►
From YouTube: SunPy Coordination Meeting 2022 - Monday
Description
Participate in the chat and the call here: https://openastronomy.element.io/#/room/#sunpycoordinationmeeting:openastronomy.org
A
D
A
E
D
A
A
Inherit
from
ndq
yeah
there's,
definitely
a
lot
of
like
2D
image,
specific
tools
but
like
I,
think
the
first
steps
with
that
yeah
is
make
it
inherent,
and
then
we
can
sit
down
and,
like
think
about
what
happens,
oh
right,
so
that
for
me,
would
that
be
top
of
my
list.
That's
what
I
wanted
reading
is
people
with
the
time
together
to.
F
F
Yeah
like
at
an
actual
start
on
that
period,
I'm
talking
about
it,
yeah,
yeah,
I
kind
of
start
to
a
point
where
you
could
like
actually
use
something
it
might
not
be
good
or
polished
or,
like
you
know,
ready
for
a
general
release.
But
you
know
actually
have
an
object
that
people
to
play
with.
You
know
in
a
happy
kind
of
way.
E
You'll
probably
go
right:
yeah
we
can
only
make
a
hacky
version
of
that
work.
On
Friday
yeah,
it's
going
to
be
inherent
delete,
a
bunch
of
map
profit.
You
know
right,
there's
a
link,
part
I.
Think
people
online
is
a
useful
conversation
because
like
well
I
did
this.
Eight
years
ago,
yeah
I
touched
it,
but
I
went
through
and
looked
at,
like
the
map
API.
If
you
hear
it
from
ND
Cube
and
you
just
strip
it
down
like
apart
from
visualization.
D
You
know
or
just
metadata
access
once
you
take
those
two
things
away:
there's
very
little
left.
You
know
so
I
think
that
would
be
a
good
discussion
to
have
as
well
I
think
your
metadata,
the
metadata
that
you
have
in
some
roster
talking
about
that
and
how
we
could
essentially
incorporate
that
into
that
too,
would
be
because
that
that
is
far
more
General
than
just
you
know
the
what
what
it's
used
for
in
sunrise
today,
yes,
I'm
glad
someone
else
brought
that
up
so
there's
interest
that
we
all
just
made.
Yes,
that's.
D
G
E
A
B
F
B
F
Like
I'm,
gonna
I'll
talk
a
little
bit
more
about
this
later
too,
but
the
like,
as
far
as
the
Affiliated
packages
goes-
and
this
is
this-
is
kind
of
goes
a
long
way.
What's
the
same
too,
it's
like
what.
How
do
we
sell
the
ability
to
packages
to
people
like
what
is
if
someone
like
someone
asked
me
at
the
the
test
meeting
a
couple
things
go
like.
Why,
like
what
is
the
the
bed
that
you
collect
becoming
an
Affiliated
package
and
I
think
I
sort
of
stumbled
through
like
why
you
know
well
like?
F
G
And
like
wealth
on.
I
Capable
with
you
know,
I
guess:
I
keep
saying
instruments,
because
that's
kind
of
the
potential
term
I'm
using
this
but
instrument
teams
coming
to
us
with
particular
needs
that
need
the
core
maintainers.
Should
you
know,
meet
those
needs
and
then
also
we
just
you
know,
ask
people
developing
Affiliated
packages
to
kind
of
practice
there.
You
know
to
use
things
like
the
package,
documentation
and
formatting
all
this
stuff,
so
I
think
coming
up
with
that
kind
of
sort
of
sales
pitch
for
lack.
I
Be
like,
like
a
specific
yeah,
so
there
is
liberals
so
to
speak,
yeah!
So
there's
a
metadata
object
that
I
wrote
in
Sun
roster
but
as
well
said,
it
is
very
much
not
specific
to
Sunrise
it's
in
there,
because
that's
the
use
case
that
it.
B
Was
written
for,
but
it
is
written
in
a
far
more
General
agnostic
way,
so
I
suppose
what
I.
A
A
I
But
I
think
yeah
I
think
having
like
an
object
or
an
API
that,
like
is
already
written
down,
hope
they
will
give
some
Focus.
B
B
A
Cool
should
we
go
around
the
room
first
time:
yeah
yeah,
so
hi,
I'm,
David,
stanstein.
G
A
In
London
and
Senpai
I'm
the
release
manager
and
maintainer
for
the
time
series
sub
package
and
the
visualization
sub
package,
hi
I'm
Marcus,
Hughes
I'm
a
lead
developer
for
punch
and
I'm
gonna
signed
by
new
car,
so
yeah,
so
I'm
Dan,
Ryan
I,
used
to
work
at
goddess
now
work
at
fhw
in
Switzerland
as
part
of
the
16.
For
those
you
know
all
through
that
time.
I've
sort
of
got
involved
in
Senpai
and
I
am.
C
The
lead,
developer
or
co-lead
developer
of
a
couple
different
Affiliated
packages,
the
main
one
that
we're
going
to
discuss
this
week,
I
think
is
MD
Cube,
but
also
Sun
raster
that
have.
B
C
More
involved
in
with
Iris
and
yeah,
so
you
know
I'm
sort
of
hit
as
a
representing
those,
but
also
been
involved
very
much
in
call
to
a
lesser
extent.
So
that's.
C
My
name
is
Shane
I'm,
a
postdoc
here
at
guys
and
also
part
of
the
16,
essentially
one
of
the
elite
Developers
for
Stix
processing
pipeline.
A
Also,
currently,
here
at
dollars,
we're
working
on
rewriting
solar,
monitor,
I
think
we'll
name
a
website
to
use
some
point
in
the
back
range,
so
some
use
cases
for
that
generally
just
complain
about
why
somebody
doesn't
do
what
he
wants.
I
wanted
to
pick
something.
A
All
right,
I'm,
I'm,
lucky
Ryan
I
help
the
scientists
still
work
with
Iris
news.
A
and
the
bill
does
everything
yeah
and
if
a
song
by
repo
exists,
nobody
else
has
committed
today.
A
C
Work
on
some
five
just
great
using
some
of
the
some
money
that
we
got
from
NASA
last
year,
which
hour
we'll
talk
more
about
later.
G
This
afternoon
so
I
work
on
Senpai
and
work
on
that
episode,
instrumentation
type
stuff
software
and
data
analysis.
Some
some
things
like
that
and
I'm.
Also
the
deputy
developer
project.
A
Work
for
my
own
company
in
the
UK
doing
software
various
scientific
software
development
stuff.
Most
of
my
time
is
spent
working
on
decas
data
and
user
tools.
But
as
of
like
a
few
weeks
ago,
I
now
get
paid
one
day
a
week
to
write
some
high
infrastructure
stuff.
So
that's
nice,
I
am
yeah,
sometimes
the
developer
and
have
been
around
for
a
long
time,
although
doing
less
and
less
it
seems.
J
Do
this
I
think
I'm
going
to
have
to
yell
out
names
of
people
who
are
on
the
call,
let
them
introduce
themselves
if
they
want
to
commonly
do
you
want
to
go
first
yeah,
so
I
am
Conor
McBride
I'm,
a
PhD
student
study
in
solar
physics
at
Queen's,
University,
Belfast
I'm,
currently
in
the
process
of
writing
up
my
thesis.
So
that's
why
I
wasn't
able
to
make
this
meeting
in
person,
unfortunately,
but
with
Senpai
I'm
the
continuous
integration
maintainer.
J
Hi
guys,
my
name
is
today:
I'm
postdoc.
F
At
Berry,
I
love,
kids,
I
know.
If
you
hear
me
right:
yeah,
yeah,
all
right,
hey
I'm,
just
a
user
I'm,
not
developing
anything
myself.
I
have
a
cubicle
next
to
Nabil
and
whenever.
A
J
A
E
Here
about,
what's
going
on,
what's
the
current
status.
A
A
You
guys
hear
me
actually
sorry
yep,
yeah,
okay,
cool
yeah,
so
I'm
I'm
here
I
used
to
work
it
as
I
work
in.
F
Okay,
I
thought
would
I
start
with
this,
which
is
what
our
website
says.
When
you
go
to
some
Bible,
you
make
it
bigger.
E
Off
as
a
one
package
and
has
grown
into
many
many
packages
as.
B
D
G
Even
like
decide
to
you
know,
weigh
in
on
the
direction
of
the
project,
what
things
we
should
like,
what
large
scale
things
we
should
work
on
and
do.
B
G
Know
we're
we're
open
and
hopefully
friendly
Bunch,
who
you
know,
welcome
contributions
of
kind
of
any
form
out
and
also
we've
written
it
all
inside.
C
The
most
active
bit
of
the
repo-
this
is
the
commit.
E
In
the
core
grouper
does
my
slides
not.
E
Too
bad
still
pretty
good
I,
always
think
that's
true:
okay,
13
people,
13
people
and
two
Bots.
E
E
E
D
Similar
signs
to
yeah
pick
up
a
pick,
a
project.
A
A
E
G
G
Mean
I
would
have
liked
to
have
got
more
stats
out
of
getting
GitHub
I
tried
to
find
like
a.
F
B
F
Like
to
retain
people
who
started
and
if
not,
why
not
yeah
yeah
is
that
the
one
thing
so
if
I
didn't
do
that,
they
needed
I'm,
sorry
yeah,
so
as
well
as
the
core
package,
we
also
have
a
quite
long
list
of
other
science
packages.
F
Packages
website
list
we
have
ND
Cube,
which
is
a
library
for
WCS,
coordinate,
aware
of
log
like
multi-dimensional
arrays,
the
RMS,
which
is
a
python
interface,
low
level,
python
interface.
E
F
F
E
Like
Danny
and
nabil's
package.
E
H
Which
is
a
library
for
getting
radios
backward
data
and
working
with
it
yeah
other
things
not
currently
listed
on
our
website
is
the
recently
adopted
some
I
saw
package
that
David
wrote,
which
is
a
Fido
plug-in
for
the
sort
of
orbital
archive
sunkit
Pi
Vista,
which
is
a
3D
plotting
Library.
Again,
that's
David
and.
E
A
Sun
expects
I'm
gonna,
throw
that
down
the
other
end
of
the
table.
I
haven't
been
paying
enough
attention
just
unexpect
recently.
H
What
does
it
do?
Well,
you
want
to
take
that
I,
don't
know.
What's
our
son
expects
is
also
getting
a
Rebrand,
so
we
should
you
know
we
should
do
that.
We
should
do
that.
No,
it
is
a
package
to
support
solar
x-ray
spectroscopy,
so
it
has
I.
Can
it
has
a
couple
functions
for
thermal
and
non-thermal
emission
and
as
well
as
that,
it
has.
H
Fitting
tool
developed
by
Chris
Cooper.
F
H
With
x-ray
Photography
in
order,
so
do
you
want
any
more?
No,
that's
fine
and
sunkit
DM
is
a.
E
A
Okay,
the
the
pre-commit
it.
H
B
D
H
E
H
Girlfriend
or
something
I
think
yes,
there's
also
sunka
instruments
which.
D
Is
a
grab
bag
of
instrument
specific
processing
routines,
but
when
instrument
teams
don't
want
to
do
Packaging?
Well,
that's
it,
but
it's
not
just
things.
D
E
E
F
D
A
Things
that
reach
the
website,
repo,
which
also
serves
as
project
documentation
to.
H
There's
also
a
package
called
a
Blog
that
some
pipe
maintains,
which
is
a
sphinx
plugin
for
writing
blogs
and
it's
one
of
the
more
active
repos
on
the
site
in
the
Senpai
org.
But
it
has
absolutely
nothing
to
do
with
solar
physics
whatsoever
and
the
bill
maintains.
A
H
B
Renamed
e-blog.
H
G
B
B
B
H
Somebody
ends
up
picking
up
the
to-do's
As
and
when
they
are
needed
to
do
they
needed
to
be
done
rather
than
either.
Somebody
may
be
actively
thinking
about
what
the
best
thing
to
do.
The
what
the
webmaster
role
is
basically
in
a
bill
at
this
point,
and
we
should
probably
just
put
him
on
the
list
and
and
yeah
there's.
There
are
descriptions
for
all.
H
A
H
H
A
F
A
Like
social
perspective
this
weekend
next
year
is
the
fact
that
we
keep
growing
this
project
into
more
and
more
repos
yeah
and
we're
doing
that
primarily
for
technical
reasons,
I
think
and
I.
Don't
think
the
I
don't
think
our
documentation
and
expectation
of
our
users
is
keeping
up
with
what
we're
doing
but
like
we
don't
want
core
to
become
massive
and
hard
to
maintain
with
a
lot
of
code.
D
Instance,
because
some
of
that
was
sort
of
like
unmaintained
or
you
know
not
hadn't
been
kept
up
with
the
same
amount
of
Maintenance
as
the
rest
of
the
package
and
yeah
I
I
wondered
now
whether
people
just
don't
know
that
stuff,
but
I
think
that's
I
I
think
that's
definitely
the
case
I
think
it
happened
recently.
Someone
actually
maybe
Alistair
might
be
online
just
looking
for
some
stuff
to
do
Siri
data
and
he
wrote
his
own,
so
we're
gonna
have
a
product,
and
now
me
afterwards
realized
that
it
was
taken
away
into
something
yeah.
D
Yeah
I
mean
and
part
of
the
problem,
because
I
did
say
after
I
did
that
I
was
going
to
go
to
Sanka
instruments
and
put
an
example
in
there
and
then
I've
not
done
that.
Yet
so
the
next
person
coming
along
is
going
to
have
the
same
problem
as
me,
where
there's
an
indication
that
you
can
read
level
1B
and
despite
them
automatically,
which
is
what
I
was
trying
to
do.
A
Over
again
like
oh,
we
could
move
this
into
a
different
thing,
right,
I
think
I
think
we
need
to
stop
until
we
have
fixed
the
other
parts
of
that
puzzle
right
like
there's
the
actual
maintenance
infrastructure
stuff
that
I'm
going
to
be
working
on
as
part
of
that,
that's
it
wrong,
and
but
then
there's
also
the
like
user
expectations
and
project
side.
That
I
think
we
also
need
to
fix
before.
B
I
H
A
H
B
A
Look
for
stuff,
if
it's
all
in
one
place
yeah,
but
how
do
we
I
mean
I?
Think
we've
tried
with
the
website
with
Dean
with
the
drop
down
you
know
to
make.
You
know
having
the
documentations
are
all
linked
to
each
other
through
the
page
like
that
was
you
know
part
of
the
idea,
the
clue
that
stuff
work
to
some
extent.
A
A
Don't
really
go
well
in
the
we
have
a
session
tomorrow.
B
D
Which
is
quite
hilarious
because
I'm
not
there,
because
I
work
in
England
remotely
I
developed.
E
B
G
B
A
B
H
F
F
A
B
B
F
F
So
this
is
a
list
of
our
all
of
our
current
sub
packages
and
then
this
they're
sort
of
like
health
for
lack
of
a
better
word
or
their
states.
As
indicated
piloted
by
these
colors.
You
know
green
basically
means
not.
A
lot
is,
is
changing
pretty
pretty
stable,
blue,
reasonably
stable,
yellow.
B
So
since
last
year
we
were,
we
actually
did
remove
completely
two
sub
packages:
the
ROI
sub
package,
which
I
think
it's
kind
of
gone,
a
bit
stale
and
unused
for
a
while
and
then
also
be
ins
instruments
possible
to
say,
instrument,
sub
package,
which
is
now
in
sunkit
instruments,
which
is
what
we
were
talking
about.
B
So
I
just
want
to
give
some
notes
on
what's
what's
new
in
each
of
these
sub
packages,
but
the
detailed
kind
of
overview
of
this
is
is,
in
our
docs
part.
As
part
of
the
our
releases.
We
go
through
kind
of
create
like
that.
More
plain,
Language,
summary
overview
of
what,
like
the
big
features
that
have
been
added,
so
you
could
go
and
read
these
on.
B
These
websites
actually
have
a
nice
working
requested
into
the
package
and
what's
been
what's
what's
changed
in
that
version,
so
for
coordinates
really
the
com
only.
H
A
Shows
how
you
can
account
when
you
do
a
reprojection,
you
can
account
for
differential
rotation
propagate
with
solar
surface
is
so
I.
Think
Albert
might
give
some
more
he's
going
to
talk
about
the
the
funding
that
we
got
later
this
afternoon
and
then
I
think
he's
going
to
sort
of
talk
about
the
the
future.
D
Coordinate
so
they'll
be
more
about
coordinates
later
on
the
afternoon
net.
One
of
the
big
things
of
this
I
think
did
you
do
this
David
was
adding
support
for
C
to
web
for
for
institute
data.
This
is
a
a
NASA
as
far
as
that
is
there
any.
This
is
the
only
kind
of
big
thing
that
stood
out
to
me
of
user-facing
thing.
There's
a
few
little
things
like
that,
URL
being
there
yeah
hall,
five
two
is
actually
quite
a
big
upgrade
in
that
business.
H
E
D
A
F
I
would
actually
yeah
that
also,
if
you
were
using
part,
5
directly,
there's
also
a
load
of
extra
API
things
that
allows
you
to
customize
how
it
works
better.
H
That's
next
math
we've
had
a
couple.
So
now
has
a
3.1.
There
is
a
reproject
two
method.
So
if
you,
if
you
have
your
exist
session
well,
these
are
you
can
sort
of
see
what
what
the
API
looks
like
this,
but
you
have
a
map
and
you
have.
You
have
like
a
sort
of
a
WCS
that
you
want
to
pre-project.
F
To
so
like
I,
don't
know
what
this
case
is,
but
you
know
if
you
have
the
like
the
wcs4,
an
observation
taken
say
from
from
Mars.
H
F
In
masterpie,
this
method,
just.
H
Makes
it
a
lot
easier
to
do
a
nitty-gritty
detail,
but
probably
an
important.
A
Are
exposed
in
their
math,
so
this
is
in
a
tends
to
be
more
convinced
for
compliance
that.
F
F
I
And
date
and
properties,
and
don't
just
collapse
at
all,
into
one
dates,
but
we
well
there's
there's
more
details.
Basically,
we
hand
we
differentiate
between
these
three
circuits
and
our
it's
working
yeah.
I
I
And
the
other
one
that
I
kind
of
like
that's-
that's,
not
a
huge
feature
but
sort
of
nifty.
As
you
can
now
do
at
the
3.1,
you
can
do
arithmetic
operations
with
maps.
I
This
is
between
maps
and
arrays
or
quantities,
not
between
two
maps.
So
we
we
still
that's
a
that
would
be
a
nice
feature
to
have
to
be
able
to
actually
like
do.
1.2.
I
F
I
A
month
type
and
divided
by,
for
example,
the
exposure
line
and
your
old
Mac
tag
is
not
valid
anymore,
and
maybe
this
goes
into
the
discussion
about
metadata
that
we're
gonna
have
later
on,
but
yeah
I
really
want
people
to
deal
with
that
and
have
somehow
you
should
deal
with
that.
Yeah!
Oh
wait!
So
you
mean
you
would
want
to
change
the
exposure
time
key
yeah.
No!
No!
No
just.
F
A
A
You
should
still
be
able
to
find
that
anyway,
yeah
yeah,
all
right
so
image.
Just
we
removed
all
the
code,
sorry
that
we
deprecated
we
deprecated
assignment
code,
that
was
in
the
image
launched.
A
C
Rotation
method,
which
is
really
about
how
I
transformed
conscious
rotation,
but
so,
for
example,
you
can
use
coupons.
C
Where
you
can
use
so
you
can
use
your.
D
And
then
we
actually
or
I
say
we
Albert
used
this
to
to
add
opencv
an
opencv.
G
E
True
yeah
I
mean
we
want
to
know
the.
G
B
G
Time
every
time
but
I.
B
Been
asking
for
for
oh
a
few
years
as
well,
especially
pushing
from
the
Ia
side
of
wanting
to
be
able
to
do.
You
know
like
how
do
I
do
rotate.
B
C
Gpu
to
do
this,
so
this
essentially
provides
like
an
entry
point
for
specifying
you
know.
How
do
you
cook
in
your
underlying
outline
transform
goes
to
like
the
rotate.
G
Time
series
I'm,
probably
not
that
qualified
to
talk.
C
D
I
was
talking
about
for
a
lot
of
Institute
data.
It
turns
out
it's
still
stored
in
and
delivered
as
CEF
file.
So
now
we
have
you,
can
you
can
stick
in
a
CEF
file
to
a
Time
series
and
you
hear
the
type
series
factor
and
you
get
back
Time
series
I,
don't
know
if
you
want
to
say
more
about
this
anything
yeah,
this
kind
of
I
guess
motivated
on
a
higher
level
by
solar
work
term
pack.
J
G
Okay,
can
you
hear
me
now
yeah,
yes,
perfect,
okay,
yeah,
so
I
just
wanted
to.
B
F
A
B
I
D
I
I
The
the
pandas
like
I,
mean
like
when
you
get
to
that
stuff.
You've
got
to
be
talking
actually
right
and
also
are
there
not
other
video.
F
G
B
A
F
C
B
H
I,
don't
think
we're
gonna
talk
about
database
right
now
come
on
yeah.
Can
we
please
deprecate
it
please?
H
H
B
H
H
D
G
F
F
B
A
Decision,
oh
God,
great,
all
right,
IO,
there's
not
much
to
say
except.
H
We
have
deprecated
the
io.fits
module,
just
not
stopped
me
from
continuing
to
use
it.
This
brings
up
another
point.
B
F
Make
it
progress
all
available?
Yeah
we've
got
to
tell
people
to
use
IO.
H
F
H
Reader
is
in
IO
yeah.
We
can't
deprecate
all
the
higher
yeah,
no
probably
private,
the
other
part
yeah
sure,
except
which
parts
jpeg
we
keep
a,
but
we
should
also
discuss
whether
or
not
we
should.
B
B
A
A
Package,
sorry,
just
before
you
understand
any
suggestions
for
an
emoji
that
can
see
they.
B
J
Hole
in
the
code
that
I
was
talking
about
earlier,
they
said
this
is
stuff
to
like
account
for
differential
rotation
in
your
code.
So
we
that's
been
deprecated
and
I.
Think
we
well.
A
Some
creating
Machinery
that
Albert
has
developed,
probably
and
probably
the
the.
A
Different
approaches,
but
it's
cross-relation
the
differential
location,
because
right
now
it's
almost
functions.
D
D
A
D
Was
probably
do.
A
Eui
AIA
and
stereo
all
reprojected
onto
the
same
Carrington.
D
H
This
before
you
move
on
to
the
next
slide,
it's
something
I
wanted
to
mention
that
no
about
some
kind
of
physics.
But
that
comes
to
mind
because
of
that
at
some
point
this
week
it
might
be
nice
to
include
a
discussion
but
to
pick
up
that
discussion,
we've
had
about
like
some
kind
of
sunflower
models-
yeah
yeah!
Yes,
so
maybe
that's
conceptually.
A
Related
to
physics,
I
found
that
note
in
some
of
our
past
notes
that
we
had
taken
about
the
coordination
meeting
and
somehow
I
got
lost
somewhere,
but
yeah
yeah,
I
I,
definitely
agree.
Maybe
either
I
don't
know
we
can
figure
out
where
yeah
I
don't
need
to
have
it
now.
But
if
he
doesn't
need
to
try
to
make
a
note
to
include
that
this
week
to
do
some
stuff
that.
A
H
Needs
to
be
private,
what
what
is,
and
what
are
things
that
are
used
like
okay,
there's,
really
two
problems.
One.
A
A
B
F
C
A
To
say
even
think
you
think
about
or
yeah
our
some
package
here,
maybe
maybe
it
seems
like
the
two
ones
that
don't
have
maintainers
are
going
to
be
one
of
them's
going
and
the
other
ones
it
was
smaller.
Yeah.
H
H
H
That
we
can
use
for
presentations,
I
mean
even
I,
don't
like
walk
mode.
Only
almost
I
mean
it
would
be
nice
to
have
templates
or
presentations
that
we
can
Version
Control.
G
H
H
It
in
the
the
chat-
oh,
you
know,
yeah.
H
F
F
A
G
Sure
package,
I
would
say
sorry
extreme.
F
Go
ahead
but
yeah
I'm
just
much
more
to
say,
you
would
give
something
image
comments.
