►
From YouTube: CasperLabs Community Call
Description
Engineering update, Roadmap update, Validator reward presentation
A
Right
after
the
update
good
morning,
everyone
good
afternoon
good
evening,
wherever
you're
joining
us
from
this
is
Mehta
parlor
car
Casco,
lab
CT,
oh
I'm
gonna
be
providing
a
development
and
engineering
update
and
then
we're
gonna
talk
about
roadmap
post
tests
net.
That's
what
the
community
wants
to
hear
about
and
then
o'nora
is
gonna
present
the
updated
senior
edge
formulation,
calculation,
I'm
gonna,
go
ahead
and
share
my
screen
again.
A
We
do
these
calls
every
Tuesday
morning
we
stream
em
on
YouTube.
You
can
always
check
out
the
community
calls
after
the
fact.
Can
you
guys
see
my
screen?
Yes,
amazing,
all
right!
So
we
are,
we
have
kicked
off
sprint
4.1.
We
are
working
really
hard
on
a
rust
implementation.
Those
of
you
that
have
been
following
these
calls
you're
aware
that
we
are
building
the
node
logic
in
rust,
because
we
believe
we'll
get
a
more
efficient,
more
performant
and
more
stable
architecture
from
doing
that.
A
As
a
result
of
doing
that,
we
are
getting
ready
to
kick
off
our
beta
test
net.
The
beta
test
net
provides
functionality
for
contract,
headers
contract
packages,
aetherium
key
support
and
upgrades
right.
Contract
packages
come
with
contract
upgrades
and
it's
a
significantly
simpler
way
to
author
smart
contracts
and
what
we
currently
have.
We
knew
that
contract
headers
was
a
long
time
coming
and
we're
happy
that
we're
going
to
be
deploying
this
to
the
Testament
soon.
If
you
want
to
get
started
on
contract
development,
we
recommend
that
you
build
and
deploy
build
locally.
A
You
can
do
that
with
darker.
You
can
also
pull
down
darker,
build
your
Dockers
locally
from
dev,
we'll
be
cutting
the
dot
20
release
later
this
week.
Once
we
cut
the
release,
docker
will
be
updated
in
Otisfield
to
pull
down
the
latest
packages
and
run
the
docker
hack.
You
can
get
to
the
docker
hack
on
github
and
that's
a
set
of
scripts
that
let
you
run
the
networking
clarity
locally
and
Ashok.
A
We
should
probably
test
and
make
sure
that
the
docker
hack
does
work
with
20
and
it
still
works.
So
we
should
probably
just
validate
that
if
we
haven't
done
so,
maybe
that's
something
you
and
I
can
do
later
tonight
moving
along.
So
we
are,
we've
got
our
2008
release
cycle
that
we
kicked
off
and
the
goal
is
to
get
to
a
feature
equivalent
of
a
Scala
test.
Net
I
think
there'll
be
a
few
things
that
this
network
will
do
that.
The
skeleton
it
does
not
with
a
minimum
of
twenty
validators.
A
B
Yes,
arias,
a
muted
us
talking,
yeah
there's
there
is
no
Noronha
show
so
far
on
a
lot
ease.
Okay,.
A
Excellent
on
highway,
so
we're
working
we're
building
out
an
integration
test
framework
for
highway
core.
The
team
really
wanted
the
test
framework
in
place
before
progressing
because
they
wanted
to
know
what
they
were
building
was
correct
as
they
built
it.
We
are
building
the
the
deploy
fetcher,
the
block
executor
and
generic
gossiping.
This
is
the
business
logic.
That's
currently
implemented
in
Scala
we're
going
to
be
implementing
that
in
rust,
using
the
reactor
model
and
then
we're
wiring
up
the
synchronizer.
When
we
talk
about
synchronizer
a
show.
What
are
we
talking
about
here?.
B
A
A
Think
they
need
a
password,
it
shouldn't
require
a
password,
it's
a
webinar.
They
should
be
able
to
get
in
it's
someone
in
telegram.
It's
Joe.
Can
you
go
see,
I
think
og
is
trying
to
get
help,
but
I
think
their
Chinese
speaker,
which
may
be
making
it
challenging.
I,
don't
know
if
Louise
is
around
okay
so
carrying
on
so
the
synchronizer.
Is
this
the
same
thing
as
block
downloading.
