►
From YouTube: CHAOSS Metrics Models Working Group Sept 12, 2023
Description
Minutes from this meeting can be found here: https://docs.google.com/document/d/1qE2duE73NbEOHzRRAbqgPSUSAfdH-AmvLruOu7uCIYg/edit#heading=h.6kvj3hfgqkjk
A
All
right
excuse
me
hi
everybody
and
welcome
to
today's
metric
model.
Meeting
chaos
metric
model
meetings.
Good
to
have
you
here
minutes
are
in
the
chat
I
put
together
an
agenda
for
today
feel
free
to
add
anything
that
you
would
like
to
on
the
agenda.
I
think
we
have
a
few
things
we'd
like
to
talk
about
today.
That
might
lead
to
some
good
conversation.
So
again,
minutes
are
in
the
chat.
I
can
share
them
here
again.
A
A
A
Also,
there
is
an
effort
called
sustain,
I'm,
not
sure.
If
you're
familiar
with
sustain
they're
a
group,
that's
been
around
for
a
long
time,
thinking
about
ways
to
sustain
open
source
kind
of
just
as
the
name
says,
and
they
have
a
new
working
group
called
the
academic
working
group
and
trying
to
organize
efforts
around
academic,
open
source.
A
A
B
A
D
C
A
D
A
C
A
A
C
B
A
So
we're
bringing
together
people
who
can
talk
about
metrics
that
might
be
relevant
to
them
or
metrics
models
that
might
be
relevant
to
them
in
their
particular
context,
but
not
having
to
do
all
of
the
chaos
work
that
comes
with
actually
building
a
metric
or
metric
model.
So
the
idea
is
that
these
groups
can
talk
a
little
bit
more
freely
about
the
things
that
they
would
like
to
see.
A
So
we
have,
we
have
these
Frameworks
that
we've
been
building
in
these
different
context
groups,
and
so
this
isn't
perfect
and
I
need
to
make
a
few
changes
here.
But
so
there
are
a
number
of
things
that
say:
a
university
cares
about
with
respect
to
open
source,
so
research,
Excellence
research,
translation.
E
A
A
broader
audience
or
how
does
open
source
helps
support
the
mission
of
research
Excellence
at
a
university,
so
they're
thinking
about
kind
of
specific
things,
at
least
in
this
case
around
research,
excellence
and
research
translation.
This
one
right
here
is
more
like
how
is
open
source
being
used
in
the
curriculum
and
how
are
we
as
an
open
source
program
office,
providing
training
for
students
and
faculty
around
around
open
source?
A
We
would
actually
bring
that
work
say
here
to
the
metrics
model
group
or
we
would
bring
that
work
to
the
Common
working
group
to
actually
do
the
development
of
the
metrics
and
metrics
models.
So,
if
we're
not
asking
these
groups
to
do
that,
work,
we're
just
asking
them
to
kind
of
talk
about
the
things
they
would
like
to
see
so
we're
trying
to
meet
people
where
they
are
without
like
showing
them
all
of
the
work
in
chaos
like
they
don't
overwhelming.
A
Okay,
okay,
so
let's
see
we
can
move
on
the
OSS,
EU
Workshop,
so
I
think
this
is
kind
of
where
we
we
settled.
I
have
reached
out
to
to
some
folks
about
whether
or
not
they'll
be
able
to
attend
a
lot
of
people
just
got
back
from
a
little
bit
of
a
break
like
their
own
vacation,
so
I
hope
to
hear
from
some
people
today,
maybe
or
tomorrow,
about
their
ability
to
attend
this
Workshop
and
participate
in
it.
C
A
C
A
C
A
C
I
I
I
I
I
kind
of
draw
something
for
that.
Maybe
we
come
first.
You
want
people
from
open
ruler
to
to
say
something
about
their:
how
to
call
how
to
manage
their
community
Through.
The
Health
evaluation
I
mean
from
that
dashboard,
and
also
we
can
discuss
a
little
bit
about
I
mean
for
the
rest
of
time.
