►
From YouTube: CHAOSS Metrics Models Working Group July 18, 2023
Description
Meeting minutes are here: https://docs.google.com/document/d/1qE2duE73NbEOHzRRAbqgPSUSAfdH-AmvLruOu7uCIYg/edit
Meeting summary is here: https://chaoss.discourse.group/t/metrics-models-working-group-meeting-summary-july-18-2023/212
A
B
A
B
Feels
longer
than
a
week
but
yeah
it
is
longer
than
a
week.
I
guess
maybe
I'd
like
to
to
put
Elizabeth
a
little
bit
on
the
spot
and
just
see
how
fussy
went.
I
know
you'll,
probably
talk
about
it
in
a
number
of
meetings
today,
but
I
think
you
were
a
representative
there,
so
you
wanna
talk
about
that.
A
little
bit.
D
D
So
it
was
the
first
year
they
did
it
in
Portland
Oregon
and
it
was
small
like
285
people
there,
but
really
great
talked
a
lot
about
chaos
and
yeah.
It
was
funny
there
because
it's
the
crowd.
There
is
like
less
commercial,
less,
you
know
corporate
it's
much
more
of
like
yeah
the
the
privacy
and
the
the
the
hardcore
open
source
folks.
So
a
lot
of
folks
hadn't
heard
of
chaos,
so
that
was
really
great
to
introduce
chaos.
D
Everyone
I
gave
a
talk
with
Justin
Flory
on
onboarding
onto
I.
Did
my
chaos
obviously
and
Justin?
Is
the
community
manager
for
Fedora,
so
Justin
who's
also
been
a
chaos
as
well,
but
so
we
did
a
double
talk
together
and
when
I
asked
at
the
beginning
of
the
talk
who
had
heard
of
chaos,
almost
everyone
in
the
room
had
by
that
point
because.
C
D
D
That
was,
it
was
really
great.
I
I
talked
about
the
metrics
models
group,
a
lot
a
lot
a
lot,
because
that
seems
to
be
kind
of
you
know
where
we've
evolved
as
a
project
from
the
individual
metrics
and
now
like
this
is
kind
of
the
heart
of
the
project
right
now
and
the
meat
of
it
and
really
getting
those
context.
User
groups,
you
know
working
with
those
to
figure
out
what
what
do
we
need
to?
What
are
we
missing
and
how
can
we
put
these
metrics
together
so
yeah?
D
D
Well,
it
could
be
anything
from
like
going
over.
You
know
developing
of
metrics
to
how
to
use
the
software
okay,
you
have
a
few
questions
from
people
who
are
kind
of
struggling
a
little
with
using
both
auger
and
gamore.
Lab
I
got
some
questions,
so
so
yeah
I
think
there's
still
work
to
be
done
with
just
getting
people
using
the
software
that
we.
C
D
D
D
Just
going
to
add
one
more
thing
so
and
Anita's
Sarma
was
there
and
I
know
she
had
a
student
from
from
China
that
was
doing
a
research
project,
so
we
were
kind
of
helping
try
to
funnel
I
I
hope
he
he's
able.
He
lives
on
the
west
coast,
though
so
I
don't
know
that
he's
gonna
be
able
to
join
these
meetings
because
they're
super
early
for
them,
but
I
was
trying
to
encourage
him
to
just
to
get
to
know
us
and
you.
B
B
Awesome
well,
I,
guess
that
that's
actually
a
good
segue
into
a
couple
of
different
things,
so
I
do
want
to
just
point
out.
There
are
a
couple
metric
models
that
have
been
started
from
the
science
context,
working
group
or
a
context
group.
So
again
we
have
kind
of
three
different
context:
groups
with
respect
to
our
metrics
models
and
the
one
of
the
metrics.
Oh
gosh,
one.
B
Groups
is
quite
brain,
is
a
corporate
I,
suppose
another
context
group
is
University
ospos
and
a
third
is
scientific
software,
and
so
the
two
that
you're
looking
at
here
are
from
the
scientific
software
context.
