►
From YouTube: CHAOSS Evolution Working Group October 4, 2022
Description
Minutes and notes from this meeting can be found on https://chaoss.community/participate.
A
C
B
B
E
D
B
E
B
B
E
C
Yeah
I
think
we
could
use
a
hybrid
kind
of
method.
People
can
choose
what
they
are
more
comfortable
with
and
things
that
are
a
little
bit
not
to
play
for
people.
We
can
discuss
that
in
a
general
working
group
like
this,
so
I
mean
just
a
way
of
saying
if
we
have
more
to
be
more
efficient
in
a
way
because
we'll
still
have
to
discuss
everything,
you
know
it's
just
if,
for
example,
Elizabeth,
if
you
look
into
this,
you
see
something
that
you
can
speed
up.
C
C
B
E
B
C
A
D
D
D
I
think
they
can
actually
be
grouped
into
different
patterns,
because
you
can
see
some
of
the
projects
and
many
of
them
they
they.
They
have
just
been
active
for
like
during
his
creation
and
then
it's
just
it
has
a
long
tail
and
it
have,
and
some
of
the
projects
you
can
observe
is
gross
and
it's
it's
declination.
So
it
has.
There
been
like
the
a
similar
study
to
summarize
this
kind
of
patterns
and
I
wonder
if
this
is
Meaningful
to.
A
A
A
That's
at
a
on
different
scales
to
see
if,
for
example,
ones
that
go
above
400
activity
items
behave
differently
than
the
ones
that
peak
in
the
100s
or
200s,
because
I
can
see
that
your
y-axis
essentially
varies
based
on
the
data
that
you're
collecting
and-
and
so
these
graphs
are
compressing
out
what
might
be
an
interesting
Insight.
That
would
be
observable
if
they
were
all
at
the
same
scale
so
that
you
might
see
projects
with
higher
levels
of
activity,
May
Peter
out
at
it
in
a
different
way
than
projects
with
lower
levels
of
activity.
C
Yeah,
you
can
also
do
the
similar
patterns
and
you
can
also
do
the
similar
project,
for
example,
just
an
example.
If
you
are
comparing
Google
Cloud
to
Azure,
that's
a
healthy
comparison,
because
the
time
the
timeline
I'm
just
giving
example
to
understand
the
direction
of
the
project,
so
the
activities
will
be
expected
and
the
kind
of
analysis,
because
if
you
look
critically
I
just
look
vividly
in
your
chat,
I
don't
really
see
any
seasonality,
which
might
mean
a
very
good
interpretation
for
some
ways
and
in
some
other
activity
it
might
be.
C
A
questionable
interpretation
that
you
don't
have
seasonal
so
might
be.
The
patterns
were
just
like
one
time
or
a
random
thing.
It
depends
on
the
interpretation
at
this
level
and
this
kind
of
analysis
is
so
healthy.
If
you
remember
last
week
in
Elizabeth,
we
were
talking
about
this
kind
of
pie,
touch
kind
of
thing.
This
is
a
practical
use
case
that
she's
bringing
up
and
I
really
like
this
way
of
analysis.
So
if
you
want
to
group
this
kind
of
project,
there
are
so
many
ways
you
can
group
and
anytime.
C
You
are
doing
kind
of
comparison
with
data.
I
hope
you
know
the
statistical
complexity.
In
doing
that,
you
always
need
to
be
adjusting
things.
It's
not
only
about
the
issue
yeah.
They
are
significant
yeah.
So
you
need
to
take
care
of
those
kind
of
correction
factors,
but
it's
an
interesting
mining,
because
one
thing
is,
we
don't
want
to
be
comparing
Apple
to
Banana
all
the
time.
B
A
So
yeah
some,
like
so
Armstrong's
point
I,
think
some
of
these
projects
is
based
on
the
line.
Graphs
illustrate
a
pattern
of
basically
dumping
the
code
and
then
doing
almost
nothing
with
it
afterwards.
So
that's
where
you've
got
your
Peaks
on
the
far
left
and
then
it
just
drops
off
and
it's
kind
of
very
flat
until
the
end
and
I
think
you've
got
other
patterns
where
you
can
see.
Obviously
the
initial
repository
creation,
but
there's
higher
levels
of
activity
as
we
go
through
time.
B
D
D
There
have
some
relationships
among
its
patterns
and
the
reasons
it
was
archived
but
I'm
not
sure
if
it
indeed
have
have
this
kind
of
relationship,
but
I
think
other
patterns
might,
as
shown
in
the
Armstrong
pointed
there
are
some
relationship
related
to
the
nature
of
this
project.
Like
is
technology
domain
and
its
Vision,
or
it's
a
programming
language
and
maybe
yeah.
