►
From YouTube: SunPy Coordination Meeting 2021 - Thursday
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
G
B
H
H
H
B
Of
things
that
I
think
are
worth
talking
about
when
around
composite
map,
which
is
kind
of
what
use
cases
we
want
to
serve
with
both
map
sequence
and
composite
map,
what
functionality
were
they
being
were
they
providing?
Are
they
providing
and
are
we
is
there
anything.
B
That
we
don't
have
anywhere
else,
and
then
you
know
if
there's,
if
they're
not
serving
any
useful
purpose
anymore,
we
can
delete
them,
but
I
suspect
that
there
is
actually
use
cases
for
something
along
the
lines
of
composite
map
somewhere
and
we're
not
necessarily
doing
a
particularly
good
job
of
it.
At
the
moment,
things
to
think
about
in
this
context
are
also
the
nd
cube
sequence
and
then
the
collection
classes
in
indie
cube.
E
Could
you
quickly
update
those
of
us
who
aren't
familiar
intimately
with
composite
map
about
what
its
intended
point
either
was
or
is,
or
is
that
kind
of
the
point
of
this.
B
That's
part,
definitely
part
of
the
point.
I
can
tell
you
what
I
think
it
does
at
the
moment.
People
can
correct
me
composite
map
stores,
a
unordered
list
of
maps.
It
allows
you
to
set
plot
settings.
E
Do
you
mean
unordered
in
the
sense
of
a
coordinate
like
it's
not
ordered
in
time,
yeah.
B
It
does
not
expect
there
to
be
a
physical
dimension
along
the
sequence.
The
assumption
in
composite
map
is
that
the
maps
are
aligned
that
are
like
temporarily.
Co
are
co-temporal
in
the
or
at
the
very
least,
you
want
to
over
plot
them
on
each
other.
E
I
I
K
It's
really
just
a
I
mean
from
that
context.
Composite
map
is
largely
a
scripting
helper
right.
It
stores
the
plot
settings
and
you
call
plot
you
just
have
to
write.
You
know
blah.plot
and
they'll
do
all
the
stuff
all
the
internal.
You
know,
sequence
of
plot
calls
and
and
retrieve
those
settings
on
the
dictionary,
and
that
kind
of
thing
are
the
attributes.
K
So
the
use
case
would
be
so
that
users
wouldn't
if
they
are
routinely
plotting
composite
maps,
then
they
wouldn't
need
to
like
do
all
those
multiple
clock
commands
over
and
over
and
over
and
have
carry
along
another
dictionary
that
carries
along
those
plot
settings
over
like
and
keep
track
of
that
all
right.
So,
yes,
you
can
replicate
a
lot
of
existing
functions.
Of
course
you
can
recreate
all
the
existing
functionality
deposit
map.
I
K
B
D
Going
to
give
another
use
case
of
times
when
it
has
been
used
is
like,
when
you're
matching
two
maps
say
like
a
lasko
c2
map
with
an
aia
map
in
time,
and
then
you
can
easily
contain
them
within
one
structure
before
plotting
so
say
you
only
wanted
a
maps
at
the
certain
times
of
c2.
You
could
create.
You
know
several
composite
maps
and
then
plot
them
easily
or
similar
with
anything
any
type
of
two
maps
that
you
want
to
match
on
time.
It's
just
kind
of
helpful.
I
guess.
I
D
D
K
And
importantly,
I
would
say
if
I
understand
what
you're
saying
well,
I
mean
map
sequence.
Is
it's
a
collection
with
which
there's
the
order
that
is
physically
motivated
and
the
maps
are
supposed
to
be
consistent,
some
manner
as
opposed
to
composite
map
or
a
map
collection
where
it's
it
could
be
overblooding
anything
or
anything
else.
It
could
be
crazy,
wacky,
stuff.
K
So
I
mean
you
know
this
boils
down
to
a
lot
of
questions
about
api.
I
would
expect
people
to
be
using
certain
things
and
again,
you
know
underlying
map
functionality.
Yeah
users
can
always
just
supposedly
have
a
map
called
explicitly
carry
around
whatever
plot
settings
and
just
copy
and
paste
over
and
over
and
over
they're
doing
this
yeah
multiple
times
or
you
know,
we
try
to
extract
that.
B
B
A
set
of
maps
that
are
not
co-aligned,
but
you
kind
of
want
to
treat
them
as
such
for
analysis
purposes
and
then
have
map
align
them
as
necessary.
I
guess
how
the
api
for
composite
map
doesn't
look
pretty
to
me.
Are
we
really
shortcutting
the
plotting
stuff,
like
sure,
there's
a
default?
If
you
do
compositemap.plot
you
get
something
which
you
don't
get
out
of
matplotlib
by
default.
D
B
K
K
B
B
K
I
mean,
I
would
say
you
know
yeah,
I
realized
yeah,
there's
a
difference
between
saying
like.
Oh
the
functionality
is
not
so
complex
or
the
you
know.
The
internal
codes
of
component
map
are
not
that
you
know
complex
that
we
can't
imagine
user
doing
that.
The
the
question
I
think
we
should
be
asking
ourselves
is
from
the
api
perspective
from
abstraction
perspective.
K
Is
there
a
reason
to
have
something
called
composite
map?
Maybe
right
now
it's
not
doing
much
more
than
plot
end,
but
it
could
be
doing
more
if
it
makes
sense
from
use
case
perspective
right.
I
think.
D
Yeah-
and
it
goes
back
to
the
question
that
people
have
asked
several
times,
you
know:
can
I
save
a
composite
map
right
and
then
they're
using
it,
obviously
for
something
other
than
plotting,
if
they're
trying
to
save
it,
and
I
think
it
goes
back
to
having
it
it's
a
nice
container
to
hold
a
collection
of
maps
that
are
related
but
not
like
it
can
be
related
in
several
dimensions
or
whatever
you
want
it
to
be
right.
Okay,.
B
B
B
It
doesn't
have
a
dick
like
interface
for
getting
those
maps
back
out
again,
based
on
some
name.
Nd
collection.
Has
that
dick,
like
interface
for
naming
your
nd
cubes
and
then
getting
well?
It
can
also
you
can
us.
You
can
say
that
the
axes
are
the
same
right.
So
a
map,
specific
subclass
of
collection
could
just
always
say
that
the
axes
were
the
same
because
like
if,
if
you
passed
in
three
maps,
it
would
be
like
oh
yeah.
These
are
all
hp,
ln,
lt
cool.
These
are
helio
projective
maps.
B
E
B
E
It's
based
on
the
pixel
grid,
so
you
say
that
you
know
in
these.
In
these
maps
the
the
zeroth
axis
are
aligned
along
the
pixel
grids.
I
don't
think
it
cares
either
way
what
the
physical
types
are.
