►
From YouTube: IPFS Q3 OKRs
Description
Agenda:
- Q2 Accomplishments
- Q2 Challenges
- Q3 Structure
- Q3 Priorities
Q3 OKRs: https://docs.google.com/spreadsheets/d/1AiNUL7vK5Jp8aa839UaMaI_AlBU5r6Bor-A40179I2A/edit#gid=1681757723
[P0] IPFS as a Service
[P0] Testing Infra
[P1] Project Operations
[P1] Package Managers
[P2] Documentation
A
Hello,
everyone
and
welcome
to
our
q3
kickoff.
Ok,
our
meeting
where
we
are
going
to
tell
you
about
all
of
the
stuff
that
was
awesome
about
q2
some
challenges
about
q2,
how
we
are
structured
going
into
q3
and
then
what
is
most
important
than
each
of
the
the
groups
will
will
present
their
own
ok
ours
from
a
parent
perspective.
A
Alright,
so
let
me
start
by
sharing
my
screen
good
place
to
start
doo-doo-doo-doo
cool
I'm,
just
gonna,
keep
it
in
this
mode
because
easier
for
clicking
sake,
but
this
is
this:
is
our
outline
accomplishments
challenges,
structure,
priorities,
starting
first
with
q2
accomplishments?
This
was
a
really
big
quarter
for
ipfs
active
exciting.
We
saw
a
lot
of
growth
growth
around
the
public
network,
growth
of
tons
of
cool
apps,
building
on
top
of
IP
FS,
more
end-users,
getting
the
benefit
from
those
apps
and
more
developers
contributing
to
the
project.
We
saw
a
lot
of
network
growth.
A
This
is
unprecedented
growth
in
the
size
of
the
network,
and
that
was
super
awesome,
super
exciting
and
also
a
challenge
for
us
in
keeping
up
with
said
network
growth.
But
it's
exciting
to
see
these
many
people
joining
the
wider
IP
fest
network.
We
also
saw
a
ton
of
apps
proliferating.
This
is
my
very,
very
happy
attempt
of
taking
this
awesome
document
created
by
the
block
projects
using
IP
of
us
and
trying
to
update
it
with
my
understanding
of
all
of
the
cool
people
who
we
got
to
chat
with
it.
A
We
also
saw
contributors
go
up
on
github.
Some
of
this
might
be
because
there
was
so
many
people
being
pulled
in
for
ipfs
can
but
I
think
in
general
kind
of
across
the
project
we're
seeing
and
as
a
we
re
bump
up
in
the
number
of
quarterly
contributors.
And
hopefully
we
can
continue
maintaining
that
level
of
engagement
and
enthusiasm
within
the
wider
community
to
contribute
directly
to
the
project
through
ipfs.
And
finally,
these
people
aren't
all
in
one
place
at
all.
A
They're
four
thousand
community
contributors
around
the
world
and
though
that's
kind
of
our
full-time
contributor
stack
and
people
have
been
participating
in
this
and
bringing
tons
of
perspectives
from
from
all
over.
So
big
thanks
to
work,
our
community
for
making
all
of
this
work
that
we
do
super
possible
so
wanted
to
look
back
a
little
bit
further
at
q1
and
the
important
things
that
were
were
happening
in
q1.
There
were
some
core
releases
of
Jas
and
ipfs.
A
So
speaking
of
IQs
camp
ipfs
camp
was
super
awesome
and
not
something.
We
spent
a
lot
of
effort
on
this
quarter
and
I
think
it
really
really
paid
off
for
us-
and
this
was
a
phenomenal
experience-
there
were
160
attendees
22
countries.
There
was
so
much
enthusiasm
and
excitement
in
this
space
and
the
before
month.
A
The
things
that
we
we
spent
our
time
on
were
creating
these
awesome
core
and
elective
courses
that
everyone
got
to
spend
some
time
in
each
each
of
the
first
two
morning's
of
ipfs
camp,
getting
to
dive
deep
on
actually
how
dad's
work.
What
really
is
the
life
cycle
of
data
and
then
building
some
cool
stuff
on
top
of
it,
which
was
super
awesome
people?
