►
From YouTube: CHAOSS.Value.May.31.2019
Description
CHAOSS.Value.May.31.2019
A
It
should
be
soon
I
didn't
like
it
started.
Recording
sorry,
yeah
I
didn't
I,
didn't
notice
that
there,
usually
when
I,
do
the
check
to
just
pull
repos
off
of
an
organization
the,
but
you
know
I
started
within
a
day
or
two
and
the
repos
are
all
there,
and
so
everything
just
works,
but
Comcast
moved
around
about
a
dozen
repos
in
the
day
or
two
between
when
I
pulled
the
list
of
repos
and
started
the
analysis
process.
A
So
it
just
façades
component
doesn't
handle
things
that
it
thinks
are
supposed
to
be
there
and
then
aren't
very
well
so
I
just
caught
that
middle.
Oh
this
week,
everything
it
is
passed
compact
Comcast,
so
it
shouldn't
be
shouldn't,
be
passed
this
weekend
for
everything
to
finish
and
I'll
keep
a
closer
eye
on
you
know.
I
was
at
a
conference
and
it
kind
of
didn't
attend
as
carefully
as
I,
maybe
thought
to
have
to
that
initial
process.
I.
B
A
Might
do
because
there
see
CTOs
I,
don't
know
what
her
official
title
is,
but
she's
on
actually
she's,
given
some
talks
of
like
chaos,
events
so
I
think
I.
Think
they're
I
mean
judging
by
the
number
of
repositories.
They
have
there's
a
lot.
No,
it's
interesting
another
thing,
that's
interesting
about
Comcast's
github
or
goes
it
contains
an
above
average
number
of
forks
of
things
that
they
use.
A
So
there
are
some
some
aspects
and
I
can
go
through
and
provide
a
list
of
the
ones
that
are
forced,
but
there
are
some
some
aspects
of
the
way
that
they're
using
github
that
are
different
than
others
are
using
it
and
I'm,
not
saying
it's
good
or
bad.
But
I'm
saying
we
should
be
aware
that
that
some
things
that
we
say
about
Comcast
looking
at
their
repos,
probably
has
to
take
into
account
that
factor
and
that's
just
they're,
just
different
than
others-
that
I've
looked
at
before
right.
I
mean
it's
not
like.
A
Nobody
has
Forks
in
their
org,
but
it's
it's
few
and
far
between
and
Comcast
has
a
lot
of
works.
Sean.
Can
you
tell
me
what
you're
searching
on
these
repositories?
For
so
what
facade
does
the?
Basically,
we
substantially
altered
facade
at
this
point,
but
what
the
core
of
facade
does
is
it
just
goes
through
a
repose
commit
log
who
committed
it
and
who
authored
it.
A
I
do
have
a
version
of
facade
that
I'm
working
on
that
or
the
facade
engine
I
think
I'll
call
it
now
that
I'm
working
on
that
will
incorporate
timestamps
as
well.
I've
done
one
a
group
of
6800
repos
with
that
approach.
The
the
only
thing
that's
broken
there
right
now
is
the
the
develop.
The
generation
of
the
cache
tables
have
a
dependency
on
matching
dates.
A
A
D
A
Right
so
promote
from
a
value
perspective,
the
the
goal
is
to
I
think
provide
a
way
to
do
kind
of
our
hourly
rate.
Calculations
on
the
fly
I
have
a
spreadsheet
that
I've
developed
that's
kind
of
the
model
that
I'm
using
where
essentially
I,
put
in
a
the
labor
rate
and
how
I
want
to
treat
complexity,
which
is
a
multiplier
and
I
multiply
lines
of
code
by
complexity
by
a
factor
for
hourly
rate
and
a
factor
for
how
I
want
to
treat
complexity.
Now
that
needs
to
be
refined.
A
That's
just
using
the
numbers
inside
of
the
COCOMO
model
deployed
by
SCC.
Okay,
you
put
that
on
the
front
end
instead
of
having
it
be
something
that's
generated
on
the
back.
Ok,
so
there's
really.
How
are
you,
okay,
so
I
think
there's
a
lot
of
analysis
that
was
gonna
have
to
happen
contextually
right,
so
none
of
this
is
gonna
stand
alone,
like
you
can't
run
this
and
say
in
this
group
of
I.
Think
we've
got
like
2500
projects.
A
We,
you
know
you
can't
just
apply
a
labor
rate
to
all
the
projects,
even
inside
of
an
org
and
make
an
estimate.
You
have
to
take
into
account
various
contextual
clues,
the
the
ones
that
I
think
we
can
account
for
best
are
our
ones
related
to
essentially
standard
multipliers
for
language
use
and
then
I.
Think,
there's
also
a
need
to
understand.
As
part
of
the
complexity
issue
is
so
complexity,
is
it
does
take
into
account
the
languages
used?