E
B
J
A
F
F
Okay,
so
the
idea
is:
we've
got
four
satellites
they're
going
to
be
producing
four
images,
white
light
images
and
we're
gonna
have
to
stitch
those
images
together
and
so
from
our
pipeline.
We're
basically
going
to
be
spitting
out
level,
one
images
which
will
be
individ
individual
images
from
each
of
the
four
satellites
and
we're
going
to
going
to
be
producing
level
3
images
which
will
be
composite
images
where
they're
stitched
together.
F
K
I
don't
know
if
that
answered
your
question
or
not.
No
supplies
might
be
a
more
extreme
example,
but
the
concern
of
the
I'm
struggling
to
think
of
an
easy
case
right
now,
but
certainly
a
single
platform
with
multiple
camera
views
in
multiple
directions.
That
you'd
want
to
treat
as
kind
of
a
unified
observation,
but
are
really
like
separate
maps.
A
A
Yeah,
I
mean
because
one
of
the
things
like
so
the
his
like
you
want
to
track
a
cme
front
out
from
hi1
into
hi2.
If
you
could
do
what
you're
saying
like
you
could
combine
the
two
mats
in
some
way
and
kind
of
treat
them
at
some
level
as
one
and
get
coordinates
out
from
here.
That
would
be
super
useful.
D
E
Yeah
or
or
not,
if
you
want
it,
it
allows
you
to
do
it.
It
gives
the
option
of
linking
them.
It
doesn't
it's
more
abstract,
you
can
just
use
it
as
a
pure
dictionary,
and
if
you
don't,
if
you
for
some
reason,
don't
like
dict,
you
know
you
can
just
use
the
end
the
any
collection
and
use
nothing
else
on
it,
but
it
gives
you
the
options
of
of
linking
things
a
bit
more
complicatedly.
What.
F
I
would
say
that
one
of
the
main
reasons
people
use
j
helio
viewer
is
for
stitching
images
together.
So
and
as
much
as
I
I
like
j
helio
viewer,
it
would
be
nice
to
actually
get
a
get
my
hands
actually
on
science
quality
calibrated
data
in
a
similar
fashion,
so
I
think
that
there's
definitely
a
use
for
it
in
with
senpai.
E
E
F
Yes,
look
I
I
I
I
I
don't
want
to
show
on
jay
leave
you.
I
think
it's
a
great
product
and
I
think
it's
wonderful,
but
I
primarily
use
it
for
just
getting
a
pretty
image.
I
wouldn't
use
it
for
science.
F
Yeah,
it
would
be
I
I
haven't
done
cme
science
in
a
while,
but
when
I
was
working
at
rob,
we
did
a
lot
of
kind
of
space,
weather
stuff
and
I
had
a
hell
of
a
time
trying
to
combine,
even
even
on
the
same
platform
core
two
and
core
core
one
and
core.
Two,
even
like
stitching
those
images
together,
it's
quite
complicated
to
put
hi
in
there,
because
it's
so
big
compared
to
core.
F
So
that's
kind
of
a
difficult
stitch,
but
just
putting
them
together
to
track
points
across
it's
difficult
and
going
from
euv
to
like
swap
to
lasko.
Again
we
I
know
people
at
rob
who
would
for
their
forecasts.
They
would
sit
there
with
kind
of
like
tracking
across
the
screen,
with
your
viewer.
C
A
B
I
I
semi
deliberately
haven't
stated
my
motivations
at
the
beginning
of
the
session,
but
I
actually
have
some
non-focus
funding
to
rewrite
composite
map.
B
B
B
B
K
So
so
it's
kind
of
like,
I
guess,
the
analogy
of
what
I
think
you're
saying
is
like
to
the
differential
of
the
diff
rot.
Whatever
we
have
it
currently
called
where
the
the
rotation
is,
is
obvious
or
is
straightforward,
the
question
is
defining
the
new
wcs,
and
so,
if
it's
one
of
the
ones
in
your
in
your
existing
composite
map,
you
don't
have
to
do
any
work.
If
it's
trying
to
figure
out
what
is
a
shared
wcs
that
naturally
fits
all
your
existing
wcs's,
I
mean
that
yeah,
that's
that's!
K
B
B
K
K
B
B
K
K
The
the
you
know,
things
that
could
be
part
of
the
consideration
is
that
you
know
like
as
part
of
your
composite
map,
you
define
how
things
are
actually
composited
right,
like,
for
example,
like
the
synoptic
map,
how
you
blend
together,
aia
euvi,
imagery
to
actually
make
the
overall
synoptic
map
that
you
know
the
gallery
example
that
we
have
for
that.
There
could
be.
You
know,
settings
and
stuff
that
would
be
used
for
the
visualization
side.
K
If
you
just
do
a
dot
plot
call
could
be
used
in
a
regrid
call
could
be
used
in
a
flattened
call
to
be
using
these
kind
of
things
and
that's
carried
along
now.
That
could
all
be
a
whole
lot
of
keyword,
arguments
on
a
function
where
you're
like
throwing
dictionaries
or
you
throw
in
all
those
other
stopping,
but
then
that
puts
on
the
user
to
carry
it
along,
and
then
you
know
to
repeatedly
perhaps
use
the
same
thing
for
different
types
of
function
calls
that
they
might
call
on
this.
K
I
think
there's
a
conceptually,
I
think,
there's
a
there's
a
space
for
a
composite
map.
You
know
concept
where
it's
the
sting
that
you
can.
You
know
just
stuff
everything
you
need
into
it
and
then
you
know
you
do
some
method
calls
on
it
and
if
that's
what
it
needs
to,
because
you've
already
stored
that
state
in
it.
B
K
Well,
I
think
the
interesting
thing
to
discuss
would
be
like
that.
Doesn't
do
anything
fancy
about
trying
to
combine
things,
especially
where
the
images
might
overlap
right
it
just
straight
overlays
them
if
you're
not
playing
those
contours,
whichever
was
the
later
one
in
the
order
of
that
they're
overlaid
is
the
one
that
shows.
One
could
certainly
imagine
a
fancy
composite
map
that
tries
to
you
know
account
for
overlaying
parts
and
kind
of
blend
them
together
or
down
transparency
or
like
localized
transparency,
or
I
don't
know.
K
B
K
Well,
no,
not
even
an
empty
map
or
just
the
mpwcs
create
the
wcas
axis
with
the
right
wcf.
Then
you
can
manually
set
your
your
axis
limits,
but
create
that
wcs
create
the
axis
limits
and
just
plot
directly
onto
those
wcs
axis.
L
K
That
that
literal
thing
I
may
not
have
done
very
recently,
but
I
know
I
did
it
when
I
was
trying
to
show
or
demonstrate
that
you
could
do
the
draw
limb
call
so
fundamentally
should
work
the
same
way
right.
I
created
an
empty
hds
grid
and
I
did
a
draw
limb
call
from
aia
and
it
worked
so
you
gave
me
just
a
little
bit.
