►
From YouTube: SunPy Coordination Meeting 2021 - Monday
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).
C
A
C
Well,
I'll,
just
I'll
just
open
up!
Sorry,
okay,
thank
you!
Everyone
for
showing
up
it's
been
a.
I
should
really
plug
in
my
webcam,
but
it's
not
gonna
happen.
I
wanna
thank
everyone
for
coming
and
hopefully
will
be
the
start
of
a
productive
couple
of
days
of
shorter
sessions
than
our
last
one.
So
today
we're
gonna
just
be
recapping
what
happened
last
time,
the
state
of
senpai
and
then
a
few
updates,
we'll
be
done
in
about
three
hours,
give
or
take
three
hours
plot.
The
math
right,
obviously
still
would
be.
C
I
guess
starting
off
with
actually
not.
Actually
me
first
mistake:
yeah,
I
don't
know
I'll
stop
rambling!
You
want
us,
you
want
to
say
anything.
A
C
So
do
we
want
to
do
the
awkward
thing
of
introductions
first,
or
do
we
just
want
to
assume?
We
all
know.
C
A
C
A
And
now
you're,
just
the
senpai
projects,
the
bill
yeah.
I
I'm
stuart
mumford,
I'm
the
developer
and
easily
distracted.
A
I'm
gonna
stop
picking
on
people
or
do
you
wanna
start
picking
on
people
in
a
bill.
I
don't
want
to
pick
on
people,
but
all
right
we're
going
to
do
this
in
order.
That
is
on
my
screen.
E
A
F
G
Hello,
I'm
david
s
if
any
other
david's
turn
up,
so
I've
been
with
some
pi
for
a
few
years
now,
I'm
at
ucl
mssl
and
do
a
lot
of
solar
or
test
stuff
day-to-day.
So
I
guess
that's
my
main
interest
in
subpac.
A
H
B
Hi,
I'm
will
I'm
the
deputy
lead
developer
for
the
project.
I'm
a
postdoc
here
at
the
naval
research
lab
in
dc
the
workouts
on
five,
for
I
guess
four
or
five
more
many
more
years
now,
but.
A
I
Hiya,
can
you
hear
me
yep
fantastic
yeah,
hi,
I'm
matt
west.
I
work
at
sweary
and
I'm
co-building
the
punch,
soc
data
root
data
reduction
pipeline
which
we're
building
primarily
in
python.
I
did
drop
into
one
of
these
meetings,
probably
about
half
a
year
ago
now
with
the
intention
to
come
into
them
regularly.
But
sadly
I
had
a
another
meeting
scheduled
at
the
same
time,
which
yeah
sure
happens
anyway.
I
We
kind
of
want
to
see
what's
happening
with
the
senpai
community,
because
we're
going
to
be
relying
on
senpai
and
ultimately
we
want
the
punch
data
reduction
pipeline
to
be
a
senpai
associated
package.
So
that's
my
interest,
that's
why
I'm
here
I
keep
getting
funny
low
signal
thing
magic,
so
I
may
be
turning
my
camera
off
periodically.
J
Hello,
I'm
monica
bobra,
I'm
at
stanford
university
and
vice
chair
of
the
board.
A
D
Thank
you
yeah,
so
I'm
danny
ryan,
who
can't
unmute
normally
can't
stop
talking.
D
So
thank
god
for
you
buttons,
but
yeah
anyway,
I'm
currently
at
fhnw
or
the
university
of
applied
sciences
in
switzerland.
I've
been
with
senpai
for
many
years
now,
and
I'm
sort
of
being
involved
in
starting
up
a
number
of
affiliated
packages
from
ndcube
which
we'll
talk
about
tomorrow.
I
think
to
sun,
expects
for
x-ray
spectroscopy
and.
D
K
Sure
yeah,
my
name
is
shane,
I'm
from
the
dublin
institute
for
advanced
studies
where
sophie
is,
and
I'm
also
in
dublin
and
yeah
kind
of
a
fringe
swim
type
person.
I
use
a
lot
day
to
day
I
like
to
complain
about
things
when
they
break
and
don't
work.
The
way
I
want.
A
Congratulations:
diaz
beats
goddard
for
representation
in
this
meeting
at
long
last
albert.
L
Yeah
well,
my
car
just
got
done
so
I'm
out
of
here
hi
albert
c:
goddard
coordinates
maintainer
and
curmudgeon.
At
times
I
don't
know
what
else
to
say:
hi
everyone.
M
I'm
chris
I'm
working
at
series
in
boulder,
mostly
with
the
suvi
instrument
and
I'm
here
basically
out
of
interest
and
because
I'm
trying
to
develop
either
a
subie
pie
or
basically
a
sun
kit
instrument
affiliated
package
for
subi.
So
that's
pretty
much
it
cool.
A
A
A
C
A
C
C
A
A
C
C
We
are
looking
at
three
days.
C
We
were
looking
at
three
days,
but
there
were
eight.
There
were
full
like
four
working
days.
The
idea
was
obviously
going
to
be
going
to
be
in
person
and
then
obviously
crave.
It
happened,
and
so
we
had
like
very
intense
days.
I
don't
know
if
people
who
were
there
last
year
remember,
but
so
the
first
day
was
essentially
bringing
up
everyone
on
to
speed
about
the
state
of
the
sun
by
project,
and
that
means
not
only
the
library
itself
with
the
affiliate
packages,
obviously
really
introduction,
and
you
know
we
went
through.
C
C
You
know,
maybe
in
the
nature
that
you
know,
served
a
in-person
meeting
but
not
necessarily
worked
on
a
zoom
call,
and
then
we
ran
it
off
today
with,
as
danny
was
mentioning
about
the
spectral
stuff
we
had.
We
talked
about
instrumental
trying
to
bring
on
instrumental
teams
into
the
sun,
powerfully
a
package
ecosystem
also
talking
about
the
future
of
some
raster
and
indie
cube
and
using
that
for
like
ice
data
and
that
kind
of
thing,
whereas
the
second
day
was
wrapping
up.
C
So
this
meeting
was
in
the
context
of
ramping
up
for
2.0,
which
is
a
memory
ago
now,
and
we
were
obviously
had
come
off
the
first
main
release
of
somebody
1.0
where
we
defined
a
bunch
of
you
know,
we
tried
to
fix
our
fix
our
api
down
I'll,
lock
it
down
sorry,
and
we
were
looking
at
2.0
going.
Okay.
What
is
what
is
actually
what?
What
do
we
actually
want?
Somebody
to
be
what
should
be
in
the
core
package?
C
What
do
we
want?
How
do
we
like
reconcile
nd,
cube
and
map
and
obviously
we're
going
to
repeat
that
topic
this
year,
because
ndq
obviously
2.0
was,
I
guess,
a
dream
back
then?
Would
that
be
accurate
to
say.
D
I
think
it's,
I
think
it's
fair
to
say
that
the
last
last
time
there
were
a
lot
more
obstacles
and
it
was.
There
was
more
aspiration
only
because
there
were
more
obstacles
to
overcome
and
it
was
further
away.
I
think
it
is
definitely
fair
to
say
we're
a
lot
closer
now,
even
if
it
has
been
a
bit
slow.
C
And
so
we
also
talked
about
how
to
organize
senpai
as
a
project
going
forwards,
steve,
who,
unfortunately
can't
be
here,
you
know,
created
the
road
map
repo
that
probably
should
be
deleted.
C
But
we
wanted
to
try
and
encapsulate
a
future
vision
of
senpai
that
would
allow
people
like
people
to
come
in
and
go
okay.
Maybe
this
might
be
interesting
to
work
on
et
cetera,
like
that,
and
you
know
probably
didn't
pan
out
that,
and
we
spent
some
time
working
on
the
vision
statement,
and
I
don't
know
if
anyone
remembers
if
we
put
that
anywhere.
C
And
I
think
this
is
the
day
we
also
worked
out.
We
hashed
out
the
affiliated
package.
So,
as
some
others
have
mentioned,
we
have
an
affiliate
package
system
that
allows
either
people
to
be
sponsored
directly
by
the
senpai
project
or
if
they
want
to
be
external
from
the
sample
project
and
they
have
control.
But
you
know
get,
I
guess,
get
the
branding
like:
hey,
hey,
pi.
A
Think
we
finalized
that
pi
astro
or
was
that
the
other
way
around?
I
can't
remember
what
order
those
two.
A
C
And
so
that
was
you
know
that
was,
I
guess,
one
of
the
big
outcomes
of
that
meeting
was
that
was
a
lot.
It
was
a
lot
of
the
work
around
that
and
then
the
final
day
was
a
lot
of
infrastructure
tooling
talking
about
how
we
actually
maintain
the
projects.
C
There
was
the
obviously
the
very
interesting
results
from
the
comp.
The
survey
that
I
think
we
want
to
keep
doing
every
so
often
now
I
don't
know
if
there
are
plans
to
repeat
that.
A
A
Thing
that
I
saw
that
came
out
of
this
survey,
this
is
just
slightly
off
topic
quickly.
Was
the
ssw
idl
talk
double
a
s
spd
a
few
weeks
ago
quoted
the
statistics
from
the
senpai
survey,
so
that
made
me
happy.
C
C
A
C
I
think
we
wrapped
off
the
day
talking
about
well,
I
guess
our
biggest
problem
as
a
project
is
a
community
trying
to
bring
on
a
larger
community
around
the
projects.
C
So
I
think
I
think
overall
we
had
we
had.
We
had
a
fun
time,
I
remember,
being
very
tired
by
the
end
of
it.
Eight
hours
of
zoom
causes
surprisingly
tiring
in
a
different
way
than
I
think,
but
I
mean
oh
wrong
tab,
so
I
have
a
bunch
of.
I
have
a
bunch
of
takeaways
that
when
I
look
back
on.
C
And
obviously,
if
people
want
to
bring
up
their
own
that
go
ahead,
so
obviously
it
was
the
first
time
we've
done
this
and
with
the
time
you
always
do
a
first
one
event
of
anything.
It's
all,
there's
always
a
learning
curve,
and
you,
like
you,
learn
on
the
job.
Essentially.
C
And
obviously
the
big
thing
was
that
kovitz
changed
it
from
an
in-person
meeting
and
obviously,
when
the
everything
was
planned,
the
schedule
was
planned.
Everything
was
scoped
out
in
an
in-person
meeting.
Everyone
was
there,
you
know
you
had
the
ability
to
break
out
into
you,
know
small
rooms
and
things
and
just
iterate
on
like
talks
and
work
in
a
way
that
you
wouldn't
be
able
to
do
in
a
in
an
online
call.
C
It
was
a
lot
more
informational.
So,
as
I
I'm
specifically
linked
like
some
of
my
talks
was.
C
That
the
example
I
talked
about
a
lot
on
the
wednesday
about
our
continuous
integration
and
our
project
template,
and
it
was
a
lot
more
informational
than
you
might
argue
that
these
talks
should
have
been.
Is
that
one
of
the?
I
guess,
a
lot
of
people
you
bring
you
bring
everyone
in
and
what
you
want
is
things
to
be
iterated
on
and
something
to
be
worked
out.
I
guess
you
know
it's
not
often
we
do
this.
You
know
we
have
our
weekly
community
calls,
but
there's
a
handful
of
people
and
we
kind
of
decide
whatever.
C
But
the
point
about
bringing
everyone
in
is
that
we
can
make
like
decisions
and
try
and
steer
things
in
a
way
that
we
can't
on
a
weekly
basis
for
the
project,
and
so
I
guess,
like
things
like
my
talk,
were
just
like
informational
for
the
most
part
and
probably
wasn't
the
most
useful
necessarily
for
these
for
this
style
of
meeting.
C
I
guess
the
I
guess
the
biggest
one
was
we
nev.
We
we
had
a
lot.
We
had
a
lot
of
arguments
about
I'm
gonna,
say
everything,
but
we
had
a
lot
of
arguments
about
say
how
to
do
off
the
shipping
code,
maybe
how
we
want
to
do
documentation,
how
we
want
to
do
package
templating
and
a
lot
of
this.
You
know
I've.
I've
reread.
I
reread
all
the
you.
N
C
You
know
we
might
have
talked
about
x,
y
and
z,
but
you
know
no,
no
consensus
was
brought.
No
consensus
was
written
down
and
much
like
the
vision
statement
from
last
year.
I
forgot
about
it
until
I
was
looking
at
the
I
started.
Looking
for
the
notes,
oh,
where
what
happened
to
that
and
things
like
that,
I
guess
I
want
to
try
and
hope
that
we
can
formulate
something
by
the
end
of
each
session.
That
allows
someone
in
say
two
months
to
work
on
senpai
or
something
related
to
somebody.
C
A
A
The
affiliated
package
thing
being
the
biggest
one
and
the
sunkit
instruments
coming
into
existence
came
out
of
the
coordination
meeting
last
year
and
those
two
are
just
the
two
that
I
can
think
of
off
the
top
of
my
head,
and
I
feel
like
a
lot
of
the
infrastructure
conversations
we
had
last
year,
fed
into
the
nasa
grant
proposal
we
put
in
whenever
that
was
so.
B
Jack
also
made
a
good
point
in
the
chat
that
I
mean
you
know,
since,
since
our
coordination
meeting
last
year
that
the
world
has
changed
pretty
dramatically,
I
think
that
that's
obviously
like
you
know,
distracted
a
lot
of
people
in
various
ways
and
especially
career-wise.
I
think
that's
made
it
hard
to
kind
of
you
know,
take
action
on
a
lot
of
the
things
that
came
out
of
that
meeting.
A
Okay,
well,
I
have
a
few
slides
on.
A
I
have
a
few
slides
on
what
happened
in
the
last
year,
see
whether
this
works.
Can
you
see
this
okay
yep.
J
A
Do
you
mean
it's
20
21
already?
I
did
the
obligatory
run
of
statistics
through
the
core
repo
with
some
rudimentary
filtering.
We
have
492,
not
backboard
prs
merged
in
the
last
12
months
and
21
different
contributors
over
a
thousand
non-merge
commits
we've
closed
nearly
300
issues
and
opened
less.
We
are
minus
40
issues
in
the
last
12
months,
which
made
me
happy
and
of
the
258
that
we've
opened
in
the
last
12
months.
We've
also
closed
167
of
them.
A
A
A
A
A
Picking
through
the
change
log
and
release
notes
the
major
things
that
changed,
not
in
order
of
majorness,
I
think
probably
the
biggest
single
change
we
made
was
in
2-1,
where
we
added
metadata
only
support
to
fido
and
massively
refactored
unified
response
to
be
table
responses,
and
it's
certainly
the
most
disruptive
change.
But
I
think
it
was
also
one
of
the
better
ones
in
terms
of
like
user
experience
and
flexibility
of
how
fido
works.