A
D
So,
fundamentally,
though,
the
data
that's
coming
out
from
these
repositories
around
files
that
are
deadlines
that
we
reach
commitment,
all
that
kind
of
stuff
mm-hmm,
then
applying
kind
of
a
value
perspective
via
labor
rate
and
complexity,
issues
right
and
how
complexity
is
determined,
has
kind
of
quite
a
bit
under
the
hood.
Yes,.
A
Quite
a
bit
under
the
hood
and
I
think
quite
a
bit
that
needs
to
be
more.
You
know.
Scc
is
a
general
tool
that
I
think
provides
some
utility
for
people
and
we
can
use
the
output
of
that
tool
which
actually
runs
on
a
repository
at
a
point
in
time,
not
on
the
commits
that
can
be
applied
to
value-added
calculations
for
each
commit
okay,
and
so
we
can
start
to
get
trend
lines
that
way
and
I
think
once
once
Andy,
for
example,
is
able
to
look
at
this
and
input
some
parameters.
A
Rate
in
certain
components
of
complexity,
delivery,
the
labor
rate
right
now
is
is
an
obvious
parameter.
The
complexity
is
I,
think
has
to
be
parameterised,
because
it
varies
somewhat.
I
would
say
wildly
from
language
to
language
and
I
need
to
dig
in
a
little
bit
more
on
on
that
and
build
build,
something
that
allows
it
to
be
parameterised.
I.
Think
at
the
language
level
initially
will
give
us
much
finer
it'll,
be
like
the
smallest
investment
for
the
most.
A
E
A
D
A
B
A
B
B
Automated
and
you
know
the
only
thing
that
kind
of
popped
into
my
head
is,
if
you
had
such
a
metric,
the
other
thing
that
I
think
would
be
interesting
to
organisations
is
to
understand
what
their
bench
strength
is.
You
know
across
the
whole
code
base,
so
you
may
have
let's
say
justjust
in
one
repo.
You
may
have
a
handful
of
developers
and
it's
it's
possible
that
for
a
critical,
you
know
high
complexity
area
of
the
code.
B
B
A
It
would
be
pretty
easy
and
what
I'm
running
down
is
a
number
of
people.
Who've
touched
each
file
and
when
they
touched
it
and
of
course
who
they
are
mm-hmm
that
would
and
then
you
could
just
sort
that
by
files
with
a
lot
of
commits
only
by
one
or
two
people,
especially
in
the
last
say
year,
yeah.
Those
would
be
your
higher
risk
areas
right,
I
think
it's
higher
risk
if
it's
over
the
life
of
a
file
or
directory
which
we
can
get
to
directory.
A
If
we
wanted
to
get
fancy,
if
it's
buying
the
file
or
by
the
directory,
then
you
know
we
can
look
at
the
life
of
it
and
if
it's
a
central
piece
of
the
code
and
has
a
small
number
of
people
over
its
entire
life
cycle,
those
probably
be,
but
it's
probably
a
more
refined
version
of
bus
factor
than
then
what's
presently
used
right
now,
we're
just
looking
at
you
know
how
many
people
are
a
quarter.
This
repository,
not
how
many
people
are
core
to
a
part
of
the
repository
and
really.
B
A
Art
there's
dozens
of
people
or
hundreds
of
people
in
some
cases
working
in
these
repos
many
of
them
actively
and
if
there's
a
piece
of
that
that
you
know
that
might
show
up
with
a
low
bus
factor
using
standard
algorithms
for
bus
factor.
But
if
you've
got
a
core
section
of
it
that
depends
on
one
or
two
people.
Then
you
have
kind
of
a
componentized
version
of
bus
factor.
A
I
guess
I
would
say
how
do
we
know
what
a
core
section
is
so
my
my
first
water,
my
first
swag
at
what
his
core
would
be
things
that
have
routine
commits
to
them
and
are
probably
you
know,
bigger
files
in
general,
so
files
or
directories
that
are
small
with
a
lot
of
commits,
probably
are
less
indicative
of
core
and
more
indicative
of
small
feature
changes.
Files
that
are
larger
in
size
and
you
know
I,
would
that's
a
guess
right.
A
D
D
A
Right,
I'm,
sorry,
really
close
this
one's
rattle
yeah,
then
then
you
can
have,
but
but
still
I
think
we're
talking
about
heuristics
and
that's
as
good
as
anything
I've
suggested.
You
know,
I
can
I
can
see
where
it
would.
I
can
see
in
each
case
where
there
are
reasons
to
be
cautious
about
just
using
that
one
form
of
assessment.
A
What
sorry?
What
is
bus
factor?
I've
never
heard
that
term
before
it's
the
number
of
developers
on
a
project
that
can
be
hit
by
a
bus
about
the
project
completely
dying,
so
a
bus
factor
of
one
is
really
bad
must
factor
of
five
ten.