I
can
show
it
for
it.
B
I
think
I
figured
I
figured
it
would,
but
it
came
to
mind
we
were
talking
about
that.
Okay,
so
we
could.
You
could
all
right
so
we're
now
thinking
about
an
object
where,
oh
god,
this
api
is
going
to
be
hell.
But
okay,
there's
an
there's,
an
object
where
you
give
it
a
collection
of
maps
and
you
define
it.
K
Yeah
we're
just
going
to
be
overlapping
kind
of
desires.
There.
B
I
E
E
B
E
As
long
as
there's
a
step
that
you
you
know
like
you
just
hit
you
just
sort
of
call
a
method
and
it
produces
a
plot.
All
that
stuff
is
done,
but
isn't
like
easily
retrievable
by
the
user
as
long
as
they
can
use
the
result
of
all
of
that
work.
In
another
step
of
analysis,
then
then
this
has
a
greater
use
as
far
as
or
usability.
As
far
as
I
can
see,.
I
So
I
just
realized
sorry,
I
was
just
gonna
say
I
realized
I've
done
a
poor
job
of
moderating
this
this
session
and
that
we
are
at
45
and
haven't
talked
about
database,
but
I
mean
if
we're,
if
we're
fine,
just
keep
going
on
constant.
I
don't
know
how
much
you
want
to
say
about
database
stuart,
but
I
just
wanted
to
make
sure
that
we
had
to.
I
A
I'm
just
going
to
ask
so
like
there's
use
cases
for
both
like
there's
a
use
case
for
using
p
color
mesh
to
just
make
a
nice
image
and
there's
a
use
case
where
I
actually
want
to
like
reproject
my
things
and
have
things
on
agreed
that
I
can.
I
can
do
some
clients
on
like
once
performance-wise
you
might
want
to
do
one
of
the
other,
depending
on
like
what
you're
like
what
you're
not
trying
to
do.
B
E
But
you
can
effectively
have
a
glorified
wrapper
function
that
does
that,
but
that
you
don't
have
to
tie
to
okay.
Now
we
have
a
composite
map
objects
for
this,
like
set
of
cases
and
a
map
sequence
for
this
set
of
cases,
but
actually
underneath
the
calculation
is
effectively
the
same.
It's
just
like
a
little
bit
of
like
processing
the
inputs
from
the
user.
G
B
L
B
B
B
If
you
want
to
stack
them
in
to
h,
if
you're,
what
you're
asking
nearly
is
stack
these
onto
the
c2
projection
or
something
that
looks
an
awful
lot
like
it
and
by
default,
you
could
be.
These
two
maps
are
both
in
hpc.
G
B
B
And
calculate
I
don't
know
how
easy
this
would
be,
but
theoretically
you
could
be
like
well.
The
pixel
scale
of
this
image
is
that
the
pixel
scale
of
that
image
is
this.
B
D
G
K
E
E
I'm
just
thinking
this
through.
I
know
if
you
keep
it
as
if,
if
you
keep
the
function
separate
and
just
say,
we've
defined
what
we've
done
is
we've
defined
like
a
container
class
and
then
we
can
have
like
a
library
of
functions.
That
knows,
you
know,
there's
a
there's
a
default
container
class
of
maps,
and
I
can
you
know,
take
that
in
return,
another
collection
of
maps
or
like
combine
them
in
some
way.
E
I
A
I
think,
as
like
a
lot
of
it,
it's
the
plotting
site,
as
albert
we're
saying
earlier
on,
like
you
know,
like
you're
gonna,
want
to
have
different
plot
settings
for
each
map,
different
levels
of
transparency
different
whatever,
and
it
seems
to
be
that's
the
sort
of
state
that's
like
user
set
and
that
would
is
useful
to
carry
around
but
not
important
for
the
calculations
or
the
underlying
you
know
stuff.
It's
just
what
you
need
to
make
a
nice
plot.
In
the
end.
B
A
B
B
B
I
think
there
are
api
changes.
I
think
there
are
api
questions
which
are
interesting,
but
I'm
probably
going
to
start
on
the
start,
with
the
more
complex
implement
a
function
that
reprojects
a
bunch
of
maps
together
stuff
and
then,
where
that
ends
up
api
wise,
we
can
decide
later.
I
think
from.
A
B
A
B
A
D
They
use
it
as
a
containerized
thing
to
like
have
like
their
radio
contour
data
that
match
their
aia,
the
closest
time
to
aia
so
that
they'll
have
for
every
second,
they
have
a
radio
image
they'll
put
like
the
closed
ai
image
to
that
every
second,
you
know
it's
a
nice
way
just
to
contain
them,
so
they
can
do
kind
of
a
loopy
thing
to
to
get
all
these
composite
maps
and
they
can
plot
them
as
a
movie
and
quite
easily
and
they're,
using
it,
basically
as
a
container
to
hold
those
similar
things
together
and
and
then
obviously
just
for
plumbing,
and
I
think,
with
anything
with
like
something
like
a
composite
or
sorry
like
a
contours
and
stuff
like
that,
it's
really
handy
in
that
way,
because
you
can
set
up
your
contour
levels
nicely
and
carry
it
around
for
each
but
yeah
definitely
worth
checking
asking
them
as
well.
K
Well,
it's
just
yeah.
Actually
I
proved
two
points,
one,
it's
a
little
tricky,
but
two
once
you
get
past
that
tricky
part,
it's
pretty
easy.
So
just
the
screen
is
just
after
the
map
is
loaded,
create
the
appropriate
header
for
the
output
wcs
using
the
header
helper
that
we
have
make
sure
to
get
the
right
projection.
L
K
And
then
down
here
just
specifies
that
it's
when
you're,
creating
the
obvious
aspects
and
the
main
trick
beyond
the
other
trick
of
getting
the
header
right
is
that
you
need
to
manually
set
the
x
and
y
limits,
because
it
has
no
idea
what
the
x
and
y
should
be.
When
you
don't
create
a
map
by
dummy
map
and
then
once
you
do,
that
then
auto
line
equals
true
is
magical.
It
does
everything
you
need
to
do.
K
I
I
I
just
say
we're
at
the
hour
right
now:
do
we
want
to
break
for
30
minutes
and
come
back
and
talk
about?
We
can
finish
up
any
any
sort
of
stuff
we
want
to
do
with
if
there's
anything
left
to
say
about
composite
map
or
we
can
just
dive
into
governance,
which
I
think
is
what
we
have
for
the
afternoon.
I
H
B
B
D
The
edges
were
all
blurred
out,
streaming
is
on,
oh,
is
it
it
was
all
blurred
out,
but
then,
obviously,
when
you
reproject
it
like
trims,
all
those
things
out,
if
you
were
to
put
it
into
hds
coordinates.
Is
that
because
the
flux
is
like
conserved
or
something
or
is
it.