We
we
can.
We
can
ask
the
questions
to
that
community
and
also
we
can
invite
some
other
guys
who
may
have
interest
related
to
the
chaos.
We
can
discard
everything.
Okay,.
A
A
E
A
A
A
What
I
know
we
have
one?
Which
is
it's
this
right?
This
is
what
is
kind
of
prompting
this
issue.
Mm-Hmm
I'll
be
to
be
totally
honest,
I'm
a
little
hesitant
about,
because
we
have
metrics
and
metrics
models
and
then
metric
model
models
or
model
chains
like
there.
There's
always
this
concern
that
people
are
a
little
confused
about
what
goes
on
in
the
chaos
project.
This
will
add
to
that.
This
will
add
to
that.
A
A
A
B
D
E
C
B
C
E
A
E
D
A
A
Metrics
and
auger
is
not
yeah,
which
is
I,
mean
that's
totally
fine,
but
and
then,
and
then,
once
we
kind
of
like
Compass
once
we
can
have
those
metrics
deployed
in
auger,
the
the
deployment
of
the
metrics
models
will
be
that
much
simpler
or
that
much
more
accessible.
Just
because
we
have
the
base
metrics
available
to
us
and
bringing
them
together
in
meaningful
ways
through
a
model
wouldn't
be
as
complicated.
D
Yeah
well,
the
metrics
will
be
available
some
percentage
more
than
half
less
than
100
percent
of
chaos.
Metrics
are
available
in
auger,
but
so
part
of
what
we're
doing
is
just
a
mapping
exercise
where
here's
how
you
find
it.
This
chaos
metric
in
auger
there
is
where
it
is
and
part
of
it
is
wherever
there's
a
gap
where
we
don't
exactly
have
that
metric
to
filling
it
and
creating
an
API
that
generates
that
metric.
D
C
So
so
max
the
question
of
course:
yeah
yeah
for
the
auger,
so
yeah
or
the
whole
August
solution.
So
it's
not
just
the
provides
as
a
metric
related
data
collection,
but
also
including
the
data
process
and
and
also
finally,
we
can.
We
can
get
the
metric
data
through
some
API
right,
yes,
okay,
so
it's
played
a
similar
I
mean
the
function
at
the
Grimlock.
D
D
Auger
is
more
focused
on
perhaps
the
data
engineering
side
of
things
so
ensuring
that
we
know
what
data
we
have
and
can
speak
to
the
level
of
completeness
within
it,
and
the
metrics
are
accessible
in
the
sense
that
you
can
follow
the
provenance
of
the
data
from
the
apis
and
the
mining
of
git
logs
through
to
the
database
and
then
generate
anything
you
want
with
it.
So
it's
not
it's
not
a
front
end
right
like
we
don't.
We
don't
really
provide
the
kind
of
front
end
that
other
tools
do,
but
we
do
have.
C
Us
something
like
a
pork
solution:
I
mean
it's
not
right
in
in
service,
so
I
mean
based
on
my
experience
of
the
compass
where
you
scream
like
by
the
background
to
collect
data
processed
it.
We
found
that
it's
really
hard
to
use
it
directly
in
production,
environment
yeah,
especially
considering
and
I,
mean
the
scope
of
data
and
and
the
pipeline.
C
We
have
to
consider
a
lot
of
things,
something
like
we
have
used
this.
The
message
bus
to
handle
in
the
different
message,
handling
yeah
from
backend
to
to
front
end
and,
of
course,
because
groom
live
I
mean
it's
a
database.
It's
a
it's
supported
by
open
search
or
elasticsearch.
So
it's
a
connector.
It
could
support
it
by
some
Frontline.
The
I
mean
write
the
inbox
solution,
something
like
kibana.
E
C
Can
access
it
from
the
graph
now
easily
so
address
the
groom
Library?