C
But
why
do
we?
There
is
one
this
guy
Sanu
for
with
software
Heritage
I
know
he
did
this
a
lot
with
with
archives
project,
because
the
reason
he
created
that
project
was
to
open
a
kind
of
museum
for
softwares
so
and
he
has
a
lot
of
tools.
He
also
does
a
lot
of
publication.
Let
me
share
the
link
here
with
you.
I
have
I'm
collaborating
with
him,
so
I
I
know
he
does
a
lot
of
work
that
looks
interesting
in
this
kind
of
domain
might
be.
Your
goals
can
be
different,
which
is
expected.
C
The
other
thing
is,
if
you
have
parents
in
your
your
quantitative
analysis
that
are
not
accountable
for
your
quantitative
analysis
might
be.
You
need
to
revisit
your
quantitative
analysis
to
see
how
the
data
were
collected,
because
things
like
this
go
to
the
initial
readme
file
and
towards
the
end
when
they
were
closing
the
readme
file.
Sometimes
the
discussions
in
those
things.
C
Have
yeah
because
of
several
reasons:
maybe
fundings
might
be
people
we're
not
really
contributing,
and
the
project
was
good
for
the
community.
They
realized
that
they
had
to
merge.
So
at
that
time,
if
you
just
analyze,
the
committee
you'll
see
that
at
some
point
this
project
cease
to
exist,
but
no
it
just
transformed
into
another
project,
which
is
a
kind
of
evolution
that
it
underwent
so
all
those
qualitative
should.
You
should
really
have
a
rich
set
of
qualitative
data.
D
And
another
another
question
I
want
to
have
is
I
I
kind
of
want
to
have
some
make
some
comparations
Evolution
wise
and
to
compare
between
the
archived,
maybe
especially
the
Dead,
the
projects
that
are
archived
and
actually
are
maintained,
not
the
projects
that
are
involved
in
other
places
and
compare
them
with.
Maybe
the
healthy
projects
so
like
like
there
are
some
machine
learning
methods
can
be
applied
like
I
can
run
a
logistic
regression
and
just
input
those
these
archived
projects
and
on
the
projects.
D
A
A
Just
it
becomes
kind
of
a
quantitative,
it's
a
dance,
because
quality
is
ultimately
or
sustainability
is
ultimately
a
subjective
thing
like
people
are
making
a
judgment,
and
if
we
can
get
people
to
make
judgments
about
a
set
of
repositories
that
they
consider
healthy.
Then
we
have
a
really
good
reference
point.
When
we're
doing
the
quantitative
analysis,
we
can
ground
the
quantitative
analysis
in
a
qualitative,
qualitative
I,
wanna,
yeah,
I'm
I'm,
trying
to
it's
not
the
word
truth.
But
it's
something
like
it's
like
a
ground
truth
in
open.
You.
B
A
A
D
C
That
yeah,
that's,
why
I
think
I
buy
Sean's
idea,
because
when
you
go
qualitatively,
there
are
a
lot
of
the
methodology
or
techniques
that
you
could
use
on
those
kind
of
data.
Even
when
you
are
learning
hearing
from
people,
for
example,
counter
factual,
you
have
content
analysis,
grounded,
Theory,
there's
so
many
of
them
that
you
can
use
the
more
data
intensive
technique.
C
So
five
stars
and
all
those
things
might
not
tell
all
the
story,
that's
why
we
have
metrics
to
measure,
but
at
the
end
of
the
day,
if
you
use
something
like
Transformers
in
natural
language
processing,
once
you
have
that
initial
value,
you
cannot
dump
in
not
huge
data.
Now
for
your
model
to
identify
some
of
those
things
that
manually,
it
will
take
a
longer
time
to
do
somehow
humanly
impossible
one
larger
repositories.
C
So
maybe
you
want
to
bring
International
Learning
when
you
start
with
the
human
aspect
of
it
to
start
identifying.
You
see
those
python,
you
see
those
things
happening.
You
cannot
train
some,
your
Transformer
or
any
other
model
using
attention
mechanism
or
any
other
thing.
You
are
more
comfortable
with.
D
A
What
are
the
signals
of
Health
like
so
healthy
could
mean
sustainable
with
an
adequate
Community
around
it.
Healthy
could
mean
both
of
that
and
attention
to
Dei
kinds
of
of
things
so,
depending
how
you
frame
what
is
healthy
or
what
is
sustainable
for
them
like
you
want
to
try
to
ask
it
in
an
open-ended
way.
But
what
can
it's
a
very
vague
question,
so
you
may
have
to
provide
them
more
context
in
terms
of
a
definition
of
healthy
or
sustainable.
D
A
But
you
could
do
what's
called
a
snowball
sample
and
I'm
just
rambling
here.