A
The
coordinate,
aware,
plotting
stuff,
including
image,
alignment
and
contours,
and
everything
that
made
it
into
3-0
at
the
last
minute,
was
a
pretty
major
jump
forward
for
that
as
well
change
tracking
term
dot.
Meta
on
map
is
a
very
nice
and
relatively
simple
change
with
nice,
like
the
small
amount
of
code,
large
impact
change
to
be
able
to
see
what
map
and
you
have
changed
in
the
meta
since
you
loaded
it
from
the
file.
A
Jsoc
cutout
support
is
nice
as
long
as
json
don't
break
their
api,
which
has
only
happened
once
so
could
be
worse
and
the
all-important
assume
spherical
screen
in
coordinates
for
when
you
want
to
assume
that
you're
looking
out
onto
the
inside
of
sphere.
A
Oh
this,
that's
too
long,
apparently
as
a
list
of
our
sponsored
packages
from
the
website.
As
we've
already
mentioned,
ndcube
is
rapidly
approaching
2-0,
unlike
core
and
d-cube,
is
still
stubbornly
sticking
to
a
semantic
versioning
scheme.
So
2-0
is
a
going
to
be
a
ap
major
api
break.
It's
at
this
point
rapidly
a
complete
rebuild
from
the
ground
up
we've
pretty
much
thrown
away
most
things
and
made
it
work
with
astropy,
814,
wcs's
and
rebuilt
the
api.
So
there's
an
open
scp.
A
Wait:
I've
put
I've,
put
the
sun
raster
thing
and
I've
put
sunk
it
in
this
is
supposed
to
be
the
sun
raster
thing.
I've
got
my
things
all
in
the
wrong
order.
Sun
raster
got
solar,
orbiter
spikes
by
support
and
then
has
gone
a
bit
quiet
sunken
image
had
one
pr
from
will
and
not
a
huge
amount
else
in
the
way
of
features.
I
think.
E
A
And
sunk,
it
instruments
came
into
existence
in
the
last
12
months.
Shane
has
been
working
on
radio
spectra
because
I've
been
seeing
the
github
notifications
not
sure
about
the
details
of
what's
been
going
on
there.
We
don't
have
a
session
where
we're
going
through
this
right.
C
A
I
was
just
won't
go
into
too
much
detail,
then
pi
f
of
ct
exists,
the
hard
part
of
that
is
done,
which
was
wrapping
the
c
code
and
making
the
c
code
compile,
and
so,
if
anybody
wants
to
do
velocity
fields
from
images
and
wants
to
write
the
nice
high
level
bits
to
make
that
return,
cubes
and
be
nicer,
the
difficult
bit
is
done
and
last,
but
by
no
means
least,
the
a
blog
package
exists
and,
dear
god,
how
is
that
so
popular?
We
get.
N
A
Issues
and
contributions
to
a
blog,
which
is
our
sphynx
plugin,
for
which
is
our
sphinx
plugin
for
blogging,
with
sphinx,
which
we
use
on
our
somepipe.org
website
and
nabil
soldiers
on
looking
after
that
and
dealing
with
the
contributions
and
the
bug
fixes
like
like
a
hero.
Well,.
A
Hey
blog
is
there
for
you
if
you,
if
you
are
a
little
bit
crazy
and
want
to
use
sphinx.
A
Anybody
got
anything
else,
they
think
has
happened
in
the
last
12
months.
That's
worthy
of
calling
out
that
I've
forgotten,
because
I
feel
like
one
of
the
things
that's
happened
in
the
last
12
months.
Is
I've
been
paying
less
attention
than
I
have
in
previous
12
months?.
C
So
I
guess
I
would
want
to
bring
up
slightly
that.
Obviously,
we've
been
running
just
every
year,
the
google
summer
of
client
people
who
are
unaware
where
we
get
google
to
pay
students
to
write
code
for
us
and
lastly,
a
lot
of
the
network
for
fido
was
done
through
that
project,
and
this
year
we
are
working
on
two
separate
projects.
One
for
nd
cube,
I
think
on
re-sampling
phone
reproduction.
C
Yeah,
you
said
reproductive
reproductive
symbol
and
the
other
one
is
working
on
with
david's
3d
plotting
for
pi
vista.
So
you
can
plot
your
images
of
the
sun
with
field
lines
in
three
dimensions,
and
the
other
thing
I
want
to
bring
up
is
that
we
also
ran
a
outreachy
project
on
benchmarking.
A
C
C
A
In
the
face,
because
the
student
who
was
working
on
it
caught
covered
halfway
through
the
project
and
basically
wiped
her
out
so
and.
A
B
B
All
right:
can
everyone
still
see
that
yep?
Yes,
okay,
all
right,
so
I'm
going
to
do
a
quick
overview
of
the
sub
sub
packages
and
what
they,
what
the
the
current
state
of
them.
So
this
will
probably
be
honestly
less
than
less
than
15
minutes,
but
feel
free
to
jump
in.
Interrupt
me
at
any
point:
okay,
so
here
here's
our
list
of
sub
packages
and
our
the
kind
of
current
health
status
of
them.
This
is
on
this
took
me
a
while
to
find
on
on
the
website.
This
is
I
can't
now.
B
B
Very
you
know,
stable
things
like
map
net
time
time
series,
the
things
that
we
rely
on
the
things
that
I
think
the
things
that
people
probably
use
the
most
out
of
the
package.
You
know
reasonably
reasonably
stable,
doing
doing
pretty
well
things
like.
I
o
data
stuff.
That's
like
tends
to
be
less
user-facing.
You
know
pretty
mature,
not
not
a
lot
going
on
there.
So
what
I'm
going
to
try
to
do
is
kind
of
step
through
the
major.
B
The
major
components
here
coordinates
net
map,
time
series
and
then
kind
of
point
out
some
some
additions
or
changes
in
the
past
year
since
the
last
coordination
meeting
and
and
kind
of
talk
about
those
and
I'll
summarize
some
of
the
other
packages
as
well
so
stuart
touched
on
this
a
little
bit
too.
I
just
wanted
to
point
out
that
we've
had
three
sub
packages
that
have
been
removed
and
or
deprecated
and
will
be
removed
in
3.1,
so
the
big
one
obviously
is
instra,
that
is
now
sun
kit
instruments.
B
It's
been
deprecated
in
version
3
and
will
be
removed
at
3.1.
The
beel
had
a
big
pr
that
removed
all
of
the
all
the
instra
code.
So
all
this
lives
in
some
kit
instruments.
Now,
with
the
idea
being
that
for
people
who
weren't
kind
of
part
of
that
discussion
last
year
that
the
instrument
specific
functionality,
that
sort
of
can't
justify
its
own
package
necessarily
or
maybe
the
you
know
things
that
an
instrument
team
doesn't
necessarily
want
to
maintain
themselves
that
stuff
goes
into
some
kit
instruments.
B
So
a
lot
of
those
goes
routines
that
those
or
the
the
fermi
gbm
stuff
that
is
now
in
in
some
instruments-
aia
prep,
which
was
heavily
used
function
in
in
probably
the
most
used
function.
I
would
guess
in
the
instra
sub
package
is
now
part
of
aia
pi
and
has
actually
been
removed,
was
removed
in
version
2.1
because
it
was
depleted
earlier.
So
now
this
lives
under
the
a
pi
calibrate
dot
register
essentially
provides
the
same
the
exact
same
functionality,
it's
it's.
B
It's
the
same
function
more
or
less,
and
we're
hoping
that
you
know
this.
This
kind
of
gives,
in
addition
to
removing
it
from
core
and
narrowing
the
scope,
accord,
the
idea
being
that
you
know
this
provides
a
way
for,
for
people
to
kind
of
dump,
now
they're
more
instrument
specific
code.
It
hopefully
makes
that
process
easier
than
having
to
to
go
the
route
of
putting
it
into
corn.
B
The
roi
sub
package
has
been
deprecated
and
also
will
be
removed
in
version
3.1.
I
have
never
used
this,
so
I
can't
really
say
exactly
what
the
utility
of
this
is
here.
I
don't
think
it's
been
used
by
many
people
in
a
long
time,
but
the
astrophy
regions
affiliated
package
is
one
possible
alternative
to
this.
I
think
the
the
most
useful
bit
of
functionality
there
was
the
chain
code.
That's
now
in
that's
now
a
net
under
under
the
under
helio.
B
Lastly,
and
probably
the
most
minor
change
is
that
the
the
color
maps,
the
cm
sub
package,
is
now
part
of
the
visualizat,
the
visualize
sub
package
under
visualize
color
maps
and
was
removed
in
version
2.1.
So
so,
since
our
last
coordination
meeting,
we've
managed
to
deprecate
and
or
remove
three
of
these
of
these
sub
packages.
D
The
other
one
thing
I'd
add
about
the
sunkit
instruments
package
is,
I
would
be
excited
if
at
some
point
in
the
future,
it
actually
acts
as
an
incubator.
So
I
think
once
things
go
in
there,
if
it
matures
to
a
point
where
it
now
does
merit
its
own
package
with
somebody
or
a
group
of
people
willing
to
maintain
it,
then
you
know
things
should
be
able
to
be
moved
out
of
that.
D
B
I
agree
yeah.
That
would
also
be
great.
I
don't
think
we've
really
seen.
I
could
be
wrong
when
nabil
can
correct
me
because
he's
been
the
one
babysitting
that
one
the
most,
but
I
don't
know
that
we've
seen
any
contributions
to
that
package
at
all.
C
C
So
we
I
forget
his
name,
there
was
a,
I
think,
a
swiss
guy
who
was
adding
functionality
for.
O
B
E
B
Yeah
and
maybe
because
that's
a
much
lower
barrier,
a
much
lower
hurdle
to
get
over
for
people
than
like
the
affiliated
package
to
say
like
okay,
you
have
a
function
or
a
few
functions
that
you
want
to
add.
Rather
than
that's,
I
mean
it's
much
easier
to
put
pr
into
an
existing
repo
than
it
is
to
say.
B
Okay,
start
your
own
affiliate
package,
you
know
for
your
particular
instrument
or
whatever,
so
maybe
we
need
to
be
pushing
that
with
along
when
we
do
more
of
our
affiliated
package
advertisement
too,
to
say
you
know
if
you
have
kind
of
a
smaller
sub
subset
of
functionality,
here's
an
easier
way
to
at
least
get
started
like
like.
Like
did.
He
was
saying,
like
the
incubator.
A
B
B
Yeah,
I
think,
that's
probably
a
discussion
we
could
have
that
the
affiliated
package
stuff
on
what
is
that
that's
wednesday
is
about
you
know
what
can
we
do
better
to
kind
of
attract
specifically
instrument
instruments
of
packages,
and
maybe
there's
people
on
this
on
this
in
this
meeting?
That
could
give
us
some
good
feedback
about
what
how
we
could
do
a
better
job
of
that.
C
C
B
All
right,
so
I'm
gonna
go
through
just
very
briefly,
the
kind
of
for
our
four
main
sub
packages
and
just
kind
of
advertise.
The
sort
of
couple
new
things
about
them
so
and-
and
albert
can
interrupt
me
if
I'm
forgetting
anything
or
or
I'm
saying
anything
wrong
here,
so
in
coordinates
at
least
the
ones
that
make
nice
pictures
are
rotated,
sun
frame,
which
came
out
in
version
two,
which
essentially
provides
a
meta
coordinate
frame
for
using
reproject
to
do
differential
rotation.
B
So
these
are
the
the
differentially
rotated,
heliographic
coordinates
or
lines
of
latitude
here
that
I'm
showing
on
the
left-
and
this
is
this-
is
from
the
gallery.
Both
these
are
from
the
gallery.
There's
also
the
assumed
spherical
screen,
which
stewart
briefly
mentioned.
This
is
also
from
a
gallery
example
for
pretending
that
your
the
sky
is
is
a
spherical
screen,
some
kind
of
more
lower
level
details.
B
The
carrington
and
stony
hill
steeler
graphic
frames
now
have
a
common
commonly
inherit
from
a
base,
healer
graphic
class,
which
changed
some
details
about
how
how
they're
instantiated
and
that
carrington
is
now
an
observer,
dependent
frame
and
so
there's
a
lot
more
kind
of
lower
level
changes
too
in
the
coordinate
stack
and
that
the
change
log.
Our
change
log
details
all
those.
But
these
are
just
some
of
the
big.
L
Ones
so
I'll
point
out
two
things:
one:
that
white
band
on
the
right
hand
side
on
that
plot
is
now
fixed.
Now
that
I
understand
what
is
causing
it,
and
so
I
actually
don't
know
which
release
it's
in
or
whether
it's
still
just
on
master,
but
it
is
now
fixed
and
the
other
thing
I'll
point
out
that
it's
not
shown
up
on
this
slide
at
least
there's
a
velocity
support
in
coordinates.
L
So
now,
there's
velocity
support
formally
and
the
informal
support
that
was
hidden
before
was
giving
the
wrong
answer
anyway.
So
hopefully
nobody
was
relying
on
it,
but
velocities
are
now
supported,
but
they're
a
little
tricky
to
work
with
the
astrophy
framework
is
not
the
most
friendly
for
that.
So
you
kind
of
have
to
know
what
you're
doing,
but
it's
supportive.
If
you
want
to
work
with
velocities.
L
L
Other
comment
just
just
jump
into
the
rotated
sun
frame.
I
think
you
mentioned
that
it's
useful
for
reproject,
which
it
absolutely
is,
but
especially
with
the
coordinate,
aware,
plotting
support
now
that
we
have
on
map,
you
don't
even
have
to
call
reaper
checked,
although
it
is
generally
a
performance
hit
to
do
it.
L
The
other
way
reprojectors
is
faster
all
around,
but
some
of
the
stuff
that
the
people
are
paying
attention
on
the
repo
or
or
some
other
stuff
that
hasn't
yet
been
immersed
in,
but
you
can
not
even
call
reproject
and
get
a
lot
of
the
different
rotation
and
other
things
happening.
L
B
That's
it
all
right
for
for
nets.
You
know
I
I
realize
that
I'm
not
watching
the
chat,
because
I'm
in
full
screen.
So
if
there's
stuff
coming
up
in
the
chat
someone
rolled
into
this
someone
just
please
shout
at
me.
A
Speaking
of
the
chat
quickly,
we
have
at
least
a
few
people
watching
on
the
stream.
So
if
you
can
use
the
matrix
channel
instead
of
the
built-in
jitsi
chat,
so
that
people
who
are
watching
the
youtube
can
see
the
questions
as
well,
that
would
be
great.
But
if
it's
just
quick
chat,
then
whatever
I'm
not
going
to
worry
too
much.