No
I
would
like
to
eventually
spend
some
more
time
because
people
really
want
to
press
loops
and
I
think
really
that
kind
of
focus
on
the
market,
but.
F
A
A
whole
bunch
of
makeings
yeah
yeah
and
also
like
that
flcc
so.
B
E
Just
wondering
no
I'm
just
wondering
this
came
up
about
doing
3D
link
version
from
different
Vantage
points,
yeah
or
like
reconstructed
like
3D
geometry,
like
U.S,
almost
stereos
yeah.
A
G
H
E
H
Come
about,
we
would
like
to
stop
being
the
truth.
We
would
like
to
stop
it
being
gone
from
green
to
Florida.
In
the
course
of
this,
we've
probably
opened
the
door
open.
The
windows,
cool.
B
A
Whether
that
was
a
good
idea
or
not
I
am
still
unconvinced.
It
was
but
I
don't
win
an
argument
so
like
like
you
know,
there
was
yeah,
it's
something
that
is
now
something
that
somehow
are
maintaining
and
like.
We
spoke
to
Helen
about
it
and
stuff.
A
G
A
G
I
I
F
C
C
I
A
A
A
From
the
other
side,
like
that,
our
team
is
only
so
big.
We
can
only.
B
F
Is
what
I'm
getting
it's
like?
How
do
we
decide
what
to
becomes
a
sponsored
packaging?
What
it
doesn't,
because
we
only
have
so
much
I
mean?
Maybe
we
should
discuss
this
well
I.
Think
it's
a
concept
of
potentially
looking
at
the
most
important
images.
Yeah
as
we
add
them,
we've
added
a
lot
and
the
question
becomes.
Should
these
be
well
a
lot
of
the
sponsored
packages?
In
fact,
if
you
go
down
the
list
of
sponsored
packages,
we
have
at
the
moment
they're
all
there
for
a
reason
right.
F
Drms
we
adopted
because
it
was
otherwise
it
was
going
to
become
unmaintained
and
it's
really
useful
to
some
play
core
and
also
the
community
at
Large
sunkit
image.
Sunkit
instruments
are
effectively.
We
decided
to
split
up
the
core
package
into
more
packages.
F
Sun
rafter
is
a
spin-up
of
a
ND,
Cube
or
I.
Guess
more
accurately
around
here,
there's
a
spin-off
of
what
was
songwriter.
F
What
else
is
on
the
list?
Obviously
you.
I
H
F
Yeah
I
just
I
mean
our
sounds
very
reasonable
and
I,
like
the
definition
actually,
which
is
on
the
page.
I
just
worry
about
the
perception
of
being
like
inside
and
outside,
but
I
think
the
the
notion
of
like.
H
F
H
Project
shouldn't
be
making
decisions
about
the
direction
of
like.
H
E
A
E
Yeah
sorry
go
ahead,
John
I
also
might
be
worth,
including
that,
like
you
know,
somebody
mentioned
earlier,
like
the
reason
not
putting,
and
why
for
like
why
we
make
certain
choices
and
I
think
the
you
know
a
good
example
here.
Is
you
know?
Why
is
eie.
E
You
might
have
someone
come
along
and
say
like
why
can't
some
pages
look
after
my
package.
B
G
B
Extrapolations
I
would
actually
say
it's.
B
E
Got
a
question
about
yeah:
this
is
a
business
question,
but
how
small
a
physical
scale.
G
Can
you
like
apply
this
Global.
G
Yeah,
okay,
someone
did
open.
A
B
G
Do
you
see
the
like
expanding?
This
is.
A
Like
you
could
do
yeah
like
a
non-linear
Force
free,
you
could
do.
B
F
E
Yeah
I've
been
affects
yeah.
B
D
B
Yeah
I
guess
that's
the
scope
for
like
extracting
some
of
the
visualization
stuff
to
be
more
generic
and
yeah.
Like
an
example
is
so
you
can
run
the.
B
D
H
H
H
D
To
maybe
just
archive
because
it
is
yeah
but
then
I
kind
of
don't
want
to
like
oh
I,
don't
want
to
like
take
them
yeah,
because
that's
my
name
and
stop
people
I,
don't
really
know
what
to
do
with
it.
But
if
it
started
doing
other
things
in
pieds,
then
the
name
doesn't
make
sense
yeah.
But
it's
like
the
improvements.
You've
talked
about
like
the
improvements
we've
talked
about
like
I.
D
B
A
C
D
Fool
me:
once
oh
yeah
I,
don't
know
it
would
mean
I
mean
that's
like
generally.
The
problem
I
think
with
a
lot
of
these
affiliate
packages
is,
they
are
like
they
do,
require
a
lot
of
very
specific
domain
expertise
and
they
tend
to
be
written
by
phds
or
postdocs
who
go
on
to
get
real
jobs,
because,
due
to
an
overwhelming
desire
to
eat
sometime
this
week
and
either
so
you
yeah.
C
A
C
C
D
F
B
It's
better
than
the
Senpai
logo
yeah,
but
anyway,
moving
on
to
the
actual
package,
it'll
already
say
it's
not
what's.
C
C
C
F
Just
I'm
not
in
control,
so
yeah,
basically
they're
the
links
that
the
the
code
base
and
the
docs
it's
you
know
some
ndq
on
GitHub,
the
big.
The
biggest
news
of
the
last
year
is.
D
And
time
from
multiple
people
has
been
released,
it
was
released
October
last
year
and
was
quite
rapidly
followed
by
you
know
subsequent
bug,
books
release,
as
is
you
know
normal,
so.
C
C
Package
which
is
designed
I,
don't
think
basically
any
line
of
code
in
the
entire
survive.
B
G
G
The
big
facilitator
of
that
was
the
release
of
astrophys
WCS.
B
B
B
I
think
the
scope
or
the
power
of
ndq
it
has
meant
that
we
have
not
backwards
skin,
so
it's
not
backwards
compatible
in
a
lot
of
ways.
So
it's
really
a.
A
New
break
a
fresh
start
and
you
know
I
think
it's
it's
a
really
powerful
tool.
Now,
in
total,
we've
had
13
people,
not
counting
Bots,
who
have
contributed
over
Justice
for
Bots
I,
think
we
had
one
bot
so
14,
that's
bad
and
then,
but
you
know,
over
the
two
years
preceding
the
release
of
2.0,
we
had
10
of
those
people
contributing
you
know.
Several
loads
are
in
this
room,
maybe
online
as
well.
I
can't
see
the
full
list
of
people
online,
so
that's
a
huge
positive.
A
Another
thing
related
to
that
is
we.
He.
F
A
So
much
of
this
code
is
like
will
now
live
in
ndq,
so
so
that's
a
successful
and
great
because
we
got
a
lot.
We've
got
some
feedback
from
The
Wider
Senpai
Community
about
what
about
that
API
and
then
the
third
thing
here
I
want
to
mention
was
that
there
is
an
ndq
paper
being
written,
which
is
actually
ND,
Cube
and
WCS.
B
The
Azure
Pi
WCS
API
and
coming
back
to
this
question
that
was
very
about
why
the
point
of
that
paper
isn't
to
you
know,
list
out
the
API
it's
to
sort
of
give
some
of
the
philosophical
and
historical
background
and
context
as
to
why
this
exists.
Why
it's
been
built
and
the
way
that
it
is
you
know
what
are
the
advantages.
B
So
it's
not
it's
not
sort
of
a
print
code
base
or
print
SCP.
It's
trying
to
be
a
bit
more
philosophical,
so.
G
A
So
so
yeah,
that's
really
the
state
of
of
ND,
Cube
and
I.
Think
we've
also
got
a
lot
more.
There
are
other
things
planned
like,
for
example,
on.
E
The
phrase
last
week
about
supporting
at
least
some
arithmetic
operations
like
math
now
does
also
some.
H
Cube
and
the
cube
already
can
do
reproject
too,
like
Matt,
but
you
know
there
are
other
ways
to
think
about
that.
If
you
don't
already
have
the
WCS,
so
there
are,
there
are
things
in
the
pipeline
and
certain
things
that
plan
so
ndq
is
not
like
PFS.
It's
not
necessarily
feature
complete
or
you
know
if
there
is
a
pathway
for
it
to
evolve
and
continue
to
get
developed.
A
2.0
is
I,
think
very
powerful
and
really
opens
up
a
lot
of
opportunities
for
other
packages,
like
some
quite
cool.
That
is
the
foundation
of
the
deacon
Force.
H
D
G
H
A
No
but
I
think
this
just
shows
like
how
how
needed
this.
This
idea
is,
and.
D
Between
ND
data
and
ndq
and
okay,
and
it
came
to
the
conclusion
that
ND
Cube.
B
B
B
Break
down
that
wider
n-dimensional,
again,
probably
something
that
should
be
documented
somewhere.
Yes,
that's
a
very:
are
there
any
questions
or
comments
online
again,
I,
just
keep
being
aware
that
it's
easy
to
talk
in
person?
Are
there
any
questions
or
requests
or
comments?
Criticism
on
this.
B
Everyone's
stunned
into
Silence
by
the
awesomeness
of
ndq
2.0,
what
does
it
even
do?
Yeah
I
guess
what's
next,
what's
like?
Is
there
again
yeah
yeah
I,
wouldn't
quite
go
that
far
I.
Think
it's
fair
to
say.
Certainly
on
my
side,
probably
a
little
bit
of
a
steward
side
is
that
the
intensity
of
development
has
been
much
lower
this
year
as
soon
as
ND
cubed.
As
soon
as
2.0
got
released,
I
think
we
kind
of
just
fell
on
the
canvas
and
you
know
needed
to
take
a
break.
B
That
being
said,
no
there
is
there.
We
still
have
many
issues.
I
mean
issue
tracker.
Many
of
them
are
suggestions
or
about
like
new
features.
So
again,
like
arithmetic
operations,
I
think
is
the
one
that
we
should
try
and
hack
something
together
this
week,
but
like
doing
proper,
coordinate
aware
Q
on
Cube
arithmetic
operations
would
be
nice
at
some
point.
At
some
point,
that's
the
honest
yeah
one
of
the
goals
is
to
be
able
to
like
add
two
cubes
together
and
it
go.
It
uses
the
coordinate,
information
and
reprojects
them
as
necessary.
B
Yeah
well
I
mean
but
like
again,
dividing
by
an
exposure
time
like
like
dividing
by
quantity
or
something
that
that
simple.
So
if
one
object
isn't
coordinate
aware
the
other
one
is,
that
should
be
like
an
easy
supportable
thing
for
us
to
do
and
should
hopefully
be
able
to
hack
something
together
this
week.
B
So
you
know
I
think
those
sorts
of
things
and
yeah
we
have.
We
have
issues
describing
potential
new
features.
I
haven't
collated
them
into
a
new
roadmap.
Yet
because
again,
I
haven't
come
back
to
it
and
thank
God
since
2.0
was
released.
That's
true!
Actually,
yes,
ndq
3.0
is
a
hypothetical
release
where
it's
just
excellent
right.
B
I'm
sort
of
a
tangent,
it
sounds
like
there's,
maybe
room
for
like
the
helper
library
for
generating
and
manipulating
World
coordinate
systems,
yeah
I
think
for
a
great
expensive,
for
example.
We
have
data
that
can
be
represented
in
that
way,
but
there
is
no
WCS
specific,
so
you
have
to
generate
them
to
make
that
easy.
So
there
are
tools
for
that,
and
you
can
do
this
in
in
ndq
right
I
wanted
to
speak.
B
So
there
is
helpers
in
ND
Cube,
which,
if
you
give
it
a
Skype
award
object
or
give
it
an
afterpi
time
object.
It
will
generate
you
lgwcs
look
up
table
based
gwcs,
but
that's
cool
yeah,
so
not
necessarily
the
most
memory
efficient
way
of
doing
it,
but
certainly
compared
to
what
was
there
before
any
any
WCS
which
isn't
lookup
table
requires
handcrafting
right,
because
it's
a
model
of
your
coordinary
transform
at
that
point
where
that's
a
fit
coordinate
so
fits
WTS
is
a
very
fixed
model
of
your
quadrant
platform.
B
Wherever
gwcs,
you
can
do
anything
you
don't
like
in
terms
of
modeling
your
equipment
platform,
but
yeah.
If
you've
got
look
up
things,
would
it
support
any
quantity?
Okay,
yeah,
yeah
and
I
spoke
to
Nadia
from
Space
Telescope,
with
maintenance,
gwcs
and
she's
up
for
those
helpers
moving
into
GW.
That
would
be
a
more
reasonable
place
for
them
to
be.
Actually
I
will
continue
the
method
of
substreaming
more
and
more
of
ND
Cube
yeah.
B
The
mdq
is
sort
of
like
just
become
this
lab
for
weird
and
different
WCS
stuff
that
slowly
builds
and
it
gets
shunted
off
into
other
places
like
Indie
Cube,
one
like
caused
WCS
API
in
a
weird
kind
of
like
there
was
all
like
parts
of
it
anyway,
we
would
cross
over
them.
I
think
another
useful
hack
in
DQ
pack
for
this
week
would
be
an
example
Gallery
like
which
I
know
Laura
has
brought
up
previously.
Are
you
are
you
that
wasn't
like
you
should
be
doing
this?
It
was
like
you've
been
banging
this
drum.
B
I
know
I
actually
had
a
little
yeah.
Maybe
that's
the
things.
Yeah
I
already
put
it
on
the
hack
list
yeah,
because
there's
there's
so
many
things
that
I
like
do
with
Indie,
Cube
and
I.
Do
them
once
and
then
they're
in
a
notebook
somewhere
I'm
like
yeah
I
forget
how
to
do
them,
because
there
are
so
many
things
which
are
just
lost
in
the
ndq
chat
case.
B
Like
figure
out
how
to
do
these
things-
and
it
really
annoys
me-
and
it's
it's
because
Indie
cube
is
like
a
hard-
it's
a
very
hard
package
to
use.
That's
not
I!
Think
that's!
That's!
Okay,
because
it's
really
complicated!
No
but
I
mean
it's
really
complicated.
It's
really
complex
and
like
the
things
you
do
with
it
are
extremely
powerful
and
it's
not
I.
B
Don't
think,
and
that's
it's
okay,
for
you
know
it
to
be
that
like
base
data
structure,
but
I
think
like
I
mean
yeah
would
be,
would
be
super
super
helpful
because
I
think
we
probably
all
have
compiled
like
you
know
ways
again.
Yeah
we've
done
weird
and
wonderful
things
with
nbcube
and
collecting
all
of
those
as
accomplished
would
make
a
would
be
a
great
reason
yeah.
It
would
make
a
lot
of
lives
a
lot
easier
yeah,
but
we
should
do
this
for
just
selfish
reasons.
B
If
it's
not
going
over
the
same
thing,
yeah
and
not
to
sort
of
you
know
not
in
a
selfish
way
of
like
trying
to
like
not
do
work,
but
I
actually
think
that
the
best
people
are
doing
that
are
like,
like
the
users,
yeah
yeah,
there's.
So
many
things
like
at
this
point,
I've
spent
so
much
time
looking
at
it,
I
think
it's
probably
true,
but
it
looked.
It
seems
intuitive
and
obvious.
Yeah
I
know
every
single
insane
compromise
that
went
into
the
design
of
crop.
B
The
high
barriers,
entry
I
think
that
also
shows
like
that
there
is
will
there
will
always
be
a
place
for
something
like
Maps
yeah.
You
know
yeah
one
of
the
big
things
this
has
just
brought
up
something
you
might
one
of
the
big
things
that
would
be
lovely
to
work
on
in
ndq.
If
anybody
fancies
doing
it
at
any
point
is
IO.
There
is
no
IO
layer
in
any
Cube
at
the
moment,
and
that
is
problematic.
B
B
I
mean
I
yielded
this
question
for
people
asking
if
like
Indie
Cubano,
no,
and
then
they
asked
why
I'm
like
yeah,
it's
not
outside
of
the
scope
of
the
package,
it
it
was
outside
your
scope
for
the
two
overlays
right
legs,
I'm,
sorry,
yeah,
fundamentally,
there's
no!
I
o
in
ndq.
It's
because
nobody's
that
time,
but
right.
B
So
they
would
even
say
something
before
I'm.
Sorry,
I
can't
remember
what
it
was
going
to
say.
Yeah,
oh
I
was
gonna:
ask
do
you
see
ndq
moving
classified
call
ever?
That's
it.
B
I
I,
don't
see
an
absolute
hard
like
reason
why
it
couldn't
I
think
that's
I,
think
probably
more
political
than
logistical,
yeah,
I
think
yeah
I
guess.
Obviously
it's
it's
package,
so
you
could
like
develop
it
and
break
it
and
release
it.
Quick
and
all
that
yeah
I
mean
the
other
thing,
I
suppose
it's.
It
was
originally
developed
for
solar
use
cases.
So
then
the
argument
is
right.
B
Do
we
want
some
pie,
the
Senpai
Community,
to
have
control
of
that
since
it
came
out
of
what
we're
doing
or
do
we
want
to
Upstream
it
to
a
much
broader
community
that
is
more
likely
to
have
more
users
and
potentially
more
developers?
So
there's
yeah
there's
trade-offs.
Whichever
way
you
go,
yeah
yeah
and
I
mean
I'm
not
against
either
there's
good
reasons
for
both.
B
So,
potentially
yes,
but
we
would
lose
solar,
yeah
influence
and
control
we'd
have
we
already
have
on
it,
so
I'm
just
going
to
say
that
we
believe
soonish,
oh
yeah,
like
seven
minutes,
I
mean
it's
only
five
minutes,
okay,
so
yeah.
If
you
want
me
to
like,
spend
two
minutes
doing
some.
B
Right
all
right,
so,
firstly,
yeah
it's
unexpects,
as
I
think
was
mentioned
earlier
on.
Oh
yeah,
let's
go!
Oh
sorry,
my
bad
things
yeah
so
10
expects
is,
has
undergone
a
rename
Rebrand
we're
going
to
call
it
sunkit
specs.
As
that's
least
that's
what
I
remember
it
being
on
the
end
of
the
email?
Again,
don't
ever
ever
name
anything
yeah.
So
in
with
some
some
good
specs,
as
we
were
two
weeks
ago,
right
right,
so
we're
gonna
have
to
follow
them
and
buying
them.
B
So
the
package
and
everything
on
the
code
base
hasn't
been
renamed,
but
this
is
what
it's
going
to
be
rebranded
as
so,
it's
still
you
know
Imports
unexpects,
but
just
let
people
know
so
in
the
next
slide,
then,
just
to
briefly
tell
people
what
it
is.
It's
yeah
it's
a
package
for
solar
x-ray
spectroscopy.
So
it's
main
thing
at
the
moment
is
that
it
provides
physics-based
models
of
thermal
and
non-thermal
x-ray
emission
and
then
the
second
thing
which
the
last
bullet
point
there
is.
B
We
also
now
have
infrastructure
to
fit
those
models
to
x-ray
count
data
that
is
that's
sort
of
developed
by
Chris
Cooper
and
there's
been
a
mammoth
effort,
but
I
think
it's
fair
to
say
it's
not
mature
at
this
point
either.
So
just
going
back
to
the
models.
Specifically,
we
have
a
model
for
thermal
emission
which,
for
any
people
who
are
familiar
with
ssw,
is
an
equivalent
of
fbth,
hugely
or
very
broadly
used
function
for
x-ray
spectroscopy.
B
We
can
break
that
down
into
whether
either
line
emission
or
Continuum
emission
and
then
for
the
non-thermal
models.
Shane
put
a
lot
of
effort
into
translating
these
again
from
ssw
and
we've
got
a
few
of
them.
The
main
ones
I
want
to
pull
out
are
thin
Target,
REM
style
and
thick
title
foreign
again
to
that
are
very,
very
widely
used
within
solar,
x-ray,
spectross51,
I,
think,
okay
and
then
yeah
next
slides.
B
B
We.
What
else
is
on
this
slide?
This
is
like
a
copy
from
the
previous
yeah.
So
there's
the
code
base,
you
can
contact,
you
know,
I,
don't
think
we
have
a.
We
have.
We
have
a
specific
Channel
within
the
Senpai,
Riot
or
element
I.
Don't
think
we
will
actually,
if
we
haven't
undergone
Affiliated
package
review
yet.
B
That's
a
very
good
point.
No.
This
is
like
we
want.
This
is
a
clear
candidate
if
once
it's
developed
to
the
sufficient
coin-
and
it
was
mentioned
earlier,
so
we
should
be.
We
should
let
people
here
be
aware
of
it.
So
yeah,
that's,
that's
all
I
have
to
say
on
that.
If
there
any
questions,
though,
like
we
can
take
them
quickly
before
we
have
to
leave
anyone
online.
B
Nope
great
well,
it's
not
that
much
functionality
in
it.
Yet
so
that's
that
makes
sense.
That's
good!
These
are
all
the
files
are
still
just
the
ones
generated
from
plus
it's
going
to
be
like
all
the
the
piano
container
data
like
it's,
not
that
much
data.
That
is
yes
there
we
go.
How
do
you
use?
Are
they
what
kind
of
file
over
there
they're
just
like
Gen,
X,
okay,
yeah.
B
Yeah,
no,
it's
not
smart
at
all,
I
mean
again
in
down
the
road
in
the
future.
It
would
be
great
to
make
this
like
integrated
with
Fiasco
or
candy
pie
or
whatever,
so
that
generating
your
line.
Emission
data
can
all
be
done
within
within
Python
and
also
like
within
the
sun
kit.
Specs
such
as
Sun
expects
like
world,
but
that
that
is
that
takes
work
and
time
and
effort
that
we
have
not
collected
yeah.
B
B
People
are
using
it
yeah,
because
Vincent
used
it
to
derive
results
that
you
published
so
yeah,
so
the
SpongeBob's
not
released
it's
already
like
contributing
to
science
and
science,
publication
I,
think
yeah
I
think
we're
going
to
make
it
to
do
the
first
release.
I,
don't
know
if
we're
going
to
talk
about
it,
but
like
just
even
having
the
functions
working
correctly.
That's
like
a
is
I
think
this
start
talking
about
the
the
models.
B
This
again
preempts
the
discussion
about
like
a
senpai
model,
sub
package
or
Affiliated
package
or
whatever
I
think
those
models
would
be
a
prime
candidate
for
inclusion
there
and
then
maybe
some
specs
could
then
effectively
just
become
or
mainly
become
the
fitting
infrastructure
and
then,
of
course,
it's
important
models,
whatever
models,
it
needs
from
Sun
pie
or
from
somewhere
wherever
you
know,
yeah
so
yeah.
That's
a
discussion
for
later
yeah.
B
I'm
gonna
turn
the
stream
off
for
lunch
time.
Okay,
do
you
want
to
post
a
message
in
the
right?
Maybe
what
time
are
we
going
to
be
back?
We
should
be
back
for
two
because
over
here.
B
So
we
also
build
a
we
also
build.
A
no
art
will
oh
yeah
without
the
sea
extension,
which
is
what
windows
will
install
by
default
and
you
can
install
if
you
don't
want
to
deal
with
a
c
compiler
because
you
know
well,
does
that?
Does
that
does
your
time
with
the
ENA
extension?
What
you're
talking
about
the
Ana
extension
well
Wheels
in
general?
But
yes,
but
the
context
for
this
was
I-
was
trying
to
compile
some
paper
product.
B
B
No
okay,
the
latest
version
of
sample
we
haven't
merged
that
change.
Where
we're
publishing.
Oh,
no.
We've
got
the
pull
request:
the
progress
to
open
as
well.
There
is
Shane.
There
is
an
environment
variable
you
can
set
to
disabled
buildings,
Anna!
Okay,
if
you
want,
if
you
just
if
you
just
don't
want
to
bother
with
having
to
compile
stuff
but
I
would
be
getting,
it
would
be
really
good
to
get
up
from
the
full
build
of
sunlight
in
highlight
yeah,
but
it's
going
okay.
B
So
far,
I
was
just
more
out
of
curiosity
about
what
I
was
doing
yeah
once
once
Conor
finishes
business:
we're
asking
to
update
the
pull
request,
but
that's
the
pure
python
wheels
and
then
what
is
the
deal
with
that
I
think
Connor
just
needs
some
time
to
there's
a
few
to-do
list
on
the
podcast.
B
I
have
I
have
some
random
slides
for
the
documentation.