C
A
So
I
got
a
piece
of
information
here,
we're
using
hack
docker
to
test
both
client
and
clarity
with
duo,
but
it
only
works
building
locally
until
we
release
there's
a
new
keygen
for
a
countdown
hash
that
isn't
on
docker
hub
as
yet.
So
that
makes
complete
sense
right.
We
supported
if
we're
supporting,
etherium
type
keys,
so
the
multi
key
support
has
changed
the
way
keys
are
handled
so
once
we
cut
the
dot
20
release,
the
docker
hack
will
work
right
until
then
you
got
to
build
locally.
A
So
thanks
Joe
for
giving
me
that
little
nugget
of
information,
that's
Joe
sash
or
one
of
our
core
devs,
you
pinged
me.
Let's
see
contract
runtime
feature
complete,
so
we'll
be
ready
for
people
to
go
full
throttle
on
contract
development.
We
do
have
our
open
sessions
that
are
every
Thursdays
at
midnight,
Pacific
and
7:00
a.m.
Pacific,
and
that
is
meant
for
all
of
you
that
want
to
develop
smart
contracts
or
learn
how
to
develop
smart
contracts,
hop
on
the
call
and
we've
got
folks
core
developers
available
to
help
get
you
started.
A
A
We're
also
working
on
documentation.
I've
done
a
lot
of
work
on
documentation
on
github,
trying
to
make
things
easier.
If
you
find
an
issue,
feel
free
to
open
up
a
pull
request
or
just
send
me
a
note
if
you
think
that
it's
confusing
happy
to
get
feedback
and
make
sure
that
documentation
is
good
and
clean,
we're
working
on
the
smart
contract,
DSL
tutorial,
we're
gonna
be
dropping
that
DSL.
In
a
few
weeks,
I've
been
working
with
gutter
anon
that
building
out
the
new
DSL
that
will
help
make
smart
contract
in
development
significantly
easier.
A
We're
also
building
support
for
solidity,
so
we're
building
a
transpiler,
basically
a
tooling
tool
chain
that
will
take
in
a
solidity
contract
and
spit
out
web
assembly
and
solidity
developers
can
also
explore
the
intermediate
rust
that
is
generated
from
the
tool
chain.
So
you
can
actually
work
with
the
rust
directly.
A
If
you
wish,
you
can
also
pass
that
rust
code
in
through
a
testing
framework
and
use
breakpoints
to
see
how
your
contract
is
behaving
and
rust
in
the
event
that
you
hit
an
issue,
we're
improving
the
Casper
labs,
signer
extension
and
clarity
to
accept
just
key
files
right,
so
you
don't
ever
have
to
handle
your
keys
in
a
perfect
world.
Clarity
drops
it
directly
in
the
signer
I,
don't
know.
If
that's
going
to
be
possible
to
do
that
kind
of
server-side
conversation
between
the
two
components,
but
it
would
be
fabulous
if
we
could
do
that.
A
Maybe
that's
future
of
the
roadmap,
we're
also
working
with
IP
we
and
building
out
their
solution.
A
coach
is
working
hard
on
building
a
data
flow
diagram.
We
had
some
great
calls
with
them
last
week,
we're
going
to
be
meeting
with
them
again
today
to
talk
through
solutions,
so
solution
architecture
is
underway,
we're
building
tech
specs
out
for
the
validator
bonding
auction
and
senior
age
specification.
A
So
we
expect
that
by
the
time
we
get
to
Delta
test
net,
we
will
have
the
full
economics
of
the
chain
implemented
in
the
next
60
days,
and
so
we
need
those
specs
done
and
then
we're
also
doing
the
auto
price
adjustment
specification,
and
this
is
to
help
reduce
the
churn
Ahn's
pricing
right.
We
wanted
to
get
to
price
stabilization
and
this
is
a
specification.
That's
going
to
help
us
get
there
exactly.
D
A
This
is,
in
addition
to
the
chain
link,
Oracle
work,
so
that's
the
update
team
is
making
extremely
good
progress.
I
believe
we're
closing
that
we're
getting
close
to
over
80%
achievement
on
our
sprint.
Commitments,
which
is
really
great
in
terms
of
being
able
to
plan
and
deliver
being
able
to
plan
and
deliver,
is
really
really
important.
