►
From YouTube: SIG Events Meeting - Feb 28, 2022
Description
For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/
B
I
think
andrea
had
a
comment
on
that
in
in
the
agenda
for
today
that
it's
not
really
announced
anywhere,
and
for
that
reason
maybe
we
should
reschedule.
B
B
No
exactly
it
would
be
a
bit
pointless,
it
would
be.
What
do
you
say
in
swedish
affirming
for
imports,
foreign.
B
B
C
C
D
B
We
were
just
catching
up,
nothing
project
related,
really.
D
Should
we
get
started,
then
sounds
good.
Okay,
okay,
singing
has
already
added
yourself
to
the
agenda.
E
D
Okay,
welcome
to
the
sick
events
today,
okay,
first
thing:
quick
review
of
the
action
item
from
last
week,
one
was
to
add
a
contribution
process
for
city
events,
which
we
did
so
the
apr
has
been
approved
and
it's
actually
now
merged
into
the
dot
gita
repo.
So
thanks
matthias
for
pointing
it
out.
So
it
means
that
it's
available
throughout
github,
wherever
there
is
a
contributing
point
like
when
you're
creating
a
new
issue,
if
you're
contributing
guidelines
that
point
to
the
contributing
guidelines,
etc
for
prs
and
so
forth,.
C
Andrea
on
that,
the
code
of
conduct
is:
is
that
the
latest
one,
the
2.1?
C
Okay,
so
so
the
right,
the
link
yeah,
that
code
of
conduct
there.
D
Yeah,
but
this
is
just
a
readme
for
the
profile,
so
this
link
is
something
that
we
right
we
added
so
I
can
either
update
this
meeting
this
link
or
remove
it,
but
I
think
it
would
be.
It
would
be
nice
to
have
it
here.
G
E
D
Next
thing,
city
count:
contributor
summit.
There
was
a
google
home
to
be
done,
and
so
I
filled
the
google
form
requested
for
less
than
10
person.
Based
on
the
response,
I
got
and
requested
support
for
remote
participants
as
a
few
of
the
responses
indicate
that
they
wouldn't
be
able
to
go
to
austin.
D
Okay,
if
there
is
any
questions,
stop
me
so
for
the
next
one
is
the
pr
document.
So
we
discussed
a
bit
about
last
time
and
I
mean
most
of
the
command
comments
there
emerged.
We
have
a
few
quotes.
It
would
be
good
to
get
a
few
more.
I
think
we
have
people
assigned,
including
myself,
so
hopefully
we
can
get
some
more
questions.
C
Yeah
and
steve
chin
from
jfrog
who's
part
of
the
cd
foundation.
He
said
he
would
do
and
I'll
just
have
to
get
his
text.
Rome.
D
Next
item
marketing
delete-
I
yeah
camera
is
not
here
today.
I
don't
know
if
there
is
anything
to
add
on
this
one.
D
In
okay
of
cd
events
at
kubecon,
there
is
a
google
form
there
that
I
set
up
where
I
was
interested
and
ideas.
D
I
got
nine
responses
so
far,
some
several
ideas
of
names
for
the
event
yeah.
If
you
have
not
replied
yet,
please
please
do.
D
Yeah,
we
need
to
just
move
on
with
the
organization
of
the
events
yeah,
that's
all
the
ad
from
the
action
items.
D
In
terms
of
updates
aromatic
groups,
so
one
thing
we
discussed
was
the
cd
foundation
and
profit
sponsorship.
Application
for
hacking
d
also
put
there
a
link,
so
you
can
go
and
have
a
look
basically,
but
he
made
the
application
and
seems
that
it's
moving
forward
nicely.
There
are
some
requirements
to
be
fulfilled
and
you
should
end
up
in
having
like
20
accounts
available
for
the
cdf,
so
yeah
keep
you
posted
on
update,
so
you
can
follow
the
issue.
D
And
the
other
thing
from
toc
is
that
baby
platform
is
now
available.
So
it's
on
community
dot,
city
foundation
and
there
is
a
cd
event
chapter.
I
think
it's
called
yeah
and-
and
so
myself
and
oleg
are
in
his
organizers,
but
there's
not
much
else
to
find
for
now,
but
yeah.
We
can
use
this
to
organize
meetups.
Basically,.
A
Is
there
any
idea
of
moving
from
zoom
to
bevy,
or
is
that
only
for
webinars.