It's
I
mean
it's
a
full
solution,
no
matter
how
mature
or
not
it
is,
but
it
can
be
a
doing.
Some
I
mean
the
concept
of
proof
in
in
the
environment.
If
you
want
to
try
to
collect
data
to,
to
make
some
Metric
evaluation
on
your
community.
D
D
You
know,
there's
a
number
of
interest
in
instances
with
over
a
hundred
thousand
repositories
in
them,
so
auger
is
just
sort
of
built
to
scale
from
the
start
as
a
data
tool,
which
is
a
little
bit
different
from
Gamora
lab,
not
saying
it's
better
or
worse.
Just
it's.
You
know
it's
designed
for
that.
That
high
scale.
A
D
C
D
It
definitely
is
I
think
all
we
need
to
do
is
provide
you
with
an
API.
In
an
instance,
I
mean
I,
think
menager,
obviously
isn't.
No,
no
individual
instances
aiming
to
quote
unquote
gather
all
the
data,
but
we
have
some.
We
have
the
public
instance
that
could
be
leveraged
to
get
data
through
our
apis
for
a
large
number
of
repositories.
Yes,
for.
C
D
C
I'm
more
interested
on
the
metric
produced
by
auger,
so
I
I
do
some
pre-starty
before,
like
like
one
years
ago,
I
noticed
that
auger
could
could
provide
some
some
more
different
Matrix
compared
with
the
gram
life
yeah.
If
auger
could
continuously
to
provide
such
a
new
metrics
I,
don't
really
care,
it's
originally
from
chaos,
but
it's
if
it's
meaningful.
You
know
close
to
the
real
case
in
in
community.
D
E
C
C
So
maybe
we
can,
you
know,
put
auger
more
focus
on
the
I
mean
the
data
scientists,
work
for
custom,
metrics
and
the
matrix
model
yeah
and
for
the
rest
of
work.
We
can,
you
know,
live
to
Grim,
live
or
any
other
engine,
and
we
can.
You
know
to
enhance
that
advantage
of
auger
to
produce
more
valuable,
Matrix
and
Matrix
models.
C
C
B
C
D
D
C
A
B
C
A
thinking
if
all
the
data
is
ready,
I
mean
we
already
put
it
on
in
some
data
data
database,
which
is
cleaned
somehow,
and
we
use
auger
API
to
create
more
sentenced
things
I'm
going
to
do
the
scientific
thing
and
no
matter
in
that
in
the
education
on
the
research
work,
we
can
do
a
lot
of
things.
I
think.
E
A
A
really
interesting
conversation
and
I
really
I,
really
do
like
it
I
we
have
talked
about
Sean
and
I,
have
talked
about
making
auger
data
and
auger
API
endpoints
available
for
people
to
query
against
and
ask
these
more
scientific
questions
against.
We
do
have
a.
There
are
a
few
concerns
that
do
come
from
that,
and
one
is
the
expectation
that
scientists
would
have
for
the
data
and.
A
I,
don't
know
about
in
China,
but
the
same
okay
so
and
the
reason
is,
is
because
they
have.
They
have
academic
outcomes
that
are
driven
largely
by
the
publication
of
papers,
not
the
contribution
back
to
to
projects
that
help
them
produce
those
papers.
So
I
think
we
would
need
to
sort
that
out
a
little
bit
too,
that
we
don't
make
this
something
like
this
available
from
a
data
science
perspective
and
then
just
tell
Sean
all
right,
you
got
you've
got
you
know:
510
new
users.
A
So
if,
if
we
could
maybe
do
something
like
this
you'll
be
listening
to
you
talk,
you
know
with
a
select
group
of
people
that
might
make
some
really
good
sense.
You
know
you
had
mentioned
some
students
might
have
an
interest
in
this
work
and
there
may
be
some
faculty
that
are
like
identified
faculty.
That
may
have
an
interest
in
this
work
as
well
and.
A
C
I
think
in
in
this
database
I
mean
the
data
story
in
this
database.