B
Okay
for
for
net
stewart
mentioned
a
few
of
these,
the
one
of
the
big
ones
in
in
2.0
that
the
the
deal
put
a
ton
of
effort
into
was
this
search
attribute
discovery.
So
you
can
do
things
like
either
autocomplete
on
on
the
on
attributes
to
see
what
options
you
have
so
an
example.
I
could
do
instrument
and
then
dot,
and
if
you're,
in
a
notebook
or
in
an
ipython
shell,
it
will
tap
complete
to
to
the
available
instruments
that
you
can
search
with
fido.
B
You
can
also
print
out
this.
This
also
works,
if
you,
if
you
are
string,
representations
for
for
a
lot
of
these
as
well
or
if
you
do
things
like
you
know,
printing
senpai.net.fido,
it
tells
you
the
available
clients.
B
So
details
like
this
that
have
made
sort
of
discovering
what
data
is
available
through
fido
much
much
easier,
and
so
that
this
works
on
kind
of
any
any
case
where
you
have
sort
of
a
sort
of
finite
set
of
possibilities.
So
this
like
for
fizzobbs
or
for
instruments.
This
is
this
is
really
nice.
It
makes
makes
searching
for
data
much
much
easier,
much
more
user
friendly.
B
The
other
one
that
serve
mentioned
was
the
metadata
search
and
the
kind
of
unifying
the
the
api
for
for
data
and
metadata
searches.
So
you
can
do
things
like
combining
an
hek
and
in
this
example
this
is
from
the
gallery
a
a
json
query.
So,
for
example,
like
you
know,
if
you
wanted
to
to
look
for
observations
of
flare,
observations
that
were
coincident
with
certain
agk
features,
if
you
wanted
to
have
the
magnetograms
for
the
corresponding
magnetogram
observations
for
for
those
results,
you
can
now
do
that
with
through
this.
B
This-
and
this
was
a
gsoc
project
last
year,
included
in
that
project
was
also
a
huge
refactor
of
generic
client,
which
also
included
a
lot
of
change
to
a
lot
of
changes
to
scraper
and
then,
as
cert
mentioned
now,
we
have
a
an
api
to
the
jsoc
cutout
service,
so
you
can
request
cutouts
through
through
fido
there's
another
there's
now
a
cutout
attribute.
B
Map
so
last
year
during
the
court
like
during
the
coordination
meeting
albert
and
I
think,
almost
fully
implemented,
all
the
the
interactive
quick
look
stuff
that
we
have
for
maps.
This
is,
if
you
print
this,
I
guess
this
works
in
a
notebook
by
just
pretty
straight
to
the
notebook.
It
also
works
in
somehow
in
a
shell,
I
believe
is
that
is
that
correct.
B
Something
yeah
it,
and
so
you
have
these
these
nice
interactive
or
you
can.
It
gives
you
a
sort
of
a
better
quick
look
on
the
data
than
something
like
peak
say
you
get
all
the
all
of
the
your
metadata
information.
You
get
a
histogram
of
your
accounts
and
then,
if
you
click
this
you
it
will
swap
between.
I
think
a
coordinate
like
world
coordinates
and
and
pixel
pixel
coordinates.
B
There's
also
draw
quadrangle,
which
replaced
draw
a
rectangle
in
in
3.0,
so
draw
a
rectangle
has
been
deprecated
which
allows
you
to
to
draw
quadrangles
along
particular
lines
of
heliographic
latitude.
So
this
is
an
example
from
from
our
from
our
gallery,
and
we
also
at
the
david
added
map
sources
for
mdi
and
hmi
synoptic
magnetograms
in
2.1.
I
had
a
pr
that
added
the
the
ui
map
source,
and
that
was
that's
a
3.0.
B
This
I
think
albert
mentioned
that
now
we
have
the
chord
coordinate.
I
forgot
to
add
this
on
here.
We
have
coordinate,
aware
plotting,
so
you
can
do
things
like
this
right
stuff,
like
the
auto
align
functionality,
where
you
can
overplot
maps
with
different
wcs's.
B
B
Okay,
then
time
series
has
not
had
a
ton
of
changes.
One
of
the
big
ones
was
laura
worked
on.
Adding
support
for
the
new
goes
net
cdf
files
and
that
was
included
in
in
2.1,
and
so
there
was
a
an
addition
in
in
net
to
also
be
able
to
to
query
these
files
from
the
new
noaa
source.
B
And
now
now
we
can
read
these
these
net
cdf
files
as
well,
and
there's
adjusted
the
the
the
time
series
source
for
xrs
to
to
account
for
the
new,
the
new
data
format
and
and
to
account
for
changes
between
13
14
15
and
then
the
ghost
1617
data
there.
B
B
David
has
recently
put
in
a
bunch
of
work
in
a
refactor
of
the
time
series
factory
which
that
pr
is
still
open.
So
I
think
that
will
be
probably
going
to
4.0
and
then
there's
been
some
discussion
around,
possibly
replacing
our
underlying
data
structure
for
time
series
going
from
a
pandas
data
frame
to
astrophy
table.
I
know
this
is
so
very
early
and
very
uncertain,
but
there
I
know
there's
been
some
discussion
at
the
at
our
weekly
meetings
about
that.
B
B
Okay,
we
have
four
days
to
do
it.
Image
has
not
seen
a
ton
of
activity,
I
think
jack
and
I
are
the
the
maintainers
for
this
sub
package.
There's
probably
a
lot
of
there's,
maybe
some
things
that
can
be
moved
to
sunken
image.
I
think
that's
a
discussion
that
we
could
have
maybe
some
parts
that
could
be
replaced
with
roof
checked.
B
I
would
say
the
the
biggest
kind
of
part
of
image
that
we've
had
to
deal
with
lately
is
dealing
with
affine
transform
so
for
people
where
there's
a
pr
that
would
allow
for
basically
more
a
more
flexible
interface
to
switching
out
the
back
end
from
what
what
package
is
doing,
the
fine
transform.
So
you
could
use
something
like
coupei
or
desk
for
the
kind
of
underlying
the
computational
workload.
B
B
There
doesn't
seem
to
have
been
a
lot
of
activity
there
if,
if
I'm
wrong,
people
should
jump
in
and
tell
me,
sunpi
physics
seems
to
exclusively
exist
to
hold
the
differential
rotation
code
and
nothing
else,
and
I'm
not
really
sure
why
we
have
a
sub
package
called
called
some
type
physics,
and
I
think,
may
we
need
to
revisit
this.
Possibly
re
renaming
it
or
you
know,
think
about,
especially
after
the
our
upcoming
unit,
refactor
of
how
differential
rotation
how
that
function
works.
B
What,
in
the
case
of
rotate
some
frame,
maybe
think
about
deprecating
this
package,
too
sun
holds
constants
and,
and
we
have
a
few
models
for
things
like
the
like
density
and
temperature
and
as
a
function
of
radius
through
through
the
sun,
and
this
I
I
would
one
thought
here
is
that
I
think
we
could
think
about
renaming
this
to
just
a
constant
sub
package
and
get
rid
of
the
models
in
there.
I'm
not
sure.
B
Who's,
maintaining
these
and
and
and
when
they
were
put
in
there,
but
that's
just
one
some
thoughts
time
seems
to
be
pretty
stable.
Util
is
mostly
kind
of
obviously
internal
tools
that
we've
been
using
and
I
I
don't
know
much
about
visualization
either,
but
anyone
should
feel
free
to.
B
C
A
N
A
E
E
B
I
think
that
leads
well
into
the
slide,
which
is
my
my
my
last
slide.
It's
just
to
say
that
we
need,
I
think
we
need
to
think
about
maintainers
for
for
all
of
our
sub
packages
and
revisiting
this
list.
I
think
this
may
be
a
little
stale,
so
if
you're
just
saying
so,
io
does
not
have
formally
any
any
sub-package
maintainers
database
doesn't
either,
but
maybe
after
thursday,
we'll
see
if
that
matters
at
all.
B
There's
quite
a
few
of
these
with
only
one
maintainer,
and
maybe
that's
okay,
you
know-
maybe
I
think,
sort
of.
If,
if
the
people
who
are
are
the
single
maintainers
for
these
sub
packages,
they
need
extra
help.
Do
we
want
to
have
two
people
and
kind
of
more
broadly?
What
is
the
po?
What
is
the
point
of
having
sub
package
maintainers?
Are
we
actually
adhering
to
these
roles
or
not?
I
think
that's.
Basically,
we
need
to.
We
need
to
think
about,
or
do
these
do
we
need
to
redistribute.
You
know
these
responsibilities.
B
I
don't
know
so
I
don't
have
any
firm
conclusions
for
this
slide.
Just
to
say
that
you
know
I
I
think
we've
had.
Maybe
a
certain
sort
of
forgotten
about
about
these
roles
and
do
we
need
to
revisit
them
and
think
about
which
which
of
these
really
really
needs?
Two
people
on
it.
C
B
H
C
I
step
in
and
say
stu
might
be
something
worth
talking
about
when
you're
talking
about
the
project,
overhaul
projects,
governance
and
yeah.
This
part.
C
E
E
A
A
C
C
A
E
A
A
A
A
Cool
sophie
then.
F
P
F
It
has
to
have
to
keep
coming
up
every
year
now
for
sure
okay,
so
I
only
have
a
couple
of
slides
because
obviously
communication
is
a
global
pantone,
because
a
lack
of
communications
in
a
global
contact,
but
first
a
community
announcement.
I
have
officially
left
central
physics
again
and
I
made
it
this
time,
so
I
I've
moved
on
from
my
research
fellow
roles,
so
still
a
job,
an
institute
for
about
studies,
but
I'm
now
working
in
more
of
a
project
management
role.
F
So
I
won't
be
doing
you
know,
writing
my
own
proposals
for
research
or
writing,
publications
or
doing
you
know,
center
physics,
conferences
or
anything
like
that
anymore,
I'll,
be
helping
everyone
else
in
the
section,
with
all
their
different
research
projects
and
helping
manage
them
and
promote
them
and
stuff
like
that.
So
definitely
not
go
back
this.
F
Is
it
I'm
sticking
with
the
project
management,
but
it
doesn't
mean
I'm
residing
all
my
soda
rolls,
so
I'm
gonna
slowly
fade
into
the
background
for
for
pretty
much
most
of
them,
but
I
will,
I
will
learn,
I'm
a
good
lurker.
So
if
you
ever
need
help
with
calm
stuff,
I
will
always
be
around,
but
I
just
want
to
have
as
much
time
as
I
used
to
not
saying
I
ever
really
had
much
time
in
the
first
place,
but
even
less
than
I
did
before,
and
so
I
obviously
started
helping.
F
I
mean
you
know.
Obviously
I've
been
using
somebody
for
such
a
long
time,
but
really
only
happened
with
the
cons
at
the
beginning
of
last
year,
which
was
a
terrible
time
to
start
something
because
obviously
covert
hit
and
very
much
a
less
is
more
approach
with
calms,
I
think,
and
education
and
stuff
in
general
in
2020,
just
because
it
sort
of
disappeared
in
the
blink
of
an
eye,
and
I
think
everyone
was
very
much
focused
on
you
know
staring
at
covert
stats
for
the
first
few
months
and
then
realizing.
F
Oh,
I
have
to
get
used
to
remote
working,
so
I
better
do
some
prep
for
that,
and
I
you
know,
I
think
anything
that
wasn't
that
sort
of
focus
or
wasn't
related
to
real
world
events
was
very
much
kind
of
put
aside.
So
less
was
more
from
an
education
and
communications
point
of
view
for
for
any
research
projects.
I
do
think
you
know
obviously
lots
of
different
things
got
sort
of
canceled
and
postponed
conference
wise,
which
is
where
you
get
to
promote
these
sorts
of
things.
F
Things
did
start
to
pick
up
and
we
started
getting
more
remote
conferences
about
the
rest
of
these.
They
definitely
then
began
sort
of
a
zoom
saturation
point,
so
there
was
almost
like
too
many
online
conferences
and
that's
not
necessarily
good
things.
You
do
have
to
be
very
careful
when
you're
targeting
promotion
for
this
sort
of
stuff
to
not
oversaturate
so
yeah.
I'm
I'm
pretty
glad.
This
is
jitsi
today,
because
it's
not
another
zoomed
in
yay.
F
This
is
like
a
nice
change.
It
looks
different
on
my
screen
for
a
change
and
the
thing
that
I
did
obviously
managed
to
get
to
a
lot
of
different
online
meetings
and
talk
to
people
about
somebody
and
what
they're
interested
in
what
they
think
of
it,
and
it
was
quite
interesting-
and
I
think
now
is
actually
a
really
good
time
to
reflect
on
promoting
project
any
target
audience.
F
It
is,
and
and
who
are
the
best
people
to
reach
out
to
and
tell
people
about
the
project,
because
for
me
just
chatting
to
people
at
poster
sessions
and
just
general
meetings
and
stuff
there's
actually
a
lot
of
enthusiasm
out
there
for
somebody
seen
you
know,
particularly
from
early
career
people,
post,
grads
and
postdocs
use
it
all
the
time
and
all
their
research.
But
then
you
get
to
the
question
of.
Do
you
want
to
get
involved?
And
then
people
aren't.
F
You
know
necessarily
as
forthcoming
with
that,
mainly
because
just
just
from
what
people
have
been
saying
is
that
it's
very
much
not
related
to
their
specific
research
projects
that
they're
working
on
you
know,
there's
not
a
there's,
not
a
spare
time
there
for
for
code
projects
rather
than
really
basic
research,
things
or
there's
a
mindset
of
maybe
perhaps
some
other
supervisors.
It's
not
necessarily
the
best
thing
to
to
devote
their
time
to
so
there's
a
there's.
Definitely
a
mindset
change
there
that
needs
to
take
place.
F
F
You
know
within
their
own
research
projects,
so
I
really
liked
that
sbd
reading
stuff
that
you
guys
did
laura
was
saying,
but
there
was
you
know,
lots
of
different
people
there,
and
particularly
some
of
the
you
know
the
people
who
would
supervise
their
early
career
or
people
who
would
write
those
ground
projects
you
know,
so
they
are
really
good
people
to
target
to
try
and
change
that
mindset
as
to
how
you
know
some
point
might
be
useful
for
them.
So
it's
something
to
definitely
think
about
in
terms
of
promoting
it.
F
I
think
going
forward
very
much
all
I
pretty
much
did
you
know
through
a
lot
of
2020
was
just
like
tweeting
and
stuff
like
that,
because
you
had
to
keep
that
light
approach.
You
don't
want
to.
You
know,
annoy
people
when
they
really
want
to
think
about
covert
and
stuff
like
that
interesting.
Looking
at
the
stats,
I
love
looking
at
stats.
These
faces
wow,
because
these
hate
stats
was
the
student
are
now
obsessed,
but
so
this
is
sort
of.