A
You
could
literally
interact
with
and
play
with,
including
a
ble
lippy
TP
transport,
which
is
awesome
and
I
can't
wait
to
hopefully
get
to
play
with
that
more
this
weekend
and
teach
G
visualizers
and
NPM
ipfs
and
a
ton
of
other
really
cool
stuff.
Super
super
awesome,
so
super
awesome
quarter,
tons
of
growth,
tons
of
exciting
stuff
tons
of
getting
the
community
together
and
pushing
things
forward,
but
also
some
challenges.
A
Definitely
one
of
the
things
was
being
able
to
deal
with
that
growth
and
having
good
network
and
Gateway
stability
and
reliability
throughout
that,
and
that
was
a
challenge
for
us
that
definitely
we
we've
noticed
having
started
working
on
but
aligning
ourselves
to
really
commit
to
this
going
forward
and
and
make
this
kind
of
a
a
core
part
of
how
we
do
our
work
is
checking
and
and
being
kind
of
like
all
hands
on
deck,
to
make
sure
that
we
have
really
good
reliability
and
scalability
there.
A
A
And
so
that's
an
area
that
we
also
think
we
can
really
dedicate
some
more
resources
to.
And
it's
super
critical
that
we
are
kind
of
listening
to
all
of
the
feedback.
We
got
from
my
PFS
camp
to
refocus
our
energies
and
make
ipfs
even
more
easy-to-use
as
a
developer
and
and
so
that's
definitely
an
area
for
increased
dedication
and
focus.
And
finally,
we
did
a
lot
of
awesome
research.
Around
package
managers
this
past
quarter.
A
And
so
that
was
definitely
an
area
where
you
know.
We
think
we
think
in
order
to
you,
know
we're
halfway
through
the
year.
In
order
to
really
hit
this
goal
out
of
the
park.
We
need
to
be
dedicating
more
engineering
resources
to
making
sure
that
that
really
happens
and
doing
it
in
a
more
tightly
coupled
way,
so
that
the
the
work
we're
doing
on
proof
of
concepts
and
research
and
understanding
has
to
be
tightly
like.
A
And
then
there
are
two
other
goals:
one
around
package
managers
and
making
sure
from
an
end
perspective
that
we
are
aligning
our
research
and
our
work
and
we're
hitting
the
school
out
of
the
parks
and
then
finally,
around
developer,
UX
and
accurate
and
usable
documentation
and
guides.
And
so
those
are
the
kind
of
needy
areas
of
problems
that
we
saw
for
this
quarter.
We're
looking
at
these
areas
and
you're
understanding
that
this
is
really
where
we
needed
to
be
focusing
our
time.
A
So
we're
we
have
reorganized
ourself
a
little
bit
from
a
from
an
org
structure,
and
this
is
not
new.
We
kind
of
do
a
little
check-in
on
org
structure
and
make
a
couple
of
changes
about
every
quarter
of
this
year.
So
to
look
back
in
q1
we
had,
and
in
this
org
structure
we
had
a
ton
of
different
working
groups,
I
think
about
10:00
or
11:00,
and
we
were
kind
of
split
into
two
teams
of
like
you
know.
A
But
at
the
end
of
q1
and
beginning
of
q2,
we
realized
that
this
kind
of
multi
membership
of
people
in
many
many
different
working
groups
and
split
across
maybe
even
three
or
four
working
groups
made
it
really
difficult
to
understand
where
they
were
spending
their
focus
and
to
have
kind
of
the
team
cohesion
of
solving
problems
together.
And
so
we
took
some
of
the
the
working
groups
that
were
kind
of
more
had
less
than
two
people
dedicated
to
them
or
we're
kind
of
more
problem
area
oriented.
A
And
so
we
kind
of
unified
into
this
like
core
core
set
of
teams
and
that
were
again
like
kind
of
a
functional
oriented
so
going
to
fest
Guiseppe
up
s,
project
or
working
group,
those
sorts
of
things,
and
but
we
condensed
ourselves
a
little
bit
in
order
to
minimize
our
planning
and
overhead
for
these
areas
and
the
condensing
part
we
thought
went
well.
