►
From YouTube: JupyterLab Team Meeting - March 9, 2022
Description
A meeting to share and discuss features, ideas, issues, and pull requests in JupyterLab and other Jupyter frontends. This meeting is open to anyone and everyone.
Join future calls via the Jupyter community calendar: https://docs.jupyter.org/en/latest/community/content-community.html#jupyter-community-meetings
Notes for upcoming meetings can be found on the agenda: https://hackmd.io/Y7fBMQPSQ1C08SDGI-fwtg
Past notes can be found on the JupyterLab team compass: https://github.com/jupyterlab/team-compass/issues?q=is%3Aissue+label%3A%22Dev+Meeting+Minutes%22+
A
Okay,
welcome
to
the
march
9th
jupiter
lab
weekly
call.
Today
we
have
what
looks
like
yeah
22
people
on
the
call,
that's
great
and
if
you
haven't
already,
please
find
the
agenda
in
the
chat
thanks
jason
for
pasting
it
in
and
if
you
would
like.
Please
sign
in.
Please
add
any
items
you'd
like
to
discuss
in
the
agenda
section
of
the
minutes
and
let's
get
started.
The
first
person
on
the
list
today
is
federic.
B
I
don't
remember
if
I
spoke
about
it
last
week,
I
think
so
so
I
started
to
add
a
pr
for
testing
some
of
our
that
are
defining
ui
things
like
shortcut,
menus
and
stuff
like
that
to
be
sure
that
in
in
core
on
the
ci,
we
can
check
that
those
data
are
not
pointing
to
comments
that
does
not
exist
and
the
side
effect
I
have
implemented
from
it
is
to
be
able
to
to
produce
a
full
list
of
the
comments
that
are
in
the
in
jupiter
app
for
creating
for
adding
them
to
the
dock.
B
B
I
get
two
questions
for
it,
so
one
of
the
things
that
the
test
is
enforcing
now
is
that
you
should
not
have
a
comment
that
don't
have
any
label,
so
we
are
sure
there
are
some
kind
of
documentation
on
every
comment,
so
I
did
a
couple
of
flowers
that
were
missing.
Let's
see
so
that's
the
first
question
is:
does
it
make
sense
and
I'll
be?
B
I
will
also
appreciate,
if
some
english
speaker
can
look
at
the
wording,
because
that's
not
my
strong
side
and
the
other
question
is
where
to
put
it
in
the
in
the
documentation.
So
for
now
I
put
it
at
the
command
panel
palette,
but
it
doesn't
make
sense
because
the
user
is
not
seeing
really
that
id
stuff.
So
if
people
have
id
where,
where
it
belongs,
that
will
be
nice
and
that's
it.
For
me.
A
Thanks
darian,
I
had
a
question
about
the
labels.
So
are
the
labels
coming
from
json
files
or
are
they
coming
from
code?
Like?
Does
translation
get
applied
to
them?.
B
So
I'm
using
what's
in
the
command
registry.
So
if
there
are
translation
available,
they
will
be
available,
so
I'm
using
playwright
so
that's
mean
that
we
load
fully
jupiter
up
and
then
I'm
injecting
some
javascript
to
say.
Oh,
give
me
the
command
list
and
from
that
command
list.
I'm
asking
all
the
label.
C
Not
every
command
is
in
the
command
palette,
though
every
command
can
be
called
from
the
keyboard
shortcut,
so
maybe
it's
better
to
group
either
commands
separately
from
command
palette
or
with
the
keyboard
shortcuts,
or
just
make
that
clear.
Even
and
I
mean
the
subtlety
here
is
not
every
command
is
added
to
the
command
palette
and
some
commands
require
arguments
and
a
specific
set
of
arguments
is
added
to
the
command
palette,
but
the
keyboard
shortcut
may
specify
other
arguments
to
a
particular
command.
A
D
Sure
all
right
well
so
thanks
everyone
for
the
feedback
that
you've
given
so
far
on
the
cell
toolbar.
Let
me
sorry,
I
can
link
the
pull
request
here,
make
sure
I
send
it
to
everybody.
So
there
is
a
binder
as
well,
which
I
can
link
in
the
chat
in
case.
Anyone
wants
to
play
around
with
it.
D
So
for
those
who
are
unfamiliar,
the
cell
toolbar
is
something
that
I've
been
working
on
for
a
few
weeks
now,
and
I
am
currently
adding
this
toolbar
here
hope
everyone
can
see
my
screen
to
every
cell
and
it's
designed
to
disappear.
The
current
behavior
is
to
have
it
disappear
in
case
it
would
interfere
with
the
content.
So
if
I,
for
example,
make
this
line
longer,
the
cell
toolbar
disappears
when
I
make
it
shorter.
D
It
appears
again
I've
been
talking
with
fernando
and
some
other
folks
about
like
ways
to
sort
of
shunt
it
out
of
the
way
in
a
way
that
you
can
make
it
appear
again.