A
A
This
is
the
document
that
I
use
for
planning
out
I
work
with.
You
know
the
team
members
to
figure
out
how
we're
going
to
get
from
here
to
there,
and
if
you
see
back
here
it's
it's
kind
of
it's
the
way
we've
been
managing
the
entire
project
right,
and
so
we
talk
about
top.
We
talk
about
our
strategic
objectives.
What
are
the
strategic
objectives?
We're
looking
for
in
each
of
the
releases
and
then
distilling
down
from
that
strategic
objective?
A
We
have
what
we
want
to
see
being
built
in
each
and
every
one
of
the
release
cycles.
So
in
each
month
one
of
the
folks.
What's
the
focus
right
and
so
right
here,
you'll
see
this
is
July.
This
is
where
we
are
right
now.
The
pieces
that
are
in
green
are
the
pieces
that
that
are
done
right,
so
beta-test
net.
We
want
to
release
contract,
headers,
multi
key
support,
host
side
transfer
contract.
We
were
targeting
July
15th.
A
We
pushed
that
out
a
little
bit
till
the
end
of
July
and
the
reason
for
that
is
we're
trying
to
get
more
big
names,
taking
pools
on
on
the
project,
and
so
we've
got
a
lot
of
engagement
from
the
community
around
becoming
a
validator
and
we're
looking
to
extend
that
with
by
adding
more
staking
pools
as
well
so
focus
is.
We
want
to
do
a
hackathon,
ideally
in
August
or
September
timeframe.
We
want
to
run.
A
My
goal
was
to
run
some
token
transfers
for
a
second
on
the
beta
test
net
and
have
the
validators
run
it.
We
want
to
increase
staffing,
developer
engagement
and
we
want
to
get
as
many
large
taking
pools
to
participate.
This
is
the
work
that
we
are
building
the
chain.
Spec
Genesis
process
improve
reactor
formalize
event
model.
That's
all
done
block
execution
is
done,
we're
in
it.
We've
integrated
the
contract
runtime
into
node,
RS
and
wired
it
up
into
a
unified
storage
and
deploy
buffer.
A
If
you
scroll
down
here,
we
look
at
we've,
got
our
s,
testing
work,
we're
doing
estás
ready
for
validators
and
we're
looking
at
building
in
get
pod
for
browser
contract
development,
and
then
we
want
to
update
the
signer
to
pull
key
files
other
than
key
data
and
that's
what
a
show
if
just
provide
a
status
on
and
then
the
contracts
DSL
and
then
support
for
solidity.
Bios
illing
moving
into
August
we're
looking
at
updating
example
contracts.
A
We
want
to
push
clarity
into
a
separate
repository
and
license
it
under
Apache
2.0,
there's
a
fair
bit
of
work.
That
needs
to
be
done
because
clarity
is
currently
kind
of
mixed
up
in
currently,
with
the
current
you
know,
Casper
labs
repo,
so
we
need
to
break
it
out
separately
and
have
a
separate
license
on
I
believe
this
is
economics
type
work
here
yeah.
This
is
what
is
this?
Your
theme
not
theme
I,.
A
A
Okay,
so
the
theme
here
is
gonna,
be
bonding
auctions,
validator,
set
rotation,
the
permission
of
this
protocol
economics
of
the
chain
right
and
so
on.
Highway
we're
gonna
start
working
on
the
adaptive
round.
Land
base
blocks,
which
flux
finalized
blocks
on
here,
the
node,
download
manager,
validator,
set
rotation
and
building
the
linear
block
chain
and
on
the
contract
run
time.
We're
gonna
do
all
the
concert,
transaction
costs
and
metering
implementation
and
then
I'd
like
to
do
the
designs
for
checkpointing.
So
this
is
still
we're,
not
sure
if
we're
gonna
do
this.
A
This
is
something
we
need
to
actually
figure
out.
So
I
want
to
mark
that
in
yellow
right
now,
they're,
not
sure
if
we're
going
to
do
that
essential
contract.
So
we're
going
to
do
some
updates
at
the
proof
of
state
contract
for
bonding
and
unbinding,
and
then
here
in
september,
proof
estate,
update
senior
rich
transaction
fee
distribution
and
slashing
100%
for
equivocations.