Oh
yeah,
there's
that
weird
act
so
I
mean
yeah.
B
B
So
I
should
just
update
my
touch
bar
to
have
an
emoji
thing
there
yeah,
because
you
know
you
can
get
the
like
insert
Emoji
thing
on
the
Mac.
Yes,
oh
yeah,.
B
B
Okay,
the
actual
it's
the
live
stream
I
mean
yes,
good.
Okay,
well
welcome
back,
fed
and
happy.
B
B
I
have
people
coming
back,
slowly,
yeah
sure
yeah,
let's
see
what
did
I
have
what
I'm
going
to
do
here,
let's
see
for
reasons
I'm
juggling
through
computers
here.
So
if
someone
wants
to
put.
B
That
link
into
something
into
the
element
chat.
Okay,
let's
see,
share
screen.
B
There
may
not
be
a
lot
to
see
on
status,
but
so
most
of
these
guys
are
actually
just
reminding
everyone
what
we
actually
propose
so
sometimes
now
funded
through
NASA
roses.
Awesome.
This
is
through
the
Roses
ostafel
program.
I,
don't
know
if
anyone
aside
from
me
actually
says
it
that
way,
but
I
insist
on
saying
it
like
that
open
source
tools,
Frameworks
and
libraries
program,
it
was
the
first
year
that
that
program
existed.
B
I
think
the
idea
is
for
it
to
keep
going,
but
fortunately
they're
funding
a
lot
of
major
libraries
through
that
we
asked
foreign
got
about
two-thirds
of
a
million
dollars
over
three
years
and
basically
the
200
and
change
each
year.
Technically
it
started
in
the
fall
of
2021,
but
basically
the
way
the
funding
Works.
B
Basically,
2024-ish
it'll
end,
maybe
a
little
bit
beyond
that,
depending
on
how
the
fund,
how
institutions
that
I'm
sending
money
out
to
actually
spend
it
and,
roughly
speaking,
very
roughly
I,
didn't
actually
go
ahead
and
check
what
we
proposed,
but
in
my
mind,
something
like
20
time
for
myself
and
Steward
and
I
have
time
for
Will
and
at
least
how
we
had
proposed
it.
B
So
I'll
have
slides
just
reminding
everyone
what
we
actually
proposed
again,
but
they
break
down
into
three
tasks:
improving
the
technical
infrastructure,
augmenting
the
scientific
functionality
with
just
two
subtasks
and
then
providing
training
and
Outreach.
B
So
first
one
is,
is
false,
squarely
largely
understood,
improving
the
technical
infrastructure.
Instead
we're
going
to
talk
more
about
how
much
this
is
still
the
plan
or
how
much
he
plans
on
attacking
in
what
sequence,
but
there's
stuff
to
shore
up
our
techno
infrastructure,
especially
as
how
it
extends
across
our
ecosystem
and
our
affiliate
packages,
et
cetera,
et
cetera,
et
cetera,
I,
don't
know
Stuart,
you
want
to
say
more
about
this.
B
Yeah
I
mean
that's
no
longer
to
do
this
good,
oh
okay,
we
can
swing
back
around
if
we
want
to
talk
more
about
this.
B
People
have
questions,
but
the
next
slide
is
the
next
tasks
or
subtasks,
or
the
second
part
is
the
sub
second,
the
first
subtask
or
the
second
task
is
supporting
multi-point
observations,
and
so
primarily,
in
my
mind
at
least,
is
about
Shoring
up
any
kind
of
difficulties
perceived
or
actual
with
working
in
some
with
some
type
coordinates,
and
so
the
first
part
is
the
part
that
is
the
the
code
part
developing
comprehensive
test
suite
for
something
coordinates,
including
benchmarking,
both
performance
and
accuracy,
to
fulfill
maintenance
and
to
optimize
performance,
and
the
second
part
is
to
actually
reach
out
and
engage
with
Mission
teams,
not
wait
till
they
come
and
and
I
say,
there's
an
issue
or
not
waiting
until
we
stumble
across
an
issue
but
to
actually
make
sure
that
everyone's
on
the
same
page
in
terms
of
coordinates.
B
Now
that
not
entirely
sure
how
to
proceed
on
that
one,
but
certainly
just
by
paying
more
attention
to
data
as
it
becomes
available
that
that
helps
already
any
questions.
Any
questions
on
this
part
I,
guess
maybe
you
what
you
just
said,
sort
of
answered
my
question,
but
yeah
do
we
have
like
a?
B
Is
there
some
sort
of
like
medium
that
we
could
agree
upon
to
engage
with
instrument
teams,
especially
about
their
coordinates
of
the
structure?
Or
is
it
thus
far
just
kind
of
been
email?
Chatter
between
like
well
I,
would
imagine
you
and
like
instruments
well,
I
think
I
think
it
would
there's
a
there's,
an
overlap
with
the
instrument
working
group
right
so
yeah.
If
we
have
an
effective
instrument
working
group,
this
would
be
a
natural
place
to
at
least
get
some
coverage.
B
Obviously,
the
main
problem
would
be
the
those
instruments
that
aren't
interested
in
Python
at
all,
but
we
still
need
to
work
with
our
data
products,
so
the
kind
of
Engagement
there
would
be
more
challenging,
I
mean
I,
don't
I,
don't
think,
there's
any
reason
that
we
couldn't
extend
the
instruction
group
to
include
you
know
that
those
uses
too
sure
I
mean
an
aspect
of
it.
Yeah
just
I
think
we
can
grow
the
scope.
I.
Think,
of
course,
the
instrument
working
group.
B
B
Go
on
I
would
say
so,
there's
a
distinction
between
what
we
want
to
do
and
I
think
that's
a
good
discussion
for
a
future
roadmap
discussion
as
part
of
this
meeting
within
the
within
the
defined
scope
of
the
staff
of
funding
it's
supposed
to
be
for
sustaining
and
not
necessarily
for
adding
stuff
and
so
a
little
bit
of
a
distinction
between
what
technical
we
are
funded
to
do
under
this
program
versus
adding
new.
You
know
significantly
new
significant
new
features.
B
Now,
that's
honestly,
we
shouldn't
do
it
I'm,
just
saying
technically,
and
this
funding
in
specific
we're
supposed
to
be
just
making
things
work
better
I'd,
be
interested
in
a
separate
discussion
to
talk
about
like
how
it
would
even
look
to
do
that
kind
of
Time
series.
Correction,
I,
I,
don't
know
how
we
would
do
that.
What
would
be
a
sensible
API
for
that?
But
it's
interesting
to
contemplate.
B
B
It's
just
you
know,
demonstrating
the
coordinates.
Stacked
with
images
is
a
lot
nicer
than
demonstrating
it
with
represent
an
ugly
numbers.
Yeah
wow,
certainly
true,
though
I
like
to
I
like
showing
some
of
the
rotated
Sun
frame
differential
rotation
stuff
for
just
numbers.
But
yes,
certainly,
images
get
the
message
across
a
hell
of
a
lot
more
effectively.
B
B
Okay,
the
next
other
subtask
under
science,
enabling
functionality
is
something
that
we
haven't
really
made
much
progress
at
all
on
I,
think
or
even
conceptually
will
might
say
more
on
this
one,
but
the
idea
was
as
proposed,
to
improve
some
time.map
to
to
take
advantage
of
parallelism
and
desk
and
other
things
that
might
be
desk
like,
of
course,
this
would
need
to
be
consistent
with
also
supporting
and
the
transition
to
ndq.
B
Yeah
I
I,
don't
know
that
we
scoped
out
like
a
very
clear
Way
Forward
on
this.
Yet
though
I
I
think
you're
I
mean
the
the
work
you
did
with
the
like
the
rotate.
You
know
the
different
rotate
backends.
The
entry
points
for
having
different
you
know,
ways
to
do
like
your
iPhone
transforms
committing
that
is
definitely
in
line
with
this.
You
know
means
that
now
it's
just
a
matter
of
reliable
desk
affine
transform
implementation.
Then
it's
just
a
matter
of
plugging
it
into
that
Machinery
yeah.
B
For
this
specific
case,
yeah
I
mean
it's.
It's
not
quite
it's
not
totally
clear.
Beyond
sort
of
that
specific
use
case,
what
are
the
other
things,
especially
in
the
context
of
like
a
single
2D
map,
how
we
would
use
desk
to
speed
up
some
of
those
operations,
but
what
about
map
Q?
Well,
certainly
yeah.
That's
that's
obvious.
Is
it
it's
in
how
how
dusk
would
make
things
a
lot
more
powerful
I
mean
you
can
already
represent,
like
with
an
indie.
B
Right
you
can,
you
can
give
it
a
WC
I
guess
probably
the
first
thing
to
do
on
this
front
would
be
just
to
go
through
and
say:
okay,
what
parts
of
ndq
or
sit
Force
like
an
eager
computation
as
opposed
to
a
lazy,
computation,
objective
and
reprojective?
Yes,
yes,
certainly
I.
Think
there's
also
the
desk.
B
B
So
there's
a
member
of
memory
and
performance,
yeah
sure
reasons
for
using
data
and
those
two
come
are
actually
separate
and
also
sometimes
conflict
yeah,
so
one
just
something
interesting
in
where
dusk.
So
if
to
paralyze
over
a
dark
array,
it
needs
to
be
chunked
into
pieces
right
yeah
that
actually
AIA
data
and
all
right
compressed
fit
files
are
chunks
on
disk,
because
the
way
that
title,
the
image
compression
Works
in
fits
means,
you
compress
tile
like
chunks
of
the
image
independently.
B
So
you
would
you'd
like
if
you
loaded
a
single
AIA
image
into
a
dark
array,
it
would
be
chunked
into
bits
and
you
could
do
analysis
in
parallel
on
each
one.
Reading
a
separate
byte
range
out
like
okay,
and
then
you
could
work
with
the
like.
However,
many
like
you're
doing
this
in
time,
like
at
four
payments
for
like
12
hours,
yeah,
if
you
like,
if
you
only
cared
about
I,
don't
know
how
the
I
don't
know
how
the
AIA
data
is
tiled,
but
yeah.
You
could
theoretically
not
read
all
of
it.
B
So
from
a
Mac
or
a
ndq
point
of
view,
what
is
is
all
that
chunking
service
at
all,
done
like
in
the
dark
gray
and
like
you
just
give
it
the
dark
gray
and
the
the
wrapper
object
doesn't
need
to
care
or
other
is
the
infrastructure.
We
need
to
build
for
map
and
Indie
Cube
to
be
able
to
handle
that
or
utilize
that
hey
I
think,
especially
in
the
case
of
any
Cube.
B
Where
you
don't
I
mean
you
don't
we
don't
have
a
lot
of
like
all
the
methods
that
we
do
on
the
map
that
are
like
doing
kind
of
Nitty
Gritty
stuff
to
the
array
for
map
that
it
might
be
yeah
there
are
things
really
good,
I
think
in
general?
No,
like
the
desk
takes
care,
like
you
specify
the
chunking
onto
that
that's
great
right
and
then
it
the
way
it
parses
out.
B
The
computation
to
your
array
is
based
on
that
just
handled
by
the
gas
Machinery,
with
the
a
is
the
Ws
use
case,
but
in
the
use
case,
which
is
great
stops
them
right.
The
other
day,
I
think
about
is
putting
Radio
Data
into
big
ndq
than
doing
analysis
on
it
very
slightly
different
concerns,
but
I
think
if
we
could
support
that,
we
so.
B
The
data
object
level.
How
is
that
not
supported
now
or
is
it
a
case
like
we
said
forward,
you
know
the
data
object
is
fine,
but
there's
other
stuff
we
need
to
do
lower
down,
like
you
know,
does
that
question
make
sense.
It
does
sorry
no
go
okay
basis,
level.
We
don't
have
a
test
Suite
that
does
a
vascara
into
all
of
this
and
share
that's
what
the
problems
are
and
what
works
already.
We
have
a
test
Suite
that
puts
a
task
array
into
ndq.
That's
called
the
decaf
user
tools.
B
Right
that
would
be
a
good
thing,
I
think.
As
a
first
cut,
it's
just
yeah
like
seeing
well
I
mean
we
have.
We
have
we
support
desk
as
an
array
type
format
like
then
it's
explicitly
like
in
numpy
Ray's
desk,
raising.
B
Oh,
oh
yeah,
yeah
I
mean
we
shouldn't
have
to
enumerate
them
right.
Anything
required
should
work,
yeah,
yeah
yeah.
Why
we
specifically
check
I,
don't
know
it
was.
There
was
a
specific
thing.
We
added
support
to
go.
I
can't
remember
anyway,
I
think
the
the
biggest
issue
with
that
from
a
usability
perspective
is
so
much.
Does
your
data
object
support
it
like
yes,
certain
amount,
routines
like
rotate
or
reproject
would
be.
It
would
be
beneficial
if
they
were
darts
to
wear
but
depending
on
what
they
are.
B
That
is
a
small
amount
to
an
entire
three-year
proposal.
Right,
like
there's
massive
complexity
in
some
of
like
in
some
of
those,
but
it's
actually
disrupting
the
dark
array.
In
the
first
place
to
hand
it
off
to
to
the
map
or
ND
cube,
is
actually
a
challenge
and
how
you
construct
that
desk
array
is
very
tied
to
the
science
you
want
to
do
on
that
array
like.
B
If,
if
you
have
say
say
you
have
like
a
4D
space
space
time,
wait
when
you
want
to
do,
and
it's
you've
got
like
three
files
for
all
wavelength
for
all
space.
So
you've
got
like
a
sequence
of
fit
files
in
time
like
Iris,
cubes
or
something
along
those
lines
you
and
you
and
you
care
about
doing
time,
series
analysis,
that's
chunking,
right
and
you're
going
to
end
up
reading
all
those
files
to
do
any
analysis
along
client,
Dimension,
so
I'm
not
convinced
that
there
is
just
one
answer
for
hey.
B
We
made
a
dark
array
for
you
because
that
it
depends
on
what
you
download
the
data
like
if
you,
if
you're
starting
off
analyzing
AIA
data
you're
stuck
with
what
you
download
right.
So
that's
that's
the
entry
point.
So
it's
partly
how
the
data
is
chunked
and
it
turns
out
that
AI
is
titled
one
row
per
tile,
which
is
really
useful
for
any
kind
of
scientific
analysis.
B
B
Well,
I
think
we
should
or
have
is
this
proposal
Scopes
tightly
to
come
up
to
some
Pike
or
no
I?
Don't
think
so,
but
I
think
that
I
well
I
mean
you
two
might
not
agree
with
me.
But
to
me
this
is
more
like
make
rope
make
the
things
that
Senpai
do
Senpai
does
have
work
better,
wouldn't
ask
if
you
give
it
a
task
array
right
and
the
starting
point
is
I.
B
Have
a
desk
array
and
I
want
to
put
it
in
map
and
EQ,
and
you
know
like
if
the
first
thing
you
do
is
concrete
that
into
a
numpy
array
to
do
rotate
on
it.
Then
what
was
the
damn
point
like
you
know,
right
I,
mean
I.
Think
the
the
kind
of
thing
that
I
sort
of
had
in
mind
I
mean
it's
a
fairly
small
stuff
when
I
like,
but
but
in
this
the
sector
like
that,
we
don't
want
to
do
anything
that
prevents
you
from
like
using
desk
in
your
workflow.
B
So
yeah
you've
constructed
the
thing,
but
we
don't
want
something
that
like,
for
example,
now
with
rotate.
You
stick
a
desk
array
in
you
call
rotate
on
the
map
with
that's
great
and
it
returns
to
a
number
array
that
immediately
kills
your
lazy
workflow
because
yeah
well
yeah.
If
you
had
like
a
like
a
10
or
20K
like
20K
image
and
called
rotate
on
it
with
your
computer,
won't
know
so
like
I
I,
guess
I
guess
is
we
don't
want
to
like
with
our
the
way
our
the
methods,
like
our
math
methods,
are
implement.
B
We
don't
want
to
rule
out
someone
using
desk,
yeah,
yeah,
but
I.
Think
yeah
a
good
place
to
start
would
just
be.
Let's
you
know.
Maybe
a
test.
Suite
is
not
the
right
word,
but
just
like.
Instead
of
doing
test
driven
about
right,
yeah
yeah,
you
see
how
he
operates
and
then
slowly
Unbreak
it
yeah
and
I.
Think
I
started
poking
around
on
this
a
while
ago.
I
think
probably
this
is
before
in
this.
This
funding
kicked
in,
which
is
how
I
realized
that
the
protein
stuff
was.
B
It
was
an
issue
there
might
be
a
few
other
places,
but
I
not
started
down
as
fat.
Yet,
but.
B
That
was
a
timing.
Yeah
sorry
I
feel
like.
Are
you
a
problem?
That's
fine,
a
important
discussion.
Okay,
that's
and
there's
more
to
say
about
what
progress
has
been
made
here:
providing
training
and
Outreach
task
more
examples,
better
examples,
more
tutorials,
better
tutorials
talk
to
other
people,
blah
blah.
B
B
The
very
fact
that
I
can
pay
attention
to
Senpai
without
I
can
defend
paying
attention
to
Senpai,
given
all
my
other
stuff
going
on,
it's
great
I'm
happy
to
have
the
stuff,
well,
it's
being
paid
through
American
University
and
he's
already
been
to
multiple
meetings
and
read
tutorials
and
that
sort
of
thing
and
Stuart.
We
finally
got
him
funded
just
a
couple
weeks
ago
and
so
I
don't
know
Stuart
if
you
formally
actually
started
using
this
funding
at
all.
Yet
yes,
so
I've
gone
like
two
days
work.
B
The
system
works
okay,
so
yeah
we
may
have
got
we
actually
I
guess
this
Monday
technically
was
awarded
back
in
September.
So
it's
almost
been
a
calendar
year,
since
this
is
a
support
that
I
did
put
in
place.
So
it's
taking
a
while
for
the
money
to
get
to
me
and
to
get
to
other
places,
but.
B
B
B
B
This
is
this
ties
into.
Perhaps
the
state
of
the
sun
type
coordinates
package
from
earliest
morning.
That
was
not
able
to
connect
to,
and
it
may
connect
to
some
of
the
roadmap
discussion
later
in
this
meeting,
but
it's
also
somewhat
related
to
the
solver
funding.
B
So
just
three
tasks:
three
bullet
groups
here
existing
GitHub
issues,
kind
of
the
major
stuff,
that's
kind
of
on
the
to-do
list
not
to
survive
myself,
but
just
for
coordinates
human
projector
radio,
because
people
keep
asking
about
it
more
and
more
on
differential
rotation
trying
to
refactor
our
code
and
make
it
perform
better
velocity
capability,
but
in
principle
it
could
be
done.
It
could
be
more
accurate.
B
Stuff
related
to
estafa
funny
looking
for
a
technical
challenge
with
that,
actually
the
most
technical
part
of
the
challenge
is
what
is
the
correct
handling
in
my
mind,
is
the
correct
handling
for
Stellar
aberration,
because
we
go
to
alt
as
technically
we're
transforming
into
the
train,
we're
still
aberration
and
making
sure
that's
consistently
done
with
where
the
sun
is
is
probably
the
biggest
part
of
the
challenge
there.
B
B
Okay,
I'm
just
curious,
because
that
would
be
really
cool
it
conceptually.
It's
fine,
it's
just
practically
the
way
that
astrophy
does
stuff
and
the
way
so
the
way
that
astrophy
does
stuff
isn't
totally
consistent,
or
at
least
it's
not
it's
assumption.
Laden
and
the
problem
with
us
is
on
the
solar
side.
We
tend
to
never
be
concerned
about
solar
operation,
because
the
sun
is
where
the
sun
is
and
everything
we
frequently
use
our
corn
system
and
index
software.
B
B
There
are
no
issues
per
se
currently
for
this,
but
it's
all
stuff
that
kind
of
needs
to
be
done:
ish,
adding
the
performance
benchmarks
to
ASV,
which
involves
me
understanding
how
that's
actually
done
as
I've
said
the
wheel
many
times,
I
still
don't
get
it.
B
Is
investigating
document
discrepancies
with
other
libraries
there's
just
some
pharmacies
about
ssw
on
GitHub
and
not
too
long
ago,
but
basically
they're
still
linear
discrepancies,
unsurprising
mostly
on
the
other
practice
side
on
the
Senpai
side.
B
But
you
know
documenting
that
and
making
it
clear
for
people
who
have
questions
about
that
of
course,
and
then,
for
my
own
part,
very
internal,
but
again
within
the
make
sense
within
the
scope
of
style
funding.
Is
that
the
transformation
code
is
a
little
bit.
B
Fiddly
I
mean
it's
a
lot
better
than
it
used
to
be
before
I
really
took
a
big
swing
at
it,
but
there's
a
lot
of
kind
of
code,
repetition
and
a
lot
of
stuff
that
could
be
probably
architectured
a
lot
more
cleanly
such
as
someone
defines
a
new
coordinate
frame
that
it
can
be
done
without
even
more
copy
and
pasting
and
stuff
like
that.
Maybe
Shane
might
be
able
to
say
something
about
whether
or
not
he
how
he
feels
about
the
structure
system.
B
It's
actually
marked
with
some
of
the
Transformations.
Now
the
last
bullet
group
is
more
Road
Mappy
than
than
the
stopper
funding,
but
based
on
some
of
the
discussions
that
have
been
going
on,
I've
been
meaning
to
look
into
supporting
other
type
of
screens
than
the
spherical
screen.
B
Ways
to
convert
vectors
that
are
not
velocity,
vectors
and
even
velocity
vectors
are
tricky
because
that's
again
assumption
laid
in
it's
they're
meant
to
transform
a
certain
way,
but
that
may
not
be
solar
friendly
and
then
this
is
almost
certainly
an
upstream
thing.
But
r,
a
ra,
declination
right
angle,
right,
Ascension
declination,
coordinate
frames
that
don't
that
aren't
centered
at
the
Earth
or
at
the
solar
system
very
center.
B
B
Adding
other
assumptions
Beyond
side
of
the
screen
would
be
super
useful,
like
especially
for
Sticks
being
able
to
just
pick
a
a
plane,
and
you
want
to
project
on
it.
That's
a
definite
thing.
I
think
there
are
other
use
cases
they're,
like
coronagraphs,
maybe
well
yeah,
I,
think
that
would
be
very
useful
for
how
for
them
exists.
B
B
Yeah,
it's
not
it's
certainly
yeah
conceptually.
It's
easy
to
imagine
other
screens,
it's
more
a
question
about
API
and
and
sensible,
and
how
to
make
the
calculation
sensible
and
what
happens
when
you
turn
them
all
on
at
once.
Yeah
don't
do
that.
B
Oh
yeah,
that's
also
a
question
about
API
how
to
make
it
so
users,
don't
try
to
do
something
like
that.
The
conception
that
I
would
have
is
to
kind
of
extend
the
existing
screen
thing,
which
would
basically
mean
it'll,
be
the
latest
assumption
right,
the
the
most
the
innermost
assumption
will
be
the
one
that
would
be
applied,
but
on.
B
Certainly,
if
that
API
made
it,
so
you
can
only
call
one
context,
monitor
manager.
That
would
be
also
that
help.
B
S
all
right
any
questions,
comments,
thoughts,
Etc,
oh
I'll,
say
as
a
it's
not
in
this
slide
deck,
but
I
did
float
the
suggestion
to
David
and
I'll,
extend
it
to
anyone
else
as
well.
Where,
because
of
the
fact
that
I
really
don't
need
another
source
of
funding,
necessarily
despite
you
know,
formally
being
paid
under
this,
and
these
are
the
way
that
will
is
getting
paid
where
he
doesn't
actually
need
money
immediately,
either.
There's
actually
some
money
floating
around
under
the
stuff
of
funding.
B
That
is
not
doesn't
have
a
clear
plan
for
usage.
So
if
folks,
like
David
or
anyone
else,
have
some
ideas
about
how
to
use
that
money
and
could
use
it
themselves,
I
mean
that's
something
that
I
would
like
to
discuss.
Getting
money
out
from
Goddard
to
another
institution
could
be
a
trick.
It
could
be
another
calendar
year
worth
of
of
mucking
around
bureaucracy,
so
don't
necessarily
expect
anything
fast,
but
you
know
if
we're
talking
about
a
major
effort
that
would
really
help
Senpai
I
think
we
should
consider
it.