If
you
can
see
this
sort
of
trend
here.
F
This
is
the
profile
visits
for
the
twitter
account
for
pretty
much
most
of
2020.
There's
a
there's,
a
massive
increase
when
I
went
on
a
bit
of
a
spate
of
following
loads
of
sort
of
physicists
in
october.
So
that's
why
you
see
the
increase.
F
It
didn't
necessarily
mean
that
they
followed
us
back,
but
they
did
like
peak
their
interest
for
a
split
second
to
like
click
on
their
profile
visit
and
but
there's
also
a
few
other
interesting
peaks
and
they're
very
much
surrounded
by
project
updates
or
events
that
were
happening
with
somebody
so
around
here
in
springtime.
There
was,
of
course,
the
the
survey
which
was
really
really.
I
think
the
community
were
fascinated
by
that
and
I've
seen
it
pop
up
and
mention
those
of
different
times
in
other
talks.
F
I
think
people
are
kind
of
getting
a
sense
of
where
we
are
in
terms
of
using
python
and
sunpower
have
been
really
really
useful.
So
I
know
you're
talking
about
earlier.
I
think
it's
a
really
good
thing
to
do
again.
Maybe
maybe
not
this
year
wait
till
people
are
a
little
bit
or
back
to
whatever
normal
is
maybe
next
year.
Think
about
doing
it,
because
I
think
that
would
be
a
good.
You
know
a
couple
years
difference
between
the
two
stats,
so
yeah
that
really
picked
up
and
then
7.20
as
well.
F
So
that
was
something
that
you
know
just
general
project
updates.
Whenever
they
happen,
people
do
actually
click
and
read
those
documents,
which
is
great,
so
people
are
following
along
with
what's
happening
with
somebody
and
also
this
you
know
this
massive
peak
was
more
to
do
with
the
scientists
the
easiest
empire.
This
was
you
know
when
you're
running
the
and
the
ask
us
anything
event.
So
people
were
clicking
on
it,
and
then
people
were
talking
about
how
they
used
somehow
on
twitter,
and
you
know,
if
you
look
at
the
top
tweets
the
top
mentions.
F
It's
really
important
to
kind
of
see
what
those
two
top
ones
are,
because
it
kind
of
highlights
it's
the
project
updates,
but
it's
also
the
users,
and
so
the
top
management
was
laura
when
she
did
that
really
cool
visualization
stuff
and
it
got
loads
and
loads
of
people
looking
at
the
twitter
profile,
and
so
you
know
you
can
post
project
updates
and
events
and
stuff,
but
I
think
a
really
key
thing
with
this
sort
of
stuff
is-
and
it
might
be
a
way
to
get
more
people
involved-
is
to
help
promote
the
stuff
that
other
people
are
working
on.
F
It's
promoting
some
by
user
output,
so
they
get
something
out
of
some
file.
You
know
they
reference
it
and
they
use
it.
They
get
promoted
as
well,
and
so
it's
something
to
think
about.
I
think
it
could
be
a
new
way
of
of
taking
sort
of
more
hands-on
approach
with
community
and
getting
them
involved.
F
You
know
the
one
thing
that's
been
sort
of
talked
about
in
more
recent
times
and
loads
of
different
projects
is
the
concept
of
nuggets
and
I'm
I'm
always
shy
away
from
them,
because
I
think
there
are
a
lot
of
effort
and
I
very
much
think
twitter
was
more
of
a
bigger
thing
in
the
pandemic
era,
because
blogs
just
take
absolutely
ages
to
write
and-
and
you
do
kind
of,
I
think
you
think
about
them
more
and
their
wording
than
if
you
just
sent
off
a
really
quick
tweet,
and
that
might
only
take
a
couple
minutes
if
you
sat
and
did
a
blog
post,
you'd
already
think
about
it,
a
lot
more
and
I
think
it
takes
more
time
on
the
creator's
point
of
view
to
do
it.
F
But
it
is
something
that
the
community
understands.
So
if
it's
understandable
concept
for
solar
scientists,
that's
not
a
bad
thing
to
do
something
like
it.
You
know
and
the
you
know
it
is
very
much
ingrained
whether
it's
you
know
haters.
You
know
this
instrument
stuff
or
or
actual
like
youtube,
center
physics
or
other
different
communities
as
well.
So
one
thing
I
thought
might
be
kind
of
a
cool
thing
to
do
something
like
a
twitter
thread
so
like.
If
someone
has
just
published
a
paper
using
some
point
and
we
see
reference,
we
asked
them.
F
Would
you
like
to
post
a
little
story
and
then
we,
you
can
just
do
something
like
unroll
and
then
it
converts
into
a
blog
post.
So
for
those
who
aren't,
because
you
know
we
can't
focus
everything
on
twitter,
you
could
then
put
it
on
the
website
for
longevity
in
the
blog
post
as
well,
so
it
enrolls
into
a
longer
post,
just
something
to
sort
of
say
thanks
for
using
it.
F
We're
really
interested
in
your
work
and
it'll
open
up
a
discussion
then
for
people
to
talk
about
how
they're
using
it
and
maybe
get
a
bit
more
help
and
involvement.
So
something
to
think
about,
I
think,
don't
focus.
Obviously
everything
on
social
media,
because
everyone
is
not
on
social
media,
but
at
least
you
can
use
them
as
a
tool
to
then
also
provide
content
for
website
blogs
and
stuff
like
that
as
well
for
for
those
who
aren't
there.
So
it's
something
to
think
about.
F
I
think
comes
as
easy
and
I
think
anyone
can
do
calms,
but
I
I
would
definitely
say
in
terms
of
a
communication
and
education
rule.
One
thing
I
definitely
noticed
this
past
year
is
that
I
definitely
do
more
comms
in
education
and
probably
because
I'm
not
as
involved
in
the
the
co-development
side
of
things.
I
think
you
need
someone
who
really
understands
the
code
base
in
order
to
run
say
a
tutorial
or
something
like
that.
But
I
also
think
that
you
need
someone
with
excellent
understanding
of
the
code.
F
Basically
someone
with
teaching
ability
and
the
time
and
the
energy
to
do
it.
So
often
the
developers
don't
really
have
that
time
right,
because
you're
volunteering
your
own
time
to
work
on
it
and
so
to
me
it
it
comes
back
to
the
inevitable
problem
of
funding,
and
so
it's
something
to
think
long
term
as
to
I
think
we
should
think
about.
Where
is
the
really
best
way
to
target
the
education,
because
I
think
people's
time
is
limited
and
it's
not
easy
to
get
education
funding?
F
My
general
experiences
education
funding
grants
will
give
you
the
money
for
workshop
materials
and
stuff
like
that,
but
not
necessarily
salary
time.
So
something
to
think
about
you
know
how
important
is
is
running
a
tutorial
at
a
as
a
side
to
accomplish
as
a
big
week
long.
You
know
summer
school
or
something
like
that.
I
think
make
sure
you
target
your
time
and
not
try
to
spread
yourselves
too
thinly
with
like
every
single
invite
that
comes
up
and
yeah.
F
I
thought
funding
was
the
next
thing
on
the
list,
so
it's
definitely
something
to
think
about.
You
know.
I
know
I'm
always
keeping
an
eye
out
for
these
things.
Sometimes
there
are
little
pots
of
money
for
these
kinds
of
things
and
that
can
help
you
to
run
workshops
and
stuff
like
that,
so
that
might
help
long
term
so
yeah,
that's
it.
That's
just
some
some
thoughts
I
had
about
the
cons
in
the
education
and
what
we
could
do
in
the
future.
F
Very
much
I
think,
trying
to
get
more
people
involved
has
to
be
trying
to
see
what
they
can
get
out
of
it
as
well
from
from
the
discussions
I've
had
with
people
just
showing
their
supervisors
or
whoever
their
grant
funders
are
that
it's
a
useful
thing
to
do,
and
I'm
gonna
hijack
this
because
I
just
posted
this
job
ad
this
morning.
So
I
didn't
want
to
say
we're
hiring
at
the
moment
and
maybe
the
person
who
gets
this
job
could
help
with
senpai
as
well.
F
So
we
we,
we
have,
you
know
we
host
tyus.
We
have
starter,
monitor,
so
we're
we're
hiring
for
a
research
software
developer.
So
if
anybody
on
the
call
knows
anyone
who's
interested,
please
send
them
to
them,
because
we'd
love
to
get
someone
who's
a
little
bit
more
into
the
code
development
side
of
things.
K
K
K
F
I
often
whenever
I
you
know,
and
I
don't
get
to
meetings
as
much
as
I
like,
but
whatever
I'm
eating.
I
think
someone
is
always
saying:
oh
that's
something
that
probably
people
don't
know
about.
We
should
do
something
about
that,
but
I
think
people
don't
really
have
the
time
to
do
something
about
it.
So
maybe
something
like
a
quick
tweet
might
actually
be
a
good
starting
point
to
highlight
something,
and
then
it
starts
off
a
conversation
with
someone
in
the
community
about
it.
You
know
yeah.
E
Definitely
I
was
gonna
say
as
well
if
you're,
based
on
like
the
twitter
things
on
how
to
get
it
into
a
blog
post
like
this,
is
a
really
good
place
for
just
the
new
discourse
channel
to
yeah
our
forum
for
things
like
that,
like
we're
trying
to,
we
have
one
post
at
the
moment,
which
is
like
a
showcase
nicole.
They
did
following
actually
albert's
tutorial
from
sbd
and
they
kind
of
built
like
a
cool
movie,
to
show
that
the
orbits
of
the
satellite.
E
So
I
think
we
just
need
to
promote
that
some
more
yeah
and
then
also
just
being
back
in
the
group
of
diets
with
students.
Again
it's
crazy,
like
you
know,
we
have
to
ask
the
python
channel
and
all
the
questions
they're.
Like
I
mean
we
need
to
find
a
way
of
getting
those
students
to
ask
the
questions
in
the
channel
or
to
document
the
issues
on
the
github
rather
than
just
asking
us
and
we're
answering
them.
E
You
know
what
I
mean
like
I
mean
because
you're
missing
and
you've
learned
so
much
about
you
know
flaws
within
the
code
base
by
listening
to
them.
So
it's
promoting.
You
know
to
get
involved
with
some
fighter
doesn't
need
to
be
contributing
code,
but
just
raising
issues
and
asking
questions
and
giving
feedback
is.
I.
F
Think
it's
really
important
to
be
a
friendly
face
for
students
as
well.
So
you
know
a
lot
of
them
are
saying
like
I'd
love
to
like
post
something
on
the
the
riot,
but
I'd
be
petrified
that
it's
a
stupid
question
or
something
like
that.
And
so
maybe
you
do
well
scientists
or
something
is
like,
but
you
know
I'm
sorry.
I
work
on
this.
You
know,
and
so
it
gives
a
facebook
you,
a
random
username
they're,
not
so
petrified
to
ask
questions
but
there's
definitely
a
student
problem.
A
I
mean
I
still
regularly
find
issues
with
somebody
every
time
we
do
any
kind
of
teaching,
and
I'm
like
how
has
nobody
told
us
about
this,
yet
I
can't
remember
what
it
was
for
the
spd
workshop,
but
I
found
something
that
had
been
hiding
in
the
code
base
for
years
and
I
was
like.
I
cannot
be
the
first
person
that
this
is
tripped
up.
K
Yeah,
I
think,
there's
a
massive
education
gap
between
like
graduate
students
from
college
who
do
physics
or
solar
physics
on
like
software,
like
they
don't
necessarily
know
about
github.
They
don't
also
know
what
issues
or
like
yeah.
It's
definitely
an
education
sort
of
thing
you
probably
address
at
like
workshops,
you'd
be
like
listening,
find
something
wrong.
Can
you
please
please
come
talk
to
us.
N
N
A
With
any
of
them
will
suffice,
it's
not
like
we're
getting
bombarded
on
twitter
with
these
bug
reports
and
not
getting
anything
on
github
right.
Nobody
is
just
talking
to
us
anywhere
about
them.
So.
B
I
think
it's
less
of
like
a
gaps
in
technical
understanding,
though
I
think
that's
definitely
true.
I
think
it's
more
about
like
people
don't
feel
like
they
own
the
software
that
they're
using
right.
Like
I
mean
often,
the
model
is
like
you,
come
in
your
advisor
hands,
you
a
piece
of
code.
It's
awful!
You
use
that
code,
you
struggle
through
it
and
then
you
get
a
phd
right.
I
think
that's.
B
The
sort
of
typical
model
people
aren't
used
to
be
able
to
say,
like
this
code
like
needs
to
be
improved,
or
this
goes
great
in
this
way
or
if
I
can
like
contribute
back
to
this
thing,
like
I
don't
think
they're
used
to
having
like
that
concept
of
like
ownership
over
the
software
you're
using
is
like.
I
think
I
think
people
are
not
used
to
that
idea
like
I
can
be
a
part
of
the
community
that
like
develops.
This
is
still
something
that
I
think
people
are
are.
That's
that's
the
missing
piece,
yeah.
D
I
think
more
accurately
they
don't
own
it
until
they
forked
it.
You
know,
like
you,
give
a
piece
of
code,
you
sort
of
copy
it
onto
your
machine.
You
get
it
to
do
what
you
need
to
do
for
your
research.
That's
your
code,
like
the
private
copy
that
you
have
and
not
shared
with
anybody,
but
made
better
for
exactly
what
you
needed
to
do.
F
Yeah-
and
I
think
I
think
it's
it's
a
culture
thing
in
a
group
as
well,
so
you
know
if
someone
was
coming
to
work
with
me.
I'd
teach
them
about
this
sort
of
stuff
straight
away,
but
if
they
were
working
with
someone
who
doesn't
really
do
that,
or
you
know,
is
working
on
their
folder
code
that
they've
been
working
on
for
20
years,
they're
never
really
going
to
have
that
experience
of
it.
F
So
that's
why
I
think
we
need
to
definitely
try
and
make
yourself
more
friendly
to
students
and
get
them
to
ask
questions,
but
it's
also
those
supervisors
and
those
people
who
are
writing
the
gram
funding
proposals
to
try
and
make
it
more
understandable
for
them
that
this
is
a
useful
tool
for
their
for
their
students
and
for
their
postdocs
to
work
on.
You
know.
K
I
think
it's
really
tough,
though,
because
I
mean,
and
in
our
group
you've
got
me
sophie
laura
and
we're
always
banging
on
about
just
ask
a
question:
send
it
like
continue
they're
still,
probably
so
sick
about
it.
I
don't
think
there's
been
a
single
issue
raised
by
anyone
and
they
definitely
had
problems
like
100
percent
of
my
problems
with
the
code
or
found
bugs,
and
so
it's
really
difficult
to.
E
A
F
Sometimes
you
can
have
too
many
channels
for
people
to
ask
questions
either.