A
We
intend
on
bring
payment
in
another
context
and
Merkel
proofs
and
then
on
note
RS
is
we
want
to
get
the
client
api's
events
and
addressing
the
replay
attacks,
so
the
idea
is
by
Delta
test
net.
We
have
basically,
we
are
pretty
close
to
feature
complete,
with
the
exception
of
two
things:
protocol
upgrades
and
checkpoints.
A
That's
fine,
no
protocol
upgrades
we
will
do
probably
as
the
first
balance
of
Delta
test
net
in
October
is
what
we're
thinking
I'd
like
to
try
to
pull
protocol
upgrades
in
earlier,
but
it
might
have
to
be
in
October
and
then
we'll
probably
create
the
social
consensus.
Api
is
somewhere
along
that
line.
A
So
the
final
final
Genesis
will
happen
in
October
time
frame,
and
then
we
open
up
for
Delta
testing
it
the
ideas
that
we
decentralize
it
kick
off
the
attack
phase
and
figure
out
the
social
consensus
mechanism,
we'll
also
start
working
on
getting
ready,
formally
getting
ready
for
maintenance,
and
that
includes
creation
of
the
final
Genesis
block
and
the
pre
mine.
So
in
this
case
we
will
have.
A
We
will
start
publishing
what
all
the
accounts
and
initial
token
balances
will
be
for
the
founding
validator
set
and
all
of
those
participants
that
participate
in
the
pre
mine
would
start
to
expect.
Looking
for
their
Genesis
account
balances
right,
and
we
would
continue
to
keep
this
up
to
date
for
any
folks
that
have
any
additional
rewards
that
are
coming
in
right.
So
the
goal
is
that
this
final
Genesis
block
is
created
transparently
in
front
of
the
community
and
the
community
sign
off
on
this
final
Genesis
block.
It's
super
important
to
get
that
right.
A
You
have
questions
on
this.
You
can
hit
me
up
and
on
any
of
my
handles
telegram
ice,
you
can
send
a
message
on
the
YouTube
channel
itself.
It's
convenient
for
you
to
do
that.
There
will
respond
there.
Of
course,
discord.
You
can
always
hit
me
up
hit
us
up
on
discord.
Twitter.
You
can
find
me
on
Twitter.
So
that's
it
for
the
questions
so
I
hope
that
was
sufficient
over
to
you
or
nor
to
talk
about
senior
age,
perfect.
D
A
D
Got
feedback
that
it
was
low
before
so
this
presentation
is
to
make
our
most
recent
scene
rage,
form
formulation
more
accessible
to
our
both
the
community
and
but
more
primarily
our
devs,
because
we
have
this
I
wrote
this
start
writing
this
paper,
which
explains
the
economics
of
the
economics
of
Casper
labs.
Basically,
but
the
notation
is
not
exactly
a
transparent
and
we
need
our
dash
to
get
yet
like
we
will
be
implementing
this
in
the
next
60
days
like
meta
sets,
so
we'll
be
recording
this
and
using
this
as
a
reference
and
I
hope
this.
D
This
is
basically
an
update
on
the
previous
version,
I
presented
exactly
6
months
ago,
and
this
is
a
much
simplified
and,
in
my
opinion,
much
more
robust
version
of
the
previous
mechanism.
So
I
think
it
will
also
be
simpler
to
understand
for
those
of
you
that
that
listen
to
it
that
word
there.
So
before
I
even
begin,
let
me
just
give
you
the
gist
of
it.
D
We
introduced
a
new
new
assumption
to
the
protocol
an
assumption
of
an
optimal
round
exponent
and
with
this
assumption,
we
are
able
to
simplify
our
formulation
like
to
an
extent
that
it's
it's.
It
feels
much
better,
so
caveat
like
before.
Weight
is
normally
an
integer
value
in
code,
but
I
will
use
percentages
in
this
presentation
and
the
parameter
values
given
I.
Given
examples
don't
represent
ya
parameters
will
have
a
maintenance,
so
we
are
having
long
running
tests
too,
and
also
as
tests
to
determine,
if
like
to
values
we
are
setting,
are
the
best
one.
D
D
D
D
To
so
I
don't
want
to
go
into
too
much
detail,
I!
Think
yeah!
You
can
watch
the
previous
presentation.
You
can
find
it
on
YouTube
on
our
YouTube
channel
for
January
14th.