Group
Melissa
had
brought
these
two
forward,
and
so
I
just
wanted
to
kind
of
put
this
on
people's
radar
that
this
is
happening
and
we're
starting
to
take
a
look
at
these
two
particular
metric
models.
B
C
B
B
One
of
the
things
that
I
did
wonder
about
is
in
the
writing
of
some
of
these
metrics
models
like
there.
There
is
some
specificity,
for
example,
sometimes
so,
for
example,
scientific
project,
sustainability
versus
just
general
project
sustainability
and
how,
from
a
metrics
model
perspective,
we
might
want
to
think
about
having
models
that
are
quite
specific
to
a
context
or
metric
models
that
are
more
General
and
could
be
applied
in
different
contexts.
Does
that
make
sense.
A
Yeah
I
mean
I.
Think
from
a
the
key
thing
is
for
a
lot
of
the
metrics
models.
I
think
we
can
instrument
what
the
values
are
to
the
particular
domain
or
context.
I
also
think
there
are
some
metrics,
for
example,
in
scientific,
open
source
software
that
wouldn't
be
relevant
for
corporatized
open
source
in
terms
of
determining
the
health,
so
it
wouldn't,
it
would
be
more
than
parameters
so
I,
yeah
I,
agree
with
I,
guess,
I,
don't
know
if
you're
proposing,
but
I
agree
that
there
are.
A
D
C
C
D
Eloquently,
which
is
a
really
hard
word
early
in
the
morning,
does
does
core
mean
the
starter
model
or
or
something
else
in
particular,
yeah
I,
don't
know,
I,
don't
know,
I
think,
maybe
maybe
we
could
like
for
this
one.
It's
it's
around
project
sustainability,
so
maybe
there's
a
core
project
sustainability
like
regardless
of
your
context.
These
are
the
things
that
maybe
are
common
across
and
then
there's
that,
like
extra
I,
don't
know
just
a
thought.
D
D
A
Yeah
I
I've
made
a
comment
in
the
text.
I
think
it's
it's.
There
are
dimensions
that
I
consider
in
these
different
domains
that
are
and
I.
So
what
you're
suggesting
Elizabeth
I
think
is
like
a
modularity
like
there's
a
core.
These
are
the
things
that
we
care
about
across
all
the
domains,
and
then
each
domain
may
have
some
Dimensions
that
are
also
important
to
it.
B
So
I
like
I,
like
that
I,
would
prefer
to
keep
these
General
like
instead
of
scientific
project
sustainability,
it
would
just
be
project
sustainability
as
an
example,
because
I
do
think
that
there's
an
interest
from
corporate
I
suppose
and
from
University
I
suppose
about
project
sustainability.
But
then,
to
your
point,
Elizabeth
anchon,
that
somewhere
in
here,
we
could
provide
some
of
that
specificity.
So
Sean
like
if
you're
in
a
scientific
software.
D
I
mean
the
easiest
way
to
do.
It
is
to
keep
it
all
in
this
Dock
and
then
maybe
just
add
a
section
around
contexts.
I
guess,
as
opposed
to
like
a
whole
other
dock.
That
would
have
to
connect
somehow.
But
then
that
means
this
is
going
to
get
pretty
long
and
like,
as
we
add
contexts,
are
we
going
to
have
to
add
it
on
every
single
metrics
model
that.
A
But
I
mean
technically,
we
need
something
that
would
let
us
easily
go
the
core
and
then,
if
I
click,
a
scientific
button
or
something
and
I
see
the
additional
or
other
things
that
are
related
to
the
scientific
domain.
For
this
metric
model,
I'm
sure
things
like
that
exist
that
we
haven't
worked
with
them
ourselves.
Yet.
B
B
I'm
gonna
put
you
on
the
spot
just
a
little
bit
so
with
respect
to
Compass
updates
and
how
things
are
going
there.