So
stop
me,
but
you
could
do
a
snowball
sample
where
some
informants
give
you
a
list
and
then
you
could
ask
them
to
recommend
other
folks
to
identify
healthy
projects
or
sustainable
projects.
And
you
know
you
could
accomplish
a
broader
survey
with
with
that
kind
of
snowballing
approach.
C
Now
I'm
just
supporting
what
you
said,
and
there
are
literature
this
thing:
how
do
they
call
it?
Systematic
literature
reviews
on
some
of
this
topic
on
health
and
playing
sustainability
in
a
way,
and
you
have
to
know
that
matrixes
that
capture
health
and
sustainabilities
are
not
like
individual
metrics,
so
sometimes
people
will
really
focus
on
a
group
of
metrics
that
are
measurable,
are
more
related
to
what
they
are
doing.
Another
project
might
not
include
all
that
same
metrics,
so
how
you
fit
in
that
line.
C
Now
you
remember
in
the
early
days
when
people
were
talking
about
sustainability,
everybody
drifted
in
towards
energy
immediately
and
they
taught
sustainability
is
to
reduce
energy.
But
later
on
scientifically,
they
said.
No
energy
is
not
even
a
an
issue
basic
because
a
data
center
now
in
let's
say
in
slow,
very
cool
areas
where
slow
it's
an
issue.
They
need
heating,
so
they
are
now
transforming
energy
from
one
form
to
another,
so
that
equation
became
the
argument
became
a
bit
weak.
C
So
they
are
now
looking
into
the
social
technical
aspect
where
human
factors
cannot
control
like
improving
the
code,
the
how
people
write
code,
best
practices
onboarding
those
kind
of
things,
that's
where
the
discussion
is
going
more
of
or
less
I
mean
the
literature
review.
Advise
you
more
on
this
kind
of
stand
to
take
that
as
Sean
says,
the
initial
place
is
to
use
snowball
sampling
because
you
identify
a
key
player
in
a
community.
C
Now,
once
you
discuss
with
those
persons,
you
cannot
ask
a
referral,
but
then
you
try
to
make
sure
you
don't
rely
on
Snowball,
because
that
persons
can
create
a
social
network
of
its
own
Clique,
which
becomes
a
bias
Network,
and
you
end
up
not
having
the
objective
truth.
So
once
you
have,
if
you
try
as
much
as
possible
to
descend
as
you
are
talking
with
people,
so
that
your
biases
should
not
create
a
link.
D
D
C
A
So
if
I
went
through
the
the
metrics
list
and
looking
at
that
issue
that
we
were
talking
about
earlier,
I
created
new
metric
lines
in
the
The
Evolution
working
group
for
issues,
updated
issues,
comma
issues,
comments
and
downloads
and
then
I
identified
an
existing
metric
for
change,
request
linked
with
issues
and
existing
proposed
metric
and
then
I.
Think
seven
and
eight
change,
request,
issue,
age
and
change.
Request.
Processing
time
are
simply
filters
on
the
change
requests
metrics.
So
those
already
exist.
A
A
A
B
B
C
A
So
I
didn't
highlight
them
because
when
I
create
different
colors,
it
drives
other
people,
crazy,
oh,
but
so
what
I
noted
is
so
change
request,
issue
mapping
this
already
exists
and
I
just
added
this
mm
related
to
the
description
and
when
I
created
issues
updated
and
issued
comments.
I
I,
just
put
mm
related
in
the
description.
A
I
didn't
make
any
note
about
the
the
seven
and
eight
other
than
in
here
I
note
that
they
are
filters
on
the
existing
change
request.
Metric
change,
request,
link
with
issues
we
highlighted
see
how
we
skipped
yeah,
so
I
think
that
covers
everything
in
that
list.
So
if
we
so
I
put
them
in
progress
optimistically,
but
we
could
go
and
like
like
issues
updated
and
issue
comments,
are
the
two
with
no
conceptualization
yet
I.
Don't
know
if
there's
any
progress
on
change
request
issue
mapping?
A
B
A
So
we
could
create
blank
ones
under
here
as
well,
yeah,
essentially
there's.
It
looks
like
that,
in
my
view,
there's
like
three
that
could
be
developed
and
parallel.
The
issue
updated
issue
comments
and,
let
me
see
maybe
downloads
is
the
other
one.
I
was
thinking
of,
or
yeah
downloads
I
put
downloads
in
here
somewhere
I
thought:
yeah.
Here's
downloads
under
Community,
I
put
it
under
Community
growth,
I,
didn't
know,
I
didn't
exactly
know
where
it
might
belong,
but
I
thought
Community
growth
made
sense
as
a
category.