B
I
B
Transforms
does
pixel
world
world
2
pixel
2.
to
get
no,
it
actually
does
that
backwards
to
to
do
the
original
reprojection,
but
then
to
check
whether
that
point
is
valid.
It
does
that
transform
in
the
opposite
direction
and
if
the
result
varies
from
the
input
by
some
tolerance
factor.
It's
like
this
point
is
nonsense
which
it
uses,
which
in
that
case
basically
is
a
check
to
see
whether
is
this
point
on
the
back
of
the
sun
or
not
like.
If
I
do
this
transform
in
the
opposite
direction,
am
I
trying
to
transform
the
point?
G
B
M
M
B
It's
really
useful,
or
it
could
be,
can
be
really
useful.
It
predates
fighter,
so
it
really
has
the
same
interface
as
the
vso.
B
B
B
B
I
won't
take
it
personally,
okay,
so
I
kind
of
want
to
have
the
same
discussion
about
database
that
we
just
had
about
composite
map,
which
is
basically
what
are
the
potential
or
actual
use
cases
for
what
is
currently
database
or
what
database
could
be
in
the
future.
B
As
a
summary
for
people
who
haven't
looked
at
database
in
a
while,
which
is
everybody
what
it
does,
is
ingests
a
vso
search,
result
and
associated
file
into
a
sql
alchemy
database,
which
can
be
backed
by
anything
sql
like
postgres,
whatever
sql
type
database,
you
want
to
back
it
with
by
default.
It
uses
a
sqlite
file
in
the
user's
sample
directory,
so
it
stores
the
whole
fits
header.
B
The
parameters
from
the
vso
search
result
such
as
time,
wavelength,
etc,
and
a
path
to
the
file,
an
absolute
path
to
the
file,
and
then
it
then
you
have
a
sql
database
that
has
all
that
information
in
it.
So
it
provides
ways
of
searching
that
database.
B
By
so,
if
you
do
the
same
vso
search
twice
through
the
database,
the
second
time
you
do
it
it'll
be
like
hey.
I've
got
all
of
these
results
in
your
date
in
the
database.
For
you
it
also
because
it
has
the
whole
fits
header
in
the
sql
table.
It
allows
you
to
do
arbitrary,
fits
header,
key
search,
so
you
could
search
all
of
the
fits
files.
You've
indexed
for
some
header
value.
B
The
thing
that
I
have
always
thought
would
be
neat
for
it
to
do,
but
we
have
never
implemented,
was
b
provide
a
index
for
a
shared
storage.
So
if
you
had,
you
could
have
20
people
all
use
the
same
database
hosted
in
postgres
or
something,
and
if
the
paths
were
relative
to
a
mount
that
they
could
all
access.
A
A
B
B
H
B
Don't
ask
me
mate
something.
B
Cool
some
things
to
notice
about
the
databases
that
is
very
old.
It
predates
fido
by
a
chunk.
Obviously
it
is
very
fit
centric.
It
is
supposed
to
be
ingesting
fizz
files
or
things
within,
and
even
more
than
that,
it
really
expects
to
be
ingesting
things
that
look
like
a
map,
but
it
is
a
bit
flexible
about
that.
H
B
The
database
has
more
metadata
to
compare
it
can
check
whether
the
row
is
the
same,
or
it
can
also
you
you.
So
what?
If
you
did
a
vso
search
or
a
phyto
search
proxy
through
the
database,
and
you
were
like
okay,
this
search
put
in
this
path,
and
then
you
did
the
same
search
later
with
a
different
path.
B
I
I
B
If
that
makes
sense,
that
always
was
on
the
roadmap,
well
has
never
been
implemented
because
it
predates
fido.
The
original
idea
before
fido
was
that,
if
you
wanted
to
use
the
database
instead
of
using
vso
client.search,
you
would
use
database.search
pass
it.
A
vso
query
see
the
results
of
that
vso
query
download
that
vso
query
through
the
database.
The
database
would
ingest
the
results
and
then
later
you
would
do
the
same
vso
search
through
the
database
and
it
would
go
hold
up
there
here
you
go
that
workflow
never
got
translated
to
fido.
B
B
So
you
start:
okay,
you
rock
up
to
a
new
computer,
green
and
beautiful.
Just
just
just
arrived
on
your
desk.
You've
got
no
data
on
it.
You
obviously
the
first
thing
you
do
is
install
senpai
and
do
a
fido
query,
because
you
know
you've
got
your
priorities
straight,
so
you
search
for
a
day's
worth
of
aia
data,
because
awesome
flare
just
happened
and
download
that
data
to
your
local
machine.
B
B
Some
of
those
files
will
be
returned
from
the
local
database
provider
in
the
local
database
table
in
the
fido
results.
Right,
like
one
of
the
one
of
the
provider
blocks
in
the
results,
would
be
your
local
database
you'd,
be
like
I
have
these
files
and
then
obviously
a
lot
longer
list
would
be
in
the
vso
block
or
jsoc
or
wherever
you
get
it
from.
I
B
B
I
B
Which
leads
to
the
question
of
in
its
current
state,
I
couldn't
extend
the
sum
by
database.
For
that
end,
I'd
have
probably
have
to
start
afresh.
L
I
I
What
at
what
point
do
you
do
the
indexing?
So
maybe
that
you,
I
don't
in
my
mind
then
like
database-
could
become
the
module
that
maintains
is,
is
maintains
that
data
source
locally
and
then
fido
is
the
interface
to
looking
at
that
data
source
right
rather
than
database.
Being
the
the
kind
of
entry
point
for
a
query.
B
B
Do
people
so
a
more
general
polling
of
thoughts?
The
database
talk
has
once
again
got
the
rabbit
to
sleep.
The
general
thoughts
about
would
people
use
database?
What
would
database
need
to
do
for
people
to
find
it
useful.
A
I
mean,
I
think
the
first
thing
you
said,
but
it
doesn't
do
right
now
if
you
could
have
a
shared
institution
like
monks
where
you're
like
this
is
all
the
data
we
have
downloaded,
don't
bother
going
to
download
it
from
the
so
again
like
we've
got
it
located,
so
you
can
just
get.
It
would
be
very,
very
useful,
at
least
for
like
little.
B
A
What
would
what
would
I
mean
the
only
issue
that
you
always
have
with
database
things
is
like
concurrency
things
like
if
someone
downloads
the
same
file
at
the
exact
same
time.
I
guess
it
doesn't
matter,
because
you
know
if
it's
in
the
database,
then
it's
already
in
there.
B
A
B
B
It's
it's
conceptually
trivial,
on
top
of
a
local
database
right,
because
the
files
are
still
local
to
the
process.
That's
interacting
because
you
mount
them
and
you
have
because
the
database
is
implemented
in
sql
alchemy,
pointing
our
remote
postgres
on
mario
db
instance
is
super
trivial,
so.
I
B
You
can
already
you
can
already
do
that.