D
Yeah,
so
I
always
ask
about
that,
and
so
at
the
moment
I
think
there
is
no
pressure
to
move
away
from
zoom
specifically,
but
it
might
be-
or
we
might
do
it
in
the
future,
to
like
move
everything
there
instead
of
using
zoom
and
baby.
Do
you
have
any
specific
points
in
favor
of
one
or
the
other
platform,
or
I'm
not
familiar
with
baby
at
all?
To
be
honest,.
A
C
C
A
D
C
Yeah,
I
don't
know,
I
don't
know
what
the
platform
looks
like
for.
I
knew
it.
I
knew
it
was
taking
place
as
the
for
organizing
the
meetups,
and
you
know
I
could
you
kind
of
join
your
local
meetup
group
through
as
bevy,
but
I
did
not.
C
I
don't
know
the
extent
past
that
you
know
on
those
on
the
web
meeting
side
what
what
capabilities
it
has.
D
All
right
I'll,
if
I
know
more
next
time,
I'll,
let
you
guys
know-
and
I
think
the
next
thing
on
the
agenda
in
fact-
is
about
where
we
have
the
city
events
webinar-
that
we
originally
planned
for
wednesday,
the
second
so
yeah
at
this
stage.
Well,
all
that
is
not.
D
C
Yeah,
I
think
we
should
reschedule
that
and
because
I
haven't
seen
anything
be
pushed
out
at
all,
so
I
think
what
we
should
do
is
somebody
should
reach
out
to
cara
to
get
it.
D
I
don't
know
if
there
if
the
current
date
I
mean
there,
there
is
an
entry
like
a
placeholder
in
the
calendar,
so
we
probably
can
just
move
that
for
now
and
yeah
there's
nothing
created
in
the
very
platform
yet,
and
I
guess
I
should
be
able
to
create
new
events
but
yeah.
I
don't
know.
D
D
C
So
I
just
have
two
quick
things:
one
is
tracy,
so
ortelius
has
a
working
group
in
australia
and
that
working
group
is
doing
a
bunch
of
work
around
events
for
the
get
ups
process
and
brad
mccoy
found
a
project
out
there
called
directive
that
has
a
huge
number
of
events
kind
of
like
predefined.
C
C
But
there
may
be
another
project
out
there
that
we
can.
It
is
an
open
source
project.
It
is
not
part
of
an
organization
from
my
understand,
so
it
may
be
able
to.
We
may
be
able
to
leverage
some
of
what
they
have
going
on.
C
C
Okay,
so
I
will
confirm
and
get
the
make
sure
that
is
the
right
and
we
can.
I
can
probably
arrange
for
either
a
demo
to
us
or
since
they're
in
australia,
maybe
even
like
a
pre-recorded
demo
as
part
of
that.
C
So
it
may
be
something
that
we
can
get
a
lot
of.
You
know
kind
of
jump
ahead,
because,
if
it,
if
it
is
what
I
thi,
if
it
from
what
they
brad
and
tracy
described
to
me,
they
it
sounded
like
they
had
a
lot
of
events
covered
for
different
tools.
C
C
See
what
what
we
can
do
get
some
more
information
about
that
project.
The
other
thing
is
for
the
our
our
cdcon
day
that
we
have
from
the
ortelius
side.
We
are
like,
I
said,
work
on
the
the
get
ops
model
with
events
that's
pulling
together,
and
I
think
we
could
probably
pull
together
a
presentation
if
there
is
a
room,
a
time
slot
available
for
us.
C
So
if
you
need
a
if
we
need
a
speaker,
I
can
round
up
some
folks
from
the
ortelia
side.
D
C
Yeah
because
it's
we're
moving
along
pretty
well
and
things
are
starting
to
get
solidified,
so
I
think
it'll
be
a
good
fit.
B
Just
one
thing
that
popped
into
my
head,
steve,
when
you
mention
get
ups,
is
in
the
eiffel
community
and
matthias
will
know
more
about
this.
There
is
an
ongoing
effort
to
see
if,
through
events,
we
can
represent
the
complete
git
graph,
so
to
say,
commits
and
trees
and
tags
and
all
that
stuff.
Just
through
the
events,
I
I'm
not
saying
that
is
something
that
that
we
need
in
cd
events,
but
it's
probably
something
that
we
could
at
least
discuss
at
some
point.
B
So
this
would
basically
mean
that
whenever
a
commit
is
created,
an
event
is
sent
for
that
commit
that
describes
the
commit
in
git's
terms,
so
to
say
to
declare
what
parents
this
has
and
what
hash
it
has,
etc,
etc.
