
practical
consequences,
I,
don't
think
it's
that
bad,
but
it
sounds
like
a
really
bad
idea,
so
probably
will
not
do
it.
So,
basically,
like
that
say,
that's
the
crux
of
the
issue:
we
need
to
have
a
status
on
server,
which
needs
to
be
hundred
percent
highly
available
that
actually
send
a
notification,
and
that's
you
know
like
we
can't
do
much
about
it.
I
Unless
we
want
to
share
this
so
and
now
it's
a
matter
of
you
know
how
much
do
you
want
centralized
things
around?
It-
and
you
know-
obviously
you
know
like
give
me
the
dusty
server-
is
gonna
be
always
up.
Then
we
want,
we
can
still
let
you
know
like
server
in
the
desktops
desktops
or
like
other
machines
like
with
the
man,
little
push
indication
service
and
actually
connect
with
the
stages,
notification,
server
or
firebase.
Maybe
this
is
a
good
at
the
same
stuff.
I
We
think
about
it,
and-
and
you
know
they
got
it
at
that
point.
We
can
lift
some
of
the
requirements
of
high
availability
from
those
sort
of
like
slave
notification
service,
and
so
that's
also,
we
determine
I'm
exploring,
and
another
thing
is
that
you
know
basically
what
I'm
doing
is
I'm
looking
at
the
pushing
patient
version
to
something
similar
to
that.
But
I
would
really
not
want
any
interaction.
I
Ideally,
so
basically,
what
happens
that
as
a
client
and
I
want
to
send
a
notification
from
A
to
B.
The
only
thing
I
need
to
know
about
B
is
his
public
key,
and
so
that
means
that
basically
I
don't
even
know
because
I
didn't
even
know
which
device
be
is
running,
I
just
send
a
notification,
and
then
and
and
that
gets
delivered,
and
so
that's
this
that's
great
in
terms
of
metadata
protection
is
not
that
great.
It
is
a
bandwidth.
I
It
will
cost
exactly
the
same
as
one
one
notification
message,
the
regardless
of
whether
the
other
end
is
using
Android
or
iOS,
which
is
good
for
privacy
reasons.
So
that's
one
way.
You
know
there
are
some
issue
with
this
and
then
I
can
invite
Italians.
Especially
we
talk
about
occasion
between,
like
my
desktop
and
stages
or
you
can
do,
is
something
that
is
a
bit
more
interactive.
So
you
you
ask
still
through
wacko
who
has
who's
taking
charge
of
these
push
notification,
and
some
node
is
gonna,
say
it's
me.
I
I
B
It
works,
I
mean
it
works.
If
people
keep
the
app
on
the
background,
which
is
Android
very
installed,
allow
or
or
kill
so
you,
it
will
be
night.
At
least
it
would
be
nice
if
we
offer
that
as
an
option,
but
that's
for
a
future
discussion.
What
I
would
like
to
ask
is,
like
I,
understand,
you're
doing
the
exploratory
work
right.
G
I
I
You
know
essentially,
like
you
two
different
things,
one
is
to
say
actually
say
say
it's
saying
to
Apple:
send
a
notification
to
dis
use
it
right
and
that
needs
to
be
centralized.
That
is
the
certificate.
Is
the
certificate
is
a
secret
that
needs
to
be
kept
as
a
secret
star,
centralized,
meaning,
like
only
stages,
should
have
that
power,
and
so
that's
one
cursor
and
the
other
concern
is
to
tell
a
server
I
want
to
send
this
notification
to
be
able
to
tend
to
status.
I
I
want
to
add
I
want
to
send
a
notification
to
this
user
right.
So
there
are
three
people
involved
in
the
conversation
there's
a
but
individually
for
because
there's
a
stateless
Apple
and
be
so
status
talking
to
Apple
is
centralized,
you
know
and
but
be
talking
to
stages.
It
could
be
that
you
know
that
that
that
that
B
is
not
centralized.
So
B
could
talk
to
any
server
that
server
talks
to
the
status
and
that's
the
way
to
be
centralized.