Oh,
they
see
the
riot,
they
see
twitter,
they
see
email
lists,
they
see
this
course.
Maybe
it's
just
pick
one
like
it's
the
discourse
that
laura's
suggesting
I'm
really
heavily
promote
that
one.
You
know,
so
you
get
more
people
using
that
or
something
yeah.
A
I
don't
know
whether
we
have
a
I
assume
we
don't
have
a
different
time
to
talk
about
this,
but
certainly
my
my
thoughts
are
that
the
discourse
could,
if
it,
if
it
works
for
us
in
the
way
that
we
hope
it
will
replace
the
mailing
list.
E
Eventually,
yeah
yeah.
We
should
really
find
it
like
try
to
put
in
some
time
for
that,
because
even
just
like
our
own
understanding
of
us,
I'm
not.
A
E
G
A
G
G
They
would
basically
refuse
to
answer
questions
privately
on
an
email
and
instead
direct
anyone
to
their
equivalent
of
an
issue
tracker
because
then
all
the
help
they
gave
was
public
and
traceable.
So
anyone
with
the
same
problem
could
then
see
that.
So
I
wonder
if
do
wonder
if
actually
having
like
helping
people
personally
is
maybe
more
approachable
but
in
the
long
term
detrimental
because
they
don't
learn
how
to
report
bugs
efficiently
and
that
help
is
then
only
available
to
one
person
instead
of
anyone.
B
Answering
emails
of
people
out
of
questions
about
senpai,
I'm
sure,
like
probably
everyone
else
in
this
call
often
has
had
that
experience
too,
because
I
there's
those
of
us
that,
like
people
in
the
community
or
wherever
we
work
like
they
know,
because
we
talk
about
it
all
the
time,
probably
that
that
we
work
on
senpai
and
that
we're
invested,
and
we
want
people
to
use
it
and
yeah,
and
I
think
that,
but
then
that
has
a
cost
right.
I
mean
just
like
you're,
saying
david,
like
that.
I
think
you
end
up.
B
F
A
A
So
I
mean
I
mean
you
can
always
make
the
documentation
better,
but
without
having.
I
personally
am
at
a
loss
on
like
other
than
just
fixing
bugs,
because
I'm
sure
there
are
functions
that
are
missing
proper
doctrines
or
whatever,
but
like
sort
of
on
a
structural
level.
I
don't
really
know
what,
where
to
go
with
the
documentation
at
the
moment.
Personally,.
B
E
Yeah,
it's
gonna
comment
on
that.
What
like
what
I'd
say?
Was
that
exact
thing
that
like
it's
either
people
like
they
want
to
google
their
problem
and
find
an
answer,
and
then
it's
like
a
lack
of
like
knowledge
of,
say,
students
that
don't
know
understand
how
to
read
documentation.
They
just
want
an
example
to
show
them
exactly
what
they
want
to
do.
So
this
goes
back
to
maybe
yeah.
I
don't
know.
A
And
that's
the
other
thing
that
I
hope
the
discourse
will
be
able
to
do
like
the
showcase
category
in
the
discourse
could
be
an
excellent
source
of
things
to
move
into
examples
like:
oh,
that's,
a
really
cool
thing
that
you've
just
put
posted
up
there.
Would
you
like
to
convert
it
into
a
python
script
and
maybe
contribute
it
over
here
or
one
of
us
can
do
it
if
they
really
don't
want
to
do
that.
F
That's
brilliant,
I
didn't
even
realize
it's
a
showcase
section,
so
I
I
think
it
could
be
the
perfect
solution
for
the
sort
of
thing
I
was
thinking
about
because,
like
I
don't
want
everything
to
be
twitter
focused,
I
think,
because
not
everyone
is
on
social
media.
So
if
you
had
something
like
that,
I'm
just
curious
where
you
could
share
like
snippets
of
different
things
that
you
can
do
with
some
player
things
that
people
are
working
on.
I
think
that'd
be
lovely.
F
A
F
A
C
Yeah
I
just
wanted
to
quickly
interrupt
because
obviously
about
the
funding
section
before
that,
obviously
I
I.
A
A
A
A
Yet
this
is
the
open
astronomy
discourse
for
those
of
you
who
don't
know
what
discourse
is
it's
a
piece
of
forum
software
astropy
have
kindly
paid
the
people
who
make
discourse
to
host
open
astronomy's
discourse
for
us.
A
So
we
have
a
professionally
hosted
instance
of
discourse
which
is
set
up
and
you
know,
will
be
reliable.
It's
not
running
in
a
server
in
my
basement
and
we
there
is
there's
still
disc,
there's
still
many
decisions
to
take
here
about
how
we
want
to
do
this.
So
for
the
moment,
the
way
that
this
is
set
up
is
we
have
high
level
showcase
and
ask
questions
categories
where
everybody
from
all
the
projects
under
open
astronomy
could
ask
their
questions
and
then
tag
questions
with
respective
things.
A
So,
if
you
had
a
senpai
demo,
you
would
post
it
in
the
showcase
section
and
not
in
the
senpai
one,
which
I
appreciate
is
somewhat
confusing
and
we've
been
laura
and
nabil,
and
I
have
been
having
some
discussions
about
whether
that's
the
best
flow
or
whether
there
are
other
ways
to
do
it.
We're
trying
to
simultaneously
learn
how
discourse
works,
what
the
limitations
are
and.
A
What's
the
best
way
to
make
it
work
for
us,
but
if
you
you
should
go
to
community.openastronomy.org
and
sign
up
for
an
account.
If
you
have
a
good
demo
or
a
good
question,
you
should
post
it
and
we
can
always
move
posts
around
getting
more
content
and
more
people
on
here
to
start
off
with
is.
A
The
best
place
to
start,
we
can
argue
about
categories
more
later.
If
you
there
is
a
way.
Where
did
I
do
this?
A
A
A
A
First,
on
the
chopping
block
for
the
mailing
list
is
definitely
some
by
dev
if
we
can
replace
some
by
dev
with
the
development
category
so
much
the
better.
E
Anyway,
I
think
this
is
great
here.
I
think
it's
awesome
and
so
much
potential
here
I
think
you're
just
going
back
to.
Maybe
this
is
a
good
time
just
to
have
a
quick
discussion
about
this
about
you
know,
should
I
would
be
worried
if
things
went
into
showcase
rather
than
into
sunplay
say
a
showcase
about
some
play.
People
won't
see
it
because,
if
you're
a
senpai
user,
you
might
not
even
really
think
you
even
use
astrophy,
even
though
you
do
and
you
click
on
the
senpai.
E
E
F
E
P
A
N
A
A
I
mean
their
online
stuff
is
very
good,
their
online
help
and
stuff,
so
I
could
just
search
for
it.
I
think
it's
really
comes
to
a
question
it
comes
down
to
like.
E
But
like
in
an
ideal
world
that
would
work
right,
but
I
just
feel
like
a
user
that
would
come
onto
this
page,
won't
understand
that
concept
or
they
might
go
into
senpai.
Looking
for
an
answer
to
the
question,
that's
actually
in
the
ask
questions,
or
maybe
this
is
just
again.
We
just
need
to
make
it
clear
when
they
come
to
this
page,
but
what
they
do
or
if
you
could
have
that
cross
category
thing
that
would
be
fantastic.
E
G
G
A
F
And
you
know
at
the
top,
when
you
click
on
somebody,
I'm
just
looking
at
it
now
it's
kind
of
like
a
description
that
I'll
talk
about
somewhere.
Maybe
you
could
like
direct
people
to
say,
for
example,
to
check
out
showcase
or
something,
and
then
it
depends
if
people
even
read
the
sentence
at
the
top,
there.
F
F
E
E
A
You
can
you
can
interact
with
each
category
subcategory
as
a
as
its
own
mailing
list,
including
creating
posts
by
email.
C
C
F
J
Hear
you,
let's
see
well
nice
presentation
sophie.
I
don't
want
to
like
distract
from
this
discourse
discussion.
I
just
I
wanted
to
say
those
that
was
awesome
presentation
and
I
very
much
appreciate
the
the
recommendation
to
shift
a
bit
towards
showcasing
examples
of
what
you
can
do
with
senpai
in
your
research,
as
opposed
to
giving
more
of
these
thoughts
of
like
what
is
senpai
what's
a
map.
But
you
know
that
kind
of
thing,
because
yeah,
I
agree
that
a
lot
of
people
are
really
interested
and
want
to
join
the
community.
J
But
one
of
the
one
of
the
major
setbacks
is
they
don't
get
support
from
their
supervisors?
So
if
their
supervisors
can
see
what
we're
doing
and
how
that
enables
the
research,
then
then
I
think
that's,
that's
probably
the
easiest
way
to
sell
it.
So
sorry
for
moving
this
conversation
back
into
that
direction.
F
I
really
appreciate
it,
I
think,
that's
it.
Isn't
it
it's
trying
to
get
into
that
mindset.
That's
the
normal
academic
scientific
mindset
that
people
are
used
to
and
and
working
off
that,
but
also
kind
of
progressing
how
you
want
to
showcase
and
play
through
that.
It's
a
tough
thing,
but
I
think
if
we
start
with
like
cool
showcase
of
people's
work
in
something
like
this
course,
that's
definitely
a
good
stepping
stone.
Isn't
it.
J
Yeah,
it
kind
of
reminds
me,
you
know,
like
a
few
people
here,
are
part
of
the
like
machine
learning
and
heliophysics
community
and,
like
the
machine
learning
talks
are
initially
like
what
is
machine
learning.
This
is
an
algorithm
just
you
know,
and
now
we're
kind
of
beyond
those
talks
and
people
are
like
this
is
what
you
can
do.
These
are
the
research
outputs.
J
So
I
kind
of
see
the
same
with
some.
I,
like
a
lot
of
our
talks
were
like
what
is
senpai.
Here's
a
map.
Here's
you
know
time
series
now
we
can
just
say
like
okay,
we
assume
that
you
you've
heard
plenty
of
those
talks
and
here's
a
lightning
round
of
examples
of
people
using
in
their
research.
Even
if
it's
like
a
minor
example
and
then
putting
that
up
on
twitter
and
blogs
and
stuff.
F
Yeah
and
you
can
always
have
that
basic
stuff,
but
it's
just
a
video
or
something
from
a
previous
thing
or
you
know,
or
links
to
resources
where
you
can
read
about
it.
So
we're
not
saying
you
can't
learn
the
basics,
but
it's
just
we're
going
to
focus
on
this,
and
this
is
how
you
would
find
and
read
more
about
the
basics.
B
This
is
almost
the
exact
opposite
of
what,
when
you
were
saying
sophie
about,
like
the
effort
that
it
takes
to
write,
a
blog
post
is
like
certainly
high,
but
I
wonder
and
like
you
know,
I
know
under
like
doing
sort
of
like
fights
more
bite-sized
examples
is,
I
wonder
if
doing
something
similar
to
like
monica
and
james
mason
had
done
with
their
helio
ml
book
of
like
illustrating,
like
with
code
like
how
do
you
write
a
paper
with
senpai
and
astrophy
and
like
the
full,
like
scientific
python,
stack
like
showing
people
like
an
end
to
end
like
here's,
my
manuscript,
here's
the
stuff?
B
To
do
my
analysis,
here's
the
stuff
to
do
my
figures
like
how?
What
does
that
actually
look?
Like
I
mean,
obviously
we
all
have
this
code
like
knocking
around
on
our
computer
somewhere
like?
Could
we
come
up
with
some
way
to
like?
I
realized?
That's
that's
a
ton
of
effort
to
just
bring
something
like
that,
but
I
think
like
showing
like
laying
out
an
example
of
like
how
do
you
complete
a
whole
project
with,
like
the
essentially
like
pie,
astro
stack,
the
lack
of
a
better
word.
N
F
J
I
think
maybe
just
like
you
know
if
we
see
a
paper,
if
somebody
comes
forward
and
says
we
use
some
pages,
just
say
like
what
would
you
like
to
do?
You
could
write
a
blog
post,
you
could
tweet
about
it.
You
could
do
your
whole
thing,
it's
kind
of
up
to
you,
but
we
would
really
want
to
help
you
advertise
your
work.
K
I
suppose
it's
been
a
bit
proactive
if,
like
yeah,
if
we
see
papers
that
are
definitely
written
with
somebody
like,
should
someone
drop
them
an
email
or
or
tweet
and
say,
listen,
we'd
love
to
advertise
this.
We
have
something
that
you
we
could
post
from
the
senpai
twitter
discourse
with.
You
know
pick
your
favorite.
I
Works
hi
everyone
yeah,
I
just
wanted
to
just
drop
in
there
and
say
for
a
number
of
years.
I
was
working
on
the
pro
2
mission
as
the
like.
The
instrument,
scientist
person
and
trying
to
get
people
to
promote
the
mission
was
was
something
that
I
had
to
do,
and
you
really
had
to
you
had
to
do
a
lot
of
the
work
for
them,
I'm
afraid
to
say
and
yeah
fine
in
a
paper.
I
They
would
maybe
like
drop
in
a
reference
in
there,
but
sometimes
I
would
even
go
as
far
as
kind
of
taking
their
abstract
and
rewriting
it
a
little
bit
advertising
how
pro
b2
was
used
in
that
mission.
Send
it
back
to
them
saying.
I
Are
you
happy
for
me
to
put
this
on
our
webpage
and
then
hopefully
they
would
agree
to
that,
and
then
you
could
release
it
as
like
a
news
article
and
tweet
it
and
all
the
rest
of
it
and-
and
the
second
thing
is
trying
to
put
something
out
extremely
frequently
was
very
important.
I
So
I
I
don't
know
what
you
guys
are
doing
with
twitter,
for
example,
but
you
should
be
kind
of
aiming
to
put
out
a
tweet
every
day,
just
something
even
if
it's
just
liking
or
re-sharing,
something
that
can
also
kind
of
help
in
terms
of
promoting
and
also
even
just
like,
highlighting
as
you
were
discussing
before,
where
you
can
actually
send
comments
and
things
like
that.
F
But
it
does
take
a
bit
of
an
effort
to
get
it
off
the
ground.
I
think
of
the
start.
J
Yeah,
I
guess
this
speaks
to
what
nabil
was
saying
in
the
beginning.
Is
that
you
know
this
is
our
opportunity
to
kind
of
make
a
long-term
plan
for
our
communication
strategy
as
opposed
to
just
haphazardly
getting
about
it
like
sophie
you've
been
doing
like
such
a
good
job,
but
we
will
need
someone
else
to
to
do
that,
so
we
should
talk
about
who
those
people
could
be
or
how
much
time
they
have
or
how
to
reach
out.