So
we
have
these
round
of
exponents
which
every
validator
announces
they
can
announce
in
any
message,
and
that
dictates
the
rounds
which
they
will
participate
at.
So
the
the
length
of
the
round
is
defined
by
2
to
the
power
the
value
they
announced.
So
here,
for
example,
they
announced
two.
D
So
you
see
this
part
as
like
four
cells,
but
in
the
second
round,
so
they
are,
of
course
the
red
lines
denote
the
round
beginnings
and
ends
during
the
second
round.
The
validator
announced
like
three.
So
when
that
round
ends
the
next
round
is
eight
cells.
Long
eight
ticks
long,
so
each
cell
is
a
tick.
D
Here
is
another
demonstration,
so
ironic
font
is
announced
here,
but
you
have
to
finish
all
the
all
the
rounds
that
that
complete
the
current
current
you
know
around
like
the
the
round,
which
the
next
one
starts
should
be
divisible
by
by
2
to
the
power
the
value
they
announced
so
like
like
I
wrote
here.
This
is
a
basic
explanation
of
how
texts
work.
D
Also
the
I
have
17
slice,
so
it's
mine,
I'm,
guessing
this
Moses
will
take
half
an
hour,
but
I'll
try
to
make
it
fast,
not
fast,
but
concise
and
nuts
hand
to
hand
too
many
too
much
on
details.
So
the
round
exponents
are
updated.
This
way
whenever
we
fail
to
finalize
leader
block
before
the
end
before
the
round
ends,
we
increase
our
round
length
and
this
is
not
a
manual
process.
This
is
automatically
done
by
the
client
software,
so
this
was
just
intro.
Now
now
the
updated
scenery
formulation
begins
C
with
the
new
formulation.
D
We
introduced
this
optimal
round
exponents,
so
we
assume
a
baseline
round
exponent,
which
the
blockchain
should
operate
under
regular
conditions.
This
is
a
protocol
parameter
reset.
We
want
to
incentivize
valid
leaders
to
operate
at
this
value
and
just
an
example.
If
we
said
the
round
Experion
to
15,
then
we
expect
validators
to
participate
in
every
30,
2007
and
68
milliseconds,
which
is
roughly
32
second
rounds,
so
they're
still
allowed
to
diverge
from
your
optimal
round
exponent.
Although
this
is
a
this
is
still
under
discussion,
they
can
definitely
go
slower.
D
It's
under
discussion,
whether
they
will
be
able
to
go
faster,
I
mean
those
grounds
would
be
valid.
But
the
point
is:
if
they
go
slower,
they
receive
less
rewards
than
the
optimal
case.
If
we
allowed
them
to
go
faster,
they
should
receive
the
same
reward
as
the
optimal
case.
So
that
means
that
their
reward
they
earn
per
unit
time.
D
The
rate
of
reward
they
receive
should
remain
the
same
because
we
don't
want
to
you
know:
go
we
don't
want
to
enable
them
to
enable
them
to
gain
the
incentive
mechanism
by
you
know
going
arbitrarily
faster
with
empty
blocks,
which
they
can
do.
So
we
have
these
measures
in
place
to
prevent
them
from
such
such
behavior
and
so
another.
D
D
D
At
the
end
of
every
era,
the
finest
helix
supplies
calculated
as
basically
initial
supply,
plus
total
rewards
that
have
been
minted
up
until
that
point
that,
in
the
whole
history
of
the
block
chain,
the
total
slashings
that
have
been
incurred
up
until
that
point
in
the
whole
history
of
the
Kasparov
block
chain,
and
that
gives
us
like
a
tangible
value.
The
the
protocol
knows
like
this
is
the
maximum
value
that
can
exist
out
in
the
wild
and
so
to
to
explain
how
I
calculate
round
rewards.
D
D
Basically,
we
make
a
proportional
calculation
and
if
we
say
let's
say
we
have
two
percent
annual
rate
of
minting
new
tokens,
how
much
tokens
will
be
meant
in
an
optimal
round,
so
that
has
a
definition.
It
will
be
more
clear
in
in
the
next
example,
so
the
base
round
rewards
the
next.
Our
next
concept
is
the
basically
a
base
value
which
we
calculate
the
round
reward
firms.
I
will
also
introduce
it
in
the
next
example.