We
can
divide
it
into
the
different
layers
and
I
mean
the
different
kind
of
data
based
on
the
precise
tips.
I
mean
we
can
provide
one
part
of
data
as
a
as
a
low
data.
It's
just
a
very
illegal
initial
data
without
any
cleaning
cleaning
process.
Maybe
some
guys
need
like
such
data,
because
that
data
would
contain
more
enriched
information
and
the
second
layer.
C
We
can
do
some
more
cleaner
process
and,
of
course
we
can
have
sir
level
and
all
those
data
available
for
this
researchers
and
scientists
and
the
students
they
can
based
on
their
requirement
to
fetch
those
data
and
and
create
their
own.
You
know
the
new
Matrix
or
matrix
model,
or
any
other
things
and
I
think
that
would
save
the
lot
of
time,
which
you
already
you
know,
take
things
more
than
I
know
this
more
than
70
percentage
of
the
time
on
the
whole
research
work.
B
D
What
are
you
thinking
on
this?
Well?
I
mean
I.
Think
yeah
I
mean
the
data
engineering
part.
That's
it's
been
it's
a
lot
of
work
to
make
sure
that
you're
handling
all
of
the
different
kind
of
anomalies
that
exist
in
data,
so
yeah
I
mean
auger,
provides
that,
like
it
deals
with
all
of
the
things
that
most
of
the
other
tooling
doesn't
deal
with.
C
E
C
A
A
D
E
C
A
B
A
Because
I'm
wondering
if
this
meeting
could
be
kind
of
that
that
it
would
be
you
know
here,
here's
how
we
go
about
identifying
the
end
points
that
need
to
be
created.
That
aren't
like
here's,
the
first
step
that
we
need
to
do
and
the
reason
we're
doing
this
step
is
so
that
we
can
do
two
things.
So
the
first
thing
would
be
so
we
can
develop
metrics
models
and
that
so
we
can
make
data
available
to
others.
A
B
D
D
A
A
D
D
I
mean
it's
give
people
examples,
I
mean
I,
think
that's
where
historically,
agar
has
done
a
good
job
as
making
data
available
for
new
things
that
we
want
to
measure
relatively
quickly
right
when
there's
a
new
data
source
for
a
new,
a
new
idea
for
how
to
shape
the
data.
We
can
move
on
that
pretty
quickly.
Just
because
of
how
we've
organized
what
we
already
have.
Okay,
because.
C
As
you
know
that
we
may,
we
may
build
and
provide
service
from
compost
right.
We
provide
some
yeah
is
a
dashboard,
but
but
people
you're
already
ask
if
you
could
provide
some
beautiful
apis
for
me,
I
can
build
more
other
things
right.
So
in
that
case,
if,
if
auger
could
provide
such
apis,
I
can
I
can
use
this
API
to
you
know,
connect
to
my
database
and
provide
data
to
to
other
guys
yeah.
A
A
Thing
for
us
to
do
this
is
what
we've
been
doing
all
along
and
for
auger
as
a
chaos
tool
to
support
many
metrics
and
as
many
metrics
models
as
it
can
is
just
a
sensible
thing.
Not
just
that.
Just
makes
a
lot
of
sense
in
doing
that,
it
sounds
like
it
will
start
to
make
more
data
available
to
others
in
the
form
of
metrics
and
metrics.
C
Models
based
on
based
on
based
on
my
understanding
that
you
know,
grow
my
power
or
many
different
background.
Sports
data
sources
like
size,
I
I,
know
all
girl
also
support
a
lot
of
different
data
sites.
Data
sources
exercise,
but
but
if
way
tree
to
grow
map
to
provide
money,
ability,
I
mean
the
capability
to
to
doing
the
data
collection
and
the
data
cleaning
work
and
I'll
tell
people.
Okay.
In
our
case,
in
our
in
our
community,
I
mean
in
chaos.
We
provide
two
main
software
components.