E
E
It
also
be
reaching
out
to
you
know,
prospective
instrument,
teams
or
people
that
are
you
know
you
can
see
they're
developing
a
package
and
be
like
hey.
Do
you
want
to
come
into
our
little
ecosystem
bubble,
or
I
don't
know
if
that
would
fall
under
that
role,
or
I
guess
you
shouldn't
think
about
it
as
well.
A
D
On
a
vaguely
related
note,
just
a
tangent,
how
much
of
this
is
like
about
something
that
struck
me
about
matt's
comment
about
trying
to
do
this
every
day.
It
seems
to
me
the
point.
One
of
the
points
of
that
is
like
creating
a
sense
of
momentum
so
like
and
then
referring
back
to
one
of
the
slides
from
stewart's
earlier
talk
is
saying:
we
had
21
contributors
and
x
number
of
pull
requests
from
like
the
core
and
we've
traditionally
just
focused
on
the
core
package.
D
That
makes
sense
it's
easy
just
to
deal
with
one
repo,
but
if
we
started
like
expanding
trying
to
expand
those
statistics
and
saying
you
know
senpai
or
the
senpai
community
or
whatever
it
is
doing,
you
know
you
know
this
number
commits,
or
this
number
of
like
issues
or
whatever
to
like
subpac,
not
to
subtract
just
affiliate
packages.
D
So
when
people
talk
about
ssw,
they
don't
just
think
about
the
gen
tree.
You
know,
and
people
aren't
only
going
to
use
sunpi
core
they're
going
to
use
like
increasingly,
you
know
affiliate
packages
like
ai
pi
and
so
on.
So
I
don't
know
if
that's
something
we
want
to
consider
is
that
redefining
our
sump
activity
as
to
include
those,
and
we
can
maybe
that
helps
with
that
sense
of
momentum
or
that
sense
of
size.
A
I
agree,
I
definitely
think,
as
we
shift
more
into
affiliated
packages
and
stuff
that
it's
important
to
try
and
always
include
those
when
we're
talking
about
senpai
with
the
capital
letters
in
here
as
the
project,
there's
a
reason
why
statistics
only
included
core
and
that's
because
I
didn't
fancy
arguing
with
github's
graph
graphql
api
at
11
o'clock
this
morning,
when
I
realized
I
hadn't
written
any
slides.
So,
oh
it
certainly
wasn't
a
criticism.
D
A
D
A
A
People
have
discoverability
issues
with
the
features
that
are
in
core,
so
I
think
we
definitely
need
to
work
out
work
work
on
the
way
we
put
things
in
front
of
people
as
we
spread
out
for,
like
the
reasons
we're
spreading
stuff
out
a
solid
and
well
reasoned
from
a
maintaining
perspective,
but
from
a
user's
perspective.
We're
not
exactly
helping
that
issue.
A
I
Yeah,
I
was
wondering
if
you
could
make
use
of
your
example
pages
a
little
bit
more
with
this
respect,
because
the
way
I've
kind
of
learned
about
senpai
hasn't
been
through
any
of
the
I
haven't
had
the
the
I've
been
to
one
of
your
workshops,
and
that
was
a
very
long
time
ago.
But
beyond
that,
I
haven't
had
the
opportunity
and
how
I've
learned
about
senpai
is
just
through
examples.
A
E
A
E
A
G
I
was
gonna
say
I
think
this
is
also
a
really
good
motivation
to
write
examples
that
are
not
specific
to
any
one
package,
which
is
what
I'm
trying
to
do
as
solar
orbiter
at
the
moment.
So
I
put
a
link
in
the
chat,
but
the
idea
is
basically
to
say
well
from
a
mission
perspective.
What
tasks
do
we
need
to
do
for
solar
orbiter
and
then
how
can
we
glue
together
different
packages
to
do
that.
A
E
H
B
I
mean,
I
think,
that's
kind
of
what
I
was
getting
at
before
with
my
comment.
It's
more
like,
like
things
that
show
sort
of
more
intense,
like
like
workflows
like
that.
Look
like
people's
everyday
workflows.
I
think
that's
been
some
of
the
criticism.
I've
heard
of
our
examples-
and
you
know
whether
it's
fair
or
unfair,
is
a
different
question,
but
people,
I
think,
see
them
as
sort
of
toy
examples,
and
they
don't
know
just
like
how
to
pull
that
stuff
out
into
their
own
workflows,
and
I
think,
having
those
like
the
more.
B
E
B
Yeah
so
maybe
the
blog,
our
blog
posts,
I
mean
we
can
do
like
the
notebook
that
you
can
pull
in
a
whole
nother
book
and
turn
that
into
a
blog
post
I
mean.
Maybe
we
should
just
aim
to
put
some
some
things
there
I
mean
and
then
that
then,
let's
enable
someone
to
just
pull
down
that
that
code
and
those
are
and
that's
a
space
where
we
can
bring
in
you
know
a
bunch
of
different.
You
know
all
the
parts
from
our
affiliated
package
ecosystem
too.
A
But
I
think
we,
I
feel
like
we're
very
rapidly
getting
to
the
point
where,
between
that
effort,
things
like
david's
effort
for
instruments
and
many
example
galleries.
We
need
a
technical
solution
like
learn.ify.org,
where
there
is
go
here,
type,
something
in
search,
box
or
c
list
of
c
summary
of
examples,
and
I
think
yeah,
we're
gonna.
B
J
I
mean
I
guess
this
goes
all
the
way
around
to
like
what
something
was
saying
in
the
beginning
is
like
the
awareness
that
no
one
really
has
time
for
this.
So
at
least,
if
you
see
something
cool
that
someone
has
done,
ask
them
to
write
online
for
twitter
and,
if
they're
willing
to
do
more,
that's
great,
but
if
not
at
least
it's
up
there.
That
does
not
require
a
huge
technical
overhaul
and
something
we
can.
We
can
afford,
and
will
I
mean,
as
sophie
demonstrated
increase
our.
F
Engagement,
something
like
learn
somewhere,
we
you'd
want
the
funded
effort
because
you'd
want
to
do
it
properly.
You
know,
I
think
it's
hard
to
do
something
like
that
on
that
big
scale
on
a
volunteer
effort
and
then
you'll
just
regret
starting
it.
You
know
because
it
won't
be
perfect,
but
I
I
don't
know
what
kinds
of
funding
and
it's
available
for
it's
very,
very,
very
hard
to
find
education
funding
for
this
sort
of
stuff.
It's
it's
very
much
a
lacking
area.
A
Well,
learnastropy.org
was
funded
mostly
by
space
telescope
and
I
think,
out
of
their
jwst
money
or
like
community
software
they've
attached
to
it.
I
think
don't
quote
me
on
that,
and
the
main
website
design
was
done
by
a
gsoul
student,
and
now
there
is
somebody
else
working
on
it:
who's
funded
a
different
institution.
F
I
do
know
a
few
examples
of
in
science,
communication
and
science
education.
I
have
some
colleagues
who
have
phd
students
who
work
on
this
kind
of
stuff,
so
they
develop
educational
resources
for
a
particular
research
project
and
then
they
actually
study
the
impact
it
has
on
their
audience
and
do
a
lot
of
evaluation
and
science
education.
So,
but
that
would
be
an
academic
sort
of
effort.
You'd
have
to
find
someone
who'd
be
interested
in
supervising
a
shoot
for
four
years
or
whatever,
but
there
are
some
weird
ways
of
doing
these
sorts
of.
A
A
I
just
jack
opened
a
issue
on
the
main
repo
entitled
learn.somepipe.org
the
other
day,
and
I
just
moved
that
to
the
senpai
project,
repo
and
added
a
few
more
details.
If
anybody
else
has
things
that
they
think
are
worth
noting
down
from
this
discussion,
please
put
it
there.
I
will
throw
the
link
in
the
matrix
channel.
A
Hey,
as
for
the
search
thing,
I'll
message,
yeah
open
an
issue
and
message
one
about
some
time,
but
the.
A
A
C
C
A
C
C
So
we
bridged
this.
We
came
over
this
topic
last
time
and
I
think
the
notes
were
what's
the
I
am
now
confused.
I
have
too
many.
C
Yeah,
that's
great
so
we've
reached
this
last
time
and
I'm
going
to
discuss
quickly
what
a
current
documentation
looks
like
what
the
feedback
I've
heard
is
and
some
cases
for
how
some
of
the
most
popular
projects
that
I
personally
use
day
to
day
and
my
work
look
like
and
whether
or
not
we
need
to
consider
not
necessarily
or
I've
used.
The
word
overhaul
here,
but
I
think
overall
was
probably
a
bit
more
dramatic
than
I
intended
it
to
sound.
C
N
C
Got
the
entire
content
just
dumped
right
on
the
front
page
and
obviously
that's
kind
of
useful,
because
you
know
you
can
find
in
search
and
you
have
access
to
the
subcategories
without
necessarily
having
to
click
on
the
sidebar,
which
obviously
would
expand.
C
A
Guides
table
of
contents
in
the
index
page
yeah,
I
think
so.
C
You've
got
the
user
guide
that
obviously
takes
takes
someone
new
I've
actually
never
read
it
myself
through,
though
the
installation
required
the
installation
part.
I
have
rewritten
so
installing
senpai
a
quick
tour
of
senpai
almost
like
I
guess
you
know
an
astropyte
learn
in
kind
of
I'm
probably
not
quite.
C
Basically
walking
through
someone
the
core
kind
of
overview
feature
the
core
showcase
features
of
each
part
of
somebody.
Obviously,
this
is
probably
should
be.
Probably
that
should
probably
be
delayed
walks.
You
through
you
know
how
to
use
fido
lots
of
examples,
each
part
of
fido
with
hilly
viewer
database
map
and
time
series,
plotting,
etc,
etc,
etc,
and
probably
should
we
really
move
that
should
that
be
in
the
user
guide,
and
obviously
this
cheat
sheet
that
I'll
come
back
to
and
I'll
come
back
to
in
future.
C
Obviously
we
have
a
code-
api,
that's
auto
generated,
build
time
and
probably
the
most
the
best
part
of
I
guess
I
could
say
this
without
any
controversy.
The
best
part
of
the
documentation
is
probably
our
example
gallery,
which
has
these
very
broad
categories,
and
I'm
going
to
bring
up
some
comments
that
I've
heard
you
know
albert
will,
I
think,
albert
might
repeat
them
later,
but
I'll
bring
him
up
as
well.
C
Oh,
we
should
rerun
this.
This
is
the
latest
anyway,
that
showcases,
you
know,
I
guess
how
to
do
things
in
senpai.
A
lot
of
the
examples
are
deliberately
short,
shorter
run,
and
I
guess
that's
you
know
what
you
want
from
an
example
is
that
it
does
one
thing
and
obviously,
when
we
come
to
the
concept
of
tutorials
and
like
the
workshop
meetings,
they
are
longer
form
pieces
of
code.
C
Obviously
the
example
gala
the
developer
guide,
but
that's
just
that's
just
for
us,
so
that's
probably
never
gonna
need
to
be
rerun
most
of
the
time,
so
these
are
probably
the
most
common.
I
would
say
these
are
the
things
I've
personally
heard
and
stored
in
my
head.
I've.
Never!
I
haven't
actually
asked
anyone,
so
if
you
have
other
things
that
they've
heard
about
how
bad
a
documentation
is,
I've
heard
people
can't
discover
things
I
normally
I
mean.
A
A
A
Seriously,
combining
the
two
things
together
and
I
think
I
don't
know
I
mean
it's,
oh
wow,
I
don't
think
our
exact.
I
don't
think
our
user
guide
is
great,
but
I
definitely
think
it
performs.
So
if
there's
a
section
where
we
want
to
say
download,
download
a
thing
and
then
map
the
thing,
that's
working
well
and
then
having
it
laid
out
all
in
one
go
like
that
is
not
necessarily
great.
C
Good
times-
and
this
is
oh-
I'm
surprised-
I
guess
the
astropod
already
have
a
gallery
because
they
have
the
learn
yeah.
So
I'm
going
to
say,
obviously
I've
heard
that
I'm
going
to
put
up,
and
obviously
we
have
had
google
jsoc
students
who
have
said
that
they've
actually
preferred
astropy's
layout
of
the
combine,
narrative
and
api
is
that
they
see
the
example
how
to
use
it
and
then
at
the
bottom
is
the
api
itself.
C
Obviously
a
gallery
organization,
a
gallery
is
split
into
broad
categories,
but
the
problem
is
that
some
of
them
obviously
overlap
categories,
because
you
know,
for
example,
acquiring
data,
but
there
are
maps
here,
so
you
know.
Is
it
really
a
map
example
a
day?
For
example,
you
know
we've
had
to
create
categories
ad
hockey
as
required
to
try
and
split
off.
You
know
the
aim
of
that
example
and
there's
no
way
to
filter
this
there's
no
way
really
to
search
it.
C
There's
no
metadata
to
you
know,
slim
down
to
find
an
example
that
you
require.
Obviously,
that's
partly
the
issue
that
soon
carried,
I
think,
has
that
been
built
and
that
would
require
an
upstream
they'll
definitely
require
an
upstream
change
in
the
gallery
which
someone
will
have
to
like
check
if
they
want
implement
it,
and
that's
probably
not
it's
probably
not
a
trivial
thing
to
be
doing,
and
so
I
guess
my
last
particular
point
was,
as
we've
already
touched,
is
that
obviously
getting
feedback
on
documentation
is
difficult?
C
You
know
you
get
the
the
general
just.
Oh,
they
didn't
like
it
for
this
or
that.
But
then
you
know
is
that
enough
really
to
be
sitting
down?
Okay,
we
should
be
burning
down
the
documentation
starting
again,
is
it
not
necessarily
just
we
don't
know
what
necessary
we
need
to
tweak,
and
I
guess
something
to
think
about
in
future-
is
how
do
we
actually
wanna?
How
do
we
get
feedback
on
our
documentation
on
our
example
gallery
in
a
way
that
allows
us
to
iterate
into
a
better
version
of
them.
A
E
G
And
also
when
I
guess,
users
aren't
experts
in
how
to
structure
documentation,
so
I
think
actually,
it's
kind
of
to
be
like
you
can
a
user
can
identify
when
it's
hard
to
find
something,
but
won't
necessarily
know
how
to
restructure
it.
So,
like
you,
wouldn't
expect
undergraduate
students
to
be
telling
universities
how
what
teaching
mechanisms
are
good,
for
example.
G
G
F
And
two
things
that
come
to
my
mind,
you
could
add
a
little
like
request
button
that
links
to
maybe
a
post
on
discourse
where
people
could
like
post
suggestions
of
I'd
like
to
see
this
example.
So
that
could
be
a
way
long
term
to
try
and
get
people
to
to
make
suggestions.