D
It
remains
constant
within
an
error,
but
is
updated
with
each
new
error
according
to
the
round
sinners
rate
and
the
supply
at
the
end
of
the
previous
round.
So
the
value,
the
the
for
me
that
I
just
introduced
here
thus
clx
supply
increases
exponentially
like
we
wanted
to,
but
also
takes
slashings
into
account
and
I,
will
also
introduce
I,
also
give
an
example
of
what
happens
with
Slash.
It's
a
very
simple
example:
we
assume
that
the
optimal
round
exponent
is
15,
that's
32,
second
round.
D
We
assume
that
the
we
set,
the
annual
senior
to
rate
to
2%
and
the
final
supply
of
the
previous
era
is.
Let's
say
we
are
really
in
an
arbitrary
era
and
the
previous
one
had
fine
like
a
final
supply
of
12
billion.
So
very
simply,
we
we
use
the
we,
so
we
use
the
relation
if
you
have
2
percent
mental
grades
in
a
year.
What's
your
mentee
great
in
32
seconds,
that's
what
we
do
here
and
it's
a
very
small
number
two
to
two
powers
on
to
something
to
the
power:
minus
8!
D
It's
it's!
Let's
I'll
just
refer
to
this
as
the
round
senior
grade
and
using
this
round
senior
grade,
we
calculate
the
base
round
reward
by
just
multiplying
it
with
the
previous
final
supplier
to
previous
error,
which
is
12
billion,
and
that
gives
us
to
240
6.9
CLX
reworded
per
round.
That
means
in
each
round
this
much
this
many
tokens
will
be
minted
and
distributes
to
the
validators.
D
Like
as
time
passes,
the
supply
will
grow.
That
means
the
rewards
will
also
grow,
and
after
a
week
here,
we
assume
that
an
area
is
a
week.
We
we
see
that
we
have.
Let
me
just
have
an
optimistic
scenario:
18,000
rounds
have
been
finalized
and
the
we
just
add
the
new
new,
like
mental
rewards
to
the
supply.
We
have
me
now
how
twelve
billion
plus
this
value,
so
that's
a
four
million
for
four
point:
five
million
more
CLX
and
the
same
when
she
knows
rate
holds
it's
the
same
throughout
time.
D
Unless
we,
unless
we
decide
through
governance,
to
change
it
so
I
threw
you
should
have
made
the
same.
Therefore,
the
new
based
round
reward
is
two
new
final
Supply
and
times
the
round
round
senior
grade.
So
before
it
was
two
hundred
forty
six
point:
nine
nine.
Now
it
has
increased
0.1
CLX
in
one
air,
so
this
increases
very
slowly
but
also
exponentially
because
dear,
it
is
a
factor
of
final
supply
which
is
designed
to
grow
exponentially
and
yeah.
This
eventually
gives
gives
us
the
desired
yield
for
the
values
and
just
to
cover
all
the
cases.
D
So
what
happens?
If
there's
a,
if
there's
a
some
slashing
and
I'll
just
exaggerate
in
numbers
to,
for
example,
let's
say
that
a
validator
hotel,
unsuccessfully
attacks
the
network,
but
they
get
slashed
and
the
slashing
amounts
to
you
percent
of
the
total
supply
and,
for
simplicity's
sake,
I'd
I,
say
there
were
no
rewards
in
this
round.
So
the
new
supply
is
just
previous
value,
minus
ten
percent
of
it.
So
here
we
have
slashed
quite
quite
a
lot.
Basically
one
point:
you
want
something
billion
tokens
the
same
round:
sins
rate
holes.
D
We
can
just
multiply
it
with
the
round
senior
trade.
To
obtain
two
hundred
twenty
two
point:
three
clx
per
round.
We
can
see
that
the
slashing
caused
24.7
clx
drop
in
the
round
reward.
So
that's
how
much
you
should
have
dropped,
because
we
know
that
that's
the
amount
of
slashing.
So
this
is
what
happens
to
the
round
rewards
if
there
is
a
slashing.
D
A
C
C
D
Okay,
that's
the
case,
but
this
is
this
is
what
will
what
we
would
implement
so
what
I'm
proposing
we
would
implement
and
like
changes
to
seniors
rate
or
changes
to?
Maybe
the
maybe
me
assume
a
different
supply
just
to
prevent
rounds
reverse
from
dropping
drastically
to
low,
because
it
will
be
harmful
to
the
ecosystem.