In
fact,
this
use
case
already
works
in
senpai
today,
as
long
as
everybody
mounts
that
path
at
the
exact
same
location
on
their
pcs.
If
everybody
was
running
a
linux
box
or
a
mac
os
box,
and
they
mount
it
as
forward
slash
diaz
data,
it
would
work
the
only
time
it
wouldn't
work
is
if
you
have
to
have
differing
group
files
for
the
for
the
to
access
the
files.
B
B
B
I
B
B
B
H
So
my
issue
is
the
the
way
it's
currently
set
up,
so
okay
use
case,
fine,
great
incredible.
The
best
thing
ever.
The
issue
is
that
the
current
database,
module
and
sql
alchemy
is
not
is
not
a
basis
for
actually
spinning
up
an
institutional-wide
database
that
could
be
managed
and
maintained.
I
B
B
A
couple
more
use
cases
than
that
it
allows
you
to
search
your
data
based
with
more
it
has.
The
local
database
would
have
more
information
than
the
web
service.
You
got
the
data
from,
for
instance,
the
entire
fits
header.
It
allows
you
to
do
more
complex
local
searches
than
is
possible
through
the
provider.
A
H
Oh
just
a
matthew
asked
question:
we
are
talking
about
the
sunpi
database
module,
which
is
an
sql
alchemy
application.
Essentially
that
allows
you
to
make
searches
and
fetches
and
it
will
process
the
data
it
downloads,
for
you.
B
H
B
H
B
I
B
B
H
As
long
as
you've
sub
as
long
as
you've
made
this
most
of
the
path
kind
of
on,
as
long
as
we
haven't
randomized
the
path.
B
B
H
H
B
B
H
B
Okay,
there
seems
to
be
general
consensus
that
that
kind
of
functionality
would
be
cool
and
useful,
which
I
guess
is
a
good
time
to
pass
over
to
the
bill.
And
let
him.
H
B
H
H
B
Fits
three
and
fits
four
have
different
conventions
for
exposure
time.
You
would
not
be
able
to
generate
a
single.
You
would
have
to
search
both
of
those
keys
explicitly
to
search
the
differences,
there's
no
abstraction
of
the
data
model.
That's
what
I'm
getting
at
you're,
just
directly
searching
the
fit
setter.
B
It's
not
like
on
ingest
we're
going!
This
key
is
exposure
time,
I'll
put
it
in
a
column
named
exposure
time.
It's
we're
just
dumping.
The
fits
header
into
a
sql
store.
H
H
I'm
just
saying
that
we
would
be
specifically
making
a
local
file
search,
that
someone
might
make
the
assumption
that
if
they
were
to
search
for
a
exposure,
they
might
get
that
on
from
the
vso
or
something
or
other
providers
is
that
I
feel
like
these
are.
These
are
adders
that
are
only
going
to
be
relevant
for
local
files,
which
will
even.
I
B
Simulation
one
thing
that
just
pops
into
my
mind
is:
if
we
want
to
contin,
if
we
want
say
a
dot
instrument
to
also
search
local
data,
we
there
is
going
to
have
to
be
a
way
to
generate
a
is
to
map
the
you
have
to
be
able
to
map
the
query
response
table
back
to
the
adders.
We
are
enforcing
a
data
model
upon
query
response
table
to
make
that
work
at
some
level.
B
I
B
So,
for
instance,
you
couldn't
search,
he
said
say
there
was
a
table
in
the
database
for
jsoc
and
a
table
in
the
database
for
vso
you
in
much
the
same
way.
You
can't
search
jsoc
with
a
dot
instrument.
You
wouldn't
be
able
to
search
local
data
held
in
the
database
with
a
dot
instrument
either
because
there
would
be
no
mapping
back
to
instrument.
B
B
B
I
think
this
is,
I
think
this
is
solvable
and
it's
an
implementation
detail.
Let's
continue
to
discuss
actual
concrete
plans
and
features
and
stuff
higher
level
stuff.
H
B
H
B
H
New
columns
like
adding.
I
H
I
H
See
I
went
back
on
this
and
someone
who
you
know
does
this
now
as
a
living
this
we
need
to
be
very
careful
about
restructuring
database
because
I
think
we
probably
need
to
sit
down.
I
think
a
new
version
of
database
probably
doesn't
look
like
doesn't
have
the
same
models
doesn't
have
the
same.
Columns
doesn't
have
the
same
rows.
B
H
Database
doesn't
have
to
be
I'm
saying
that
you
know
a
lot
of
the
columns
as
it
stands.
Right
now
is
tied
and
in
a
world
where
you
know
we
want
to
support
as
different
I'm
not
sure
how
as
diff
deals
with,
I
actually
don't
know
I've
never
loaded
it
as
a
default,
but
the
metadata
for
asdiff,
I
assume,
is
just
easily
we
just
index
the
ammo
file,
which
videos.
I
B
H
H
B
H
B
H
H
B
B
H
H
H
B
H
H
H
H
H
B
H
I
What
yeah
it'd
be
nice
to
have
a
smallish
pot
of
money
for
someone
to
do
this,
I
mean
this
is
a
pretty
significant
effort
and
sort
of
also,
but
also
fairly
limited
in
scope
too,
like
I.
I
B
H
N
My
mic
is
working.
If
you
can
hear
me,
I
think
that'd
be
great,
like
I'd,
be
totally
happy
for
them
to
spend
some
time
left.
H
H
H
That
you
know
your
web
app
uses
the
database
to
store
its
settings
or
whatever,
but
I
think
at
that
point
once
you're
once
you're
writing
your
own
database
to
do
that
stuff.
Then
I
think
in
a
pipeline
point
of
view
from
that
you
would
probably
not
use
sunpower
to
manage
your
data
like
that.
But
well
I
guess
yeah.
H
B
H
Should
we
move
on
to
the
project,
then,
unless
someone
else
has
thoughts
on
database
as
a
concept.
H
We
are
basically
writing
and
we
are
basically
writing
a
database
of
our
our
own,
we're
just
we're
just
encapsulating
sql
coming
out.
That's
what
we're
gonna.
B
Do
okay
all
right
project
governance,
then
so
I
am
under
prepared
to
discuss.
This
has
been
something
that
has
been
in
the
back
of
my
mind
for
a
long
time
and
have
not
wait.
Did
I
not
write
any
notes
about
this?
No,
I
didn't
oh
cool,
okay.
I
really
am
doing
this
off
the
top
of
my
head
yeah
and
I'm
a
bit
under
prepared
to
think
about.
I
B
However,
I
have
been,
and
the
board
to
some
extent
have
have
have
discussed
the
fact
that
we
think
that
the
current
setup
of
the
sunpi
governance
is
maybe
lacking
a
bit.
So
for
those
who
don't
know,
I
will
quickly
summarize
how
the
project
currently
runs.