So,
right
fundamentally,
you
could
get
a
perfect
understanding
of
what
commits
there
are
without
even
talking
to
the
repository.
C
Yeah
and
that-
and
I
think
that
would
be
a
really
nice
thing
to
have
to
have
that
happen,
so
you
don't
have
to
plumb
together
a
bunch
of
web
hooks
to
get
that
information.
B
A
Yeah,
I
guess
maybe
if
we
have
some
things,
because
it's
not
there
were
a
couple
questions
left
over
there
yeah.
So
I
guess
we
have
to
settle
a
little
bit
more
before
we.
We
can
take
the
presentation
forward,
but
yeah.
That's
a
good
idea
to
sync
between.
C
D
Yeah
one
of
the
things
and
for
the
metrics
poc
that
I
started
looking
into
is
like
you
need
to
to
to
be
able
to
say
what
is
the
the
time
to
or
change
to
get
in
production.
So,
basically,
you
need
to
know
when
a
specific
change
gets
into
a
release
and
then
what
that
releases
in
production
and
so
to
be
able
to
correlate
when
a
change
is
in
a
release
you
need
to
have
to
access
to
some
kind
of
changelog
or
releases
or
git
history
of
some
kind,
so
yeah.
D
I
think
it
would
be
relevant
for
this
kind
of
problems
as
well
yeah.
But
one
question
that
pops
to
mind,
though,
is
what
what
do
you
do?
If
you
start
listening
to
events
at
a
certain
point,
you
don't
have
all
the
history
from
what
happened
before
that
point
in
time.
You
still
need
probably
to
have
some
access
to
some.
C
Yeah-
and
I
think
I
I
keep
on
coming
back
to
this-
that
I
think
down
the
road
that
we're
going
to
need
to
have
a
persist,
the
events.
So,
if
you
jump
in,
like
you
said,
if
you
plug
in
and
start
listening
to
events,
you
want
to
go
back
through
the
history
that
somewhere
persisted
out.
There
is
a
way
to
get
the
the
history
of
the
events.
C
You
know
that's
a
lot
of
data,
but
I
can
foresee
us
being
it
needing
that
down
the
road.
A
way
to
persist
this
the
event
graph
for
lack
of
a
better
better
word.
A
Yeah
yeah,
you
know,
for
we
have
it's
basically
as
a
principle
that
it's
yeah
there
is
an
event
repository
which
you
basically
will
listen
to
all
the
events,
so
either
job
or
some
tasks
and
listen
directly
to
the
event
story.
And
then,
if
you
need
some
more
data,
you
can
actually
traverse
this
event.
Repository
where
we
have
the
events
persisted.
A
Currently,
using
mongodb
okay
yeah,
but
I
guess
if,
if
there
is
other
alternatives
out
there,
especially
if
there
is
something
high
performance,
if
you're
sending
a
lot
a
lot
of
events,
and
then
I
don't
know
if
one
db
is
the
correct
one
to
use.
But.
D
Yeah,
I
know
the
the
four
keys
project
from
google.
They
use
bigquery,
of
course,
because
that's
what's
available
in
google
cloud,
but
I
think
it's
a
similar
type
of
technology.
I
mean
you,
you
store
json
documents
right,
you
can.
You
can
query
on
specific
fields.
I
think
so.
C
Yeah
we're
looking
at
a
graph
database.
I
just
put
it
in
in
the
chat
we're
looking
at
this
graph
database
because
graph
databases
manage
relationships,
really
well
million.
You
know
million
a
million
types
of
relationships.
It
can
query
very
quickly
and
that
may
fit
into.
C
It
is
it's
like
whether
you
get
commercial
support
or
not
commercial
support.
I
think
from
I
remember.
C
But
it
is
a
so
one
of
the
things.
So
it's
a
it's
a
graph
database
which
is
different
than
graphql,
so
just
you
just
have
to
make
sure
that
you
don't
get
the
two
mixed
up.
So
the
when
you
start
looking
for
these
make
sure
you're
looking
at
a
graph
database
and
it's
similar,
it
is
similar
to
like
a
the
way.
It
persists
things,
but
also
it
builds
additional
pathways
to
do
the
relationships
really
quickly.
A
Yeah
but
yeah
in
ericsson,
it's
actually
is,
since
we
are
dividing
things
into
different
domains.
We
kind
of
like
avoid
the
issue
a
little
bit.