So
anytime
talk.
C
D
Well,
the
the
ones
who
get
slashed
if
they're
attackers
and
they
were
actually
malicious,
guys.
It's
it's
debatable,
I
think
I
think
we
should
we
could.
We
can
discuss
it
another
in
another
meeting,
but
yeah
I'll
just
continue
with
the
formulation.
That's
macro
level.
Stuff
I
just
want
to
explain
to
how
how
the
system
work.
So
that's
yeah
the.
D
So
the
assignment
coming
up
to
the
protocol,
of
course,
if
you
have
been
tuning
tuning
in
to
our
previous
calls,
you
know
that
we
have
three
phases
in
around
the
first
leader
sends
a
lambda
message,
basically
a
leader
message,
which
contains
a
block
the
values
as
soon
as
they
had
received
that
message
sand.
They
ran
the
response
message:
meat
saying
that,
yes,
we
receive
your
block
and
then
once
a
defiant,
definite
amount
of
time
passes.
D
Everybody
sends
a
third
message:
the
Omega
message
sighting
everyone
else's
with
all
the
messages
that
they
have
seen
in
that
error
and
the
with
the
theory
we
have.
It
ensures
that
if
enough
enough
validators
do
this,
the
block
will
be
finalized,
which
means
that
it
will
not
possible
to
they
will
reach
only
they
release
on
consensus
on
the
leaders
message,
which
is
the
block
which
contains
the
transactions,
and
here
I
just
demonstrated
the
concept
of
assigned
weight.
D
So
we
know
that
violators
are
participating
in
round
rounds
on
their
own,
so
they
can
choose
the
ticks
which
their
rounds
start.
Basically,
here,
C
is
participating
like
two
times
slower
than
the
other
ones.
So
here's
C
just
participates
in
three
rounds,
but
and
B
participate
in
six
rounds
so
the
rounds
so
in
this
example
I
assume
I.
We
have
only
three
values
which
are
which
have
equal
weights
roughly,
so
the
ones
where
they
all
participate
like
hundred
percent
assigned
weight,
but
the
ones
where
C
is
missing
has
had
only
six
to
six
percent.
D
D
So
what
do
we
use
assigned
wait
for?
We
use
it
to
scale
the
base
round
reward.
Now
we
use
this
to
reduce
to
incentivize,
where
leaders
to
converge
on
the
same
round
exponent,
and
we
also
have
the
concept
of
participative
weight
or
also
the
crop
size
of
the
summit.
I
referred
as
forum
size
later
on,
but
a
participating
weight
is
the
actual
weight
that
ends
up
participating
in
a
year
round
or
block
so
I
introduced
further
limits
on
these
definitions.
D
D
Be
me,
on
the
other
hand,
sends
the
santa's
messages
after
the
ground,
so
these
should
have
been
inside
these
realm,
but
they're
after
so
these
are
out
of
these
round.
In
this
case
c
sends
the
omega
l.
Sorry,
the
land
responds
message
in
in
Israel,
but
the
Omega
message
out
of
his
round.
So
there
is
this
is
the
no
and
for
for
finality
our
notion
of
finality.
D
These
messages
should
all
be
in
their
respective
realms,
not
out
yeah,
so
I'll
just
introduce
the
notion.
So
do
this
is
highway.
This
is
a
terminology.
The
quorum
size
of
the
level
1
summit,
achieved
by
in
round
messages,
should
be
greater
than
50%,
plus
a
value
of
fault
tolerance
threshold
which
we
choose
like
1,
so
50,
plus
something
something
or
we
just
say
refer
to
it
as
51%,
and
that
means
that
so
the
quorum,
so
the
chrome
is
the
validators
whose
see
each
other
and
also
who
see
each
other,
seeing
each
other's
messages.
D
So
basically
there's
a
oh
yeah.
This
is
not
the
highway
press
presentation,
so
I'll
just
hope
that
you
will
take
the
definition
at
face
value
and
move
on
from
that
for
round
2
be
allocator
rewards.
A
sign
way
must
be
enough
for
on
time
finalization
if
it's
smaller
than
that
values
can
now
receive
any
rewards
from
that
round.
D
So
here
is
an
example:
successful
versus
failed
on
time,
finding
finalization
these
Omega
messages
out
of
his
realm,
but
the
curb
size.