C
First
is
Google
Map
help
us
to
connect
with
different
many
different
data
sources
and
help
us
to
collection,
doing
the
data
collection
and
data
data
cleaning
and
process,
and
we
have
another
wonderful
software
component
to
provide
some
very
good,
just
API
or
any
other
apis
to
doing
the
data
scientist
work.
So
yeah
things
happen.
E
So
yeah
sorry
can
I
just
ask
a
question.
So
what
he
is
describing
right
now
is
what
my
understanding
of
the
difference
between
the
two
was
was
that
auger
was
really
focused
on
things
you
might
pull
from
GitHub
and
grimoire
lab
was
focused
on
like
being
able
to
go,
get
things
from
lots
of
places
and
so
like.
If
you
needed
to
get
your
stuff
out
of
jira
or
discourse
or
something,
then
you
were
looking
at
your
more
lab
and
if
you're,
just
looking
yeah,
so
I
guess
so
I
guess.
E
C
Or
yeah
I
mean
I
mean
currently
actually
grew.
My
also
support
for
the
GitHub
data
collection-
yes,
I
know
I
know.
We
know
that
auger
also
support
for
that,
but
I
think
the
the
advantage
of
vulgar
through
the
provides
the
most
wonderful
apis
to
create
more
high
level
application
of
data
scientists
work
and
to
like
the
rest
of
work
to
the
groom
life
yeah.
D
There's
a
lot
of
data
engineering
involved
in
getting
this.
These
metrics
actually
built
and
I
think
the
issue
that
you
people
run
into
with
grimoire
lab
is
just
when
they
hit
scale
when
they
go
above
a
thousand
repositories.
They
there
are
challenges
with
an
open
search
database
that
sort
of
cascade
through.
A
A
So
so
on
this,
this
is
really
a
great
conversation
and
has
taken
a
lot
of
good
time.
It
sounds
like
the
thing
that
we
need
to
think
about
is
kind
of
this.
This
bottom
part
here
with
the
points
that
Jen
had
brought
up
and
the
points
that
you
Hui
and
Sean
are
making,
which
is,
if,
if
we
make
these
these
endpoints
available
to
a
lot
of
people,
it
will
attract
people
who
have
an
interest
in
consuming
those
endpoints
yeah.
E
D
A
But
that
that's
cool,
because
I
think
to
Yahoo's
Point
way
early
on
that
starts
to
create
Community
around
these
tools,
like
people
who
have
an
interest
in
contributing
back
Upstream
to
the
tools,
because
on
one
end
we
could
basically
just
say
here
are
the
end
points
you
use
them
at
your
own
risk.
We
don't
we
we
don't
take
contributions,
and
this
isn't
one
extreme.
We
we
have
developed
these
and
we
make
them
available.
You
use
them.
Good
luck,
On,
The,
Other
Extreme.
A
E
A
A
A
A
A
A
A
A
C
Do
you
need
actually,
in
the
in
the
last
meeting
I'm
talk
I'm
talking
about
that
for
the
intro
of
the
imported?
The
matrix
model
would
would
provide
a
chaos
block
for
that
with
the
declared
the
implementation
for
that
and
I'm
continue
working
on
that,
okay,
so
yeah
after
that,
we
maybe
we
can
post
the
blogs
for
for
those
matrix
model
that
make
things
expandation
more
easier.
Okay,.
A
C
A
A
D
B
B
B
So
right
now,
in
order
to
tell
whether
a
product
has
done
all
the
diligence
to
be.
You
know
to
be
ready
for
the
next
level
in
in
a
CNC
of
yeah,
so
yeah
this
one
yeah,
if
you
could
go
down
to
the
almost
the
middle
I,
think
that
it
has
some
that
measurement
yeah
so
yeah.
So
so
what
what
you're
developing
yeah,
starting
from
here
so
yeah?
So
so
it
will,
it
will
use
the
GitHub
Dev
stats,
I,
think
yeah.