A
Yeah
precisely
but
yes,
that's
it's
one
of
those
things
that
we
should
get
better
at
so
yeah,
interesting
with
the
orango
db.
C
Nice,
so
those
are
just
a
couple
things
I
stumbled
across
the
last
week
or
so,
and
that
that's
basically
all
I
had
for
today.
D
One
thing
that
just
came
to
mind
as
as
I
mentioned,
the
metric
plc
just
wanted
to
mention
that
I
started
working
and
in
this
action
d.
So
if
you
want
to
contribute
or
have
ideas,
feel
free
to
to
comment
or
add
comments
on
it
and
yeah,
so
the
the
architecture
is
again
with
like
multiple
platforms
on
top
sending
events
to
an
event
broker,
and
then
we
have
this
event.
Storage
that
we
were
just
talking
about.
D
So
this
discussion
is
very
relevant
for
this
pse
as
well,
because
yeah,
we
need
to
to
pick
one
kind
of
storage
that
we
can
use
and
maybe
yeah
mongodb
could
be
an
option
or
around
good
to
be
so.
I
will
look
into
these
options
and
for
the
poc
we're
not
trying
to
build
specifically
pipeline
visualization,
but
it's
more
about
the
metrics
and
the
idea
is
to
to
be
able
to
provide
the
four
dora
metrics
for
ocd
events.
D
So
what
I
did
I
started
looking
at
the
different
metrics
and
what
kind
of
events
we
have
and
what
kind
of
events
type
we
could
use
yeah,
and
I
think
one
area
is
like
in
the
change
failure
rate
specifically
that
we
don't
have
any
event,
possibly
also
in
the
time
to
restore
service
so
yeah.
We
need
to
discuss
what
what
to
do
there,
or
maybe
I
can
make
a
proposal.
D
D
So
you
could
have
like
incident
start
and
incident
finished,
and
you
could
imagine
if
you're
tracking,
your
incidents,
for
instance
through
github
or
jira,
that
event
could
come
from
an
issue
a
ticket
created
there
or,
if
you're,
something
like
captain
where
you
have
like
sli
or
slow
monitoring,
and
if
there
isn't
some
of
these
that
are
failing,
because
after
ruby's
performance
is
not
good
anymore,
then
automatically
an
event
is
an
incident.
Type
of
event
is
created,
so.
C
Look
at
I
just
was
looking
at
that
directive
and
they
have
a
it
says,
cloud
events
from
logs,
so
that
may
I
can't
tell
they
look
like
they
have
stackdriver
and
cloud
watch,
but
I
think
they
may
plug
into
prometheus
as
well.
But
I
would
I
would
poke
around
at
that
directive.
They
may
you
may
be
able
to
uncover
some
stuff
that
will
help
you,
along
with
that.
D
D
Well,
maybe
I
would
like
to
get
some
feedback
on
this
specifically
whether
folks
make
things
that
make
sense
to
expand
kind
of
the
the
current
stages
that
we
have
into
that
area,
because
today
we
have
like
chorus,
icds
cd,
what
else
change
management
systems,
but
we
don't
have
anything
on
the
kind
of
operation
side
of
things
or
like
incident
and
so
forth,
but
I
think
it
would
make
sense
to
to
have
this
category,
because,
even
just
for
having
satisfying
these
kind
of
use
cases
for
the
dora
metrics
you
would
have.
D
C
Yeah
and
kind
of
circling
back
around
to
when
we're
talking
about
the
the
eiffel
get
events,
one
of
the
things
that
we're
running
into
from
the
artelia
side
is
being
able
to
get
the
in
order
to
do
some
of
the
door
metrics.
You
need
to
know
like
the
when
the
commit
was
created,
the
time
the
commit
was
created.
C
So
you
can
then
do
the
time
difference
between
when
the
commit
was
created
and
when
it
was
actually
deployed
to
like
production,
and
that's
gonna,
be
when
your
door
metrics
is
the
the
time
between
commit
and
deployment
that
that's
some
of
the
information
that
you'll
you'll
need
and
things
other
things
like
lines
of
number
of
lines
changed
between
certain
commits
so
and
and
between
multiple
commits
as
well.
C
Because
one
of
the
weird
things
is
the
the
get
the
git
commit,
you
can't
query:
there's
not
like
a
an
easy
way
to
get
the
timestamp
for
a
commit
without
having
access
to
the
repo.
So
you
can't
like
hit
a
restful
api
endpoint
and
get
the
timestamp.
You
actually
have
to
have
the
repo
local
to
inquire.