B
Spent
more
money
through
non-focus
to
make
that
easier
it
it
could
help
the
the
the
funding
that
we've
set
up.
The
grant.
That's
set
up
is
kind
of
a
fixed
cost
cap.
B
So
had
we
thought
ahead
of
time,
maybe
we
could
have
done
been
that
a
little
differently,
but
it
might
be
possible
the
mod
deck
Grant
otherwise
setting
up
through
I
exaggerate
about
a
calendar
year,
I
think,
even
if
even
if
it's
not
an
unfocused
on
organization
instead
of
a
new
Grant
at
most
in
principle,
would
be
something
like
three
months.
Something
like
that.
So
it
shouldn't
be
that
bad.
B
B
That
could
be
a
much
faster
change
getting
started
because
that
that
is
actually
being
paid
through
a
grant
with
a
much
higher
cost
cap.
That's
not
directly
tied
to
the
stoffile
funding
and
it's
not
directly
tied
to
Will.
B
Or
a
plan
to
reduce
or
paying
someone
else.
Well,
that's
why
I
mean
like
right
now:
Will's,
not
getting
okay,
technically
speaking,
right
now
will
is
not
getting
paid
a
dime
out
of
this
ground
right
and
then
I
am
charging
some
very
much
lower
than
the
level
that's
proposed.
B
So
it's
just
the
nature
is
already
available
for
myself
and
well
that
it's
just
so
does
that
mean
you're
working
on
other
stuff
instead
of
working
on
that
they're
sorry.
Does
that
mean
you're
working
on
other
stuff
for
some
high
ground?
B
It
means
that
the
time
I'm
currently
spending
on
the
Senpai
Grant
is
probably
not
commensurate
to
the
amount
that
I'm
actually
charging
I
am
under
charging.
The
grant
for
the
mathematically
spent
time
she's
spending
on
something
okay,.
B
B
B
B
B
It
makes
sense
one
of
the
interesting
things
I
point
out
that
I
never
put
up,
but
it
was
it
came
from
a
helium
knots.
Discussion
was
that
you
can
use
the
spherical
screen
in
principle.
B
You
can
hack
it
to
use
it
for
to
be
the
Thompson
scattering
surface,
so
the
Observer
is
not
actually
any
Observer,
but
if
you
specify
the
observer
in
the
right
location
with
the
right
radius,
it
becomes
you
can
make
it
the
top
and
scattering
sphere
instead
of
the
a
sphere
centered
at
The
Observer,
which
is
kind
of
a
neat
way
to
hack
it.
B
But
that
would
use
the
existing
functionality
just
then
it
would
not
require
new
functionality,
but
if
we
wanted
to
officially
support
that
in
a
way
that
would
be
usable
for
easily
usable
that
user,
you
know
you,
you
would
call
it
something
else,
maybe
or
make
it
easier
to
say.
You
know:
I
want
a
server,
that's
half
the
distance
of
the
actual
Observer
distance,
or
something
like
that.
Yeah
I
guess
maybe
even
just
having
a
screen
object
that
represents
a
plane
or
a
sphere
with
a
center
somewhere
arbitrary
in
a
certain
radius.
B
B
Oh
yeah
functions
to
define
a
screen
based
on
a
create
a
screen
object
based
on
certain
other
things
like
I
want
the
screen
to
go
through
prefer
to
go
to
this
guy
going
through
this
point
or
something,
and
then
you
create
a
screen
function
that
creates
a
screen
based
on
that
information,
or
something
like
that
yeah
it
could
be
really.
It
could
be
a
nice
way
to
do
it.
B
B
B
B
Town
hall
is
the
documentation
is
difficult.
Would
you
say
difficult
useless?
No,
no
I
think
that's
extreme.
How
useless
it
is.
Okay,
so.
B
B
This
will
probably
have
to
end
up.
This
will
end
up
shaping
all
of
those
documentations.
They
live
together,
even
if
they
are
separate.
So
we
have
the
user
guides
which
exists
more
as
a
overview.
So
this
is
the
first
problem.
We
have
a
user
guide
that
hasn't
seen
any
changes
since
1.0
I
think
any
changes
that
have
happened
to
the
guide
since
one
by
the
primary
just
to
fix.
Oh
we've
changed
this
paper
all
right.
It
breaks
that
that
is
what
that
has
changed.
B
B
I'm.
Sorry,
that's
one
of
the
problems.
So
they
ask
when
you
say
it's
not
a
guide.
What
is
a
guide?
Well
yeah.
This
is
this
is
something
we
need
to
bring
me
on
to
this
way.
Yeah.
B
B
B
Part
of
the
problem
with
this
is
that
the
United
search
for
the
gallery
is
using
the
search
box
on
the
left.
There's
no
like
yeah,
there's,
no
like
tags,
there's
no
way
to.
Maybe
you
only
care
about
certain
instrument,
there's
a
way
to
you
can
search
for
the
instrument,
but
then
you
also
get
all
the
API
documentation.
If
you
search
through
the
left
browser
here,
such
as
technically
it
should
search
all
of
the
sub
packages
as
well
Affiliated
packages
Patrol.
They
are
linked
emotional,
moisture.
B
So
yeah
there's
no
way
to
search
through
there's
no
way
to
filter
for
the
gallery.
You
would
argue
that
maybe
having
this,
this
example,
which
generates
a
magnet
set
of
data,
doesn't
belong
next
to
following
the
parts,
for
example,
into
a
histogram
there's
no
kind
of
organization.
We
group
them
together
by
the
time
organization
within
them
and
how
they
should.
B
You
know
if
you're
searching
this,
this
organization
doesn't
feel
very
useful
to
a
new
user,
but
that
okay
I
mean
I'm,
not
disagreeing
with
what
you're
saying
I'm
just
thinking
that
means
I'm
thinking
like
would
having
tags
seriously.
Aid
discoverability
I,
don't
know,
but
I
feel
like
until
I
would
like
to
see
them
and
then
do
my
very
simply
word
I.
Think
if
you
had
a
list
of
tags,
I
think
a
fixed
list
of
tag
is
either
based
off,
like,
like
I,
think
an
instrument,
Source
type
of
relationship.
B
B
B
You
get
this,
you
get
the
very
nice
cut
out.
This
is
the
this
is
probably
my
my
favorite
example,
because
it
has
the
nicest
it
has
the
nicest
image.
Well,
yeah,
it's
not
useful.
In
that
sense,
maybe
tags
would
maybe
tags
with
AIDS
describe
like
another
like
the
first
time,
music
and
play
tag.
That's
just
like.
What's
math
and
doing
things
and
then
kind
of
advances,
I
would
argue
that
the
gallery
is
probably
not
the
entry
point.
Somebody
who
wants
Senpai
explained
to
them.
It
should
be
landing
well,
I,
guess
yeah.
B
The
gallery
has
very
short
examples,
but
it's
like
okay
I
want
to
make
a
random
array,
create
a
fake
WCS
and
we're
looking
at
Mac
and
the
next
one
is
like.
Oh,
how
do
you
reproject
it's
very
long
form
like
reprojection
that
we
have
yeah?
The
example
is
not
just
examples.
It's
like
some
are
very
depth.
Some
are
very
detailed
and
it's
like
just
together
and
it's
an
example
collaboration.
If
you
don't.
B
We
actually
want
one.
How
do
we
organize
that?
That's
something
we
maintain
a
project
website,
but
it
lacks
kind
of
I'm,
usually
Synergy,
because
I
think
we
have
a
documentation
where
we
have
a
project
website.
But
where
is
it
here?
It
clicks,
you
click
getting
hypnosis.
For
us,
this
is
a
rehash
of
half
the
half
the
text
that
is
in
the
subpart
application.
The
same
for
the
GitHub
page
like
this
is
a
project
documentation,
and
this
is
kind
of
useless
in
college.
B
B
A
project-
and
we
have
some
links
to
partly
going
back
to
Pokemon-
is
that
almost
reviewed
and
useful
part
of
religion
is
our
development
documentation,
which
you
know
a
lot
of
us
changed
up
by
change
it
all
the
time-
and
you
know
it's
really
good,
but
you
know
well,
that's
because
we're
users
of
our
developer
documentation
we're
not
using
the
user
documents
exactly
yeah,
exactly
which
is
the
problem,
and
then
we
discussed
this
this
morning.
There's
no,
why?
What
is
the
design,
the
philosophy
of
somebody?
Why
do
we
do
the
things.
B
Oh
yes,
my
grand,
unified
theory
of
documentation
was
I
think
this
has
been.
This
idea
has
been
rolled
out
by
a
lot
of
high-level
packages
that
paid
for
like
actual
technical
rights.
That's
the
place.
The
paper
tell
everyone
to
sit
down
work
out,
what
to
do
for
a
documentation.
Obviously,
working
on
weekends
afford
one
is
a
different,
a
different
topic,
but
I
think
until
we
get
one
we
kind
of
want
to
decide
what
what
documentation
will
want
to
look
like
before
the
entire
technology?
To
do
that,
because
sprinting.
B
B
B
Could
you
click
into
and
Show
an
example
of
that,
or
is
that
a
bit
too
on
the
play
I'm
trying
to
think
what
would
be
a
good
example?
I
mean
so
I've
openly
just
showing
yeah,
but.
B
So,
even
so,
our
landing
page,
for
example,
has
this
is
how
we
this
is
how
it's
structured,
but
James,
already
describe
here's
the
first
step.
Are
you
literally
knew
from
General?
This
is
this?
Is
the
link
you
click
on
all
right?
Are
you
more?
Do
you
want
a
tutorial?
These
are
this
tutorial?
These
are
the
parts,
do
more
advanced
things,
I've
been
getting
help
and
then
it
explains
you
like
how
the
documentation
is
actually
organized
on
the
landing
page.
B
B
Obviously
it's
that
theme
I
think
it
helps
this
because
I
don't
see,
they've
got
a
sidebar
where
they
have
like
getting
help
where
we
kind
of
have
to
have
like
a
section
of
a
page.
They
have
a
sidebar
that
has
that
kind
of
getting
help.
Information
right.
It
separates,
there's
a
clear
separation
between
it's
using
the
whole.
Obviously,
that's
a
limitation
of
our
thing.
We
can't
we
can't
quite
do
the
same
with
that
with
our
families
Maybe
something
to
look
into
the
future,
but
so
the
problem
is
I
had
to
lay
up.
B
How
do
we
lay
out?
What
is
this,
the
answer,
but
the
axis
of
documentation.
B
Guides
the
explanation
on
our
weapons
and
what
we
have
most
application
holds
an
app.
We
have
this
in
some
guides.
We
have
well,
we
don't
have
the
explanation.
We
have
the
reference,
which
is
our
API
Pages.
We
have
essentially
how-to
guides,
which
I
think
are
examples.
B
B
They're,
all
there
I,
don't
think
really
any
even
the
long-form
examples
in
the
gallery
to
me.
Don't
really
cross
the
the
zero
point
of
the
x-axis
on
that
bra
publishing
tutorials
for
people
how
to
learn
how
to
to
learn
how
to
do
sometimes
and
how
to
make
a
difference,
because
it
will
actually
exist
on
how
that
should
exist.
But
obviously
the
thing
we're
missing
is
the
explanation
is
the
understanding
of
why
things
are
probably
like.
B
Yeah
I
mean
I,
think
I
I
didn't
find
it
like
you're,
not
likely
to
stumble
across
that
page.
Put
it
that
way
and
I
also
I
don't
get
another
reference
documentation.
It's
got
extra
stuff
in
there
like
it's,
not
just
a
reference.
It's
like
this
is
this.
This
is
how
you
use
this.
It's
not
just
yeah,
it's
the
glasses.
B
Documentation
to
me
does
not
have
to
be
API
documentation.
Reference
documentation
is
documentation
that
you
are
using
as
a
reference.
It's
like
you
know.
How
does
this
keyword
argument?
Work
is
an
obvious
example,
but
it
could
also
be
like
you
know
what
does
coordinates
mean
when
it
says
Observer
yeah
right,
like
no
I'm,
just
I'm,
not
saying
that
it
has
to
be
just
the
API
and
nothing
else,
but
as
some
place
seems
to
have
a
lot
of
other
stuff,
that
probably
belongs
in
one
of
those
other
boxes.
B
B
B
B
B
Who's
opened,
he
was
kind
of
gone
through
our
guides
for
map
and
time
series
of
Europe,
because
I
left,
like
100
comments.
Thoughts
in
progress
I
have
been
making
changes
to
the
installation
page,
which
I
think
things
are
partly
one
of
the
reasons.
What
is
the
goal
of
the
project?
Sorry,
what
is
the
goal
of
the
call
libraries
documentation
versus
the
project
if
we
have
an
installation
knowledge
which
we
replicate
all
of
our
all
of
our
packages?
B
Something
would
be
different
across
anyone,
a
label
so
with
the
idea
of
having
a
generic
one.
That
was
on
our
website,
one,
which
is
we
replicate
across
everything,
it's.
What
is
the
yeah?
What
is
how
do
we
make
our
project
website
more
useful
to
everyone?
So
I
think
the
conversation
to
be
had
but
I
think
information
that
potentially
is
repeated
across
packages
could
be
really
up
a
level
into
the
project
website
by
everyone
you
mean
all
the
ability.
B
Yeah,
maybe
it's
the
project
website
is,
and
you
call
that
installing
I
know
as
an
example
but
they're
getting
help.
For
example,
I
think
almost
the
same
yeah
I
mean
the
the
initial
thought
I
have
on
that
is,
for
example,
if
you're
looking
at
you
know,
paper,
ffi
or
ndcube,
or
even
somebody
called
it's
nice.
B
And
then,
if
you
want
to
use
a
certain
package,
you
know
go
to
the
specific
like
documentation
for
that
package
and
stuff,
like
installing
I,
think
it's
sort
of
very
low.
It's
the
same
across
all
packages.
It's
still
specific
to
I
want
to
know
about
the
a
package
and
therefore
installing
is
you
know
a
package
specific
task,
but
yeah
I
mean
maybe
going
back
to
this.
Why
thing
and
talking
about
the
ecosystem
of
philosophy?
B
Maybe
that
project
level
is
the
way
that
lives
I,
think
maybe
there's
also
I
think
the
four
different
types
of
documentation.
Some
of
them
can
be
project
level
and
some
can
be
package
level.
So,
like
your
reference,
documentation
is
all
on
a
package
by
package
basis,
the
API,
but
perhaps
tutorials,
which
are
learning
oriented.
We
don't,
oh
sorry,
how
to
guides
which
are
problem
oriented.
We
don't
want
to
distinguish
between
packages.
B
We
like
we
want
to
have
an
example
gallery
that
is
how
to
roll
AIA
or
how
to
register
AIA
images
in
with
all
the
rest
of
those
how-to
guides.
So
someone
who's
doing
solar
physics,
kind
of
doesn't
see
the
boundaries
between
packages
but
they're,
just
naturally
using
all
the
packages.
Yeah
yeah
I.
Think
in
the
examples
in
the
how-to
guys,
it
would
be
great
to
have
like,
like
search
functionality
like
cross
cross
package
search
functionality
across
like
multiple
or
multiple
example,
galleries.
That's
what
I'm
trying
to
say
so!
B
Yeah
I,
like
I'm,
looking
how
to
do
like
no
I
might
not
be
looking
how
to
use
Senpai
I'm.
Looking
at
like
how
to
do
a
tillage
translation
from
an
HMI
magnetogram
that
involves
you
know
a
lot
of
different
steps
and
maybe
I
want
to
like
then
lay
those
on
top
of
an
AIA
image
that
I've
registered
I
mean
that
involves
all
these
pieces
of
the
ecosystem.
How
do
we
enable
someone
to
get
all
that
information.
B
B
B
Because
if
you
start
veering
over
between
any
of
them,
then
where
does
someone
go
yeah,
it's
more
like
if
you
read
the
background
on
this,
it's
more
it's
kind
of
come
out
of
who
are
the
people
who
are
using
your
documentation
right?
What
what
do
the
people
who
are
using
the
documentation
want
to
get
out
of
it
like
you?
Can
you
can
imagine
some
different
people
or
using
the
documentation,
like
you,
can
imagine
one
person
walking
up
the
documentation?
B
Who's
never
used
python
before
doesn't
know
where
to
start
doesn't
even
really
understand
like
what
zombie
is
right,
they're
going
to
have
very
different
questions
towards
the
documentation
versus
like
me
when
I
forget
how
to
use
reprojective
right,
one
is
very
kind
of
tutorials
and
explanation
and
the
other
one
is
like
the
deepest
dark.
It's
called
a
reference,
you
know,
and
if
you
mix
them
up,
then
that
what
page
doesn't
match
for
any
one
person
right
you
you,
don't
you
can't
like
find
the
right
page
for
you.
B
Yeah,
so
just
to
quickly
site,
so
the
other
two
progress
is
that
we
have
a
checklist
for
a
few
developers
who
contacted
Senpai.
What
should
I
check?
What
should
they
do
when
they're
putting
progress
to
the
first
line?
What's
on
these
reviews,
please
review
so
the
last
progress
is
me
going
crazy,
so.
B
I
went
through
the
development
documentation
again
and
the
about
documentation.
I
tidied
up
some
of
it.
Some
of
it
is
missing
which
I
planted
at
the
back.
For
example,
I,
have
moved
Albert's
exploration
in
the
coordinates
that
are
nested
in
the
reference
in
the
API
documentation
of
each
of
the
guide.
For
now
with
the
idea
that
eventually
we'll
be
moving
back,
but
that's
so
my
plan
was
to
the
problem,
would
do
the
doxary
factor.
Is
that
there's
a
lot
of
files,
the
little
text?
So
the
question
becomes?
B
B
So
we
have
a
clean,
API
documentation,
though
this
bear,
and
then
you
can
decide
where,
where
the
information
should
be,
what
should
it
be?
Where
should
it
be?
The
problem
is
as
a
consequence
that
it
makes
the
rest
of
documentary
slightly
worse,
because
it's
got
it.
There's
a
lot
of
scaffolding
around
it
until
we
finish
it,
and
so
that's
a
discussion
I
think
to
be
had
maybe
obvious
PR
maybe
hear
about
also
how
we
learn
about
this
because
yeah
we
have
a
branch
on
Main.
B
We
can
brush
off
master's
name,
have
a
brands
that
we
have
to
just
keep
on
placing
with
Coach
engines
requirements
all
again.
Eventually
we
merged
that
back
in,
but
there's
only
one
massive
pull
request.
Unless
people
are
reviewing
the
individual
changes
we've
made
to
our
Branch,
depending
on
people's
level
involvement.
B
B
The
aspects
I
think
that
kind
of
goes
back,
because
I
think
also
remember
that
I
think
what
we're,
how
we
automatically
also
eventually
inviting
everything
sponsored
package
under
the
GitHub
repository
list.
It's
something
we're
going
to
have
to
eventually
well
I,
think
yeah,
I
I
think
it
is,
but
from
both
the
technical
and
a
philosophical
perspective,
we
need
to
document
the
whole
project,
not
seven
different
packages.
Yeah
like
like
we
were
talking
about
earlier
with
the
messaging
about
the
multi-package
project.
B
B
Of
course,
assuming,
like
you,
still
want
each
other
packages
to
have
their
own
documentation
and
the
bulk
of
the
detail
of
what
that
package
does.
Is
there
right
or
am
I
misunderstanding
yeah
I
think
so?
There's
a
bit
of
a
conflict
there
in
my
head
between,
like
technical
maintenance
and
end
result,
yeah,
like
so
I,
think
for
our
own
sanity,
yes,
but
for
a
user
perspective
coming
into
the
documentation,
not
as
much
it's
like.
You
were
saying
with
the
example
Gallery
earlier
yeah
yeah,
it
would
be
lovely
to
have
one
unified
example
Gallery.
B
It
would
also
be
enlightenmentation
well
yeah,
because
if
you
have
different
or
conflicting
version
requirements
for
all
those
differences,
all
your
versions,
yeah
I
mean
that
you
I
mean
don't
get
me
wrong.
There's
always
Technical
Solutions,
but
it's
so
like
that's
a
massive
Undertake
and
also
moving
that
somewhere
else
increases
the
contributor
workload
but,
like
oh
I've,
got
an
example
to
add
to
ndq,
well
yeah,
but
I
don't
feel
that
2020,
oh
I
wrote
the
example
of
a
different
person
of
this.
B
Could
some
random
dependency
solution
where
each
sub
package
has
its
own
example
gallery
and
then,
when
you're,
building
a
project
documents
like
you
pull
in
all
of
the
examples,
obviously
there's
still
the
same
part
of
the
versioning.
That
would
be
what
I
would
like
to
have
to
happen.
Yeah,
because
then
you
can
have
the
tags
that
you
talk,
that
you
mentioned
earlier
you're
on,
so
I.
Don't
think
that
works,
because
what,
if
you
say,
have
a
okay,
it
doesn't
exists
now,
but,
let's
imagine
there's
a
HMI
Pi.
B
B
Well,
at
that
level,
like
you
call
it
the
project
level,
the
project
can
have
its
own
example
gallery
and
then
you,
but
then,
if
I
want
to
do
a
task
and
do
I
go
to
the
project.
How
do
I
know
which
of
the
10
galleries
to
go
to
no
yeah
I
was
saying
you
would
merge
them.
Yeah
I
mean
they
have
to
present
separately
right
yeah.
So
it
would
be
the
ideal
you
like
the
yeah
from
a
user's
perspective.
B
B
You'd
have
like,
like
direct
links
from
like,
say
to
ND
Cube
and
when
you
click
on
that
link
to
go
to
that
page,
it's
actually
like
in
the
ndq
repo,
but
from
from
a
user
point
of
view,
I've,
like
click
on
this
project
website
and
I,
just
click
the
link
to
that
documentation
and
click
back
or
whatever.
No
actually,
I
can
still.
There
still
has
to
be
a
home
for
examples
that
you
and
loads
for
packages
I.
Think
it's
in
a
different
part
of
this.
B
Like
the
can
you
go
back
into
the
slide
that
we
all
like?
Oh
yeah,
like
the
example
galleries
for
me,
live
in
the
the
how-to
guys
like
in
the
in
the
right
quadrant
and
then
like
what
you're
talking
about
this
thing?
Is
this
no
I'm
talking
about
how-to,
guys
you're,
so
you're
exclusively
talking
about
how
see
for
me
I,
think
of
like
when
we're
when
we're
starting
to
glue
multiple
parts
of
the
ecosystem
together,
that
is
that's
traveling,
leftward.
B
You
know
like
if
you
yeah
yeah
say
there
was
your
HDMI
Pi
example.
Maybe
you
just
use
AIA
pipe
for
this,
but,
like
you
you
know
you
before
you
do
your
extrapolation.
You
want
to
update
the
coordinate
information
right,
like
all
you're
doing
is
import
HMI,
Pi
dot,
you
know,
update,
coordinate
information
running
it
and
carrying
on
I
would
argue
that
that
gallery.
That
example
goes
in
the
Creator.
Yeah
interaction
is
minimal,
but
you
know
whatever
that's
like.
B
I,
yes,
okay,
yeah
I
think
so
I
think
definitely
I
I,
don't
I
think
it's
true
sometimes,
but
like
no
you're,
both
right.
Like
sorry,
it's
like
you
I,
know,
you're
about
to
say
something.
No
I,
think
you're
gonna
say
the
same
thing
like
that
example
is
like
hyperbo
sets.
You
don't
really
care
about
the
things.
The
examples
you
show
you,
though,
and
then
somewhere
you
could
have
a
tutorial
which
is
Project
orientated.
We
could
use
all
of
them
and
explain
all
of
the
details.
I
guess
I
think
yeah
yeah.
B
But
the
thing
is
you
don't
need.
You
can
have
an
example
that
use
multiple
packages,
but
doesn't
explain
why
yeah
yeah
you
just
okay,
I'm
gonna
load
this
HMI
image
now
to
make
it
more
accurate
I
need
to
do
this,
coordinate
fix
first
and
then
I
carry
on
with
the
example
or
like,
and
what
you
can't
have
is
a
link
to
the
explanation.
Yeah.
B
This
is
how
this
is.