The
formal
governance
for
the
project
is
laid
out
in
scp-01,
which
describes
how
the
project
is
set
up.
The
project
is
set
up
following
the
general
flavor
of
a
u.s
non-profit,
and
so
what
we
have
is
a
advisory
board
which
is
self-electing.
B
B
The
board
asks
people
if
anybody's
interested
and
selects
people
to
join
the
board.
The
lead
developer
is
nominated
and
elected
by
the
board.
It
has
been
me
for
the
lifetime
of
the
board.
I
think
I
think
the
board
came
after
albert's
tenure.
B
B
B
Things
and
that's
the
way
it
works.
The
only
exception
to
that
is
the
financial
situation
of
senpai.
The
project,
as
in
the
money
that
non-focus
manages
for
us,
is
still
the
remit
of
the
board
at
the
moment.
B
And
that
that's
that's,
our
current
governance
structure
doesn't
already
have
any
questions
about
that.
Have
I
done
a
reasonable
job
of
summarizing
that
I
have.
H
B
Yes,
because
non-focus
are
actually
the
us
non-profit,
so
for
the
purposes
of
being
a
u.s
non-profit,
it's
the
non-focus
board
that
exists,
and
so
and
and
the
non-focus
board
ultimately
has
authority
over
our
financial
spending
in
such
that
we
can't
breach
the
rules
for
a
u.s
non-profit
and
it's
the
non-focused
board.
That
is
tasked
with
ensuring
that
none
of
the
projects
under
non-focus
act
in
a
way
that
is
not
allowed.
H
I
B
G
B
Very
little
historically,
I
think
the
major
expenses
of
the
project
so
far
are
pretty
much
limited
to
buying
stickers
and
I
think
we
paid
for
somebody
to
go
to
the
non-focus
summit.
I
think
it
was
laura
we
paid
to
go
to
the
non-focus
summit
a
couple
of
years
ago.
I
I
think,
that's
pretty
much
it
in
terms
of
what
we've
ever
spent.
We
have
no
ongoing
running
costs,
we
make
money
through
gsoc
and
some
donations
and
it
sits
in
the
non-focused
part
waiting
to
be
required.
G
G
B
If
I
mean
if
that
was
going
to
give
us
some
concrete
benefits,
then
I'm
what
I'm
I'm
open
to
looking
into
it.
I
personally
pay
for
our
read
the
docs
and
have
done
for
many
years.
I
H
B
This
is
this,
is
this
is
worthwhile
covering
because
of
where
I
want
to
go
next.
So,
okay,
so
say
we
okay,
the
way
the
current
governance
is
currently
structured
as
in
terms
of
what
has
been
merged
in
the
sap
repo.
The
following
things
would
happen.
If
we,
the
community
on
this
call,
decided
we
wanted
to
spend
a
hundred
dollars
a
month
on
having
all
the
azure
builds,
we
could
ever
possibly
want.
B
I,
as
executive
director
in
the
saps,
would
go
to
the
board
hey
board.
We
should
spend
the
money
on
this
and
they
would
go
okay,
cool
and
then
one
of
us
would
email
non-focus
and
be
like
non-focused.
We
need
to
pay
these
people
and
they
would
go
okay,
here's
a
get
them
to
invoice
us
or
here's
a
credit
card
and
they
would
be
paid,
it
would
come
out
of
our.
It
would
come
out
of
our
pot
at
non-focus.
Non-Focus.
Give
us
regular
reports
on
how
much
money
we
have
in
said
pot.
B
B
D
B
B
The
makeup
of
the
board
tends
to
be
more
senior
people
in
solar
physics
or
with
experience
of
solar
physics,
and
the
objective
of
the
board
is
to
guide
and
advise
in
that
regard,
and
that
scp
is
the
most
the
only
real
formal
thing
the
board
has
for
that,
but
the
sap
process,
but
the
board's
involvement
in
the
sap
process
have
shifted
a
little
bit.
The
wording
has
shifted
a
little
bit
in
the
saps
over
the
years
and
it's
now
more
a
yeah
final
approval
deal
than
it.
B
B
G
B
In
terms
of
like
the
code,
development
and
management
of
the
project,
the
board
advises
the
lead
developer
in
terms
of
other
high-level
stuff
like
funding
and
involvement
with
spd
and
those
sort
of
things
the
board
has
just
done
it
like
people,
members
of
the
board
have
taken
those
projects
on.
B
I
B
Steve
has
opened
an
scp
to
move
financial
responsibility
for
the
project
into
the
remit
of
the
lead
developer,
so
that
that
would
that
would
give
me
and
anybody
I
delegated
it
to.
B
Final
decision
on
how
money
is
spent
the
board
would,
I
think
I
can't
remember
exact
wording
in
that
svp,
but
it's
basically
like
the
board
would
advise
and
consent
on
a
spending
plan
for
the
year
or
whatever
right
like
the
beginning
of
the
year.
The
lead
developer
would
go
to
the
board
and
okay.
This
year
we
think
we're
probably
going
to
spend
three
grand
on
ci
and
50
grand
on
stickers
and
the
board
would
go.
That
seems
like
a
lot
of
stickers.
B
That
that
is
a
short
term
and
practically
motivated
change.
That
probably
needs
to
happen
faster
than
it
has
done.
So
that's
that's
the
first
thing
and
there
is
an
sap
pr
open
if
somebody
could
dredge
up
the
link.
That
would
be
great,
and
I
encourage
people
to
have
a
look
at
that
if
they're
interested
in
it.
B
The
main
issues
that
I
see
personally
and
I'm
actually
no
I'm
going
to
do
this
differently.
Does
anybody
on
this
call
have
opinions
about
things
they
see
as
problems
with
our
current
governance
structure
or
things
they
would
like
to
change
with
it.
G
Completely
agree
with
the
statement
that
the
board
appears
to
exist
in
a
vacuum.
I
yeah
I
mean
I
don't
know
if
what
the
board
does
or
discusses
or
is
documented
anywhere
and
I'm
just
missing
it,
but.
H
B
D
B
B
B
Yeah,
that's
kind
of
discussed
from
a
more
community
engagement
perspective
in
some
ways
like
something
that,
relatively
recently
in
terms
of
board
2020,
has
not
been
good
for
the
sample
board.
The
things
like
instrument
packages
and
engaging
instrument
teams
that
that
has
been
like
a
topic
of
conversation.
I
In
particular,
the
instrument
team
engagement
seems
especially
given
like
the
people
who
are
on
the
board,
you
know
they're
much
more
senior
people.
There
are
people
that
are
a
lot
of
them
directly
involved
with
instruments.
You
know
people
on
those
instrument
teams
it
would
be.
It
would
be
great
to
have
them
take
more
of
a
lead
on
on
that
front,
and
I
know
that
was
sort
of
started
at
the
end
of
2019
and
yeah,
as
with
many
things
was
derailed
by
the
dumpster
fire
of.