The
log
at
that
point.
A
And
then
we
have.
We
don't
have
a
time
step
in
this
one,
but
the
kind
of
idea
with
with
a4
is,
of
course
you
send
the
event
as
close
as
possible.
So
it
might
not
be
the
exact
timestamp
when
the
when
they
was
commit
was
created.
But
the
idea
is
that
if,
if
you
create
something,
you
should
also
send
an
event
for
it,
and
so
that
way
you
actually
know
when
the
the
commit
was
created
by
looking
at
time
stamped
event
right.
C
D
Yeah
yeah
indeed,
so
I
think
we
we
have
like
change,
created,
change,
updated
and
change.
The
change
merge
defense
would
be
the
cd
event,
one
nothing
corresponds
to
what
you
were
measuring
materials
and
then
we
we
have
a
timestamp
there
as
well
so
yeah.
So
I
think
for
the,
for
that
part
is
we
do
have
that
information,
but
so
yeah
we
can
just
rely
on
time.
I
mean
if,
let's
let's
say,
if
we
have,
we
know
that
a
certain
point,
a
release,
happens
or
a
deployment
happen.
D
But
if
you
want
to
to
know
that
is
in
fact
in
you
need
to
have
like
more
even
more
information
to
be
able
to
correlate
changes
to
releases
and
then
releases
to
deployments
but
yeah,
so
for
the
poc,
depending
on
how
far
we
we
get,
we
can.
We
can
get
that
information
or
we
can
just
really
rely
on
on
the
timestamps
to
begin
with
and
then
at
more
events.
D
Yeah,
thanks
for
all
the
the
input
on
this,
I
put
the
link
to
the
the
pse
d
into
the
agenda
and,
if
you
want
to
browse
for
it
and
leave
comments
and
extra
comments
in
there
and
yeah
very
welcome.
C
D
C
A
One
question:
when
it
comes
to
the
plc
here,
so
we
have
a
product
for
keys
and
which,
as
I've
understood,
basically
does
some
events
and
actually
those
sort
of
kind
of
like
heavy
lifting.
D
So
what
four
keys
does
is
is
collecting
events
from
existing
sources.
Is
writing
adapters
to
make
assumptions
about
what
the
events
mean
and
is
creating
a
dashboard
that
basically
do
the
query.
On
top
of
the
events.
D
To
show
you
a
dashboard
basically,
and
so
definitely
that's
the
stuff,
we
can
reuse
from
from
four
keys
yeah,
but
the
I
think,
one
of
the
one
of
the
problems
in
one
of
the
amount
of
work
that
they
had
to
do.
At
least
that's
my
understanding
from
looking
at
the
presentation
is
that
because
they
use
a
lot
of
tecton
events
and
there
is
no
real
information
in
there
beyond
like
tasks
and
pipelines.
So
there
is
no
like
higher
level
abstractions
about
what
what
the
pipeline
is
doing.
D
So
they
need
to
rely
on
some
kind
of
naming
conventions
and
then
logic,
post-processing
logic
to
say.
Okay,
this
pattern
was
actually
doing
a
deployment
into
this
pipeline
was
actually
doing
a
builds
and
then
correlate
this
information
so
that
they
can
create
the
the
final
metrics
and
the
idea
here
would
be
because
in
cd
events,
we
have
more
much
more
semantic
to
show
that
we
can
do
that.
D
But
yeah
I
would
love,
for
instance,
to
be
able
to
use
the
the
same
dashboard
showing
the
four
metrics
with
a
green
light
or
red
light.
D
Yeah
also,
I
think
they
they
use
bigquery
there,
but
I
think
yeah,
the
back
end
might
be
interchangeable.
So
if
we
like
have
a
local
mongodb
or
some
other
kind
of
back-ends.
D
A
D
So
someone,
I
think,
asked
us
sorry,
just
the
name
of
the
lady
escaping
me.
D
Yeah
the
engineer
for
from
google
who
worked
on
the
project,
but
during
a
presentation
at
cdcon
someone
asked
about
how
strict
the
dependencies
to
be
query,
or
you
know
she
said
that
yeah
there's
nothing
that
there's
really
bigquery
specific,
so
you
could
use
another
track
and
then
change
some
of
the
logic
in
the
project
to
run
it
on
top
of
something
else.
C
And
there
may
be
other
projects
that
we
can
reuse
like
when
we
go
to
like
visualize
this.