But
then,
when
looking
at
these
teams
and
thinking
about
these
problems
we
needed
to
solve
in
q3,
we
realized
all
of
these
problems.
A
We're
gonna
be
spread
across
all
of
the
different
working
groups
and
add
additional
overhead.
So
when
looking
at
what
we
decided
to
do
at
the
end
of
q2
going
into
q3,
this
was
actually
orienting
ourselves
into
kind
of
task
force.
Oriented
groups
where
we're
focused
on
a
single
problem.
And
so
everyone
still
has
membership
in
a
in
a
particular
one
particular
working
group
and
not
spread
across
many
of
them,
but
really
dedicated
to
a
particular
problem
area.
B
We're
still
coming
up
with
a
name
I
business
as
a
service
get
raised
as
a
service
infrastructure
as
a
service.
We
don't
know
yet.
But
the
basic
idea
of
this
task
force
is
to
bring
gateways
into
shape
in
a
way
that
we
can
run
this
infrastructure
in
a
more
serious
production,
product
fashion,
having
canary
deployments,
having
nightly
deployments
having
service
level
agreements,
service
level,
objectives
and
monitoring.
All
of
that
in
a
consolidated
way.
B
For
example,
you
cannot
use
them
to
offer
content
with
I,
get
ways
that
we
do
not
agree
with
by
policy
and
at
the
end,
to
have
this
as
a
sustainable
effort,
meaning
to
have
a
team
around
this
to
have
team
waste
around
this,
and
you
have
to
have
inertia,
how
we
are
rendering
all
of
this
together.
So
it
is
sustainable
in
the
future
and
we
can
get
creative
ways
to
to
have
this
moving
forward
in
the
future.
I
hope
that
is
OK
for
this
for
description.
C
Me
so
so,
yes,
basically,
we
took
a
decision
based
on
you
know
over
observations
of
the
performance
numbers
and
a
few
other
regressions
that
the
latest
releases
contained
the
latest
go
ipfs.
Release
has
contained
to
basically
hold
the
release
train
until
we
can
actually
characterize
releases
and
their
impact
on
first
of
all
in
terms
of
performance,
in
terms
of
the
local
environment
of
the
node,
and
also
in
terms
of
how
it
affects
how
a
particular
release
affects
the
network
and
the
network.
C
So,
of
course,
these
this
level
of
characterization
is
you
know
you
can
ask
a
lot
of
questions
here,
a
how
much
has
the
behaved
behavior
of
my
node
improve
or
deteriorated,
but
this
particular
change
that
I'm
committing,
or
this
particular
PR,
that
introduces
honorific
changes
that
introduces
protocol
changes
and
so
on.
How
does
the
network
behave
improve
or
deteriorate
if
we
deployed
this
particular
King
set
to
the
entire
network?
How
would
this
particular
change
affect
the
current
overall
emergent
behavior
when
the
current
composition
of
the
network?
C
What
we
are
calling
live
canary
testing
so
basically
a
particular
change
set
a
particular
release,
a
particular
commit
we're
gonna,
be
deploying
it
to
to
the
net
where
we're
gonna
be
running
it
against
a
live
network
and
running
a
master
test
plan
and
capturing
a
lot
of
results
and
a
lot
of
metrics
from
each
of
those
test
cases.
There's
gonna
be
a
test
coordination
service.
This
we're
gonna,
integrate
all
of
this
with
a
software
engineering
tooling
that
we
have
in
place.
This
is
a
key
aspect.
C
We
want
this
tool
to
be
integrated
in
day
to
day
lives
of
ipfs
engineers,
otherwise
is
gonna
stay
in
an
island
users.
Engineers
are
going
to
be
able
to
manually
trigger
jobs
from
from
github
we're
gonna
publish
all
results
in
a
dashboard
that
would
be
accessible
as
well
to
the
community
so
that
you
can
engineers
will
be
able
to
compare.
You
know
how
did
this
particular
commit
affect
the
behavior
relative
to
this
other
commit
right,
we're
gonna,
have
this
party
we're
gonna,
build
a
basic.