Of
course,
you'd
have
to
encourage
people
to
do
that,
and
but
one
thing
that
I
found
really
useful
in
the
past
and
something
we
did
last
week,
which
is
why
it's
in
my
mind,
it's
just
a
user
group.
F
If
you
want
some
like
first
feedback
of
how
people
have
found
us,
you
could
get
a
group
of
you
know
just
some.
Some
post
grads
from
your
own
research
groups
or
post
docs
or
you
know
just
a
variety
of
different
people-
who'd-
be
willing
to
just
chat
for
an
hour
on
a
zoom
meeting
or
a
jitsie
or
whatever,
and
get
them
to
have
a
look
at
it
beforehand
and
and
give
you
some
direct
like
discussion
feedback
on
it.
F
But
it
requires
time
and
effort
to
do,
of
course,
as
everything
but
user
groups
are
very
handy
for
that
sort
of
thing.
It
might
just
give
you
a
general
direction
of
where
to
go
next.
C
So
when
I
think
about
what
we
should
do
to
documentation,
I
guess
in
my
mind
I
you
know,
after
after
my
postdoc
didn't
pan
out
the
way
I
wanted
it.
I
became
a
python
developer,
a
python
web
developer
and
I
got
intimate
with
django
x-ray,
fast
api
and
I'm
and
I'm
going
to
basically
go
through
the
api
documentation
on
a
surface
level
about
how
they
structure
and
how
they
present
their
information,
as
necessarily
as
a
way
to
demonstrate
what
we
could
potentially
do.
C
It
yeah,
so
that's
what
that's
what
I'll
be?
Comparing
I'm
not
necessarily
talking
about
the
structure
because,
for
the
most
part
the
structure
is
the
same.
They
have,
they
have
examples,
they
have
tutorials.
They
have.
You
know
the
code
api,
obviously
because
you
need
that
code
somewhere,
and
so
I'm
not
going
to
talk
about
azure
buyers,
documentation
because
I
yeah,
I
don't
really
use
it
a
lot,
so
I'm
gonna
start
with
django's
first.
C
So
this
is
the
landing
page
for
django's
documentation
and
it's
obviously
a
very
different
style
completely
to
how
we
like
how
our
website
even
looks
like.
But
you
can
see
the
landing
pages
first
steps
I
think,
links
to
overview
a
tutorial
advanced
tutorials,
getting
help.
It
then
even
explains
like
how
the
documentation
is
itself
laid
out
and
then
goes
into
django
itself.
The
model
layers,
the
view
layers,
the
templates,
the
forms
and
then
et
cetera,
et
cetera,
and
this
is
a
very
long
page
because
it's
a
lot
of
django
and
at
the
bottom.
C
I
guess
thought
about
since
is
that
a
lot
of
design
choices
that
we
make
in
some
part
or
have
made
necessarily
maybe
without
any
thought,
django
try
to
write
them
down,
but
you
know
in
a
way
that
you
know
new
people
can
come
back.
Okay.
Why
was
this?
Why
was
this
done
like
this,
and
there
is
at
least
some
text
you
may
not.
You
may
not
agree
with
the
the
text
you
may
not.
C
K
I
I
think
one
of
the
key
sections
from
that
first
page,
though,
is
that
one
how
the
documentation
is
organized.
We
don't
even
tell
people
where
they
should
go
to
look
for
certain
things,
which
I
I
think
is
one
of
the
biggest
hurdles.
Is
that,
like
you,
have
to
tell
people
where
they
should
be
looking
because
they
can't
know
a
priori
where
things
are
also
the
fact
that.
C
C
A
C
Style
because
there
are
parts
of
the
user
guide
that
I
think
aren't
necessarily
useful
within
the
user
guides
yeah.
Obviously
you
have
to
remind
you
have
to
remember
that
I'm
going
to
show
django's
recommendation
and
I'm
going
to
show
fast,
fast,
api's
documentation
that
they
are
obviously
a
library.
C
You
know
it's
a
project
that
you
you
import
and
you
create
a
very
specific
kind
of
program
with
it
you're
not
unlike
say,
pandas
or
x-ray.
You
aren't
creating
a
website.
You
know
not
creating
a
a
service
that
other
people
can
use,
so
obviously
their
documentation
is
laid
out
in
whether
okay.
This
is
how
you
write
a
django,
app
you're,
not
gonna.
You
know
we're
not
gonna.
Be
writing
an
example
like
how
do
you
you
know,
write
a
senpai
app,
because
that's
not
how
you
know
people
would
use
senpai.
C
C
The
other
thing
I
was
going
to
talk
about
is
x-ray,
so
I'm
going
to
talk
about
this
in
a
second.
Is
that
so
our
senpai
theme
came
about
through
a
gtop
project?
Is
it
three
years
ago
now
was
that
project
2016?
C
C
You
know
a
theme
layout
how
we
should
to
do
everything,
and
one
of
the
issues
is
that
the
underlying
theme
is
now
quite
old.
It's
a
sphinx,
it's
a
things.
Project
frame,
that's
push
up
three,
it
hasn't
seen
an
update
since
2019,
and
so
I'm
currently
in
the
process
of
working
on
replacing
that
with
the
pi
data
theme
underneath,
which
is
what
xray
looks
like
and
pandas
as
well.
C
Years
ago,
so
this
this
will
change
in
some
form
in
future,
but
obviously
xray
have
a
slightly
different
landing
page,
as
you
can
see
from
here,
though,
I
wonder
if
I
should
look
at
the
latest,
just
probably
look
at
the
latest
just
in
case
it's
obviously
just
getting
a
touch,
their
focus
and
their
history,
and
then
you
get
right
stuck
into
their
okay,
that's
a
very
short
guest
article.
It's
not.
C
So
but
obviously
from
the
sidebar,
you
can
see
the
layout
here
so
obviously
a
clear
denomination
between
developers
and
contributors.
The
community
like
links
to
like
a
stack
overflow
job,
but
one
of
the
things
I
wanted
to
point
out
was
the
they
actually
directly
link
to
tutorials
and
videos
that
we
have
on
our
so
on
our
project,
page
on
our
which
is
separate
from
the
documentation,
but
you
can
obviously
get
to
it.
C
We
have
presentation
posters
on
the
org
website,
but
they
aren't
linked
to
on
the
project
on
the
actual
documentation
website
and
whether
or
not
we
want
to
control
that
you
know.
Maybe
we
want
to
merge
the
two,
but
they
have
a
very
similar
to
us.
They
have
to
get
started
which
I
basically
just
talked
about
installing
and
then
the
user
guide,
which
I
think
has
a
nice
terminology
page
and
then
breaks
it
down
the
same
way
that
you
know
our
musical
works.
C
Yeah,
so
I
think
this
is
the
part
I
wanted
to
focus
on,
for
x-ray
itself
was
that
they
have
a.
How
do
I
page
and
I'll
say?
Okay,
you
want
to
add
a
corner
value.
You
want
to
change
the
order,
you
want
to
stack
them.
You
want
to
remove
this,
you
want
to,
you,
know,
shrink
down
and
they
have
an
action
and
the
functions
that
you
would
probably
tweak
or,
of
course,
or
tweak
the
functions.
C
You
would
call
methods
you
would
call
whatever
you
would
want,
and
some
of
them
aren't-
and
obviously
we
have
something
like
this,
but
for
s
for
basically
translating
from
idle
to
python,
and
this
is
probably
what
I
would
say
that
we
need
to
steal
from
x-rays.
Documentation
is
a
how
do
I
page
it
goes.
Okay,
what
are
the
most
common
features
and
we
just
like
put
the
string.
So
if
someone
were
to
search
differential
rotation,
they
might
actually
be
linked,
they
might
the
first
step
might
actually
be
do.
B
One
of
these
is
like
much
easier
than
say,
having
like
a
whole
gallery
example,
for,
I
think
one
recent
example
I
think
laura
brought
up
a
while
ago,
is
that
people
weren't
aware
of
like
how
to
save
something
to
just
to
a
fix
file
right
or
like
dumping,
a
map
to
a
fixed
file,
because
because
of
how
our
documentation
is
structured,
that
wasn't
necessarily
obvious
and
then
obviously
that's
just
one
call.
You
know
a
map.save
call,
that's
not
necessarily
worthy
of
a
whole
example
gallery
entry.
B
A
E
K
One
thing
that
was
that
for
x-ray,
because
they're
very
it's
a
very
focused
package
on
doing
things
with
you
know:
n-dimensional
arrays
it'll
be
a
little
bit
harder
for
somebody,
because
it
covers
a
much
broader
range
of
things,
not
impossible.
It's
more
difficult
to
be
as
concise
with
the
questions
like.
K
A
A
L
A
C
C
And
so
obviously,
I
want
to
say
that
I
think
another
thing
about
http
related
projects
on
there,
so
I
think
one
of
the
issues-
I
guess
when
I
when
I
look
at
the
main
website,
is
that
we
have
obviously
a
strict.
You
know:
separation
between
the
project
and
our
documentation,
we're
obviously
xray
like
okay,
you
know
if
you
want
to
get
find
a
list
of
affiliate
packages,
you
have
to
go
to
the
projects
website
to
find
them,
but
obviously
x-rays
like
okay.
It's
just
right
here.
C
A
C
A
C
C
C
Yeah,
so
we're
gonna
still
have
a
top
bar
and
we're
gonna
probably
override
okay.
This
is
the
wrong
page
to
scroll
down
on,
but
yeah.
I'm
not
gonna
talk
through
the
theme,
because
I'm
gonna
do
that.
So
my
last
example
is
pandas
and
balance.
It
takes
a
completely
different
route
to
all
the
two
we've
seen
so
far,
though
more
in
line
to
django,
but
as
you
see
from
their
landing
page,
they
have
four
big
buttons:
four
big
cards
with
buttons.
Underneath
them,
you
have
get
started.
You
get
the
user
guide.
C
C
So
obviously
panelists
are
very
different,
but
a
different
layout
and
a
lot
of
like
collapsible
html
elements.
That's
like
okay,
how
do
you
install
like
this
sales?
Don't
condon
the
sales
right
in
pandas,
obviously
with
in-depth
and
then
you
know
an
intro
with
links
to
tutorials
and
then,
if
you're
coming
from
any
of
these,
you
know
this
is
how
you
might
you
know,
compare
with
r,
for
example,
and
much
like
our
idl
page,
which
maybe
could
do
with
updating.
If
I'm
honest,
maybe
there's
a.
C
B
Setting
it
up
as
a
how
do
I
versus
as
as
opposed
to
a
you
know,
I
yelled
to
python
translation.
I
think
how
how
do
I
search
for
the
same
purpose
except
lifts
it
up
like
a
like?
You
know
one
level
higher.
So
how
do
I
do
this
task
in
python
rather
than
how
do
I
you
know,
and
I
think
it's
sort
of
the
same
question
that
may
be
phrased
a
bit
more.
C
Yeah,
so
once
you
get
past
the
opening
part
of
panda's
docks,
it
then
turns
into,
for
example,
our
exam.
Our
user
guide
is
like.
There
are
examples,
ten
minutes
to
pandas
intro
to
basically
most
of
their
structures,
cetera
et
cetera
and
their
api
documentation
is
just
their
api
documentation,
there's
no
narrative
docs
with
it.
C
So,
in
terms
of
like
the
overall
structure,
we
follow
how
other
projects
do
it
and
I
don't
think,
there's
necessarily
a
change
to
we
don't
necessarily
might
need.
We
don't
think
we
necessarily
need
to
merge
our
api
documents
with
a
user
guide
or
parts
of
the
user
guide
which
need
to
lock
out,
I
think,
primarily,
on
a
layout
and
presenting
on
the
first
instance
the
basically
our
index
pages
need
to
be
tidied
up
to
be
useful.
C
C
C
And
this
is
all
in
this
is
all
marked
down.
This
uses
mk
dock
so
far,
calm.
I
Hi
yeah
just
one
quick
question:
where
do
you
guys
actually
stand
on
having
examples
within
the
code.
I
Because,
when
we've
been
working
on,
transform,
we've
had
a
couple
of
small
examples
actually
within
the
the
code
itself.
How
do
you
feel
about
that?
Because
it
means
you
always
have
an
example
transported.
I
C
Sorry,
so
we
let
me
we,
we
don't
enforce
it,
but
we
do
have
that
in
our
try
to
think
of
one
that
actually
has
a
useful.
C
Yeah,
so
I've
just
opened
open,
transform
with
sunsender,
which
has
an.
A
I
Okay-
and
I
I
I-
should
I've
been
a
bit
lazy-
I
haven't
actually
looked
through
in
terms
of
your
kind
of
your
test,
suite
itself,
how
many
kind
of
like
tests
do
you
normally
kind
of
implement
for
each.
I
I
A
D
D
A
Up
to
2
300
unit
tests
collected
by
pi
test
when
you
run
it
now,
so
it's
a
decent
number.
A
I'm
trying
to
find
an
example
hang
on
package
jsoc
each
doc
test
item
is
collected
as
a
separate
item.
A
A
K
But
I
think
we
talked
about
this
last
time
as
well.
Is
that,
while
all
of
the
content
that
is
there
is,
is
actually
pretty
good,
that
there
is
some
stuff,
that's
kind
of
missing,
and
that's
like
the
the
the
thing
that
neville
showed
like
the
philosophy
like
so
like.
What?
What
is
a
map
and
what
it's
supposed
to
do,
but
not
in
terms
of
like
the
api,
but
in
terms
of
like
what
was
it
intended
to
do
in
the
first
place
or
like?
Why.
K
And
what
are
they
supposed
to
do
but
again,
like
not
necessarily,
you
know
code,
but
just
like
yeah
kind
of
a
high
level.
I
mean
this
is
what
it's
supposed
to
do.
This
is
what
it's
not
supposed
to
do,
because
a
lot
of
times
people
you
know
they
as
we
claim
misinterpret
what
math
does
and
doesn't
do
and
don't
know
exactly
what
chords
it's
doing.
Don't
do
and
some
of
that's
outside
the
scope.
K
At
some
point
you
know
it's,
you
know,
go
and
read
the
papers
and
stuff
like
that,
but
some
of
it,
I
think,
could
be
handled
by
like,
like
more
in-depth
articles
about
a
piece
of
you
know.
C
K
C
C
K
Yeah,
I
mean
I
don't
even
know
if
it's
philosophy
there's
just
like
longer
articles
or
longer
bits
of
information
that
don't
belong
in
the
code
record,
because
it's
not
code,
they
don't
belong
in
they're,
getting
started
because
it's
not
getting
started.
It's
kind
of
like
a
separate.
I
think
in
one
of
the
things
that
we
linked
to
before
there
was
like
four
sections:
there's
like
getting
started,
how
to's
and
then
like
articles
or
discussions,
and
then
a
code
record
was
like
the
other
section.