Those
are
all
better.
Higher
numbers
are
better,
it's
the
opposite
of
golf.
A
So
what
you
want
is
high
bus
factor,
yeah
and
also
there's
a
corresponding
element,
which
requires
the
resolution
of
whom
people
work
for,
but
elephant
factor
is
the
number
of
organizations
contributing
to
a
project
so
like
when
you
look
at
repositories
owned
by
really
any
tech
company
in
many
cases
like
Facebook.
Is
this
way
Twitter?
It's
this
way,
Red
Hat!
A
So
you
can
have
something
really
useful
like
Twitter.
Bootstrap
is
a
web
framework,
that's
very
popular
and
has
almost
entirely
twitter.com
contributors
and
it's
widely
used.
So
you
can
have
something
that
has
a
low
elephant
factor
and
have
it
still
be
widely
you
same
with
Facebook's
libraries
that
are
now
widely
used.
So
it's
not
necessarily
indicative
of
anything.
That's
negative
in
terms
of
business
viability,
unlike
bus
factor,
which
is
well
again.
D
A
Mean
when
I
think
about
the
these
measurements
that
we
can
do
I
think
about
kiosk
on
last
August,
and
you
know
there
is
a
desire
for
these
red,
yellow,
green
dashboards
and
they're
her
to
arrive
at
because
there
is
so
much
context
around
all
these
projects.
So,
on
the
one
hand,
chaos
is
agnostic
and,
on
the
other
hand,
I
think
at
least
I
have
one
away.
I
think
I
have
to
talk
about
the
ways
that
things
can
be
used
and
the
potential
utility
and
also
the
potential
lack
of
utility
or
the
I'm
always
concerned
about.
A
You
know
so
even
within
the
organizations
that
we've
cloned
I
think
we'll
see
different
kinds
of
projects
that
we
might
want
to
instrument
in
different
ways
and
I.
Think
that's
where
the
work
that
Vinod
and
I
are
doing
right
now
and
Kevin
actually
around
genres
of
open
source.
I.
Think!
That's
where
that
some
of
that's
going
to
have
a
lot
of
utility
for
chaos.
D
D
Right
well,
then,
there's
work
that
needs
to
be
done.
So,
if
you
so
here,
here's
the
document
there's
some
mismatch,
that's
going
on.
So
this
is
the
spreadsheet
that
we're
using
to
track
potential
metrics
mm-hmm,
and
these
are
the
focus
areas
from
value
and
they're
they're,
not
necessarily
commensurate
with
one
another.
A
D
B
D
A
A
A
B
How
about
if
we
said
our
metrics
are
going
to
be
just
the
top
four
and
in
addition
to
those
top
four
metrics
you
know
we'll
have
some
way
of
exporting
the
JSON
or
the
CSV.
You
know
into
a
spreadsheet
and
then
you
know,
maybe
maybe
we
could
produce
like
a
one-page
guide.
Saying:
hey
yeah,
you
know
here's
your
version,
zero.
You
know
entry
into
the
world
of
parameterised
metrics,
you
know,
you've
got
your
CSV.
You've
got
your
spreadsheet.
Here
are
some
examples
of
things
you
might
want
to
do
to.
A
A
So
SCC
the
the
software
that
I'm
using
for
the
Kokomo
also
does
labor
estimates
and
there
are
built-in
Koran.
There
are
built-in
values
for
the
parameters
that
they
use
for
standard
parameters
and
okay,
so
that's
accessible
and
I
said:
that's
accessible,
I'm,
just
skeptical
about
the
standard
parameters,
because
it
shows
me
some
things
that
look
a
little
wild
and
so
I,
don't
trust
it
and
right
left.
You
were
saying,
like
Al
Gore's
worth
nine
million
dollars,
it
would
cost
a
crate.
Auger
I
know
that's
wrong
or
we
are
amazing.
B
Well,
I
think
it
depends
a
lot
on
context.
Okay,
you've
got
Shawn
and
in
your
world
you
know
people
who
come
with
intrinsic
motivation
and
if
you
are
in
a
corporate
environment-
and
you
had
to
recreate
what
you've
got
with
people
who
are
just
motivated
by
you
know,
dollars
and
cents
very
well
could
be
that
the
cost
of
you
high.
Let's
say
if
it
was
part
of
a
government
agency
and
you
had
to
go
and
hire
contractors
or
consultants,
mm-hmm.
F
A
B
F
A
Sure
that's
right,
it's
probably
more
than
three
times
by
still
think
it's
close
to
nine
million
dollars.
Unless
it's
like
a
really
bad
consulting
company
like
if
I
had
a
giant
pot
of
money,
then
an
unlimited
amount
that
I
could
spend
then
I'm
sure
somebody
could
find
a
way
to
charge
nine
million
dollars
to
build
it.