D
C
Of
initial
test
scenarios
and
which
is
going
to
be
expanded
over
the
next
quarters,
we
expressly
taking
a
minimal,
viable
approach
to
this.
We
know
that
the
test
lab
and
like
test
bed
and
so
on
efforts
have
been
going
on
or
have
been
in
conceptualization
stages
for
a
long
time
that
been
several
iteration.
So
we
really
wanted
AI
dumb
it
down
this
time
and
get
to
the
essentials
such
that
we
can
unblock.
You
know
the
rainy
stream.
That
is
the
core
objective
right
now.
C
Another
type
of
tests
that
we're
gonna
be
working
on
will
be
calling
private
private
network
tests,
and
these
are
basically
running
an
IP
FS
release
against
a
inside
a
private
network
that
we
control
will
different
that
also
simulation
and
different
levels
of
control.
So,
on
the
first
on
a
first
iteration
we're
going
to
be
focusing
on
non-record
usable
test
test
scenarios,
this
is
basically
running
a
test
without
full
control
of
you
know
the
topology,
the
network
and
so
on,
and
probably
run
it
a
few
times.
C
So
you
can
get
different
observations
and
inform
judgments
based
on
that,
then
we're
going
to
be
evolving,
that
towards
reproducible
test
scenarios.
Where
will
you
actually
run
a
word
code
within
text?
Pictures
fixtures
that
will
be
able
to
provide
a
fixed
environment
for
those
tests
to
get
more
accurate
and
repeatable
results
all
the
time
and
then?
Finally,
what
I'm
calling
her
a
bit
hermetic
test
cases
which
is
a
bit
of
an
esoteric
term,
but
basically
it
means
we
control
every
single
aspect
of
that
test,
including
the
network.
C
So
we
need
the
until
we
control
the
environment
that
that
particular
test
is
running
on
in
terms
of
I/o
in
terms
of
CPU
in
terms
of
network
latency,
jitter
and
packet
loss
and
well
everything
that
you
can
think
about
such
that
time
over
time.
Every
single
iteration
of
those
tests
we
expect
them
to
produce
if
run
against
the
same,
commits
like
very
little
jitter
in
the
results,
if
anything
very
much
situations,
but
they
should
be
completely
repeated
and
I
think
I
did
consume
about
five
minutes.
A
A
D
All
right
we
were
coming
in
from
from
D
webcam,
where
there
are
various
projects
being
set
up
around
this.
It
looks
very
cold.
It's
a
little
chilly,
it's
not
freezing
so
project
operations.
This
was
the
group
that
was
basically
tasked
with
keeping
via
the
nuts
and
bolts
of
the
operation
running
through
the
quarter,
serving
the
needs
of
the
other
task
forces
and
making
sure
that
several
high
priority
initiatives
that
we
do
want
to
keep
pushing
along
that
by
bit
are
not
lost.
There
are
a
couple
of
well.
There
is
a
cast
of
characters
involved.
D
Each
one
of
these
different
areas
in
the
project,
Operations
Group,
could
be
a
task
force
of
its
own.
Some
of
them
are
several
working
groups
combined
to
some
extent,
so
they,
you
will
probably
hear
their
voices
in
the
future,
but
I
don't
do
a
the
best
that
I
can
do
in
summarizing
what
they're
going
to
be
doing
this
quarter.
The
first
is
the
ipfs
implementations
groups,
both
Jas
and
go
are
shipping
quality
releases
on
a
schedule.
D
If
you
look
into
the
details
of
the
key
results
for
this
quarter,
there
are
some
interesting
overlaps
with
the
test
group
and
it's
great
to
see
all
of
these
groups
focusing
on
quality
and
stability
in
their
core
work.
This
quarter,
as
well
as
a
number
of
changes
and
improvements
to
the
process
of
how
the
releases
are
constructed,
formulated
and
shipped
that
the
second
major
segment
is
the
activist
community
work.
A
lot
of
this
work
is
really
focused
on
improving
our
communications
infrastructure,
making
sure
that
the
community
is
well
supported.