I
did
want
to
let
you
know
that
I
had
kind
of
shared
the
that
kind
of,
but
I
had
shared
the
site
and
the
repo
with
folks
from
GitHub.
They
were
asking
about
projects
that
have
deployed
chaos,
software
through
like
dashboards
and
interfaces.
So
I
was
talking
to
some
folks
at
GitHub,
just
about
Compass
as
well,
not
to
mention
auger,
and
they
were
already
familiar
with
auger,
more
lab.
E
E
We
would
select
a
data
site
and
the
metric
sites
and
compose
a
new
matrix
model,
as
I
mentioned,
they
have
more
than
70
more
than
70s
metrics,
already
supported
in
compass,
which
originated
from
chaos.
So
right
now
the
background
designing
work
is
it's
almost
finished
and
the
implementation
work
just
to
start
this
week.
So,
hopefully,
to
two
weeks
later.
Maybe
we
can
show
our
first
solution
of
this
Compass
like
to
all
of
you.
B
E
Yeah,
if
you
remember
correctly,
I
have
had
once
some
some
sort
of
way
some
kind
of
survey
or
chaos.
That's
that's!
Documentations
to
you,
like
you,
tell
us
which
metrics
it's
got
highest
priority
to
be
implemented
as
an
independent
metric
model
in
compass.
That's
what
you
used
to
that
is
intended
used
by
Compass,
like
that's
what
you're
talking
about
exactly
yeah
so
right
now
the
the
background
works
I
mean
including
this
independent
matrix
model
component
development
work.
B
Then
the
question,
a
follow-up
question
I
have
and
I
I
seem
to
recall
this
conversation
from
earlier.
But
if,
if
somebody
deploys
a
new
metric
model
for
example-
and
they
bring
together,
say
whatever
four
or
five
different
metrics
that
are
here
on
this
list,
the
waiting
of
those
metrics.
How
is
that
handled
on
the
compass
side,
because
I
know
that
there
is
kind
of
an
aggregate?
That's
provided.
E
E
Okay
and
of
course,
you
can
provide
some
some
some
some
other
new
algorithm
to
calculate
a
bit.
But
that's
really
depends
on
your
your
requirement,
but
basically
we
provide
some
default
with
default
threshold.
Of
course,
both
threshold
and
words
are
configurable
and-
and
we
provide
some
default
algorithm
to
go
to
the
final
metrics
score,
but
that's
also
as
just
the
provider
as
a
default
solution.
E
B
E
B
B
E
Here
about
Compass
lab,
let's
all
and
I
would
like
to
share,
maybe
in
the
future
two
weeks
two
weeks
later,
and
we
could
provide
as
a
demo
presentation
because
and
we
have
doing
a
lot
of
work
on
the
designing
phase.
So,
finally,
we
got
into
a
step
into
the
implementation
phase.
This
week,
that'd
be
great,
okay
sounds
good
yeah
and
the
other
things
about
sustainability.
E
Of
the
Developers
that
are
demoed
in
last
meeting
about
developer
model,
we
are
still
working
on
light
and
because
we
got
a
lot
of
feedback
from
all
of
you
in
in
the
last
meeting,
we
also
got
a
lot
of
feedbacks
from
our
friends
in
China
and
then
so.
Finally,
they
would
provide
a
final
aggregated
developer
developer
model,
including
the
profile
I
noticed
that
in
the
last
matrix
model
you
show
with
us
about
Santa's
scientist
model.
That's
mentioned
something
about
Armenia
or
leadership
in
the
in
the
community.
E
E
E
B
B
B
B
B
I
had
just
responded,
I
think
it's
about
both
the
way
I
had
understood
it,
so
it
could
be
the
business
Readiness
of
a
dependency
or
okay,
or
it
could
also
be
the
use
of
a
piece
of
software.
That's
just
you
know
kind
of
a
piece
of
software
by
itself,
so
I
had
to
try
to
get
that
into
the.