B
F
D
D
D
A
D
A
D
D
D
D
B
D
A
D
A
A
It
may
not
be
like
we
may
think
of
response
time
in
a
different
way
than
the
other
working
group,
and
so
then
we
would
need
to
define
our
metric
and
probably
reference
the
other
one
and
identify
a
section
for
how
they
are
distinct.
Okay.
That
would
be
my
like
guess,
because
I
I
can
see
where
we
would
call
something
the
same
and
actually,
when
we
get
to
the
details
or
have
a
pretty
different
okay
and.
A
And
I'll
try
to
arrive
I'll,
try
to
deploy
what
we've
used
on
evolution,
oh
yeah,
and
if
I
mean
if
they
can
cross.
If
it's
the
same
thing
use
it
well,
I
mean
a
naming
conventions
as
well,
so
I
may
refine
the
the
names
for
some
of
these
things
according
to
some
kind
of
standard.
Okay,
all
right,
for
example,
because
this
is
in
a
markdown
document,
we
reference
it.
We
probably
will
have
dashes
instead
of
spaces.
So
if
nothing
else
changes
that's
likely
to
change
in
both
places.
A
D
D
A
D
A
I
think
that
would
be
great
and
also
you'll
have
I,
can
guarantee
you'll
have
auger
stuff
to
look
at
before
that
as
well.
Okay,.
D
A
If
you
like,
are
you
wanting
is
so
I
mean
I
have
docker
configurations,
there's
a
doctor
and
the
in
the
auto
repo
probably
needs
to
be
updated.
Okay,
but
if
you
I
think
if
you
created
a
nod
like
it
once
I've
got,
if
I
could
point
you
at
a
version
and
give
you
a
docker,
config
I
think
right
now,
we're
still
we're
I'm
still
doing
something.
I'm
still
have
I'm
halfway
in
each
world
right
now,
halfway
in
the
in
halfway
in
the
old
so
well.
B
C
C
B
B
C
A
A
C
A
And
I
I
believe
I
think
you
may
have
missed
the
very
beginning
of
the
call
when
I
said
that
I
had
some
issues
with
the
Comcast
repos
not
being
there
and
got
stuck
on
that,
because
I
didn't
attend
to
it.
I
just
assumed
that
everything
was
gonna
work
because
it
usually
does
the
first
time
through,
but
Comcast
moved
around
about
a
dozen
repos
and
when
they're
not
there
and
the
facade
engine
chokes
I
think
was.
Is
this
related
to
a
release.
A
That's
gonna
stay
there
just
because
there
are
useful
endpoints
that
people
use
and
useful
front-end
graphics
that
people
use
to
do
comparisons
with
any
github
repo
and
so
I
think
we'll
keep
that
alive,
but
not
make
it
part
of
our
deployment.
Make
it
something
that
people
just
get
access
to
a
shared
gh
torrent
database
that
they
can
access
if
it
starts
to
become
a
problem
with
performance
than
all
I'll
solve
that
later.
But
right
now
that,
when
is
your
next
release?
Do
you
think
a
little.
D
A
B
B
So
I
had
just
a
little
bit
of
feedback
marketing-wise
when
you
do
a
Google
search
for
the
word
to
augur
or
marker
software.
The
blockchain
people
come
up
the
blockchain
people.
Maybe
it
would
be
helpful
to
refer
to
the
project
as
augur
metrics
and
then
you
know
tweak
the
the
home
page.
So
it
you
know
if
people
are
trained
to
look
for
auger
metrics.
Maybe
that
would
be
helpful.
That's.
A
A
good
suggestion
I
mean
we're.
Certainly
you
know
it's
like
they
came
up
with
their
name
like
two
weeks
before
we
came
up
with
ours
and
we
had
no
way
of
I
mean
we
probably
should
we
were
aware
of
it.
We
probably
just
should
have
gone
with
something
else,
but
it
took
my
team
so
long
to
come
up
with
a
name.
They
all
agreed
on
that
I
didn't
want
to
I
just
went
with
it,
which
I
think
in
hindsight
is
a
marketing
error.
A
D
D
A
D
I
mean
it's
just
it's
just
basically
what
it's
gonna
come
down
to
Andy
is
the
code
of
conduct.
Document
will
still
be
in
the
value
repository
it'll
still
be
there
that
markdown
file,
but
it'll
just
be
instead
of
the
actual
full
code
of
conduct
and
will
just
be
a
small
link
that
says
here's
the
code
of
conduct.
Please
click
here,
sure
you
don't
see
it
that
way.
We
don't
have
like
400
codes
of
conduct
across
all
repositories,
and
if
we
do
a
small
tweak
to
the
overall
chaos
code
of
conduct,
you.