D
The
last
couple
of
quarters
we
had
some
things
that
we
needed
to
communicate
to
the
community,
everything
from
network
scale
stuff
to
how
we're
responding
to
operations
and
and
Gateway
down
times,
and
things
like
that
to
also
some
of
the
changes
that
we
made
in
in
the
project
and
prioritization.
So
the
some
of
the
work
here
surrounds
making
sure
that
we
have,
and
also
things
like
IP
FS
camp,
that
we
have
so
much
to
share
all
the
things
that
came
out
of
camp
the
community.
D
That's
really
probably
six
months
worth
of
stories
that
we
could
we
be
telling
so
having
an
organized
and
and
clear
schedule
and
infrastructure
to
support
that
communication
in
an
ongoing
way.
Keeping
the
community
both
informed
and
engaged
and
aware
of
what
we're
focusing
on
the.
D
My
anti
hacked
hot
spot
got
me
again
locked
my
own
screen,
the
third
area,
as
the
collaborations
group,
so
Arkady
and
I
spent
some
time
last
week,
credit
sitting
down
and
kind
of
designing
how
we
can
manage
the
really
vast
and
diverse
set
of
partners
that
we
have
in
the
relationships
that
we
have
with
them,
and
also
a
system
for
understanding
who
our
future
partners
might
be.
Who
is
the
total
set
of
people
that
are
using
idea
fests
implementing
their
projects?
It
really
dat,
adding
ipfs
to
new
whole
industries
that
that
we've
been
saying?
A
E
The
the
experience
from
performance
and
kind
of
the
the
root
for
it
not
being
able
to
to
kind
of
set
those
users
up
to
fail
is
really
key
for
this,
because
we
want
people
to
to
kind
of
be
able
to
self-serve
and
be
able
to
take
the
tooling
that
we
have
and
run
with
it
and
start
to
kind
of
be
able
to
bring
in
all
the
package
managers
into
ipfs,
without
necessarily
needing
us
to
be
pushing
that
directly.
Next
slide.
A
E
E
They're
very
cute,
so
it
kind
of
balances
out
the
the
first
thing
that
we're
gonna
do
is
develop
a
number
of
been
like
baseline
benchmarks
to
give
us
visibility
on
the
current
state
of
ipfs
and
how
that
how
it
handles
importing
these
very
large
directories
of
files.
Looking
at
up
to
a
million
files
in,
say
the
Ubuntu
repository,
which
is
significant
and
also
kind
of
unknown
as
to
the
different
ways
that
this
can
be
done
and
then
from
the
end
users
perspective
as
well.
Once
you
have
a
mirror
on
ipfs.
E
I
get
lost
a
little
bit
yes
a
little
bit,
so
those
benchmarks
will
give
us
the
kind
of
the
path
to
then
go
down
for
the
rest
of
the
quarter
and
actually
try
and
improve
those
things,
and
we
already
have
a
fairly
good
idea
of
the
process
of
setting
up
a
package.
Manager'
they're
documented
across
every
different
kind
of
linux,
distro
already,
and
we're
literally
going
to
be
running
those
things
and
working
out
like
where
are
the
pain
points
in
ipfs
right
now?
Next
slide.
E
The
speed
of
adding
data
to
ipfs
is
one
of
the
key
things
we're
going
to
be
hitting
and
that's
like
the
majority
of
the
work.
So
it's
it's
very
laser
focused
on
being
able
to
successfully
import
these
giant
file
systems
into
IP
FST's,
and
then
we
can
move
on
to
actually
how
we
go
about
spreading
that
across
the
internet
and
alongside
that,
we're
actually
kind
of
slightly
experimenting
with
the
way
that
we
work.
E
F
Yes,
start
a
timer
here
too,
that's
like
a
good
idea.
Yeah
one
of
the
three
big
areas
that
Molly
mentioned
were
surfaced
in
last.
Quarter
was
indeed
documentation.
User
experience,
no
ipfs
is
is
hard
to
use
and
we
want
to
make
it
a
little
bit
more
accessible
and,
of
course,
Docs
is
like
a
small
word
which
contains
the
universe.