Why
yeah
that's
fair,
that's
fair,
not
to
mention
the
thing
that
we've
often
complained
about,
but
like
this
project
level
stuff,
it's
we're
kind
of
moving
towards
modern,
like
an
ssw
conception
in
this,
in
the
sense
that
you
know
I
just
like.
Well,
we
all
have
this
thing
and
I
just
want
like
I'm,
a
scientist
I,
don't
care
where
it
is.
B
I
just
want
a
workflow
if
I
have
to
import
it
from
different
packages
again
I,
don't
care
I
want
to
I
want
to
do
these
tasks,
and
then
this
project
documentation
is
like
sounds
like
it's.
It's
allowing
a
mental
map
along
those
lines
for
users.
You
know,
and
then
technically
you
know
how
we
divide
it
up
is
is
a
slightly
more
developerative
thing,
but
you
know
like
I
want
a
time.
Sorry
go
ahead.
B
No,
no
sorry
I'm
just
going
to
build
on
that
like
what
is
the
scope
of
the
documentation
like
do
we
have
responsibility
to
explain
what
WCS
is
and
why
you
why
this
is
important
and
sometimes
kind
of
becoming
maybe
just
going
a
bit
off
topic,
but
sometimes
becoming
a
place.
We're
also
sharing
knowledge
of
some
point.
How
those
impact
knowledge
that
you
know
just
talking
to
people
that
are
working
on
it
well,
yeah,
yeah,
I,
think
yeah.
Why
is
HMI
up
today,
yeah
seriously?
Why.
B
The
fact
we're
missing
that
explanation,
quadrant
almost
entirely
from
our
documentation,
is
probably
really
hurting
those
people
who
are
like
the
big,
the
beginners
I,
think
it
absolutely
agree.
It's
easy
to
it's.
If
you
follow
a
guide
that
doesn't
have
an
explanation,
then,
as
soon
as
you
have
to
deviate
from
the
gate,
you
can
make
fundamental
misunderstandings
in
how
you're
applying
certain
tools,
whereas
if
you've
got
the
explanations
for
someone
that
just
wants
to
load
an
HMI,
they
can
follow
their
follow.
B
The
you
know
the
how-to,
but
if
someone
then
wants
to
be
like
okay
now,
how
do
I
do
this
next
step?
They
might
they
need
those
explanations,
so
they
don't
misuse
the
tools
that
that
we've
provided
or
like
worse
yet
instantly
get
stuck
and
like
this
result
is
too
hard.
It
can
only
do
the
things
that
you
know.
They
explicitly
tell
me
to
do,
and
I
can't
do
anything
else.
Yeah
yeah,
I
was
gonna,
say
I,
think
also.
B
If
we
had
really
good
quality
extra
excellent,
all
started
they're
more
likely
to
like
fall
in
love
with
using
zombie
and
stick
around.
They
are
if
they
get
like
confused
and
frustrated.
They
might
be
more
related
to
ask
a
question
as
well
like,
because
once
you
get
like
a
bit
of
understanding,
you
start
knowing
the
gaps
and
then
then
you
know,
okay
and
now.
I
know
that
I'm
missing
this
knowledge
to
ask
a
question:
who's
done.
You
can
have
different
levels
of
tutorials.
B
There
are
places
that
have
like
brand
new
to
the
package
tutorial
that
literally
just
do
something
and
and
then
you
like
intermediate
like
a
little
bit
more
complex
and
then
like
really
Advanced
tutorials,
where
you're
going
into
a
whole
workflow-
and
you
say,
while
you're
doing
it
whatever
and
link
to
the
apis
and
stuff,
but
so
that
someone
can
go
from
I,
don't
know
what
a
map
is
I
just
up
here.
Here's
a
map
to
I
want
to
differentially
rotate
and
extract,
coordinates
and
do
something
that's
an
advanced
tutorial.
B
But
if
you
have
it
in
a
tutorials
page
and
it's
labeled
that
with
internet
and
find
it
yeah,
and
then
you
link
the
explanations
of
go
and
read
our
coordinate
stock
explanations.
Understand
why
the
coordinate
systems
work
this
way
and
see
a
better
map
decision
to
understand
why
Maps
work?
This
way,
that's
like
actually
really
interesting
because
at
the
test
meeting
last
week,
not
actually
after
the
session,
but
just
talking
to
people
telling
them
to
come
to
the
session.
B
There
is
a
huge
gap
between
starting
to
learn,
Supply
and
knowing
what
you're
doing
and
the
documentation
so
someone's
saying
like
I,
don't
like
someone
coming
from
ssw,
you
never
really
used
twice
or
some
part
before
it
said
they
could
use
the
very
first
start
of
it
and
then
the
documentation
went
way
over
their
heads.
So
there's
not
as
good
of
a
scale
if
you're
kind
of
doing
intermediate
things
at
some
point,
you
need
to
stick
your
head
into
the
source
code.
B
Don't
you
to
understand
really
and
then
that's
a
step
too
far,
but
what
Shane
was
saying
is
like
there's
so
much
scope
there,
potentially
there.
If
you're
talking
about
beginner,
you
can
even
link
to
if
you're
talking
about
these
explanations
that
go
in
a
bit
in
more
depth
as
to
why
things
are
done.
That
way,
you
could
link
them
at
relevant
points
in
the
tutorials
say
like.
Oh,
it
might
help
you
to
understand.
If
you
read
this
explanation
on
WCS
yeah
yeah.
B
B
A
deeper
level
of
explanation
like
structure,
it's
not
even
a
vaccinations,
but
you
know
for
the
simplest
level
what's
wrong
with
on
the
landing
page.
Basically,
you
haven't
done
headings
like
you
can
go
and
pause
them.
Jacob
calls
them
topic
guides,
which
I
think
is
quite
a
nice
little
pricing
Maps
coordinates
timing
series.
They
could
each
other
instruments
specific
things.
You
should
know
the
coordinates
of
the
really
the
first
things
because
they're
like
basic
underlying
about
something
and
these
decisions
are
made
about
how
they
work
and
why?
B
What
certain
ways
that
aren't
explained
unless
you're
going
to
read
it.
So
it's
what
they
need
to
be.
Do
that
doesn't
tell
you
why
it's
that
way?
It
tells
you
the
way
it
is
but
enough
that
this
way,
because
we're
probably
10
other
ways
and
there's
problems-
and
this
is
this-
is
the
way
that
it
works,
because,
historically,
you
have
to
leave
it
this
way
to
make
something
else.
B
First
I
wonder
one
of
the
reasons
we've
got
this
Gap
that
you've
mentioned
is
because
we've
up
to
now
have
thought
that
our
responsibility
to
explain
like
some
type
of
stuff
and
conceptually
a
lot
of
that
is
actually
an
astrophy,
so,
like
a
sky,
coordinate
the
idea
of
like
the
fact
that
in
software,
you've
got
to
think
in
three-dimensional
coordinates,
whereas
again
it's
so
much
of
like
it's
a
double
user
right.
Suddenly,
Got
Away
through
my
whole
PhD.
B
If
I
thought
about
an
image,
it
was
an
image
with,
like
you
know,
a
Cartesian
grid
of
like
you
know
seconds,
and
there
was
no
sense
of
like
where
the
Observer
was
like
the
idea
of
like
reprojecting
for
different
Observer
I.
Never
really
thought
about
that
until,
like
you
start
talking
about
like
big
words,
so
like
we're,
still
repeatedly
confused
senior
solar
physicists
with
the
fact
that
Helio
projective
cares
about
the
Observer
yeah.
No,
it's
not
right.
B
I,
like
the
fact
that
you've
got
to
put
a
time
into
a
scope,
I
just
care
about
space.
I.
Think,
like
you
know
this
sort
of
stuff
right,
so
so
I
think
that
maybe
one
of
the
reasons
the
Gap
exists
is
oh
well,
that's
Astro
quiet,
so
people
should
go
to
astrophy
and
learn
that
stuff.
But
again
you
sort
of
have
this.
B
You
know
this
like
welcome
to
senpai
or,
like
you
know
this
huge
big
app,
which
is
basically
all
the
stuff
that
some
people
Builds
on
Azure,
quiet
and
then
back
to
like
you
know
what
Senpai
does
building
on
top
of
that.
So,
like
maybe
that's
one
of
the
reasons
you
don't
have
to
recreate
their
documentation,
but
you
can
certainly
outline
what
it
is
and
then
point
point
to
their
documentation
in
certain
parts
where
it
makes
sense.
B
B
Hey!
Can
you
guys
hear
me
yeah,
yeah,
I,
I,
just
been
listening
to
this
and
I
must
admit
it's
a
little
hard
to
listen
and
it's
a
little
bit
muffled
on
the
on
my
end
but
yeah,
one
of
the
things
I've
I've
kind
of
been
thinking
about
you
know,
sometime
in
terms
of
approachability
of
our
code
and
in
terms
of
like
researchers
right
like
I
would
say,
the
difference
is
there's
like
kind
of
Baseline
IDL
right.
B
Let's
say
a
researcher
has
an
ideal
background
right
and
there's
kind
of
like
Baseline
IDL,
and
then
there
are
more
complex
packages,
for
example
like
the
resi
package,
which
I
don't
know
if
anybody
is
familiar
with
it,
I
I
know
some
of
you
are
familiar
with
it,
but
it's
very
object-based.
It
is
a
very
dense
package
and
I
think
that's
where
we
are,
unfortunately
like
a
normal
user.
B
You
know,
can
use
our
code,
but
as
soon
as
they
need
to
sort
of
do
something
that
is
custom
or
dig
into
the
source
code.
It
is
no
longer
possible
for
them,
and
you
know
with
resi
what
we
would
do
is
like
we
would
go
find.
You
know
Richard,
Schwartz
or
others
to
like
try
to
fix
things
for
us
and
add
capabilities,
because
it
was
just
impossible
to
do
it
on
our
own
and
I
I.
B
Wonder
if
that's
something
we
should
consider
like
try
in
a
long
term,
moving
making
code
easier,
simpler
and
more
approachable
such
that
you
know
a
researcher
can
go
into
the
code
and
and
sort
of
be
like.
Oh
okay,
I
see
what
they're
doing
here.
I
just
need
to
make
this
change.
Does
that?
Does
that
make
sense?
B
B
You
know,
use
less
complex
python
approaches
that
kind
of
stuff.
You
know
like
just
make
it
more
approachable
and
try
to
keep
it
simpler,
I
think
of
two
things,
I
think
in
general
we
probably
try
and
keep
our
code
like.
B
We
don't
write
arbitrarily
complex
because
we
don't
write
complex
codes
for
no
reason
generally
like
there
are
bits
of
senpai
that
are
very
technical
python,
but
they
exist
to
provide
a
nice
easy
user
interface
right,
like
let's
take
the
most
gnarly
bit
of
somebody
in
terms
of
Python
Programming,
which
is
how
the
Atta
system
works,
with
registering
values
of
assets
and
the
tab.
Completion
on
assets
right
right
to
really
understand
that
bit
of
code.
B
B
Think
it's
important
to
differentiate
between
the
things
that
your
standard,
your
spherical
chicken
researcher,
is
likely
to
actually
want
to
change,
like
it's
unlikely
that
that
person
is
like
going
to
want
to
change
how
the
Fido
atas
system
works
right.
B
They
should
be
able
to
write
a
new
Fido
fight
and
without
understanding
how
the
depths
of
the
Fido
actor
system
works,
or
even
more
likely,
they
should
be
able
to
add
their
image
processing
routine
to
sunk,
your
image
move
out
understanding
any
of
the
really
complex
stuff,
so
I
think
there's
a
place
for
both
like
I
I.
You
know
you
have
to
think
about
the
more
common
contribution
paths
and
you
know,
there's
always
going
to
be
a
scale
of
from
easy
to
hard.
B
Probably
a
number
of
times
at
me
is
that
people
don't
know
where
to
find
out
what
attributes
are
on
a
map.
So
when
you
go
in
look
at
a
map-
and
you
look
at
the
if
they
go
to
a
map
and
a
map,
they
see
stuff,
but
it's
only
if
you
know
to
go
to
generic
map
which
it
inherits
from
it.
You
see
all
of
the
possible
things.
B
I,
don't
know
how
many
times
that
happened,
that
people
are
asking
me
so
as
opposed
to
where
are
they
looking
to
start
with,
and
it
ended
up
in
the
reference
documentation
yeah
on
AIA
map?
Well,
no,
if
you
just
go
to
map
yeah
and
well,
it's
obvious
to
me
right.
I
would
just
point
out
that
AIM
app
does
list
every
attribute,
so
it
doesn't
have
a
nice
table.
Okay!
Well,
actually,
that's
been
on
Facebook,
but
is
that
all
of
them?
We
didn't
change
that
at
some
point.
Okay,
right!
Yes,.
B
Sorry,
I'm
not
quite
sure
which
part
you
want
me
to
respond
to
about
about
how
we
write
our
code
and
what
you
know
how
we
approach
it.
Yeah
I
was
just
wondering
if
you
wanted
the
opportunity
to
say
anything
else
on
that
before
we
diverge,
oh
yeah,
I
mean
I,
agree
with
you,
Stuart
I,
don't
think
we're
writing
code
that
is
unnecessarily
complex.
B
I
do
think
that
we
frequently,
you
know,
use
approaches
and
techniques
that
are
that,
for
us
are
perhaps
not
Advanced,
but
for
a
beginner
python
user
are
Advanced
and
we
don't
really
think
about
that.
But
I
mean
that's
just
the
nature
of
code.
I
mean
I
I,
don't
know
that.
There's
really
a
good
solution
to
this.
Besides,
just
accepting
it,
you
know
I
think
Stephen,
I,
don't
know
what
you're
what
the
problem
is
that
you're
talking
about
like
what.
B
The
problem,
I
think,
is
that
we
expect
that
our
users
will
be
good
enough
to
kind
of
look
through
the
code
like
like
we
have
in
the
past
and
kind
of
make
changes
or
contribute
or
or
you
know
or
get
the
stuff
that
they
want
to
do
like
I've
heard
complaints.
You
know
I
couldn't
figure
out
how
to
do
this
in
Senpai
and
so
I
kind
of
hacked
it
my
own
way.
B
You
know
why
didn't
they
kind
of
contribute
back
to
Senpai?
Why
did
they
hack
this
I
think?
Maybe
we
should
just
Embrace
that,
like
people
really
think
of
senpai,
it's
kind
of
like
a
black
box,
like
you
know,
I'll
make
like
oh
specs,
for
example,
but
to
me
oh
specs
is
a
black
box.
If
it
doesn't
do
what
I
needed
to
do,
then
you
know
you
know:
I
have
to
go
top
to
people.
B
That
know
enough
about
uzbeks
to
add
a
capability,
and
so,
if,
if
we
think
of
some
pie
that
way,
then
we
have
to
work
with
our
users
differently,
I
think
than
how
we
have
it
an
example
that
you
just
provided.
If
they
aren't
going
to
look
at
the
Cub,
is
they
treat
it
like
a
black
box?
They're
gonna
have
to
talk
to
us
regardless,
like
I,
don't
know
what
you
know,
writing
more
approachable
code
internally
does
if
people
aren't
actually.
B
I
wouldn't
expect
most
normal
users
to
open
the
source
code
of
some
kind.
I,
certainly
wouldn't
like
to
think
that
we
write
our
package
and
documentation
expecting
people
to
need
to
do
it
yeah.
We
think,
but
I
think
this
brings
up
a
point
like
I
do
go
like
I've
got
to
get
up
and
look
at
the
source
code,
but
I
think
also
I
mean
like
I.
Don't
know
this
object
not
like
from
a
developer
point
of
view
from
a
user
point
of
view.
B
Yes,
no
I
was
just
going
to
say
I,
don't
think
I
think
Steven's
Point
is
totally
correct,
but
I
do
think
you
end
up
looking
at
the
source
code
as
soon
as
you
go
out
with
the
kind
of
the
Black
Box
thing
and
then
there's
there
are
parts
of
the
codes
that
are
I.
Don't
think
you
should
change
them
because
they,
like
students,
said
they
are
there
to
accomplish
certain
tasks
like
all
the
auto
class
magic
stuff
like
that
is
like
unintelligible
to
anyone
that
isn't
a
python
developer.
B
B
B
I
mean
our
answer
to
that
is
probably
not
something
I'm
available
on,
but
you
will
potentially
be
enabling
they
have
GitHub
account.
It's
enabling
the
discussion
feature
yeah,
like
I,
think
like
the
Advent
of
Helio
nuts
has
been.
You
know
whether
you
know
fracturing
the
the
kind
of
communication
channels
that
we
have
is
good
or
bad
I.
Think
helium
nuts
has
been
like
a
very,
become
like
a
very
informal
book
like
place
to
people
for
people
to
discuss
their
python
issues
or
some
high
issues.
B
That's
where
a
lot
of
that
discussion
has
come
up
and
I
think
because
people
see
that
as
more
of
like
an
informal
kind
of
like
safe
space
is
not
the
right
word,
but
I
mean
in
some
ways
yeah
it's
something
they're
not
intimidated
by
and
I
think
they
are
intimidated
by,
like
our
maybe
our
issue,
tracker
or
like
even
the
Google
group
or
the
discourse,
because
they
see
that
as
like
a
senpai
space
and
Helio,
not
more
of
like
a
community
space.
So
I
think
the
question
is:
how
do
we
turn
the
r
like
spaces?
B
I?
Think
we
see
as
Community
spaces
I,
don't
think.
Like
necessarily
people
outside
of
like
the
the
contributors,
see
that
I
think
Steve
wanted
to
yeah
yeah
I
mean
yeah
I.
Think
that's
a
that
is
a
helpful
thing
like
creating
a
user
space
where
it's
really
just
users
trying
to
talk
to
each
other,
trying
to
figure
things
out,
but
I
think
the
other.
B
B
B
And
providing
the
solutions
as
soon
as
possible,
yeah,
but
as
soon
as
possible,
depending
on
the
feature
request
could
be
years
like
we
don't
have
a
lot
of
like
resources.
If
people
every
time
people
have
come
claws,
we
can
see
relatively
quick
to
understand.
We
might
not
be
necessarily
quick
to
release
the
thing,
because
if
you
responding
to
a
problem
and
getting
emerged
and
released
is
up,
it's
not
an
immediate.
It's
not
an
immediate
timeline.
B
But
I
feel
like:
if
people
encounters
with
problems,
we
have
been
responsible,
I,
don't
yeah,
I,
think
I
think
we
do
a
good
job
of
responding
to
things.
Obviously,
if
people
come
to
us
with
a
large
feature
request
that
can
feature
request,
I
think
that
difference
because
we
don't
have
the
developer
resources
to
do
it
yeah
exactly
I
think
we
have
some
bugs
specifically
or
problems.
They've
had
I
thought,
I.
Think
we've
slightly
veered
off
the
discussion.
Well,
I
think
this
is
the
problem
with
the
documentation
is
that
it's
so
wide
scale
work.
B
Oh,
we
don't
have
leave
it
until
we
start
to
yeah
and
I
mean,
as
Albert
pointed
out,
there's
a
couple
of
things
in
coordinates
in
topic
guides.
Even
if
that's
all
that
there
is
there
for
the
moment,
so
I
don't
think
it
would
be
like
literally
empty
ing
to
Japanese.
Tell
us
what
you
want
to
hear
about.
I.
Don't
know
yeah
yeah
like
what
you
said
on
the
on
the
document.
B
This
is
like
way
like
five
six
years
ago,
like
on
the
documentation
page
on
the
GitHub
page,
there
was
like
a
little
IRC
box
that
you
could
just
like
start
typing.
It
was
like
the
chat
window
was
in
Wisconsin.
Oh
really
yeah
on
the
website,
Yeah
well,
yeah,
unfortunately
I
mean
I.
Can
I
can
rant
about
this
over
here,
but
unfortunately,
element
betrayed
me
on
that
one
I
am
so
yeah
okay,
so
we
can
add.
We
can
add
the
four
sections.
B
I'll
drop
pill
request
I
want
to
tidy
up
the
document,
the
API
documentation
to
a
point
where
we
can
it's
in
a
state
where
we
can
start
adding
stuff
to
it.
It's
not
I,
don't
think
it's
currently
I
want
to
get
that.
Obviously
that
means
some
of
that
text
100
of
the
guide,
which
will
make
it
worse,
but
well,
no
I
think
if
we
add
I,
think
if
we
add
a
Johnny's
doing
that
pull
request.
B
B
That
just
exists
somewhere.
This
is
written
work
that
was
just
sitting
there
being
unused
unadvertised
and
we
end
up
rereading
tutorials
every
like
I
mean
often
you
sort
of
pull
in
this
stuff
from
other
notebooks
area.
That's
it
yeah
I
mean,
as
you
say,
in
somewhere
on
one
websites.
You
said
they
just
they
just
go
there
and
they
Rock,
because
they
fall
on
so
I.
Think
I
have
I.
B
I
think
I
think
the
topics
they
cover
and
those
notebooks
could
be
used
as
a
really
good
foundation
for
writing
long-term
tutorials,
but
I
think
from
like
pedological
perspective.
You
need
to
take
a
different
approach
when
you're
not
standing
in
front
of
somebody
talks,
so
I
think
there's
like
still
a
significant
amount
of
work
there
to
turn
those
into
a
good
website
document
taking
content.
B
So
we've
basically
started
that
by
saying
you
know
the
four
sections
in
the
landing
page
I
think
maybe
another
thing
would
be
to
write
up
a
description
of
how
this
documentation
is
or
in
realistic
will
be
arranged,
and
that's
going
to
clarify
a
lot
of
our
thinking
on
this
so
just
to
I
want
to
just
before
I
answer
the
student's
question.
B
Obviously
this
brings
us
to
the
point
of
our
sidebar
and
having
all
of
these
subsections,
whether
or
not
we
need
to
reorganize
this
completely.
Obviously,
yeah.
Well,
another
thing
that
so
this
bit,
but
we
have
installation,
music,
release
history
emerging
citing
known
issues,
we're
not.
This
sidebar
needs
to
be
reorganized.
Actually,
after
Danny
mentioned
any
Cube
ducks
earlier
in
this
discussion,
I
went
and
looked
at
them
and
specifically
at
the
sidebar.
The
sidebar
for
the
Indy
Cube
doc
seems
very
quick.
B
I
want
to
like
use
ndq,
whereas
for
on
our
our
docs
for
our
sidebar,
it
seems
just
more
like
an
organizational
tool
right,
but
this
is
like
this
it.
It
seems
like
it's
meant
for
you
to
like
sort
of
Step
through,
like
the
section
like
each
each
kind
of
point
on
the
sidebar,
like
it's
a
guide
to
the
documentation.
B
B
Documentation
is
basically
just
a
tutorous
like
a
tutorial.
Has
the
tutorial
category
of
so
the
ndcube
one
yeah
yeah
yeah
and
maybe
API
yeah
okay,
so
you
have
API
I,
think
it's
basically
a
big
tutorial
and
then
reference
yeah,
so
I
think
yeah.
It's
a
big
tutorial.
Api
reference
developer,
documentation,
change,
yeah,
so.
B
Yeah
yeah
yeah,
yeah,
yeah
and
just
students
a
question:
do
we
have
much
feedback
from
users
that
don't
lose
senpai
calibration,
wasn't
good
enough
and
I
guess
the
question
was
this
something
that
came
up
in
tests
because
we've
had
a
big
problem
with
getting
feedback
from
users
in
general,
and
was
this
something
that
came
up
in
test
yeah
I?
Guess
it
was
I
mean
it
didn't?
It
was
hard
for
new
users
and
two
was
people
kind
of
I
think
we're
talking
about
like
it
was
like
the
thing.
B
Oh
someplace
can't
do
this,
and
then
they
stopped
there
without
thinking,
even
though
it
can
do
it
right.
Yeah
that
was
their
oh
I
can't
Theory,
it's
not
the
documentation,
it
doesn't
exist,
whereas
there
is
actually
a
lot
of
stuff
that
does
exist
or
recurrence.
That
is
oh,
but
it's
just
so
basically
the
same
okay
right.
B
This
is
whether
it
failed
is
probably
one
thing
would
be
running
the
survey
again
and
seeing
like
I
think
because
I
think
like
on
the
last
survey,
the
Jew
and
Monica
ran.