H
This,
how
would
this
select
hypothetically-
let's
imagine
that
there
are
two
arias
apps
with
which
are
published
by
two
different
legal
entities.
Could
you
imagine
that
the
legal
entity
yet
operating
there,
this
curse?
That
would
run
that
because
it
has
to
controller
certificate
and
that
would
be
the
same
sort
of
status
server.
But
then
someone
creates
the
different
personal
status,
that's
published
under
a
different
Apple
account
and
they
would
run
their
own
sleuth
service
and
then
these
other
notes
would
be
able
to
of
add.
Add
those
notes.
Would
such
an
architecture
be
possible?
I
I
mean
if
you
go
for
separation,
concern
right.
You
know
that
those
client
application,
we
would
have.
No
knowledge
of
you-
know
no
difference
between
talking
to
our
servers
or
the
server
from
the
second
link
identity.
Of
course,
the
notification
would
want
to
be
received
by
they
are
that
has
access
to
these
certificate
like
so
it's
a
different
that,
but
essentially,
if
you
want
to
create
a
dictation,
you
change
the
app
ID.
You
register
any
certificate
just
on
it
anymore.
H
Cool
yeah
because
I
think
something
that
that
would
be
useful
also
to
get
around
this
with
certain
Android
version.
It's
not
being
killed
as
well
as
different
times
in
the
future
that
that
different
ways
right
so
cool,
yeah
and
I
will
stick
with
that
with
what
would
each
actor
knows
because
it,
because
the
biggest
concern
with
metadata
leakage
and
whatnot
and
yeah
all
this.
I
Yes,
instead
of
a
spanner
was
that
probably
I
would
gladly
discuss
because
I
think
we
need
to
make
some
decisions
about
piece.
So
I
don't
have
to
set
on
my
mind
and
I.
Think
I
need
someone
else
to
actually
help
me
out
will
be
the
decision
because
you
know
for
say
you
know
pros
and
cons,
and
you
know
I
can
wear
them
all
on
myself,
but.
E
I
H
H
H
H
H
H
So
yeah
realize
we
don't
really
have
like
a
data.
Science
is
rolled,
but
I
do
think,
there's
a
bunch
of
people
that
are
looking
into
like
performance
metrics
and
these
types
of
things.
So
maybe
one
of
them
would
have
enough
to
the
database
food
to
be
able
to
pull
this
out
and
that
would
inform
their
analysis
as
well,
potentially
low
cost
for
install
and
if
I
don't.
A
H
F
Have
another
question
just
about
education
is
regarding
battery
usage
in
because
now
you
actually
do
both
of
the
messages
and
when
you
open
they're,
all
the
measures
are
already
downloaded.
So
maybe
the
battery
uses
is
is
kind
of
the
same
or
not
much
higher,
and
how
also
that
that
affects
on
the
performance
of
the
app
app
like
you
open
and
now
the
messages
are
already
there.
You
don't
have
to
wait
to
synchronize
the.
B
The
percentage
the
the
battery
the
battery
added
usage
is
measured
in
a
difference
of
percentual
points
like
points
of
a
percent
in
hours
of
use.
Okay
I
know
that
we
looked
at
it
in
QA,
has
done
extensive
testing
on
that
I
can
look,
I
can
look
back
at
the
data
and,
if
you
give
you
that
after
the
call
and
added
to
the
notes,
but
what
I
remember
from
it
was
was
that
it
was
negligible.
A
A
So
yeah
two
things
I
do
actually
want
to
check
if
everyone
who
would
want
to
join
the
discussion
is
still
available
because
I
was
actually
thinking
given
where
we
are
now
on
the
agenda
and
meetings
longer
than
an
hour
to
turn
off,
usually
production.
We
could
also
set
up
a
separate
call
either
like
later
today,
Oscar
they'll,
probably
get
late
for
you
or
tomorrow
same
time,
so
that
we're
all
fresh
to
discuss
this
I'm
open
to
both
I'm.