Why
it
matters
part
seems.
B
B
Good
so
I
think
we've
been
through
this
several
times
right
now.
I
think
this
was
something
that
the
nod
had
led.
If
you
recall,
if
you'd
like
to
add
yourself
as
a
contributor,
please
feel
free,
I
think
anybody
on
this
call
is
certainly
a
contributor
so
or
if
you'd
like
to
add
yourself
just
put
yourself
in
the
minutes
and
I
can
add
it
here
as
well.
So
either
way
is
all
good.
B
I
think
so,
and
you
know
our
our
process
for
metrics
models
is
considerably
simpler,
just
because
they
do
move
a
little
bit.
You
know
what
I
mean
like
versus
metrics,
so
I
think
we're
good
to
go
here.
Awesome,
okay,
I
think
the
only
my
only
hesitation
with
this.
If
anybody
cares,
is
we
usually
like
to
keep
metrics
models
that
a
fairly
limited
set
of
metrics?
This
does
have
eight
metrics
associated
with
it.
A
A
Essentially
it's
just
the
organizational
affiliation
of
a
contributor
that
changes
the
basically
the
elephant,
Factor
like
if
I'm
the
committer
and
if
I
change
my
organizational
affiliation
and
my
commits
in
the
future,
change
to
that
new
organizational
affiliation
which
Alters
the
elephant
Factor.
But
it's
I
mean
there's.
So
these
are
I
guess
what
I'm
saying
is,
unlike
some
of
the
other
ones,
where
we've
had
more
metrics
than
we
prefer.
C
B
A
Test
coverage
remains
I
mean
there's
test
cover
test
coverage
is
a
largely
aspirational
metric.
It's
not
one.
That's
easily
deployed
across
a
wide
range
of
repositories,
there's
a
lot
of
language,
specific
considerations
for
test
coverage
checking,
but
that's
been
the
case
from
the
beginning
and
I.
Don't
think
we're
going
to
solve
it.
B
And
as
we
I
think,
as
we
say
in
a
lot
of
these
metrics
models
like
here
the
here
in
this
case
eight
like
pick
and
choose-
which
ones
are
important
to
you.
But
this
is
kind
of
the
base
we
set
out
and
to
your
point
Sean.
Maybe
this
one's
a
little
tricky
to
get
to.
But
if
somebody
wants
to
take
the
time.
C
A
B
C
E
C
E
Because
the
whole
contributor
things
or
developer
things,
it's
it's
quite
complicated
people,
as
I
mentioned
earlier.
In
the
last
meeting
we
we
would
like
to
set
up
the
full
profile
for
the
developers
from
different
dimensions
like
from
the
ecosystem
or
leadership.
That's
that's
Prospect
review
or
from
the
contribute
frequency
and
from
the
contribution
type.
B
E
C
E
You
know
we
we
use
the
email
suffix,
that's
the
really
and
it's
a
easiest
way
to
distinguish
the
people
from
which
Corporation,
but
you
know,
and
not
all
the
people
use
their.
You
know.
Corporation
email
account
so,
except
for
that,
and
if
this
member
is
a
it's
a
it's
a
call
member
or
early
mean
role
in
from
the
communities.
E
E
There
are
other
means,
so
all
those
things
behind
some.
You
know
contribution
actions,
no
matter
from
issue
broadcast
or
call
commit
that's
only
from
the
GitHub
that
sources,
but
also
we
have
some
other
data
sources,
like
slack
like
Twitter
how
to
distinguish
that.
That's
all
the
questions
we
have
installed.
It.
E
A
I
mean
I've
I've
been
through
it
and
the
only
times
that
I've
had
what
I
think
are
really
good.
Resolutions
of
organizations
are
when
I'm
inside
the
walls
of
that
organization
and
I
have
their
list
of
all
the
emails
that
a
person
uses
auger
like
if
so,
in
the
case
of
auger,
we
can
like,
if
somebody
uses
their
Corporate
email
one
time
we
can
Alias
all
the
other
emails
associated
with
that
same
user
together.