F
What
she
did
for
cluster
in
the
last
quarter,
so
we
had
that
as
kind
of
a
pattern
to
iterate
on
and
follow,
and
you
know
the
it's:
it's
gonna
serve
his
problems,
its
surface
problems
already,
in
fact,
some
I
believe
someone's
already
issued
eight
or
posted
a
PR
for
a
fix
for
one
of
the
things
that
was
surfaced.
So
we
are,
you
know,
we're
we're
building
the
plane
while
we're
flying
it
and-
and
it's
awesome
fantastic.
F
So
then
we
want
to
already
talk
about
the
features
and
that
we
want
to
look
for
in
a
in
a
tool,
and
maybe
we
end
up
where
we
are
now,
but
we're
thinking,
perhaps
something
more
along
the
lines
of
a
Gatsby
and
people
are
slacking.
Some
ideas
like
is
it
docu,
Soros
or
something
was
thrown
out
yesterday
and
slack.
If
people
have
ideas,
please
please
bring
them
on.
We
have
a
dock
that
were
that
we're
putting
together
with
that
lays
out
the
different
features
and
advantages.
F
A
big
one
is
translation,
ability,
the
internationalization
and,
surprisingly,
not
a
lot.
Not
a
lot
of
tools
really
have
that
as
sort
of
a
core
functionality,
and
we
want
something
that
we
don't
have
to
you
know
do
a
lot
of
customers
that,
like
you,
know,
custom
building
for
so
I
think
docu
source,
maybe
maybe
does
have
that
and
then
down
to
understanding
our
users
and
their
needs
where
we
have
some
some
super
awesome.
F
Ux
talent
on
the
team
now,
for
example,
Jessica,
and
we
want
to
just
be
methodical
about
who
our
users
are
and
in
one
way
we
can
get
some
metrics
surrounding
that
is
to
quickly
create
kind
of
a
question
feature
on
the
homepage
of
IPF
SEO.
Where
people
can
they
can
pick
from
a
list
of
you
know
what
what
do
I
want
to
do?
You
know
what
am
I
trying
to
accomplish.
You
know
I'm
I'm,
an
app
developer.
How?
How
can
I
BFS
make
my
life
easier
or
something
like
that?
F
And
and
these
would
this
would
kind
of
be
kind
of
a
double-edged
sword?
Hopefully
it
would,
it
would
make
the
content,
it
would
drive
these
people
to
maybe
a
a
specific
area
of
the
current
documentation
site
that
has
that
has
some
content
that
might
be
useful
to
them,
but
at
the
same
time
it
also
gives
us
some
metrics.
F
We
want
to
see
about
getting
a
documentation
specialist,
whether
that's
a
scope,
contributor
or
or
whatever
someone
who
is
good
at
words,
and
you
know
that
that's
the
most
important
thing
and
that's
going
to
take
some
work
as
we
know
to
bring
on
anyone.
You
always
takes
a
little
time
and
we're
not
seeing
that
as
a
blocker,
though,
to
to
improving
our
content
right
now,
because
we
have
content
right
now.
F
That
was
graded
by
a
lot
of
us
and-
and
we
have
we
understand,
a
lot
of
us
understand,
ipfs
better
than
anyone
else
in
the
world,
and
so
a
be
leaning
on
I
might
be
leaning
reaching
out
to
folks
to
contribute
and
also
feel
to
feel
your
pain
points.
You
know
I
we
want
to.
We
really
want
to
hear
what
you
have
heard.
You
know
if
you've,
if
you
yourself,
have
experienced
confusion
throughout
documentation
or
you
are
aware
of
issues
that
people
have
posted
or
just
conversations,
maybe
the
ipfs
camp.
A
A
Round
of
applause
for
all
of
the
the
folks
who
spent
their
hard
time,
I'd
worked
getting
in
their
classic
a
mutant
class
could
be
that
an
awesome,
awesome,
awesome,
yeah
andrew,
is
also
mentioning
in
the
chat
he
got
a
fix
in
for
the
docs
already
I
I
created
a
PR
on
the
dogs,
so
we
we're
all.
We
can
all
chip
in
and
help
I'll
make
that
one
a
lot
better.
It's
not
not
only
the
docs
team
is
gonna
write
docs
for
all
the
great
Cox
awesome.