B
G
G
B
I
agree
again
that
is,
that
has
not
always
been
the
case.
The
people
on
the
board
have
moved
away
from
many
people
on
the
board,
have
moved
away
from
being
active
developers
but
have
maintained
their
seat
on
the
board.
G
B
Minute,
well
so
it
should
be.
It
should
be
noted
that
last
board
meeting
we
put
on
hold
putting
new
people
on
the
board,
because
we
were
going
to
have
these
conversations,
and
that
has
been
true
for
pushing
a
year
at
this
point.
So,
yes,
you
are
correct,
but
rather
than
worry
about
getting
people
onto
the
board
as
it
currently
existed,
we
were
like,
let's
stop
and
think
about
what
we
want
this
to
look
like
and
then
we'll
get
people
where
we
want
the
people
to
be
so.
B
B
B
B
B
B
Rarely
use
executive
power
in
like
any
meaningful
way.
I
still
maintain,
like
the
only
time
I
ever
really
have,
is
decided
that
our
community
meetings
are
on
uk
time
and
not
us
time.
H
H
Yeah,
so
I
see
yeah
to
me
about
formulating
the
positions
it.
You
know
it's
ultimately
something
that
goes
on
the
website
and
it's
something
I
annoy
you.
If
I
need
something
that
I
can't
do
and
that's
about
it
right.
I
don't,
you
know,
call
a
lead
developer,
but
you
know
it's
not
in
the
ground
scheme
of
things:
it's
not
like
you're.
Actually
in
the
development
of
senpai,
you
just
hold
the
keys
to
senpai
and
that
that's
worked
and
it
still
works.
I
don't
know
I
don't.
H
But
seriously
give
me
admin
rights,
I
would
say
the
only
other
okay,
okay,
actually,
I
will
slightly
clarify.
I
think
there
is
one
problem
is
that
we
we
suck
at
recruitment
like
ignoring
the
board
issue
for
a
moment,
is
that
we
have
a
lot
of
you
know
open
roles
on
the
projects
that
would
allow
more
people
to
be
involved
if
they
had
time,
for
example,
the
affiliated
package
liaison
there's,
we
don't
have
a
webmaster,
we
don't
have
a
continuous,
we
don't
have
a
release
manager.
H
Who
would
do
already?
We
don't
have.
I
forget
what
the
other
positions
of
the
website
are,
but
we
have
a
lot
of
responsibilities
that
ultimately
fall
onto
primarily
stuart
that
we
could
move
offshore
if
we
had
more
people
to
to
officially
assign
them
roles
that
to
cover
aspects
of
the
project
as
a
whole.
H
B
So
drilling
down
into
I
mean
that's
a
difficult
problem
to
solve
yeah
and
there's
a
little
all
to
what
I'm
after
but.
H
I
would
argue
that
something
the
board
should
be
more
in.
You
know
in
theory,
if
a
more
active
board
so
well
not
necessarily
solved,
but
at
least
be
proactive
on.
If
there
is
going
to
be
a
board.
D
Oh
sorry,
no!
No!
No!
No!
No!
No!
Just
the
fact
that
it's
like
say
with
those
roles
right
like
saying
the
release
manager
and
the
continuous
integration
maintainer
and
they
fill
it
in
package
liaison
like
I
think,
to
have
those
real.
Maybe
this
is
just
my
concept
of
it.
I
was
like,
oh
god,
I
don't
even
know
like
when
there's
like
a
ci
problem
or
things
aren't
working,
I'm
like
I
don't
know.
What's
going
on,
I
asked
one
of
you
to
or
like
I
asked
somebody
else.
D
I
think
it's
like
limited
and
who
knows
the
inner
workings
of,
or
maybe
I'm
just
I
don't
know,
I
guess
they're
just
hard
things
to.
A
I
All
right,
oh,
I
was
just
going
to
say
yeah
for
the
especially
for
the
more
you
know,
less
technical
I
mean
like
they
can
the
ci,
maintainer
and
release
manager
more
are
more
on
the
technical
side,
maybe
the
webmaster,
but
especially
for
things
like
lead
newcomer,
develop
the
newcomer
mentor
or
the
affiliated
package
liaison.
I
mean,
I
think
those
are
things
that
I
mean.
Those
are
pro.
Those
are
kind
of
high
level
project
project
level
roles.
That's
the
kind
of
thing
that
the
board
should
be
thinking
more
about
like
it's.
I
Not
it's
not
necessarily,
it's
not
really
the
nitty-gritty
day-to-day
stuff
that
they,
but
it's
you
know
the
it's
stuff
that
affects
the
project
as
whole,
as
it
is
much
especially
in
terms
of
the
affiliated
package
liaison
is
a
much
broader
scope
than
the
core
package.
I
think
those
people
on
the
board
are
well
positioned
to
to
kind
of
think
about
those,
those
higher
level
decisions
and
maybe
they're
not.
N
B
The
board,
role
to
some
greater
or
lesser
extent,
was
always
considered
to
be
advisory
right.
The
the
I
am
not
the
the
way
the
government
is
set
up,
I'm
not
lead
developer
of
senpai
core.
I
am
the
executive
director
of
the
sunpower
project.
Recruiting
people
into
those
roles
falls
under
the
remit
of
the
league
dev
right,
because
the
board
is
supposed
to
be
advisory.
Yes,
they
may
be
able
to
advise
on
good
people
or
sources
of
places
to
look
to
recruit
people
into
higher
level
roles
that
need
better
experience
of
the
community
or
whatever.
B
I
I
would
say
that's
something
that
it's
changed
then,
okay,
so
in
my
mind.
B
Okay,
so
I
think
this
is
a
reasonably
good
point
to
bring
in
like
thoughts
of
I
I
had
I
had
so
last
year
at
the
beginning
of
last
year.
Right
at
the
point
where
all
the
covid
set
the
world
on
fire
was,
the
astropy
project
went
through
a
governance
restructure,
which
is
just
now
coming
to
fruition.
B
That
effort
was
funded
by
their
more
grant
and
motivated
by
the
fact
that,
when
they
applied
for
a
large
nsf
grant,
the
nsf
basically
turned
round
to
them,
and
when
you
are
not
an
organization,
that's
run
in
a
way.
We
can
trust
with
five
million
dollars.
B
I'm
paraphrasing
a
lot
there,
but
one
of
the
main
reasons
they
didn't
get
the
large
nsf
grant
was
because
nsf
were
like
what
are
you
can't
give
you
five
million
dollars?
We
don't
know
who
you
are
or
what
you
are
or
who's
responsible
for
anything.
B
B
B
I
was
thinking
of
making
the
lead
developer
into
a
group
of
people
instead
of
a
person,
and
I
was
also
which
me
hopefully
means
that,
like
to
some
extent,
I
don't
think
a
governance
restructure
is
ever
going
to
solve
the
lack
of
time
problems
right
like
there
is
a
serious
lack
of
time,
bootstrapping
problem
and
refactoring.