E
A
Exactly
no,
it's
it's
and
people
I
mean
I,
think
increasingly
open
source
contributors
here.
Are
they
don't
want
to
be
associated
with
an
organization?
They
want
to
be
viewed
as
free
agents,
and
so
the
the
in
there's
an
almost
there's
a
growing
I
would
say:
intentionality
of
ambiguating,
my
organizational
affiliation
when
I'm
contributing
to
open
source,
so
that
I
think
the
problem
is
becoming
more
challenging
as
time
passes.
B
E
A
C
B
C
F
So
a
lot
of
a
tendency
is
to
for
people
who
has
not
digged
so
deep
into
the
Matrix
is
tend
to
come
up
with
the
things
like
creating
a
like
a
a
model
to
just
just
kind
of
a
high
level
things
to
do
to
reach
a
certain
either
a
maturity
model
or
whatever
stage
open
source
Community
need
to
be
so
matrix
model.
Definitely
is
the
Matrix
chaos.
Matrix.
F
Definitely
is
awesome
to
go
deeper
that
nobody
else,
I
think
has
gone
as
deep,
so
it'd
be
great
to
go
back
up
from
from
the
foundation
how
to
use
this
detailed
Matrix
to
summarize
and
go
back
to
to
measure
those
level
maturity
whatever
it
is.
You
know
in
a
so
that
can
be
consumed
Easier
by
the
by
the
after
management,
because
usually
after
Management
in
in
corporations
tend
to
want
to
read
one
page
things
right,
rather
than
lots
of
Matrix.
C
B
C
B
Open
source
as
connected
to
research,
Excellence
how
to
identify
open
sources
connected
to
research,
translation
and
so
on,
and
so
forth
down
the
list
and
then
within
each
one
of
these
goals
at
the
top
or
functions
across
the
top,
we
would
have
kind
of
smaller
sets
of
practices
that
an
ospo
can
do
and
then
within
each
one
of
these
practices,
for
example,
so
like
if
I
was
to
pick
that
box
of
the
the
one
that
I
highlighted
right
there
like
there
would
be
oops.
Let
me
pick
one:
there
would
ultimately
be
metrics
and
metrics
models.
B
B
F
F
B
F
F
Yeah
so
yeah,
yes
yeah,
so
this
table
so
I
think
this
group
is
doing
a
great
job.
The
the
team
that's
been
building.
This
initially
is
instead
of
just
relying
you
know,
common
knowledge
to
come
up
with
Matrix.
This
group
really
go
deep
into
what
it
may
just
need
to
be
me
mattered
in
when
it
comes
to
Merit
like
success
of
a
platform,
so
so
I
think
this
could
be
I
guess
with
chaos.
I
guess
maybe
a
two
week,
maybe
like
from
from
your
perspective
anything
else.
F
F
A
F
B
F
So
this
is
more
that's
the
thing
so
so.
Defining
a
mature
model
is
really
hard
because
how
to
really,
because
each
factor
may
now
reach
the
same
level
at
the
same
time
right.
So
one
thing
that
this
team
has
been
the
original
author
of
this
model.
I,
like
is,
are
they
the
factor
they
each
factor
doesn't
has
to
reach
a
particular
level
at
the
same
time,
so
it's
kind
of
independent,
so
yeah
and
so
how
to
yeah
how
to
effectively
sorry
I
have
to
ask
I
forgot.
F
C
C
F
B
B
F
F
If
you
see
the
the
the
actual
factors
the
using
that
in
that
model,
it's
not
something
that
you
have
to
do.
This
technology,
you
have
to
you
know,
increase
it's
more
of
a
in
in
a
principal
way
like
if
you
follow
the
principle,
regardless
of
what
technology
you
use,
it
might
be
there,
you
may
get
there.