You
know
you
have
like
kyoli
and
even
prometheus,
and
you
know
istio
all
have
some
sort
of
visualization
of
the
of
relationships,
so
some
of
those
may
be
able
to
you
can
grab
things
from
those
projects
as
well.
C
C
So
the
the
two
visualization
libraries
I
end
up
using
a
lot
is
d3s
is
one
of
them.
C
D
just
d,
the
letter
d,
then
three,
the
number
three
s,
the
other
one
is
viz
v,
I
s
dot,
j
s.
C
Yeah
they're
both
javascript,
but
the
more
more
on
the
note.
Those
are
the
two
primarily
they
are.
They
have
favorable
licenses
for
us
to
include
some
of
the
extensions
may
be
in
like
a
a
new
license,
but
the
core
libraries
are
all.
I
think
mit
or
apache.
C
Yeah
we
use
them
in
ardelius
both
of
those.
D
Okay,
cool
yeah,
I
mean
if
we,
if
we
get
to
a
point
where
we
we
want
to
build
some
visualization.
Maybe
if
you
have
someone
on
a
tv
site
that
has
a
few
cycles
to
spare,
maybe
we
can
work
together,
yeah,
definitely
something
I
have
very
limited
experiences
on
javascript.
So
definitely.
C
A
lot
of
it
is
just
getting
the
data
if
you
can
get
the
data
in
the
right
format
that
the
library
is
expecting
usually
some
sort
of
json
format
that
you
can.
You
know
they'll
give
you
a
sample
and
if
you
can
follow
that
pattern
and
get
the
data,
your
data
in
that
format,
it's
pretty
plug
and
play
to
drop
in
and
and
get
the
something
to
render
pretty
easily
cool.
C
But
usually
they
the
input
is
a
a
json
format.
C
Use
a
transposed
data
set
to
do
the
rendering,
but
once
you
have
that
base
json
file,
json
format,
then
it's
it.
That's
the
the
key
part.
D
All
right
thanks
I'll
yeah
I'll,
add
some
notes
about
this.
Also
in
the
in
the
poc
document
then-
and
I
mean
it
will
be-
I
think
it
would
be
great
if
we
can
get
some
some
of
these
as
well.
I
mean
like
having
a
dashboard
that
shows
you,
the
the
metrics.
C
C
C
D
Yeah
same
same
here,
but
yeah
thanks
a
lot
steve
and
I
guess
then
see
you
next
time.
A
Andrea,
yes,
just
a
short
one.
I
dropped
in
a
link
there
in
the
chat
I
saw.
You
talked
about
releases
and
so
on.
Yeah
right.
G
A
Quickly
in
april,
we
we,
basically,
if
you
have
a
release,
and
you
want
to
group
like
events
together.
We
have
this
flow
context,
defined
events
and
that
one
event
would
then
other
events
would
link
to
this
one.
A
So
then,
you
would
actually
define
your
release
or
something
there,
and
you
would
link
all
the
other
events
to
that
event.
That
is
how
we
have
solved
it
in
eiffel
to
group
events
together
to
something.
E
A
D
A
A
Backwards
in
some
cases,
but
usually
as
you
define
like
a
track
opening
or
something
you
define
now,
we
have
a
certain
we're
preparing
for
a
certain
lease,
and
then
you
have
all
the
events
connecting
to
that
one.
Instead,.
A
Yeah,
if
you're,
if
you're,
defining
the
release
afterwards,
then
we
need
it,
but
then
maybe
you
could
create
enable
it
would
be
a
composition
defined
to
create
a
composition
to
group
things
together,
and
then
you
can
refer
to
that.
Okay,
the
in
this
composition
is
the
one
that
defines
this
release.
D
D
A
Yeah
yeah,
I
just
wanted
to
point
it
out
because
I
saw
when
I
looked
at
the
you
talked
about
releases
yeah.
A
So
yeah
because
you
said
there
that
it's
yeah
optional
repo
attributes
are
we
talking
about
like
a
release
event
or
what's
that
something?
Maybe
I
haven't
read
this.
I
probably
should
read
it
before.
I
ask
too
much
if
you
need
to
go
now.
A
But
you
can
escape
it,
I
can
read
better
and
you
can
go
on
to
your
next
meeting.
I
don't
want
to
hold
it
too
much.
D
All
right:
well
thanks,
thanks
again
that
yes
yeah,
I
think
there's
lots
we
can
benefit
from
from
eiffel
and
I
need
to
to
go
inspect
into
the
links
as
well
to
to
see
make
some
proposals
around
that.