We
did
get
a
lot
of
like
negative
feedback
and
then,
like
people
saying
negative
things
about
package
which
I
think
is
that's
I
mean
it
kind
of
what's
up,
but
I
think
that's
that
that
is
helpful
and
sometimes
you're
not
going
to
get
that
in
any
other
way.
Yeah,
that's
good.
B
B
B
Think
we've
got
to
the
point
where
we
needed
to
stop
and
have
a
think
about
structure
whole
scale
structure
of
the
documentation,
but
once
we're
happy
with
that
and
have
put
that
in
place,
we
can
go
back
to
making
more
incremental
improvements.
Writing
more
content,
reorganizing
things
a
little
bit.
It
doesn't
have
to
be
like
burned
down
and
start
it
again.
So
I
think
what
might
be
helpful
like
not
as
a
group
in
this
discussion
but
like
if
people
want
to
think
about
this
on
their
own.
B
Since
you
mentioned
you
looked
at
the
end
of
the
cube
dogs,
I
think
at
the
end,
I
think
anyway,
as
opposed
to
wrote
them
so
not
a
very
good,
independent
judge,
but
I
think
the
only
Cube
does
a
better
job
at
some
point
of
trying
to
explain
the
why?
But
it's
mixed
in
it's
not
separated
from
like
the
tutorials
of
its
sites,
so
any
Cube
doc.
Certainly
I've
now
realized
they're
far
from
perfect,
but
thinking
about
well.
B
What
does
Senpai
do
well
and
what
does
ND
Cube
do
well
and
the
inverse
of
those
just
having
two
points
of
reference
might
help
people
think
about
how
to
improve
both
or
specifically,
the
sunpi
ones,
sort
of
take
another
pause
to
see
whether
anyone
online
wants
to
chip
in
with
anything
I'm
conscious.
B
That's
done
much
like
learning
some
fire
from
the
docs
at
this
point,
yeah
I,
don't
have
much
much
of
a
problem
like
I
think
that
the
API
references
nice
to
follow
but
I,
do
think
of
colleagues
who
encounter
a
little
friction
and
then
stop
because
they
don't
know
what
to
do
next.
B
I
was
thinking
about
so
we
use
prefect
for
punch.
It's
another
python
package
and
they've
done
a
fabulous
job.
Their
docs
are
pretty
easy
to
read,
I
think,
but
then
they
also
have
a
slack
Channel.
That's
super
active.
They
have
people
paid
to
go
answer.
Questions
I,
don't
know
how
you
do
that
with
the
limited
support
of
only
a
few
people,
but
I
think
the
problem
we
have
with
I
mean
certainly
for
me,
I
think
we're
relatively
good
at
answering
questions
and
we
have
a
few
forums
in
which
you
can
do
that.
B
I
feel
like
we
don't
get
as
many
questions
as
I
would
expect.
We
should
do.
Is
it
because
of
the
four
I?
Don't
know
this
course
I
didn't
realize
you
had
one
to
be
honest,
but
that
last
three
years
so
it's
like
taking
a
while
for
that
so
I
mean
that's
a
pretty
standard,
but
the
people
have
hesitancies
of
using
like
Matrix
element,
stuff
or
I.
Don't
know,
maybe
something
that
they're
not
commonly
outside
of
sunpai
I.
Don't
think
I've
ever
seen
it
I'm
sorry.
B
B
I
believe
the
biggest
reason
is
because
I
don't
want
to
fragment
the
community
like
you
can
Bridge
it,
but
it's
not
perfect
and
why?
Why
not
for
our
element
in
the
blend
and
you
slack
because
I
care
deeply
about
using
open
tools
for
open
projects,
github.com
isn't
open.
I
wish
I
can
move
away
from
github.com
as
well.
B
If
it
wasn't
such
a
absolute
pain
at
the
bottom
and
if
there
was
an
actual
good
alternative
because
gitlab
isn't
perfect
either
whereas
I
mean
I
appreciate
that
most
people
don't
care
as
much
about
this
as
I
do,
but
I
do
care
a
lot
about
it
and
yeah
I
think
you
have
an
opportunity
with
helium
Arts,
though,
because
there
are
so
many
people
going
to
healing
or
not,
but
if
you
can
make
it
obvious
that
you're,
the
presence
is
there
and
then
you'll
answer
questions
there
then
yeah
so
I
would
be
very
happy
to
pick
up
alien
or
as
like,
an
official
Senpai
Channel
if
it
was
open
registration,
but
it
isn't
and
it
won't
be.
B
B
He
directs
people
at
the
Senpai
stuff
from
here
and
also
if
he
needs
to,
but
yeah
I
think
that
the
fact
that
NASA
have
like
decided
to
flat
out
high
level
it
can't
be
open
registration,
because
it's
a
NASA
server
really
kills
it
from
our
perspective,
like
we
can't
like
so
Steve
has
made
a
point
here
as
well,
which
is
more
probably
more
General,
but
to
minimize
the
time
and
energy
needed
for
someone
to
discover
what
somebody
can
do
then
would
definitely
help
that
user
foreign.
B
Discourse
file
was
because
I
hope
that
is
an
easier
place
to
post
messages.
I
think
the
thing
there's
work
to
do
in
unhiding
it
from
behind
the
open
astronomy
wizard,
like
honestly
I,
think
just
putting
it
in
open
astronomy
discourse
is
really
confusing
and
I
think
we
need
to
link
their
into
the
sump
height
there
really
clearly
in
the
sunlight,
I
mean
putting
it
on
having
an
open
astronomy.
B
One
was
ultimately
cost
saving
exercise
we
pay
or
after
Buy
pay
for
the
open,
astronomy
District,
and
we
and
like
a
lot
of
questions
like
there
is
a
bunch
of
questions
which
are
relevant
to
like
find
some
pipe
together
and
stuff.
B
So
I
feel
like
we
can
build
more
of
a
community
around
that,
hopefully
that
that
was
one
of
the
things
that
I
would
like
to
say
is
like
we
build
a
more
diverse
and
interactive
Community,
where,
if
solar
physicists
have
afro
questions,
they
get
answers
right,
I
think
you
just
need
to
push
these
more
so
like
even
at
the
example
Gallery
we're
like.
B
Is
this
not
exactly
what
you're
looking
for
click
on
the
chat
and
ask
us
here
or
click
on
the
Discord
and
ask
you
because
I
just
think
I
think
it
is
just
visibility:
people
don't
realize
they're,
either
chat,
yeah,
yeah,
I,
agree,
yeah!
That's
really
good
idea
and
also
I
wish
I
think
we
talked
about
this
before
on
the
simplified
discourse.
B
We
can
change
that
yeah,
so
I'm
very
happy
to
adapt
the
open
astronomy
discourse
to
what
people
think
is
a
good
way
of
doing
it
like
if
we
we
just
yeah
and
I,
have
a
little
experience.
Admining
a
discourse
I
set
it
off
and
very
little
else
has
happened
since,
but
because
you
know
it's
yeah.
If
we
want
to
change
that
round,
then
we
can
talk
about
doing
it,
run
it
past.
B
They
ask
for
like
the
anthropod
people
and
go
for
it,
but
we
can
add
questions
foreign
I
think
in
a
in
a
perfect
world
it
would
work
better,
like
I
said
after
yeah
I
think.
Maybe
that's
not
a
good
way
of
encouraging
people
to
ask
questions
so
I
think
I,
think
you're,
right,
I
think
it
is.
It's
definitely
good,
but
I
just
think
I'm,
a
selling
music
perspective,
they're
gonna
watch
open
astronomy
like
what
just
your
pie,
I,
don't
know
yeah
in
the
chat.
B
So
yam
says
what
about
the
discussion
boards
available
there
in
GitHub
projects,
Stephen
says:
maybe
she
uses
don't
necessarily
have
GitHub
accounts
and
which
I
think
is
a
good,
really
good
point
that
I
think
the
interesting
suggestion
not
familiar
with
that
feature.
Leon
said
it's
new
I
think,
but
because
the
periods
of
GitHub
I
think
it's
a
more
obvious
Place
compared
to
hearing
August
foreign.
B
B
I
think
this
course,
as
a
platform
is
more
flexible
But.
Ultimately,
maybe
we
aren't
putting
the
effort
into
need
to
really
take
advantage
of
that
and
again,
the
GitHub
account
thing
and
open
tools
for
open
projects.
Again,
this
port
is
and
open
source
largely
Community
project,
and
it's
really
powerful.
It
does
everything
we
needed
to
do
now.
B
At
some
point,
you
can
have
20
different
ways
to
contact
people
that
can
help
you,
but
you
just
need
what
you'd
really
need
is
every
Pi
that
has
a
student
starting
on
a
project
that
uses
some
pie
to
see.
If
you
need
help,
you
know,
ask
come
to
me
or
go
to
these
place
or
go
to
this
place,
or
you
know,
and
that's
you
can
have
all
these
different
availabilities,
which
they
don't
know
about
them.
B
Does
it
matter
yeah,
but
because
you
definitely
can't
argue
with
something
that
not
not
me
or
anything,
but
you
guys
are
so
helpful,
like
anytime,
anyone
goes
into
the
the
channel
for
help.
You
get
an
unbelievable
amount
of
effort
put
forward
by
people
to
help
them
out
so
same
with
the
GitHub
issues.
There's
definitely
no
question
that
when
people
come
for
help
they're
getting
it,
it's
just
that,
like
Stuart
says
people
probably
aren't
coming
for
help
very
often
I
think
that
when
I
come
back
to
what
Laura
suggested
is
like
throw
out
the
jobs.
B
B
Yeah,
we
should
add
the
discourse
to
that.
That's
true.
Do
you
think
if
we
sculpted?
Maybe
this
is
a
bit
random
but
like
having
an
anonymous
someone
where
someone
can
ask
something
Anonymous,
so
I
guess
I,
don't
know
they
didn't
want
to
ask
yeah.
Ideally
that
would
be
lovely.
Unfortunately,
a
non-annuity
and
the
internet
tend
to
mix
badly.
B
Maybe
you
need
to
have
like
a
register
to
pay
to
use
it,
and
then
you
can
post
anonymously
and
then
you
can
flag
it,
and
that
usually
forget
I
mean
you
can
make
an
account
on
the
discourse
that
isn't
really
identifiable
to
you.
If
you
want,
like
you,
don't
need
to
provide
an
actual
name
yeah
of
the
Discord.
If
you
don't
want
to
I
suggest
we
refocus
this
a
little
bit
yeah.
We
have.
B
We
have
dived
into
tomorrow
morning's
session
on
the
community,
yeah
right,
so
quite
hard,
yeah
yeah,
from
what
I'm
hearing
what
I've
heard
of
before
we
got
into
that's.
The
video
discussion
is
the
way
forward
with
documentation
to
start
the
way
forward.
Documentation
is
to
basically
work
on
the
structure
and
that's
creating
those
four
sections.
B
They
don't
have
to
be
full
of
stuff,
but
you
know
move
what
what
we
do
have
and
is
they
should
be
in
those
sections
move
them
there.
Think
then,
about
like
in
the
future
sort
of
adding
stuff
there
going
forward
and
then
also
writing.
This
is
how
our
documentation
is.
Slash
will
be,
slash,
is
intended
to
be
structured
and
if
we
can
write
something
like
that,
that'll
Force
us
to
focus
our
minds
and
think
about
this
more
critically
than
just
a
group
discussion
that,
in
the
end,
can
be
like.
B
B
And
yeah:
are
we
okay,
we're
just
going
straight
into
the
roadmap,
or
do
people
want
to
it's
ready
for
four
minute
break
yeah
yeah?
We
should.
We
should
have
like
we're.
Gonna
come
back
like
at
five
after
closer
to
10..
Let
Stevo
are
you
still
good
to
do
the
roadmap
discussion
today?
Yep?
Okay,
are
you
okay,
with
starting
like
five
five
after
yeah,
that's
kind
of
thing?
B
Okay,
great!
Let's
do
that
thanks
for
all
the
input
online
I'm!
Sorry,
if
it
was
a
bit
hard
at
times,
because
we
were
talking
over
each
other
here,
we
will
try
and
clock
over
each
other
a
little
bit
lean
towards
the
mics
a
little
bit
more
yeah
and
we
yeah.
We
have
the
chat
box
up
now,
so
we
can
see
what
you're
typing
yeah
yeah,
also
yeah.
B
B
B
B
Have
an
Irish
as
a
tax.
You
know
what
that's
a
good
question
to
me
faces
here:
yeah,
that's
what
right,
because
she's
a
hair
to
avoid
not
employees
yeah
avoid
yeah
of
what
that
is.
Yeah,
avoid
if
age
is
the
illegal
one,
that's
right
avoid
is
perfectly
legal.
Well,
no!
Actually
they
do
actually
avoid
tax
here
as
well,
but
yeah
yeah
they're,
just
in
a
lower
tax
jurisdiction
and
they
pay.
They
don't
even
pay
that,
but
maybe
not
having
they
don't,
have
like
an
actual
business,
because
then
they'll
be
liable
for
other
times
I'm.
B
Just
guessing,
but
I
mean
like,
like
people
get
stuff
like
Amazon,
like
brands
have
gone
around
the
city
like
oh
no,
but
they
don't
have
like
the
IE
website.
Yeah
I,
don't
know
what
that's
I
suspect.
It
was
simply
because,
let's
assume
by
the
day,
especially
with
like
the
border
and
everything
it
was
with
the
common
and
the
common
like
travel
area
and
everything,
it
was
all
just
effectively
one
jurisdiction
with
the
two
currencies
yeah
you
know
and
then
again
minus
the
the
tax
consideration
yeah.
So
it's
just
easier
to
have
it
all
in
written.
B
But
no
it's
a
good
question.
I!
Don't
know
why
for
sure,
especially
now
it
makes
it
so
much
easier
if
there
was
a
DOT
IE,
one
yeah,
because
sometimes
it's
hard.
It's
like
you
have
to
work
to
try
and
go
to
the
doctor
E1
and
get
stuff
delivered
from
Germany
and
then
sometimes
like
you
accidentally
end
up
doing
through
code.uk,
and
then
your
stuff
gets
like
delayed
by
two
weeks
and
you
get
a
few
big.
You
know:
custom
Suite,
import
fee
control.
B
You
should,
if
Ireland
brexits,
oh,
it's
joined
the
United
Kingdom
again
yeah
yeah
I
I,
just
can't
get
over
the
fact
that
there
were
British
politicians
who
thought
that
was
a
seriously
like
the
solution,
and
some
like
there
was
something
that
we
honestly
thought
we
would
because
I
don't
it
doesn't
make
any
sense
for
them
to
be
separate
from
us
like
they
would
show
them
were
generally
shocked,
taken
as
an
axis
yeah,
yeah
yeah.
No,
no
I
know.
B
B
B
B
I
did
what
I
was
gonna
mention
the
road
map
repo
which
I
know
Stuart
doesn't
like,
and
it's
been
sort
of,
abandoned,
I.
Think
at
the
end
of
this
we
should
decide
whether
we
are
using
it
or
not,
and
if
we're
not,
we
should
just
delete
delete
it
because
that
question
for
you
already.
No,
the
roadmap
is
added
to
the
website.
B
B
Thank
you
come
on
Steve,
so
I
did.
If
you
put
under
the
notes,
I
kind
of
put
a
kind
of
a
structure
of
how
I
was
thinking.
We
could
do
this
I
thought
we
could
review
what
we
think
we're
doing
well,
what
we
think
we're
doing
poorly.
Therefore,
where
can
we
improve
and
what
do
we
want
to
stop
doing
so?
I
was
going
to
kind
of
open,
have
those
as
open
questions
and
then
from
there
move
on
to
kind
of
the
roadmap,
which
I
know
is
like
very
broad.
B
You
guys
have
already
talked
about
a
number
of
things.
Where
did
you
put
it?
These
notes?
I,
don't
see
this
yeah,
it's
in
the
Hedgehog
thing,
yeah,
that's
what
I'm
looking
I
think.
Did
you
put
it
in
the
agenda
one
from
before
it
is
underneath
the
Spectra
affiliate
package
notes,
do
I
need
to
publish
this,
or
is
this
live
I
think
you
might
be
in
a
different
document
to
the
restaurant?
Can
you
show
the
link
I'll
put
the
link
to
the
running
notes,
we're
using
in
the
element,
coordination,
Channel,
I
just
published.
B
B
B
B
I
shouldn't
both
microphones
stop
working
now,
I
need
to
undo
undo.
My
changes.
I
mean
we've
preserved
that
that
version
of
the
2020
notes
in
the
wiki
I
think
so
it
doesn't
matter
those
can.
The
Hedgehog
version
can
in
dying
sorry
I
was
confused,
I
think
so.
I
was
gonna,
say
it's
on
the
website.
I'm,
pretty
sure
we
I
linked
to
the
demo,
I'm
sure
I
at
least
copy
country,
Oh,
no
you're
right,
sorry,
you're,
right,
yeah,
yeah!
B
So
don't
answer
me
I'm,
sorry,
but
again,
if
you
can
delete
that
just
delete
that
I
mean
I
was
gonna
say
who
cares?
No
one
will
find
it,
but
clearly
this
always
happens.
A
lot
of
history
documentation.
You
used
to
get
down
the
reads
old
at
some
point,
doc,
film
from
like
2013
or
something
there's
like
0.2
Steve,
since
we're
still
waiting
for
a
couple.
B
B
B
B
B
No,
it
talks
about
WCS,
pretty
much
right
away,
I
mean
that's
the
problem.
You
don't
understand
what
what
WCS
is
right,
which
I
think
a
lot
of
people
then
you're
like,
and
it's
like
in
the
first
paragraph
you're
like
what
what
is
happening
right
like
yeah,
what
do
I
care
about?
Wcs
can
I
use,
ndq
I'm,
not
sure.
B
Yeah
I
guess
you
I
guess
you
want
some
kind
of
explanation
along
the
lines
of
like
images
or
in
pixels,
but
that
pixel
represents
a
real
coordinate
and
how
do
you
map
from
pixels,
to
coordinates
that's
kind
of
what
that
first
paragraph
we're
trying
to
do
I
mean?
Maybe
it's
not
doing
it
very
well.
Yeah
he's
doing
it
terribly,
but
that's
what
it's
trying
to
do.
B
I
mean
I
would
also
argue
that,
like
the
audiences
for
the
ndq
documentation
and
the
audiences
for
the
Senpai
documentation
at
least
I-
don't
know-
maybe
maybe
you
guys
disagree
with
this,
but
I
I.
Think
in
my
mind,
like
those
are
very
different.
Like
I,
the
someone
who's
looking
at
the
ndq
documentation
is
probably
not
like
I,
don't
know,
yeah.
That
being
said,
you'd
hardly
call
Steve
Green
and
if,
if
something
you're
having
trouble
with
it,
then
that's
that's
very
important
to
know
well
yeah.
B
That
is
true.
That's
true
anyway.
Well
thanks
thanks
for
that,
we've
we've
got
a
full
compliment
here.
So
thanks
for
that
feedback,
and
if
you
have
any
more
detail,
you
want
to
give
at
any
points.
Yeah
I'd
be
able
to
hear
more
thoughts
on
that
even
more
thrilled.
If
those
thoughts
came
with
Associated
pull
requests,
yeah
yeah,
all
right
should
we
should
we
get
started
yeah
all
right,
so
so,
first
of
all,
I
wanted
to
make
sure
people
knew
there
was
some
some
time
ago,
and
maybe
it's
all
deprecated,
but
there
was
a
repo.
B
There
still
is
a
repo,
a
road
map
repo
which
can
we
just
agree
are
we?
Are
we
not
using
this?
Are
we
getting
rid
of
it?
B
For
example,
yeah:
okay,
that's
fine
with
me:
yeah
I,
don't
know
whether
you're
about
to
dive
into
this
Steve
or
not,
but
one
of
the
things
that
I
would
like
to
like
make
sure
everybody
is
thinking.
The
same
here
is
like
you
know:
what
is
the
purpose
of
this
roadmap?
Why
do
we
want
one.
B
Basically,
any
project,
we
want
a
long-term
vision
of
where
we're
headed
right.
We
want
to
be
able
to
understand,
you
know
kind
of
the
Arc
of
the
project,
and
you
know
what
what
are
our
long-term
projects,
whatever
short-term
projects?
What
are
midterm
projects
I
mean
I,
think:
aren't
we
required
to
have
a
road
map
by
num.
Focus
required
might
be
a
strong
word
encouraged.
It
might
be
so
I,
just
like
the
numpy
roadmap
is
really
interesting
to
me,
because
it's
it's
the
kind
of
low
effort
roadmap
that
I
can
really
get
like
yeah.
B
B
But
but
nobody
I
think
is
a
bit
different
than
us,
because
they
are
a
very
stable
package.
Like
I
I
see
we
still
have
gaps.
B
B
B
Think
a
roadmap
with
time,
because
we
do
not
have
the
resources
to
hit
deadlines.
Yeah,
no,
but
I-
think
I
I!
Disagree
that
it
is
it
a
useful
tool
for
us
I
think
it
should
be
like
something
that
guides
our
kind
of
maybe
not
day
to
day
but
sort
of
like
you
know,
month-to-month
decision
making
because
I
I
think
it.
It
helps
us
not
just
to
not
just
do
a
bunch
of
kind
of
incoherent
make
a
bunch
of
sort
of
incoherent
decisions.
B
I
think
like
big
decisions
like
should
we
base
you
know
in
you
know,
map
on
a
DQ
or
or
even
even
slightly
smaller
ones.
We
should
be
thinking
of
you
know.
How
does
this
fit
in
with
our
roadmap?
Is
this
taking
us
where
we
want
to
go
so
we're
just
going
to
deprecate
this
in
six
months,
because
we're
going
in
a
different
direction?
Yeah,
okay,
yeah
I,
mean
I.
B
That
we've
set
out
like
I
I,
mean
obviously
it's
great
for
funding
and
showing
people
that
we
have
that
Vision
I
think
we
probably
all
have
that
kind
of
same
sort
of
collective
ideas
about
what
we
want
the
project
to
be,
but
I
think
if
we,
because,
like
write
those
down,
say
them
out
loud
like
that,
you
know
helps
to
to
kind
of
make
that
thought
process
much
more
popular
foreign
who
wants
to
say
anything
against
having
a
road
map.
B
Otherwise,
maybe
we
should
proceed
with
making
it
to
be
clear:
I,
don't
think
we
shouldn't
have
one
I
think
we've
just
I
just
wanted
to
make
sure
that
we
grew
up
a
little.
We
were
all
writing
the
same
document
for
the
same
reason
before
we
started
doing
it
I'm,
along
with
Will's
lives.
The
only
other
thing
I
would
say
is
what
do
we
think
like
the
time?
B
B
Unless
there's
a
specific
Grant,
no
I
I,
don't
mean
I,
don't
mean
in
terms
of
we're
going
to
do
this
thing
by
next
year,
I
mean
like
it's
a
thing:
that's
achievable
in
yeah
a
year
versus
I,
don't
know
maybe-
or
maybe
it
doesn't.
Maybe
that
doesn't
matter
but
I
would
say.
We've
got
to
be
thinking
about
rewriting
this
document
at
every
coordination,
yeah
I
think
Steve
had
a
comment.
B
Go
ahead,
Steve,
yeah,
sorry,
yeah,
I,
agree,
I,
mean
Steward
said
what
I
was
going
to
say
that
I
think
it
should
be
reviewed,
which
I
think
is
it's
very
different
than
what
you
said.
We
should
not
be
rewriting
the
roadmap
every
Sun.
We
should
be
reviewing
it
right.
Presumably
our
long-term.
Our
long-term
aims
do
not
change
that.
Frequently.
I
need
to
rewrite
the
whole
thing,
but
reviewed
once
per
year.
I
think
is
the
right
approach.
B
Maybe
roadmap
isn't
the
right
word
at
that
point,
but
I
think,
like
the
time
scale
of
the
thing
you're
putting
on
a
roadmap
right?
How?
How
does
it
change
like?
We
could
put
smaller
things
on
a
roadmap
where,
basically,
we
would
have
done
most
of
it
in
a
year
or
we
could
put
massive
things
on
a
roadmap
where
we
wouldn't
have
done
most
of
it
in
a
decade
like
you
know,
there
are
like
we,
for
instance,
right.