Our
governance
is
not
going
to
fix
it.
B
B
I
was
toying
with
the
idea
of
largely
leaving
the
boards
alone,
in
terms
of
it
being
self-electing
and
advisory
and
potentially
having
an
elected
committee
of
people
running
the
project.
Instead.
B
It's
fixed
or
not,
but
I
can
tell
you
what
astrophy
have
done.
Astropy
have
created
a
group
of
people
called
voting
members
who
are
active
in
the
project
and
have
shown
a
willingness
to
contribute
to
the
project
and
like
move
the
project
forward.
Basically,
I
should
be
able
to
remember
the
list
of
criteria,
but
I
can't
those
voting
members
vote
new
members
into
the
voting
members,
so
they're,
a
self-governing
self-growing
organization
of
voting
members
and
those
voting
members
elect
the
coordinating
committee
that
run
the
project.
B
B
Now
you
said
that
that's
interesting,
I
mean
if
we
were
to
okay-
let's,
let's
put
this
back
this
up
again.
If
we
were
to
just
ape
the
astrophy
governance
structure,
pun
intended
like
we
do
many
other
things.
B
The
people
in
roles
and
the
people
who
commit
regularly
to
senpai
would
become
the
voting
members
and
the
lead
developer
would
become
the
coordinating
committee
and
the
board
does
not
have
an
equivalent
in
after
pi,
but
in
this
situation
I
would
envisage
the
board
becoming
entirely
advisory
and
guiding
and
helping
the
development
committee
as
they
need
it.
Development
committee
is
a
bad
idea
because
it
needs
to
be
more
than
developer.
It
needs
to
be
project,
as
we
have
already.
B
D
D
B
D
B
B
We've
brought
in
a
few
new
people,
but
that's
what
it
is.
But,
yes,
I
I
feel
like
I
I
feel
like
the
people
who
do
the
work
are
not
represented
in
the
governance
structure
anymore.
A
B
But
not
really,
most
of
us
people
heavily
involved
in
astropyth
and
the
governments
of
astrapai,
certainly
when
the
project
was
set
up,
were
quite
junior
and
are
therefore
not
as
junior
anymore,
but
still
relatively
speaking,
and
I
think,
having
a
board,
which
is
a
large
and
b,
doesn't
have
to
make
many
decisions,
because
having
a
large
group
of
people
that
have
to
make
lots
of
decisions
is
always
a
recipe
for
a
disaster.
B
And
to
be,
I
mean,
obviously
I'm
going
to
take
this
back,
so
I
asked
I
asked
a
while
ago
for
people
who
are
interested
in
talking
about
this
stuff
and
haven't
been
able
to
do
much
with
it,
but
yeah
a
few
board.
Members
are
interested
in
working
on
making
this
problem
better,
and
I
am
not
necessarily
aiming
in
fact,
when
I
say
not
necessarily,
I
am
literally
not
aiming
to
come
up
with
a
new
governance
structure
out
of
this
discussion.
D
I
guess
it's
like
an
important
thing
to
think
about,
like
just
from
masterclass
experience
of
nsf.
Coming
back
to
them
and
saying
you
know
you
don't
have
a
proper
structure.
Who
are
you?
It's
not.
Everybody
should
think
about
a
senpai
girls
right
because
down
the
line,
the
one
to
be
able
to
apply
for
big
grants
and
not
have
that
problem
right
yeah.
So
you
think
about
yeah.
K
B
Anyway,
thank
you
for
your
thoughts
and
contributions.
Anybody
have
anything
else
to
add
in
terms
of
goals
for
running
the
project
or
how
the
project
is
run.
I
guess.
H
Yep,
that
is
the
this
is
the
final
session.
So
again,
I
want
to
thank
everyone
for
showing
up
and
not
everyone
who
could
be
here
on
the
final
day
and
to
thank
them
as
well.
I
will
be
pinging
the
room
with
a
really
specific
questionnaire
at
the
end,
so
people
would
like
to
have
a
look
at
that
folder
and
if
they,
if
they
want
to
it's
anonymous,
I
won't
I
won't
be
collecting
your
your
identity
and
tracking
you
down
later.
You
don't
have
to
worry
about
that.
B
H
H
Available,
hopefully
others
will
be,
I
will
be
around
as
well.
I
will
be
probably
working
on
a
few
things.
C
B
G
I
put
down
just
go
through
and
copy
edit
the
database
guide
because
I
remember
going
through
it
and
it
being.
B
That's
fire,
if
you
is
that
another
patch.
B
H
B
Me
right
access
to
all
repos
everywhere
and
we
were.
H
G
C
J
I
Like
hours
like
co-working
hours,
like
same
as
from
the
meeting,
it's
been
all
week,.
B
B
On
friday,
on
a
friday,
but
presumably
people
will
yeah,
we
can
have
a
bit
of
overlap
and
stuff
coordinate
in
the
coordination
meeting
room.
B
If
I
mean
I
think
we're
really
down
to
the
point
where
I'm
addressing
this
directly
at
matt
but
matt.
If
you
want
to
have
somebody
to
chat
to
and
do
a
senpai
contribution
tomorrow,
I'm
very
happy
to
sit
with
you
on
the
call
and
like
help
you
through
that
process.
If
that's
interesting
to
you,
yeah.
F
Yeah,
thank
you
I'll
try
and
have
a
think
about
something
kind
of
constructive.
I
might
just
drop
in
to
see
what
you
guys
are
up
to
in
any
case.
B
F
B
F
Sure,
yeah
absolutely
yeah.
I
I'm
not
sure
if
we're
kind
of
like
mature
enough,
yes
in
our
kind
of
programming
cycle
for
me,
to
be
able
to
ask
ask
for
suggestions
on
coding
but
I'll
enjoy,
seeing
how
you
guys
are
actually
working
there,
and
one
thing
that
I
have
done
is
I've
been
asked
next
week,
there's
a
flow
tracking
meeting
for
punch,
and
I
think
solar
orbiter
and
I
was
asked
to
present
the
status
of
the
punch
data
reduction
pipeline.
F
F
But
let's
leave
it
fluid,
let's
see
what
see
what
happens
tomorrow.
A
D
B
D
All
right
both-
I
guess
it's
like
seeing
if
it
is
going
to
really
slow
down
things.
This
is
dave
I
was
saying
yesterday
yeah,
rather
than
actually
pulling
her
fully
like
implementing
it,
and
I
think
girls
talks
about
making
some
blocks
remember
like
introductory
coordinates,
rather
than
like
codes,
specific
stuff.
B
D
I'm
actually
giving
a
tutorial
tomorrow
for
the
classified
hack
week
is
on
this
week.
It's
a
nice
tutorial,
so
that
should
be
good
interesting
to.