C
K
B
Well,
we
have,
in
the
I
mean
in
the
user
guide,
we
have
the
data
type,
we
have
a
whole
date
type
section
and
we
have
a
map
and
a
time
series
section,
and
I
mean
it
does
sort
of
go
through.
What
a
map
is,
what
are
the
one
of
the
different
parts?
What's
the
what's
the
point
I
mean
it
could
probably
use
a
bit
more
but
yeah
so
this.
Well,
this
is
the
tour.
B
C
A
C
A
E
I
don't
know
people
just
think
it's
an
array
and
then
you
just
change
your
axes
around
it.
Rather
than
having
to
think
about
what
your
observer
is
doing,
or
you
know
even
the
fact
that,
like
you
know,
people
are
like.
Why
am
I
getting
a
bug
and
I'm
trying
to
convert
this
coordinates
into
something
else,
because
you
need
an
observer
and
I
don't
know.
D
I
think
this
potentially
speaks
to.
You
know
the
weakness
of
modularization,
although
you
know
we've
pitched
its
many
advantages,
we're
trying
to
explain.
Like
I
yeah
I
mean
I
think,
sort
of
people
like
to
go
back
to
something
we
said
earlier.
It's
like
I
want
to
work,
for
example
like
these
are
all
like
baby
examples
and
like
how
do
I
actually
do
my
work
and
people
like
they
don't
care
that
it's
senpai
or
nd
cube
or
astro
pi?
You
know,
that's
like
I
have
aia
data
or
I
have
whatever
it
is
data.
D
N
C
Well,
I
think
this
brings
up
the
kind
of
the
concept
of
what
is
an
example
and,
to
me
an
example
should
be
a
small.
You
know
when
someone
says
I
want
a
workflow
exam.
You
know
I
want
to
work
through
like
a
workflow
example
like
that's
a
tutorial
like
that
is
like
a
longer
form
piece
like
obviously
when
I
say
we
shouldn't
put
them
in
a
documentation,
but
I
think
that's
a
separation.
That
seems
slightly
arbitrary
to
example
gallery,
but
I
feel
like
example,
I
should
do
like
one
thing
and
that's
like.
D
D
D
I
think
the
majority
of
users
don't
care
that
it's
in
core
or
not
in
core,
they
say
or
something
it
doesn't
have
good
enough
documentation,
because
it
doesn't
tell
me
how
to
you
know,
prep
ai,
because
that's
now
an
ai
pi
or
it
doesn't
tell
me
how
to
get
the
ghost
temperature,
because
that's
now
in
some
kit
instruments
like
I
said,
I
don't
have
a
solution,
and
I
recognize
that
the
the
very
sensible
reasons
why
it
has
been
so
modularized,
oh
yeah,
I'm
just
complaining,
yeah.
E
But
I'm
sorry,
just
there's
like
just
a
coordinate
stuff
like
the
coordinates
documentation
is,
is
really
clear
when
you
understand
the
coordinates
framework
right,
but
there's
like
a
missing
bridge
between
it
and
needs
like
an
instructor
page
to
you
know
how
does
the
coordinate
stack
work
within
senpai
and
how
does
the
asterisk
work
with
like
solar
frames
and
just
in
introductions?
Then
people
appreciate
it
and
then
can
use
it
rather
than
being
like
this
is
really
complicated.
But
it's
once
you
kind
of
understand
the
kind
of
philosophy
change
behind
it.
E
It
becomes
a
lot
more
clear.
But
again,
where
does
that
lie?
Is
the
question,
but
I
think
there's
a
is
it
kind
of
scoped
I
think,
sort
of
the
place
for
that
to
go
somewhere.
O
E
D
Nothing
to
do
with
mdq,
it's
talking
about
astropywcs
or
wcs
like
in
general,
but
just
like
something
somebody
said
about
x-ray
is
that
ndq
is
fairly
like
directed.
You
know
like
some
pi
is
a
lot
more
sprawling,
so
there
are
much
greater
consequences
for
having
that.
Taking
that
same
idea
and
it's
putting
everywhere
in
some
fight,
but
I'd
say
that's
what
users.
D
D
N
Want
to
build
on
laura's
perspective
here
that
right
I
mean
I'm
coming
in.
I
haven't
done
any
sunpi
development,
I'm
moving
from
pbl
into
the
python
arena
and
the
biggest
problem
I've
had
as
a
noob
user
is
the
plethora
of
of
modules
that
are
all
sort
of
interrelated
have
overlapping
functionality.
N
It
has
not
been
clear
at
all
when
I
want
to
reach
for
an
astro
pi
object
or
a
sunpi
object,
and
so
having
these
these
larger
blocks
of
documentation
that
tie
in.
Why
bother
with
a
map
as
opposed
to
you
know,
an
astro
pi
fits
object
that
will
really
help
speed
adoption
rates
and
lower
the
the
entry.
E
K
Yeah,
oh
yeah,
I
I
definitely
think
it
would
be
good
to
ask
people.
I
think
it's
also
good
to
decide
where
it
goes,
because
I
don't
think
it
really
belongs
in
code
reference,
because
it's
not
really
that
so
yeah
you
might
want
to
have
a
separate
place,
but
I
guess,
if
you
have
content,
you
can
just
move
it
around
afterwards.
I
I've
got
a
question
who,
who
who
writes
the
examples?
That's
that's
one
thing
and
I'm
going
somewhere
with
this.
I
I
Okay
and
something
else,
one
thing
I
was
going
to
say
as
a
kind
of
I'm
kind
of
learning,
as
I'm
going
with
senpai
feel
free
to
use
me
as
a
sounding
board.
If
you've
got
a
new
example,
you
want
somebody
who
has
got
nothing
to
do
with
it
to
look
at
and
just
see
if
it
makes
sense,
feel
free
to
shoot
me
a
message
and
I'll
happily
see.
If,
if
I
can
follow
it.
I
Like
when
it
comes
to
these
example
pages,
I'm
not
sure
whether
you
you're
never
sure
when
someone's
developing
these
pages,
whether
they're
very
specifically
chosen
or
because
you
could
imagine
a
scenario
where
you
can
get
thousands
and
then
suddenly.
The
page
is
not
useful
anymore,
because
there's
too
many
examples,
we're.
D
How
much
kicking
and
screaming
resistance
would
there
be
in
trying
to
come
up
with
something
that
craig
suggested
that,
like
it's
something
maybe
along
the
lines
of
like
what
are
the
main
things
that
somebody's
actually
built
on?
And
I
know
that's
not
our
package
and
we
don't
maintain
it.
We
don't
do
it,
but
we're
going
to
write
like
you
know,
an
article
on
it
to
you
know.
So,
when
people
land
in
senpai
within
a
click
or
two,
you
know
they
can
sort
of
get
readings
like.
D
I
Do
you
run
the
risk
of
running
into
like
deprecation
issues?
Yes,
sir,.
D
D
D
A
D
D
You
know
folks
in
this
particular
example
like
what
wcs
is
and
how,
like
the
astropy
pack
of
wcs
package,
then
like
implements
the
concept
of
wcs,
so
that
people
are
just
like
they
see
how
the
pieces
fit
together.
I
don't
want
to
have
to
like
maintain
it,
but
I
also
appreciate
that
I.
D
N
N
Even
you
know
if
there
were
a
catalog
somewhere
of
these
are
the
packages
we
know
about,
and
these
are
the
roles
they
might
fulfill.
This
is
why
we
like
this
one
and
not
that
one
and
that
would
go
a
long
way
toward
getting
noobs
sort
of
ramped
up.
N
If
you
think
about
like
solarsoft,
you
get
a
package
deal
right,
everything
is
within
idl
and
it's
all
sort
of
allegedly
mutually
compatible,
and
somebody
coming
into
the
python
community
right
is
faced
with
with
this
wild
jungle
of
things
that
work
together,
and
there
are
sort
of
there
are
clusters
of
packages
that
work
well
together,
but
but
it
it's
not
all
guaranteed
to
integrate,
and
literally
just
having
a
directory
of
these
are
the
things
that
that
stuart
likes
is
is
a
big
step
in
the
right
direction.
A
A
N
A
A
D
K
Yeah
we
worthwhile
making
like
a
really
small
pipeline
package
that
just
installs
a
whole
bunch
of
versions
of
rfid
packages
and
stuff
that
we
know
work
together
in
a
nice
way
so
that
you
know
they
could
install.
I
don't
know
what
you'd
call
it
but
yeah
it
just
basically
installs
some
pi.
You
know
ndq
map
but
live
sci-fi,
all
those
sort
of
good
things.
K
C
C
Yeah,
so
I
I
put
a
bullet
point
of
things
that
I
thought
based
off
of
review
that
we
could.
This
is
when
I
was
writing,
think
their
landing
page.
We
obviously
have
some
like
get
help
text
on
our
website
that
isn't
in
our
documentation.
That,
maybe
should
be,
I
think,
maybe
advertising
our
workshop
material
somewhere
in
the
documentation.
C
C
A
C
A
C
A
Okay,
we're
down
to
our
last
like
eight
minutes
or
something
so
just
people
have
other
things.
They
want
to
talk
about
specific
to
what
we've
chatted
about
already.
C
I
guess
if
we
could
talk
about
that,
if
people
don't
talk
about
documentation,
so
I
guess
the
my
point
is
that
I've
been
out
of.
I
mean
I've
never
had
yeah.
I.
A
In
this
section
right-
and
I
I
don't-
I'm
not
convinced
it's
worth,
starting
actually
with
seven
minutes
left
to
go.
I
think
we
should
bump
it.
C
A
A
C
A
A
Q
I'll
just
throw
in
that
the
roses
proposal
is
still
pending.
I
guess
five
plus
months,
no,
no
announcement.
K
And
these
these
google
server
docs
as
well,
which
we
hastily
and
it's
unsuccessfully
applied
for
last
year,
but
that's
something
that
we
could
try
and
put
a
bit
more
time
into
it.
When's
that
october
or
something.
A
It's
true:
actually
it's
not
that
far,
probably
not
that
far
off
again.
Is
it
once
again,
we
ask:
how
is
it
2021.
A
I
mean
there
is
github
sponsors
if,
in
all
seriousness,
if
you
do
want
to
throw
dollars
senpai
we
take,
we
do
take
donations
either
through
github
sponsors
or
directly
to
non-focus
and.
A
G
A
B
N
A
I
A
I
don't
know
about
like
the
beginning.
Well,
I
think
my
understanding
was
the
beginning
of
solar,
software's,
all
instrument,
funding
and
okay,
but
it's
been
a
long
time
since
I've
spoken
to
anybody
about
that.
So
I
could
could
be
wrong.
I
It
might
be
interesting
to
see
if
whatever
missions
are
funded
in
future,
whether
there
can
be
a
I
don't
know,
a
senpai
contribution.
N
A
Well,
I
mean
there's
if
you
can
prove
that
a
lot
of
research
and
mission
instrument
teams
are
using
it.
It
does
become
an
interesting
question
for
the
funding
agencies
on
how
they
want
to
fund
that
right.
What
is
it
nasa
gives
you
money
every
few
years
when
you
ask
for
it
and
then
all
the
science
uses
it
or
is
it
yeah
a
few
thousand
dollars
on
each
grant
proposal
that
goes
in?
You
know.
I
But
I
I
was
a
will.
He
was
talking
about
the
cable.
Perhaps
that
is
the
way
to
go
because
it
from
what
I'm
hearing
from
this
discussion
is:
there's
not
a
kind
of
legitimate
funding
source
to
actually
go
to
to
get
this
actually
funded.
D
K
And
we
don't
there's
nothing
on
the
european
side
either
like
there's
no
equivalent
of
the
hde
grants.
Yes,
isa
does
have
money
and
with.
A
I
think
yeah
okay,
as
the
last
thing
we're
out
of
time.
So
I
think
one
of
the
interesting
questions
is
how
to
get
contributions
at
all
money
or
otherwise,
from
different
sources.
Mission
teams,
researchers
right,
whether
we're
talking
about
money
or
time
is
almost
irrelevant.
It's
a
real
struggle
to
get
the
priorities
of
those
people
aligned
with
the
priorities
of
the
senpai
team,
who
are
currently
at
least
almost
entirely
volunteers,
so
cool!
Well,
thank
you.
Everybody
for
coming
and
tomorrow
we're
talking
about
map
entirely
cool
today.
A
Tomorrow
is
a
day
about
map
and
array
data,
we're
going
to
talk
about
metadata
and
nd
cube
and
how
how
map
is
going
to
be
upended
by
ndcube.
D
A
D
Cool
I'd
like
to
just
also
specifically
thank
matt,
and,
although
he's
just
left
craig
for
showing
up,
because
I
think
you
know
influence
in
this
perspective-
that
they
bring,
you
know,
is
really
valuable
and
that
we're
not
just
all
talking
in
a
fishbowl.
A
I
I
Would
be
really
interesting
for
us
because
we're
having
to
combine
with
punch
because
we're
having
to
combine
in
our
like
nominal
data
pipeline,
we
have
to
combine
four
images
for
four
data
sites
for
arrays
and
we
were
originally
looking
to
kind
of
build
our
own
little
bit
of
code
to
do
this
and
obviously
ndcube
is
perhaps
going
to
be
the
way
we're
going,
but
we
need
to
make
sure
that
it
can
carry
non-wcs
extra
metadata
and
things
like
that.
So
for
me
this
is
going
to
be
the
interesting
one.
A
The
focus
for
tomorrow's
session
is
specifically
it's
probably
a
bit
more
technical
than
general
on
indie
cube.
We're
gonna
be
talking
about
yeah.
C
C
D
Was
just
gonna,
we
should
make
maybe
separate
to
this
meeting.
You
know
if
that's,
if
that's
what
you're
excited
about,
then
either
myself
or
steward,
or
both
should
make
time
to
have
like
a
very
specific
conversation
with
you
guys
about
that,
because
I'd
love
to
get
more
specific
feedback
from
instrument
teams
on
what
your
needs
specifically
are
so
sure.
I
Just
a
question
to
think
about
for
tomorrow
and
it-
and
you
may
not
have
an
answer
for
this
and
we're
going
into
producing
our
software
soon,
and
I
just
heard
you
say
that
you're
perhaps
going
to
be
releasing
ndqv
2.0
yep.
That
scares
me
a
little
bit
because
I
want
to
know
how
this
is
going
to
fit
in
with
the
production
of
the
pipeline,
because
if
the
release
of
this
and
then
all
the
debugging
means
it's
going
to
be
fully
released
in
two
years,
then
then
that's.
I
A
A
I
Yeah:
okay,
that's
excellent,
but
yeah
that
was
gonna
be
one
of
my
questions
for
tomorrow,
but
yeah
well.
Look
forward
to
it
then.