B
B
B
Okay,
do
you
guys
want
to
start
that
way,
or
do
you
guys
want
to
start
with
kind
of
the
questions
which
which
I
had
listed
like
we
could
start
with
just
writing
down
the
things
we
kind
of
know
are
supposed
to
be
we've
already
made.
Some
sort
of
agreement
are
going
to
happen,
I
think
well,
I
think
you
may
be
answering
your
questions.
Will
lead
us
to
writing
down
okay,
yeah
all
right.
B
So,
let's
start
with
that
I
mean
so
the
first
one
I
wrote
was
you
know
what
what
do
we
think
we're
doing
well
and
the
idea
there
is,
if
we're
already
doing
it.
Well,
then,
it's
not
necessarily
an
area
that
we
want
to
focus
too
much
of
our
energy
in
and
I'll
just
leave
it
open
and
take
notes,
as
as
people
throw
things
out
there.
B
Could
you
get
us
started
with
an
example
to
just
just
so
we
could
have
a
trigger
on
how
what
you're
thinking
I'd
say,
coordinates,
coordinate,
transforms
and
using
those
with
maps
like
being
able
to
read
the
map
onto
another
map.
It's
really
cool
and
is
relatively
new
foreign.
B
B
I
think
the
visual
like
visualization,
easy,
easy
visualization
that
like
map,
enables
I
mean
this
is
sort
of
and
because
it's
coordinate
aware.
So
maybe
it's
just
like
wrapped
up
at
the
first
one
I
mean
I.
Think
if
we're
gonna
just
go
everything
with
touching
coordinates.
We're
done.
This
is
going
to
be
like
a
one
element
list
right,
like
some
players,
all
core.
Yes
at
this
point,
maybe
we
should
just
go
around
and
everyone
should
choose
one
thing:
yeah
that
sounds
right
and
then
everyone
gets
to
say
people
online
who
have
a
thing.
B
B
Okay,
yeah
I
suspect
we're
gonna
get
a
lot
more
participation
on
the
next
version.
B
Yeah
I:
let's
not
move
to
the
next
one
too
quickly,
because
I
think
we
do
want
to
kind
of
agree
on
the
areas
that
that
yeah
we
do
well
before
we
start
complaining.
I've
got
one
so
yeah
I.
Think
Senpai
is
a
good
example
of
scientists,
embracing
open
source
development
methodologies
and
doing
it
really
well,
it's
kind
of
a
a
lesson
to
everyone
else.
B
B
I'm
going
to
suggest
one,
but
if
people
may
disagree,
I
think
coming
from
coming
from
a
scientist
angle
rather
than
a
private
sector
angle
I
think
we
do
a
good
job
at
testing
code.
You
know
we
have.
We
have
a
big
test
Suite
we
have
pretty
good
coverage
and
again
as
a
cultural
thing
of
testing
and
solar
physics,
I
think
that's
a
great
contribution
to
the
to
the
community.
Even
if
maybe
you
know,
computer
scientists
or,
like
professional
programmers
might
say
maybe
but
I
think
I've
been
in
the
context
of
science.
B
B
B
And
I've
heard
all
those
said
earlier:
I
think
the
Community
engagement
is
really
something
else.
I
think
somebody
does
a
really
good
job
of
us
like
the
animal
chat
and
wherever
it
is
anytime.
I've
noticed
myself
or
literally
anybody
have
a
problem.
It
is
dealt
with
very,
very
quickly
and
very
like
in
a
very
friendly
manner.
There's
nobody
ever
like
saying.
Oh
my
god,
get
into
this
simple
thing.
I
think
everybody's,
very,
very
friendly
for
lack
of
a
better
word.
B
You
know
like
those
people
coming
back
and
coming
part
of
me,
but
I
think
because
I'd
be
interested
apart
from
just
on
the
chat
for
hell
approachable.
Are
we
to
new
people?
How
friendly
are
we
to
new
people
and
that's
hard,
I
suppose
for
people
who
are
like
re-engage
all
the
time
and
have
personal
like
good
personal
relationship?
That's
something
that's
a
bit
hard
to
judge
the
closer
to
the
center
of
some
fire,
as
you
like,
you're
already
friends.
B
So
if
you
have
any
comments
on
that,
yes,
like
in
a
broader
sense
and
I,
think
it
helps
that
I
know
some
of
you
in
the
room
already.
Definitely
I
think
that's
what
sort
of
got
me
to
make
the
first
message
or
make
making
the
first
message
easier.
But
then
another
part
of
it
was
like
seeing
the
Matrix
shot.
Wasn't
always
just
people
talking
about
the
nitty-gritty
of
whatever
it
is.
B
There
is
genuine,
like
people
out
there
who
are
totally
new
or
just
trying
to
get
something
done,
I
think
I,
think
and
that
was
sort
of
shows
itself
to
be
you,
your
real
people,
who
are
quite
helpful
I
should
only
in
my
video,
I'm,
sorry,
hi
and
I.
Think
yeah
I
think
you
just
you
make
a
good
impression
on
The
Matrix,
just
I,
don't
know
how
to
describe
that.
Unfortunately,
yeah,
it's
just
like,
like
the
actions
speak
for
themselves,
basically
I
think
everybody's
working
with
their
and
the
betterment
of
everybody
else.
B
B
Okay,
give
given
the
time
should
we
just
do
one
more
anybody
else
have
one
more
and
then
we'll
move
to
the
I.
Think
I
think
I'll
release
Cadence
and
our
release
notes
are
good,
like
I
think
we
have
got
to
the
point
where
we
have
a
strong
deprecation
policy,
clear
documentation
on
changes
to
the
packages
and
a
regular
release
schedule
we've
hit
our
release
deadline
for
the
last
two
release,
major
releases,
so
in
fact,
3-1
was
early.
B
I
think
yeah
well,
I
was
when
I
was
making
the
slides
for
the
sub
package
sets
earlier
today,
I
was
reading
through
the
what's
new
I
was
like
this
is
a
really
nice
resource
and
the
fact
that
we've
decided
to
you
know
ensure
that
for
every
all,
these,
like
big
features,
significant
features
that
we're
writing
that
were
as
part
of
the
pull
requests
were
requiring
and
what's
the
inventory
to
like
describe
it,
you
know
a
high
level,
more
plain
like
like
plain
Language
summary:
it's
like
what
does
this
thing
do?
B
B
B
Y's
policy
only
excuse
me
in
an
effect
for
corn
yeah
AI-
probably
definitely
not
it's
sort
of
whatever
bugs
need
to
be
well
or
it's
whatever
new
features
get
added
and
we're
launches
yeah
indicate
with
some
roster,
don't
have
regular
cages.
They're
kind
of
like
ndq
doesn't
need
one
right
enough,
doesn't
change
in
the
last
six
months.
It's
finished
no,
but
I
mean
I.
Think.
B
B
Being
you
know
open
and
welcoming,
we
all
agree
on
that.
You
know
we
could
do
a
better
job
of
getting
more
of
the
community
engaged
on
that
score,
but
I
wouldn't
say
we're
doing
poorly
yeah
I
mean
I
wanted
to
focus
on
the
poorly
just
because
it
could
constantly,
and
the
road
map
is
a
way
to
focus
our
energies
on
the
areas
that
need
energy
focused
on
I.
B
Think
all
the
other
stuff
should
be
happening
all
the
time
we
should
be
continuously
improving,
all
the
other
stuff,
too
yeah
I
guess
there
are
some
things
that
are
doing.
Okay,
but
like
it's
also
still
a
priority
to
improve
them,
but
like
that's
good,
to
clarify
what
you
mean
by
those
two
questions
being
coupled
so
I
think
that's
the
thing.
That's
fine,
okay,
all
right!
So
what
what
are
we
doing?
Partly
and
where
can
we
improve
ment?
B
B
Can
you
hear
me:
okay,
yeah,
yeah,
okay,
first
of
all,
before
I
start,
complaining,
I
agree
that,
generally,
the
documentation
is
very
well
written
and
there's
a
lot
of
help
when
you
have
issues
so
you're
doing
a
really
great
job
there
and
that's
really
appreciated.
B
However,
I
did
contribute
for
the
first
time
recently
for
subie,
and
that
has
been
a
really
frustrating
experience
and
it
really
discouraged
me
to
contribute
something
in
the
future.
The
reason
being
I
went
through
that
developers
guide
and
I
really
spent
weeks,
trying
to
clean
up
my
code
according
to
that
documentation
And.
Yet
when
I
submitted
it
there
were
like
weeks
and
weeks
and
weeks
of
complaining
what
I
did
wrong
and
like
back
and
forth
like
the
variables
need
to
go
there,
they
can't
be
there.
B
You
need
to
split
up
your
code
by
functionality,
and
none
of
that
was
in
the
documentation
for
the
developer,
and
that
was
really
frustrating.
So
if
you
want
to
encourage
people
to
contribute,
more
I
would
suggest
to
find
a
way
to
make
that
less
frustrating
would
that.
Do
you
think
you
suggested
that
a
solution
is
that
the
solution
you
want
to
see
as
in
the
developer
docs
need
to
be
more
explicit?
Yes,
they
need
to
reflect
what
actually
needs
to
be
done.
B
I
mean
there's
a
lot
of
stuff
in
there
already,
but,
for
example,
like
to
break
up
The
Code
by
functionality
and
like
not
putting
everything
into
one
big
python
code
or
like
to
split
out
the
variables
and
all
of
that
stuff
to
the
beginning
of
the
code.
None
of
that
is
in
the
documentation
and
after
spending
weeks
of
trying
to
get
it
right
and
you
get
the
code
rejected
multiple
times.
That's
where
you
start
to
get
like.
Okay,
I'm
done
with
this
Chris
I
tried
to
write
up
your
comment
in
the
notes.
B
If
you
could
take
a
look
at
it
and
either
add,
or
you
know,
it
would
be
very
helpful
sure
we'll
do
that.
Thank
you.
B
Yeah
thanks
so
much
for
that
comment.
I
I
can
identify
actually
when
I
first
entered
Senpai
I,
whatever
reason
I
just
kept
going,
but
you
know,
a
lot
of
people
wouldn't
have
said
happened
so
that
that's
a
fair
point.
Yeah
it's
difficult
to
balance
like,
especially
when
first-time
contributions
are
large.
It's
difficult
to
balance
like.
B
Like
you
know,
yeah
yeah
that
goes
PR
was
yeah
very
similar
to
your
super
one
I
think
documentation
be
like
if
we
want
to
open
a
few
and
maybe
start
with
a
very,
very
small
one
so
that
you
know
the
process.
I
think
that's
a
very
good
point
or
just
pull
requests
super
early,
yeah
requests
to
draft.
That's
really
really
rough.
Yeah,
yeah
I
think
the
Small
Things
yeah.
B
I
think
there's
a
difference
between
like
unfamiliarity
with
the
process
versus
like
anybody,
a
big
feature:
we've
got
a
coach
and
getting
like
lots
of
feedback
on
restructuring
the
code
and
things
so
we
have
both
of
those
together
I
think
it's
completely
Fair
what
Chris
is
saying:
yeah
yeah
absolutely
so
like
just
the
thing
on
that
there
is
that
lead
newcomer
role
as
anybody
in
any
thoughts
into
like
anybody
that
would
like
to
take
on
that
role.
B
To
help
with
these
type
of
things
was
any
of
the
hospital
funding
for
that
kind
of
stuff.
Yeah
I
mean
I,
think
a
lot
of
it.
There
is
some,
the
funding
is
kind
of
Outreach
focus
and
so
effectively
had
written
in
was
I,
guess
well,
I,
guess
I
was
gonna,
say
instrument
you
folks,
but
that's
not
that's
not
true.
I
mean
again
thought
of
it
in
those
terms,
but
I
mean
my
provocative
question
is
you're
being
paid
at
newcomer.
Steve
wants
to
jump
in
as
well.
B
B
Looking
at
by
or
playing
without
without
going
down
a
solution,
route
I
think
that
is
tied
in
with
ndq
with
map
yeah,
yeah
I
think
plotting
multiple
maps
on
one
page
can
get
very
complicated
very
quickly,
and
maybe
that
is
the
documentation
issue,
but
it
isn't
straightforward
compared
to
the
day,
if
you
like
have
the
two
months
three
months
and
you
want
to
do
Contours
on
the
mall
and
they're
kind
of
slightly
different
times,
but
you
have
to
understand
the
coordinate
framework
you're
so
into
it
and
be
like
plotting
three
maps
in
one
on
one
app
set
of
activities
yeah
and
that
featured
there
yeah
yeah
I
would
agree,
I
think
doing
anything
more
than
map
Dot
Plot
yeah
yeah,
it's
non-trivial
yeah.
B
What
would
you
say
that
yeah
I
suppose?
Is
it
a
documentation
issue
or
is
it?
The
code
needs
to
be
refactored
to
be
makes
a
more
implicit
assumptions
to
support
certain
use
cases
or
yeah
yeah?
Maybe
maybe
it
is
I
think
I
think
from
a
from
a
base
user
point
of
view
isn't
fully
aware
of
the
whole
intricacies
of
WS,
yet
WCS
axes,
I.
B
Think
a
lot
of
helper
functions
could
be
written
somewhere
has
grown
well,
math
has
grown
quite
a
lot
of
overporting
support
like
with
overall
also
a
line
through
and
Contour
now,
I
think
it's
still
difficult
to
discover
and
do
that
yeah
and
maybe
there's
a
documentation,
but
that's
just
something
yeah,
but
it's
also
great.
It
has
lots
of
good
things.
It
does
too,
but
this
that's
a
related
to
the
two
map
problem
is
I
would
say
spectral
data
like
spectroscopic
data,
we
yeah,
there's
basically
I
mean
there's
so
Mica
is
online.
B
Now
they
have
the
ice
ice
pack
package,
there's
Sun,
roster
and
Iris
pie
and
the
spice
tools
that
other
folks
are
working
on,
but
it's
very
very
much
kind
of
an
influx.
You
know
effort,
none
of
it's
been
cool
and
none
of
it,
which
is
a
table
Yeah.
So
something
core
has
no
support
for
spectroscopic
data.
That's
why
I
like
I
mean
you'd,
say
it
only
supports
images
and
simple
time
series:
that's
yeah!
B
B
B
Can
you
hear
me
but
yeah?
Can
you
hear
me
yeah
great?
This
is
Michael
wieberg,
just
I
guess
in
addition
to
the
we
were
talking
about
spectral
data
and
multitational
does
Senpai
support
time.
Distance
diagrams.
B
Saying
something:
sorry
go
ahead:
I
guess
I!
Guess
the
fundamental
question
there
for
me
is:
can
you
make
a
map
just
as
a
senpai
map?
Where
one
axis
is
time
and
one
is
you
know
some
spatial
coordinate?
You
can
make
an
ND
Cube.
Do
that?
Yes,
the
answer
is
no
to
your
questions.
Right,
there's
not
enough
about
them
right,
yeah!
B
B
But
you
can
you
can
have
that
data
in
Python
and
like
but
yeah.
It's
just
not
designed
to
do
that.
So
I
think
a
lot
of
the
fundamentals
are
there,
but
there's
no
like
a
high
level
document
like
there's
no
tutorial
on
how
to
do
it.
Yeah
or
any
helpful
functions.
But
yeah
map
is
basically
implicitly
two
space
dimensions.
It
would.
It
would
never
be
a
map,
but
there's
still
no
support
for
it
anywhere.
B
B
B
Just
to
comment
on
that
so
see
just
release
their
data
and
if
I
just
read
it
into
math
and
works,
fine,
that's
cool.
B
B
I
mean
there's
probably
some
weird
brand-based
instruments
that
we
don't
have
map
sources
for,
but
everyone's
awesome
for
them.
A
lot
of
weird
ground
bikes
instruments,
don't
have
metadata
so.
B
B
B
B
B
I
don't
feel
like
we
focus
that
much
at
all
on
the
computer
directly,
but
we
could
you
elaborate
on
that
Steve
by
what
you
mean
well
like
there
is
this
kind
of
like
fuzzy
region
and
time
series
where
we
could
take
on
a
bunch
more
stuff
and
in
fact
there
is
a
little
bit
now
right
in
in
some
High,
where
we're
using
I
forget
which
CDF
reader,
you
know
we
could
also
easily
add
you
know
downloading
from
from
the
spdf
into
Fido.
B
So
it's
it
sort
of
feels
like
we're
like
slow
kind
of
inching
into
kind
of
space
physics
a
little
bit
and
I
worry
that
we
have
kind
of
this
fuzzy
boundary
that
we
haven't
defined.
Well
enough
does
that
does
that
I
mean
I?
B
Think
we
need
to
find
we've
often
defined
core
as
providing
like
search
and
download
functionality,
providing
like
based
data
structures
and
then
doing
the
coordinates
and
I
think
that
providing
the
ability
to
search
for
and
download
helium
and
read
it
into
a
data
structure
as
well
within
the
scope
of
core,
as
we
as
we've
defined
it
for
years
now.
So
I
don't
know
if
I
agree
with
that,
I
mean
that
covers.
Basically
everything
though
right
I
mean
we're
not
so.
B
B
It
doesn't
include
planetary
physics,
it
doesn't
mean
so
I
think
yeah,
I
I,
would
say
that's
kind
of
driven
by
the
community
of
users,
so
they
say
well
what
day
are
we
going
to
support
if,
if
they're
stung
by
using
community
says
like
this
is
important,
then
you
know?
Presumably
we
want.
We
want
to
support
that
I.
Think
but
I
mean
do.
We
need
a
hard
Edge
to
push
back
on.
That
I
mean
there
are
results
who
care
about
the
Earth's
magnetosphere
as
well,
but
there
are
also
great
actors
out
there
already
that
are
providing.
B
You
know
like
interfaces
to
more
magnetospheric,
focused
data
and
we're
not
going
to
start
all
of
a
sudden.
You
know
like
I,
don't
know
no,
but
in
that
case,
what
we
do
is
we
point
those
users
towards
that
tool?
We
don't
we
don't
just
redevelop
everything
under
some
fire,
but
if
something
isn't
doesn't
exist,
yeah
and
our
user
Community
need
it
and
not
more
than
just
like
one
person,
I
mean.
B
Presumably
that's
then
in
scope,
isn't
it
I
mean
even
if
we,
but
then
there's
a
there's,
a
category
of
software
where
you
know
oh,
we
might
want
to
duplicate
this
thing
to
add
it
to
Fido,
because
then
you
can
search
it
all
once
or
something
I.
Don't
know.
Go
on
to
sorry,
I'm
just
going
to
say
that
this
conversation
is
evidence
for
the
statement.
That's
written
down.
B
Yeah,
but
is
that
a
problem
that
needs
improving
I,
don't
at
the
moment
so
far
unless
we're
duplicating
code
in
other
Helio
packages,
I,
don't
so
I,
don't
see
that
as
something
that
needs
to
be
improved,
well,
understanding
the
scope,
our
package
I
think
is
important
right
it
it
limits
what
the
roadmap
should
cover
right.
There
are
some
areas
we
shouldn't
be
going
into
because
we've
agreed
that
is
not
appropriate.
B
So
I
don't
know
how
to
define
the
scope,
but
then,
within
that
scope,
you're
saying
we
need
a
hard
boundary
within
that
right.
Yeah,
because
otherwise,
we'll
accept
we'll
accept
things
like
if
additions
to
the
package.
That
are,
you
know
and
we'll
work
on
those
because
they're
being
provided
that
shouldn't
be
there
and
then
we
have
to.
We
have
to
support
them
and
support
it
and
everything
so
I
mean
I,
guess
the
concrete
example.
B
Yeah,
yes,
it
doesn't
exist
anywhere,
yeah,
yeah,
okay,
but
you
can't
then
integrate
it
with
the
Sun
data
that
you
want
to
do
so.
Let's
say
you
want
to
correlate.
X-Rays
looks
with
my
guitar,
probably
not
a
good
idea,
but
let's
just
say
you
did
right.
B
I
agree
with
Steve
that,
for
the
purposes
of
this
discussion,
I
think
that's
a
valid
point
to
have
on
my
list
and
we
should
keep
it
there.
Okay,
yeah,
in
that
case,
we
should
probably
move
on
yeah
yeah
I
would
I
would
suggest
we
move
on
to
the
next
one,
which
I
already
added
one
thing:
what
should
what
do
we
want
to
stop
doing?
B
I'm,
not
convinced
we
didn't
want
to
stop
doing
database.
We
just
want
to
not
everybody,
we're
doing
it
at
the
moment,
but
as
it
exists,
we
want
to
stop
it
database
and
start
doing
database
too
yeah.
What
do
we
do
badly?
Yeah
yeah,
so
I
didn't
know
something
by
the
database
until
a
few
weeks
ago,
where
I
was
searching.
The
issues
I
found
an
issue
from
like
last
year
or
two
years
ago,
where
someone's
like
the
reply,
was
basically
like
we're
not
going
to
fix
this
because
we're
going
to
redo
databases
yeah.
B
So
if
you're
going
to
redo
databases,
then
you
know
I
mean
probably
don't
keep
it
in
limbo.
For
years
and
years
and
years
yeah,
we
discussed
earlier
today
like
getting
Community
feedback
on
yeah
yeah,
exactly
yeah.
B
The
current
Database
Top
Model
is
on
The
Chopping
blog.
Unless
somebody
goes,
no,
please
don't
delete
it.
I've
written
my
entire
Pipeline
on
it.
If
you
look,
if
you
search
GitHub
for
some
pie,
dot
database
there's
like
83
repos
use
it,
but
a
lot
of
them
are
like
the
automatic
like
file
listing
ones.
So
there
are
some
people
using
it,
but
time
series
is
like
what
800
or
500
in
maps
is
like
3000,
fair
enough
composite
map
yeah.
B
Well,
do
you,
though,
because
like
a
lot
of
the
functionality,
that's
in
composite
map
now
is
actually
being
built
into
map.com,
yeah
I
think
it
I
think
having
a
theory
is
still
confusing
for
a
lot
of
people.
As
far
as
I
can
tell,
it
doesn't
work,
at
least
not
not,
for
me
certainly
doesn't
work
well
yeah.
We
should
probably
pull
composite
map.
B
All
I
meant
well
yeah,
I
guess
we
could
argue,
that's
now
a
some
image
problem.
So
that's
only
the
organization
of
the
co-alignment
code
in
some
well.
Okay,
wait!
Wait!
We're
on
that.
What
do
we
stop?
Doing?
No
I
think
it's
now,
I
guess:
we've
core
has
stopped
doing.
The
work
map
is
not
just
caught
yeah.
No,
so
I,
don't
know,
I
mean
no
I.
Think
people
use
the
co-alignment
I.
Don't
think
we
do
want
to
stop
these
yeah
yeah.
B
B
I
I
have
a
controversial
one
version
of
Time
series
and
replace
it
only
with
time
with
Astor
pi
time
series.
Readers
I
know
that
we
get
rid
of
all
that
meta
stuff
that
we
have
to
join
multiple
time
series
together,
well,
I
think
we
could
keep
that
and
do
what
you
said
like
there's,
there's
a
strophy
time
series
astral
pie
table
in
particular
is
pretty
flexible.
You
can
do
different
things
with
it.
We
could
keep
a
lot
of
our
method.
We
could.
We
could
even
Upstream
the
metadata
handling
yeah
exactly
so.
B
There
are
options
there,
so
that
would
mean
like
we
stopped
supporting
our
own
time
series
object,
which
is
already
confusing,
since
the
astropai
is
also
called
time
series.
So
now
there's
two
there's
a
senpai
time
series
and
there's
a
astrophyte
Time
series
and
it
will
become
even
more
confusing
when
the
sunpied
time
series
is
using
the
Astro
python
series
too
many
time
series
I,
think
I
think
the
first
stage
of
that
process
would
be
to
yeah
I
mean
yes,
that's
a
long-term
project
but
yes,
sure
yeah.
B
So
so
what
do
we
want
to
keep
we
want
to
remove?
Are
we
saying
we
want
to
remove
our
own
time
series
class,
because
that's
right
I
would
phrase
it
a
little
more
Woolly
and
say
something
along
the
lines
of
remove
duplication
with
astral
Pi
Time
series
I
would
agree
with
Steve
I
I
think
we
should
I
think
long
term.
B
B
What
do
the
sources
actually
do
things
there
are
more
to
do
with
reading
requirements.