So
you
already
have
a
and
C
participating
in
the
core,
so
the
quorum
size
is
66%
which
is
greater
than
51%.
So
you
have
successful
on-time
finalization
on
the
right
hand,
side
both
B
and
C-
did
not
send
their
messages
on
time
and
also
AIDS
message
is
not
does
not
cite
them
so
the
quorum
size
is
actually
33%.
D
D
D
So,
like
I
said
we
have
the
base
round,
we
worked.
We
first
multiply
it
with
the
weight
of
two
values
that
want
to
participate
in
the
round,
so
the
assigned
weight.
So
this
is
a
scalar
value
between
zero
and
one.
So,
let's
say,
assign
weight
was
like
90%
just
because
of
that
the
base
round
reward
is
scaled
down
to
90%
and,
let's
say
the
weight
of
the
values
that
end
up
participating
in
the
round,
which
is
calculated
by
using
the
quorum
size
divided
by
total
weight.
D
B
D
B
D
D
So
we
first
multiply
by
a
sound
weight,
multiplied
divided
by
total
weight
and
what
we
also
multiply
by
quorum.
Sighs
do
I
assign
weight
which
gives
us
quorum
size
divided
by
total
weight
here,
and
the
second
term
is
actually
we
have
the
optimal
round
exponent
as
I
introduced
in
the
beginning.
If
the
leaders
round
exponent
is
smaller
than
the
optimal
round
exponent,
which
means
leader
is
too
fast,
we
divide
by
powers
of
two
to
ensure
that
they
the
same
reward
per
unit
time,
the
same
rate
of
reward.
D
D
D
Happens
so
this
needs
to
be
take
place,
I'm,
sorry
for
holding
you
ups
for
so
long,
but
yeah.
This
is
for
later
reference.
We
just
distribute
parada.
So
every
the
realm
we
word
we
calculate
here
is
just
distributed.
So
weight
of
the
validator
I,
divided
by
the
total
weight,
gives
how
much
we.
So
this
is
I,
think
simpler.
So,
in
conclusion,
we
in
send
away.
We
try
to
incentivize
validators
to
converge
on
the
same
round
exponent.
D
We
incentivize
values
to
operate
at
the
optimal
round
length
and
we
also
use
itemize
values
to
finalize
it
round
with
as
much
as
great
as
possible
before
it
ends.
So
this
was
the
first
point,
the
most
important
point
yeah
this.
This
is
a
much
simplified
version
of
the
one
I
presented
in
January
and
I
hope
yeah.
It
makes
sense
to
you,
I'm,
open
to
feedback.
I
will
share
this
presentation.
You
can
you
can
send
a
mail
to
own
or
at
Casper
labs
at
I/o.
D
A
Yeah,
it's
a
lot
to
digest,
but
I
think
I'm
interested,
particularly
in
the
mathematical
paper
that
at
the
beginning,
because
it
it's
very
much
like
the
ethereal
paper,
the
idea
being
that
this
specification,
which
should
be
implementable
in
multiple
languages
and
we'll
be
putting
it
out
here
shortly.
What
does
that
get
we're
gonna
be
ready.
This
is
already
done.
I.
D
Mean
the
scene
which
part
is
done:
we
will
gradually
partial
transaction
fees,
but
this
papers
a
higher
a
little
paper,
whereas
the
yellow
paper
is
a
spec
for
that.
So
the
paper
I'm
reading
is
more
or
less
a
guideline
for
economic
design.
I'm
also
specifying
this
in
the
simulator,
so
the
simulator
I
wrote
in
Python
will
act
as
a
reference
and
also
the
today.
We
are
also
writing
in
the
tech,
new
tech,
spec,
we're
updating
tech
spec,
not
a
new
tech,
spec,
updating
the
tech
spec,
so
that
will
that
will
be
the
yellow
paper
here.
A
Yeah
very
technical
tech,
spec
good
well,
that
was
a
good,
a
comprehensive
discussion
today,
loss
to
think
about
so
for
those
of
you
that
stuck
around
to
the
end,
remember
weekly
workshops
on
Thursday
and
we'll
do
this
again
next
week.
Catch
me
on
telegram
in
to
this
cord
or
Twitter.
If
you
have
more
questions,
Cheers
have
a
great
day.
Thanks
can.