You
could
then
transform
into
find
distance
calculations.
Oh
I
mean
yeah
I
think
we
need
to
support.
Yes,
everything
to
do
a
lot
of
stuff
like
that
yeah
yeah,
there's
also
plotting
support
on
somebody
time
series.
It
doesn't
exist
on
actual
five
times.
B
I,
don't
know
yes
put
it
on
the
list.
I'm
gonna
put
a
question
mark
next
to
it.
Since
I
I
agree
with
the
broad
concept,
I
think
I
am
reluctant
to
phrase
it
delete,
Senpai
time
series
because
I
think
there
might
be
some
technical
reasons
why
we
want
to
keep
a
very
cut
down
version
of
that
around
I.
Think
we've
been
something
that
goes
like
curves,
yeah,
so
I
think
I
think
there
is
some
functionality
there,
but
it
like.
Yes,
an
awful
lot
of
what
we
have
in
Time
series
should
be
replaced
by
iPhone.
B
B
It
has
problems
with
seconds
there's
times
so
many
leap
seconds
anyway.
Yes,
it's
on
the
list.
We
should
move
on,
or
else
we'll,
never
finish,
yeah.
How
much
more
time
do
you
guys
we're
on
the
hour
I
think
I
think
everyone
in
the
room
I
think
carry
on,
but
we're
going
people
online
yeah.
B
I
would
definitely
roll
on
for
at
least
a
lot
of
a
little
while
to
get
some
more
of
this
down.
I
think
so
this
is
useful.
B
If
you're
up
for
it
yeah
yeah
I'm
fine,
do
you
guys
want
to
keep
adding
to
these
lists
or
what
I
was
thinking
we
would
do
next
is
I,
guess,
yeah
translate
these
into
actual
things.
We
would
do
yeah
and
then
and
then
add
add
priorities
to
them
and
kind
of
what?
What
is
the
kind
of
the
time
frame
foreign
I
was
gonna.
Just
suggest,
there's
a
category
that
we
haven't
captured
here
and
that's
what
new
things
do
we
need
isn't
that
captured
in
what
we're
doing
poorly?
B
B
B
B
So
where
is
this
list?
Oh
you're?
Actually,
except
let's
use
the
other
thing,
yeah,
yeah
and
I
kind
of
wrote.
You
could
see
below
that,
but
some
of
my
preliminary
notes
on
these
kind
of
long-term
projects
which
have
yeah
so
so
Dan.
Let
me
let
me
explain,
sort
of
push
on
you
in
this
I
mean
once
we
once
map
and
Harry's
from
a
d
Cube
I
mean
that's
enables
a
lot
more
other
stuff,
but
you
are
not
suggesting
we
do
that
or
are
you?
B
B
Sorry
I
was
just
going
to
say
that
I
think
that
map
should
stay
space-based
forever,
I
think
other
com,
other
data
could
be
or
maybe
not
forever
forever
is
a
strong
word.
I
think
Matt
will
be
staying
space-based
for
a
very
long
time.
B
Okay,
but
I
mean
we
could
have
I
think
a
good
way
to
start
from.
In
a
lot
of
different
ways
is
to
make
math
as
a
2d
object,
inherit
from
ndq
we're
going
to
learn
a
lot
from
that
it
might
change
bits
of
the
API,
hopefully
not
or
we
would.
We
would
learn
up
more
about
what
that
well,
keeping
the
same
API
means.
But
another
item
could
be
I,
don't
say
it's
the
right
term,
but
you
know
create
a
map.
Cube
object,
you
know,
which
is
space
based
time
or
space
space.
Anything
all
right!
B
Well,
yeah
I
have
a
real,
lingering
fear
that
I've
had
for
years
at
this
point
that
this
is
cracking
door
open
to,
which
is
every
single
combination
of
axes.
You
can
think
of
gets
its
own
object,
there's
an
object
for
you
and
you
and
you
and
you,
and
it's
just
I,
don't
know
if
Albert's
walked
away.
Yes
yeah,
he
has
yeah
he's
telling
him
hypercube.
B
Space
space
time
is
a
cube
space.
Space
is
a
map,
a
spectrum,
a
Spectra.
There
are
certain
things
and
that's
it
if
you
want
to
do
something
weird
and
wonderful,
but
you
got
to
go
yeah,
but
what
happens
when
you
have
space
space
time,
Spectra
and
stocks,
then
you
then
you
that
wouldn't
mean
to
implies.
You'd
have
to
have
your
own
package
to
do
that.
Sorry,
but
like
that
data
exists,
oh
it
does
yeah,
but
yeah.
So
like
I,
don't
look,
that's
the
problem.
B
B
So
if
Maps
is
going
to
stay
2D
and
to
be
a
bit
provocative,
what
what
are
the
advantages
of
having
inheritance
from
ndq
just
because
I
have
forgotten,
there
are
several
of
him.
The
first
one
that
jumps
to
mind
is
that
the
API
you
can
have
a
consistent
API
across
different
data
types,
whatever
objects
you're,
using
whether
it's
a
specialized
one
or
a
cdq.
You
can
you
that
you
can
the
things
that
you
would
do
common
between,
say
like
a
map
or
a
map,
queue
of
some
kind
or
a
you
know
a
spectrogram.
B
You
know,
then
you,
if
you're,
dealing
with
spectrograms
or
images
the
same
source
of
operation
if
you've
done
the
same
way
plus
then
the
stuff
that
is
coming
a
lot
of
that
code
base
will
live
in
ndq.
So
when
you
do
want
to
make
a
spectrum
class,
you
know
you
just
inherit
that
straight
away
and
it's
the
same
as
it
would
be
in
that
yeah
I
I
think
we're
driving
off
topic
of
it
here,
like
I,
think
we
agree.
B
This
is
a
good
thing
to
do
and
we
can
talk
about
the
Practical
upshot
of
it
later.
We
should
put
it
on
the
roadmap
and
put
more
things
on
the
road
right
yeah.
So
so
I
wrote
just
to
be
clear
and
not
to
limit
things
as
much
but
begin
to
integrate
into
ND
Cube
into
sunpakore,
and
do
we
need
to
begin
to
on
the
roadmap
like
well
long
term?
We
just
want
to
do
that
right.
Okay,
fine,
yes,
so
we're
moving
again
good
point,
but
does
that
mean
moving
ndq
the
code
base
into
core?
B
B
Another
thing
right
right
exactly
if
you
have
another
suggestion:
meditator
objects,
yes,
I
mean
that's.
We
need
so
just
to
just
clarify
this
list
of
points
we
want
on.
The
roadmap
are
influenced
by
what
we've
discussed
before
yes,
yeah
I
was
just
Let's
Travel
for
everything.
Sorry
I'm,
yeah
and
ndq
mapped
ndq
position,
and
that
comes
with
probably
replacing
that
sequence
with
the
PQ
sequence.
B
Metadata
I
know:
Steve
has
a
very
old
having
allele
project
an
issue
in
getting
back
like
2016
on
the
metadata
new
metadata
class.
That's
something
I'm
a
bit
torn
on
putting
metadata
on
the
roadmap
because
there's
so
much
to
it,
but
maybe
that's
what
we
need
to
do.
We
just
need
to
put
a
big
mess,
the
data
heading
and
then,
at
the
moment
like
we
can
figure
that
out
later
well,
no
one
said
we're
doing
metadata
poorly
yeah,
exactly
yeah
we're
not
addressing
it.
The
shortcoming,
no
but
I,
think
sorry
I.
B
B
B
I,
don't
know
that
it's
necessarily
A
roadmap
item,
though
I
do
agree
that
probably
a
whole
class
would
be
nice,
but
I
yeah
I
I
tend
to
agree
that
we
never
identified
this
as
something
we're
doing,
particularly
poorly,
so
yeah
I,
even
though
I'm
a
big
proponent
of
this
I,
don't
know
that
we
should
be
dumping
a
bunch
of
energy
into
it.
B
But
one
thing
is
that
if
you're
going
to
support
more
arithmetic
operations
on
your
numbers
and
map
sequences-
and
there
are
header
keywords
that
are
derived
from
the
data-
you've
got
to
be-
it's
not
careful
in
what
you
do
otherwise
you're
gonna
screw
people,
there's
lots
of
metadata
things.
We
could
improve.
I
know
I,
didn't
say
earlier,
but
like
that
I
have
a
whole
list
in
my
brain
and
on
various
issues
like
there
were
definitely
metadata
things.
We
could
do
better
in
terms
of
segregate
how
we
Translate
from
metadata
to
python
objects.
B
We
don't
have
really
have
a
way
of
modifying
the
python
object
and
going
back
to
metadata.
We
don't
have
a
way
of
separating
different
classes
of
metadata
keywords
into
ones
that
somebody
understands
versus
just
extra
metadata
that
we
don't
ever
modify
and
touch
like
stats
keywords.
We
don't
have
a
way
of
supporting
other
metadata
representations,
other
than
fits
WCS
or
just
fits
in
general,
but
I
think
that
I
think
that
we
should
identify
a
user
need
for
each
of
those.
B
Instead
of
saying,
I
want
my
message
to
be
python
objects
because,
yes,
we
already
translate
the
metadata
into
python
objects.
That's
what
that
does
yeah.
We,
the
user,
need
for
things
like
stats.
Keywords
exists
like
we're
just
marking
them
as
stale
or
just
telling
someone
that
this
is
no
longer.
You
just
done
something
else.
Right.
B
Yeah
I've
got
a
question
so
like
a
a
cube
of
maps
where,
like
one
of
the
dimensions,
is
time,
is
that
easy
to
handle
in
ND
Cube?
Yes,
you
could
immediately
if
you're
gonna
replace
map,
then
could
you
replace
map
sequence?
B
Yes,
okay,
so
do
you
want
to
put
that
in
there
as
well
yeah,
which
is
on
the
list?
I,
just
put
it
in
have
map
sequence
in
here
from
ND,
Cube
sequence:
okay,
yeah,
sorry,
the
alternative
to
that
is,
make
it
a
map
Cube,
but
that
would
then
needs
the
data
to
be
reprojected
to
the
same
WCS
but
like
either.
Both
are
valid
ways
to
go
and
both
can
be
done
in
terms
of
the
method
on
nbq
sequence
right,
then
you
could
just
do
it.
Yeah
yeah,
yeah,
yeah,
yeah,.
B
B
Yes,
that's
a
grenade,
because
it's
not
really
what
that's
designed
for
it's
not
right.
So
it's
for
everything
except
one
dimension.
Well,
like
it's,
okay,
ND
Cube
people
don't
expect
to
load
a
Time
series.
B
Objects
like
you
know
and
like
have
the
ndq
API
in
my
opinion,
I,
don't
know,
I,
don't
use
time
series
data
very
much,
but
you
know
they
expect
it
to
act
like
a
table
and
then
the
cube,
just
kind
of
doesn't
I
think
the
I
think
the
biggest
argument
is
that
there's
so
much
support
in
like
either
pandas
or
or
time
series
like
four
specific
time,
series,
analysis
and
I
think
duplicating
all
of
that
on
the
ndq
API,
probably
unnecessary,
given
how
mature
tools
exist.
For
that
specific
case,
that's
my
okay.
B
So
right
time,
series
are
special
is
what
you're
saying?
Yes,
you
can
represent
time
in
an
ND
Cube,
that's
fine,
because
you
can
represent
in
a
WCS
but
I
think
in
a
lot
of
use
cases.
People
don't
have
the
time
as
WCS
and
there's
like
a
lot
of
like
Club
sampling
and
down
sampling
and
various
like
things
that
there's
so
many
tools
to
do
that
sort
of
analysis,
specifically
with
time
that
there
are
more
mature
tools
that
people
are
already
used
to
so
yeah
I
would
say
it's
from
that
in
practice.
B
Time
series
is
a
bit
special,
but
if
you
want
to,
you
could
do
it
with
anything
so
just
before
Stephen
jumps
in
so,
if
I
take.
If
I
sit
on
one
pixel
in
a
map,
Cube
I'm
gonna
get
out
something
which
is
not
an
ND
Cube
and
so
I'm
gonna
have
to
switch
API.
Then
no
you
get
and
slicing
an
ND
Cube
gives
you
an
ndq
always
so.
I'm
gonna
have
to
recast
that
into
a
Time
series.
B
If
I
want
to
hand
if
I
want
to
use
it
in
all
these
nice
time,
series
tools,
if
you
want
to
use
a
Time
series,
then
de
facto,
yes
as
an
electron
to
this
object,
if
you
want
all
the
functionality
of
the
time
series
object,
gives
you
then
yes,
but
maybe
there's
like
two
times
these
methods.
B
Okay,
is
that,
like
sorry,
even
yeah,
so
sorry
just
to
just
to
follow
into
what
you
just
said,
Jack
I
mean
can't
you
just
have
the
ND
Cube
like
when
you
slice,
and
it
sees
it's
just
time,
just
I'll
put
a
Time
series.
So
that's
that's
Albert's.
Hypercube
concept
is
that
you
know,
depending
on
which
slice
or
which
angle
you
look
at
the
queue.
It
appears
to
be
a
map
or
a
map
Cube
or
slash
sequence
or
a
Time
series.
So
that's
that
that's
how
that's
conception-
yeah,
that's
still
a
heated.
B
There
are
different,
different
views
on
that
for
for
valid
reasons
on
both
sides.
Okay,
so
so
what
I
was
going
to
say?
The
reason
why
I
raised
my
hand
was
you
know,
ND
cube,
is
really
great
when
you
have
a
WCS
object,
otherwise
otherwise
you're
kind
of
stuck.
So
what
data
types
are
come
with.
Wcs
objects,
I
mean
it's
mostly
Imaging
type
stuff
right
like
not
time
series,
of
course,
as
we
just
said,
but
what
about
Spectra
yeah
lots
of
spectral
data
has
WCS
yeah.
B
Okay,
so
that
naturally
kind
of
provides
a
scope
for
where
ND
Cube
would
be
applicable
and
where
it
isn't
now
now
remember.
Remember,
though,
that
you
know
once
you
leave
solar
physics
and
astrophysics
I
guess:
WCS
objects
are
not
used
so
doing
a
hard.
You
know.
Integration
of
ND
Cube
into
sunpi
basically
means
that
you
can
no
longer
use
some
pie
easily
for
other
data
sets,
because
you
know
people
have
no
idea
what
a
WCS
object
is
in.
You
know
space
physics.
B
Like
it,
I
I
feel
like
with,
with
we've
lost
the
yeah.
The
thread
of
writing
a
road
map
here,
but
it's
easy
enough
to
make
a
WCS
like
eventually
two
from
x-ray
can
happen,
Elementary,
okay,
to
derail
this
one.
In
Jack's
case
you
will
just
have
a
methods
on
ending
Cube.
This
is
a
one-dimensional.
There
is
a
two-time
series
that
does
something
odd,
that
they've
output
a
completely
different
thing
than
you
could
use
in
that
case.
That
would
be
something
to
maybe
discuss
about
any
cubes
IO
API
in
the
future.
B
B
B
B
How
about
implementing
a
visualization
object
or
like
images
like
breaking
it
out
from
map
s
under
the
ndq
list?
Okay,
all
right
it
sort
of
is
he
may
be
just
improving.
Plotting
is
actually
basically
a
roadmap,
but
it
is
super
nice.
It
says
to
not
to
have
the
opportunity
to
not
have
to
use
not
that
label
which,
which
yeah
rolling
out
a
plot
of
class
too.
B
New
framework
Club
do
we
want
to
have
the
documentation
stuff
on
the
roadmap,
or
is
that
yeah
somewhere
else,
yeah
well,
I
think
we
all
these
so
far
are
very
like
technical
things
and
I
think
I
think
one
of
the
most
important
things
that
we've
seen
we've
done
poorly,
at
least
in
my
opinion,
is
Chris's
comment
about
the
contribution.
Yeah
I
mean
yeah,
yeah
I.
Think
a
lot
of
people
have
versions
of
this
experience
and
I
I'm,
not
really
sure
what
the
yeah
I
mean.
B
I
think
a
maybe
a
big,
clear
expectations,
as
Chris
said
that
we
had
certain
things
in
the
docks
yeah,
but
that
didn't
cover
all
of
the
expectations.
Yeah
I
mean
clarified
these,
like
unspoken
yeah,
like
sort
of
not
even
rules
but
like
expectations
that
exist
like
yeah,
also
not
just
what
you
we
expect
from
the
code,
but
what
you
should
expect
from
a
review
process.
Well,
yeah
like
what
that's
a
good
point,
yeah
I
think
having
maybe
a
concrete
thing,
maybe
well
the
thing.
B
This
is
too
specific
for
roadmap,
but
it
is
a
like
a
clearly
defining
what
a
code
review
is
or
like
what?
What
what
are
they
like
or
viewer
checklists?
Not
just
a
contributor
checklist,
improving
the
contributor
yeah
pipeline
yeah?
That's
the
roadmap
item
Steve
yeah
I
was
going
to
say
that
the
contributor
issue,
I
think,
should
be
at
its
own
item.
It
shouldn't
be
buried
under
documentation.
I
just
broke
it
out.
B
B
B
I
think
in
this
is
connected
to
that.
Maybe
not
quite
the
same
thing,
though,
is
is
maybe
we
just
put
this
running
running
the
survey
more
at
a
more
regular
Cadence
to
have
like
taking
the
temperature
of
the
community
as
far
as
use
of
senpai
and
and
like
attitudes
towards
the
project
towards
the
developer
community.
B
B
B
I
was
just
gonna
say,
like
I,
think
a
roadmap
pattern
could
be
a
more
clear,
more
clearer
messaging,
around
the
structure
around
the
structure
of
the
project,
and
that
includes
the
division
between
core
affiliated
sponsored
okay,
never
mentioned
mentioned
again
in
public.
Apparently
there's
a
lot
of
bullet
points
that
go
underneath
as
well.
B
I'm
going
to
like
just
run
like
giving
talks
and
Outreach
tutorials
like
it
would
be
nice
to
have
a
set
of
ready-made
talks
and
ready-made
tutorials
so
that
you
could
go
and
be
like
anyone
could
go
and
give
a
tutorial
because
you
don't
need
to
develop
the
whole
notebook
or
a
whole
slide
deck.
It's
just.
It
was
nice.
It
would
be
nice
to
be
able
to
like
take
a
tutorial
out
of
the
documentation,
download
it
as
a
notebook
yeah.
It's
it's
fun.
B
B
Think
the
broader
heading
for
these
is
like
improve
our
branding
like
yeah
yeah
I,
think
that
that
includes
yeah,
the
tutorial
stuff,
the
the
reduction
Edition
so
and
then
also
just
like
basic
things
like
having
like
a
sense
of
templates,
for
we
were
talking
about
this
last
night
here
earlier
today.
Sets
of
templates
for
for
giving
presentations
like
a
Google
slide
template
a
keynote
template,
yeah
or
mode
whatever.
B
We
should
brand
knitting.
Maybe
it
is
so
I
included
in
what
we're
doing
fully
I
added
a
few
things.
Now
what
we
were
saying
so
communicating
some
Pi
Project
capabilities
are
spread
across
multiple
packages.
Improving
people's
understanding
of
what
sponsor
depleted
packages
are,
as
well
as
the
scope
of
core,
and
also
something
we
discussed
earlier,
is
communicating
or
having
a
pitch
for
the
benefits
of
becoming
a
sponsor
Affiliated
package
which
came
up
this
morning.
So
just
to
let
people
know
that's
there.
Yeah.
B
B
Are
we
saying
that
that's
going
to
be
the
base
class
is
going
to
be
in
D
Cube
as
well?
Yes,
sir?
Well
I
guess,
ultimately,
that's
a
good
point.
Yes,
that's
a
version
of
an
implementation
of
a
spectral
object
is
in
Sun
roster
and
that
is
based
off
of
ND
Cube,
so
I
mean
one
potential
Way
Forward
is
to
have
as
a
long-term
roadmap
or
whatever
term
roadmap
item
is
to
move
some
of
sun
raster
into
core.
B
B
It's
in
there
because
effectively,
that
was
like
an
incubator
for
that
stuff.
Yeah,
it's
broken
out.
It's
going
from!
Oh
I
need
slit
scanning
object
and
then
realize.
Well,
actually,
you
know
I
need
a
spectrogram
object
and
then
sort
of
inherit
from
that
and
make
it
more
specific.
So
it's
there,
because
that
was
the
incubator,
absolutely
yeah.
What
about
it
when
a
slitless
spectrograph,
also
you
can
scope
for
or
somewhere
else.
B
Yeah
is
that
captured
in
the
roadmap
listing
there,
so
the
bottom
special?
Yes,
yes,.
B
Steve
says
what
about
spec
utils,
unfortunately
Steve
you're,
not
in
the
pub,
so
you
won't
hear
my
opinions.
I'll
expect
you
build.
B
We've
got
one
minute.
Is
there
anything
missing
from
our
approaches?
Just
bring
it
back
to
the
Spectra
radio
Spectra,
but
I
guess
there's
more
work
to
be
done.
Oh
God,
yes,
support
from
Radio
Data
separate
for
support,
suspect
for
later
I
mean
ideally,
yeah
would
just
inherit
from
a
broader
spectrogram
object.
B
B
B
Subclass
so
right
you
can
figure
out
what
it
does
and
doesn't
do
for
what
we
needed
to
and
work
backwards
from
that
like
I,
think
there
are
some
issues
with
it
for
what
we're
going
to
need-
and
you
know
it
might
not
might
be
easier
to
add
through
things
it
might
not
be
depending
on
you
know
how
jwst
is
going
I
guess,
but,
like
you
know,
it
definitely
needs
evaluating
carefully
if
nothing
else,
I
think
this.
If
the
solution
ends
up
being
we
just
in
the
project
level
documentation.
So
he
has
a
solo
physicist
physicist.
B
If
you
want
to
do
Spectra
imports
from
spec
utils
like
if
that
that
you
know
I
think,
is
a
valid
to
result
from
that
that
inspection
or
that
this
task.
It
probably
won't
be
that
simple,
but
that's
an
acceptable
solution
in
his
current
form,
is
probably
too
focused
on,
like
1D
Spectra
and
not
well
suited
for
spectrogram
compiled
higher
stuff
in
comparison.
But
again,
this
is
not
a
time
for
this
discussion
like
we
do
right.
B
Right
just
to
go
back
to
the
time
series
stuff
like
do
you
have
a
plan
to
support
vectors
at
all.
I
know
that
we
can
plot
like
these
ADB
through
all
the
access
stuff,
like
that,
but
I'm
sure
there's
other
things
that
you
want
to
do
with
Speculator
that
you
know
don't
help
me
sort
of
like
impact
the
object
just.
B
Things
that
you
might
want
yeah,
a
really
common
thing
saying,
is:
transforming
like
magnetic
tailback,
so
between
different
coordinate
systems.
I
think
Albert
had
that
a
lot
yeah
transforms
for
Vector
quantities.
Oh
I
feel
like
we're
missing
a
whole
coordinate
section
from
this
roadmap
as
well
and
I'm.
Sure
Albert
would
like
to
put
some
stuff
on
there
or
something
should
we
just
just
put
a
coordinate
and
just
let
Albert
out
and
then
people
you
will
100
put
that
in
as
soon
as
see.
B
B
B
But
okay,
it
should
end
up
somewhere
and
I
will
I
will
warn
you
I
knew
we
can
argue
in
the
comments
on
the
pr
that
I
think
some
of
these
things
are
maybe
not
worthwhile
to
be
on
the
roadmap,
because
they're
just
areas
of
continual
Improvement
sure
so
yeah,
yeah,
I,
think
yeah,
probably
like
the
really
high
level
stuff,
and
probably
a
lot
of
this
can
slide
like
since
light
is
bullet
points
under
the
kind
of
larger
headings.
B
B
Improvement
I
hear:
are
you
around
for
the
roadmap
section
the
governance
session
tomorrow,
yeah
yeah
sweet,
it's
a
long
session.
Anyone
else
online
have
anything
else.
Anything
I
want
to
add
before
we
wrap
up.
B
Irish,
summertime
IST
when
we
can
argue
about
chat
platforms
and
you
all
can
lose
someone
quickly,
gonna
make
a
slack
something
one
did
exist,
should
be
a
Hellion
of
slimes.