D
I
don't
have
anything
to
show
you
today,
so
this
is
the
binder.
I
also
have
another
concept
here.
Oh
this
is
not
working.
Let
me
try
reloading
my
browser
tab
here
so
based
on
some
feedback
I
got
from
lane.
He
was
suggesting
that
we
move
the
move
up
and
down
buttons
to
the
left
and
lane,
as
we
have
some
to
show
you
just
yet,
but
it
was
intended
to
kind
of
move
those
functional
buttons
to
another
place.
They
wouldn't
get
overwritten
by
the
content.
D
So
this
is
what
things
would
look
like.
So
with
with
my
alternative
branch
here,
which
is
not
yet
available
on
github,
the
in
cell
sorry,
the
cell
toolbar.
I
have
to
stop
saying
in
cell.
That
means
something
I
shouldn't
say
on
a
recorded
line,
so
there's
a
four
button
toolbar
and
then
the
up
and
down
buttons
get
moved
to
the
left.
D
Lane
had
another
concept
which
I'm
curious
to
get
folks
opinions
on
where
there's
this
little
invisible
and
we
use
the
term
grabber
that
if
I
grab
this,
I
can
drag
and
move
it
up
and
down
the
fact
that
it's
invisible
makes
it
very
hard
to
for
for
a
novice
user
to
understand
that
there's
anything
here,
because
it's
just
invisible.
D
D
E
To
review
it-
and
I
haven't
because
I
have
a
lot
of
questions
about
accessibility
in
a
lot
of
places
so
and
mike
is
asking
a
good
question
with
that
in
the
chat.
How
would
that
work
with
the
mobile
layout
is
like
a
good
way
to
test
it?
But
so
I
guess
what
I'm
saying
is:
is
there?
E
Is
there
either
a
like
list
of
things
that
have
been
added,
so
I
can
easily
review
or
like?
Would
it
be
possible
to
schedule
like
a
call
with
you,
that
forces
me
to
sit
down
and
start
pressing
buttons
like?
I
don't
know
what
your
time
commitment
is,
but
I
really
do
want
to
review
this,
because
this
is
like
there's
so
much,
but
I
just
have
another
chance.
So.
D
Oh
sure,
so
sorry
for
cutting
you
off
there
momentarily
so
isabella.
I
appreciate
the
concern
about
accessibility
in
the
the
pull
request,
frederick
added
an
interesting
question
that
I
don't
have
an
easy
answer
for
which
is
about
the
accessibility
of
it.
So
if
I
put
my
if
I
put
this
into
command
mode
and
I
hit
tab,
it
will
select
the
first
button
on
the
toolbar.
D
If
I
hit
tab
again,
it
selects
the
second
button
on
the
toolbar
frederick
made
a
good
observation,
which
is
that
the
the
I
think
it's
the
w3c
guidelines
are
that
tab
should
focus
the
toolbar
and
then
left
and
right
will
then
navigate
within
the
toolbar.
However,
left
and
right
are
already
bound
in
this
mode
to
navigate
between
headings.
D
So,
if
I
hit
right,
it's
going
to
jump
me
to
heading
one
and
if
I
hit
left,
it's
gonna
collapse
the
heading
and
that's
what
left
and
right
are
meant
to
do
in
jupiter
lab,
so
we're
going
to
either
need
to
unbind
left
and
right
when
the
cell
toolbar
is
focused
or
find
some
other
way
to
navigate
between
them
or
just
leave
the
tab.
Navigation
as
it
is
yeah.
B
No
just
to
say
that,
normally,
if
the
toolbar
is
focused,
you
can
just
stop
the
propagation
of
the
event,
and
you
should
be
fine.
D
D
Made
before
it
gets
it's
up
to
me,
frederick,
so
I
gotta
jump
the
queue
or
or
something
or
unbind
it
temporarily.
I
also
noticed
that
mr
grout
said
this.
The
heading
navigation
is
very,
very
recent.
I
think
it's
a
good
feature.
To
be
honest,
I
don't
know
if
I
should
undo
it
in
favor
of
this,
but
it
it
creates
a
sort
of
a
deadlock
for
me
and
that,
like
I,
I
see
the
merit
and
be
able
to
navigate
using
the
arrow
keys.
D
We
do
have
an
alternative
which
is
to
simply
tab
from
button
to
button
which
I
know
is
not
strictly
in
agreement
with
what
the
w3c
recommends,
but
it
does
work.
C
That
heading
navigation,
I
think,
was
introduced
very
recently
and
I
don't
think
had
a
lot
of
overall
review
about
adding
that
shortcut.
So
I
wouldn't
be
opposed
to
taking
a
second
look
at
that
shortcut
and
if
it's
complicating
things
in
other
places
where
we
want
an
overall
picture
of
how
our
shortcuts
to
behave,
I
I
think
we
can
take
a
second
look
at
it
and
reconsider
it.
D
Oh
for
sure,
and
to
be
honest
like
I
don't
particularly
dislike
that
heading
navigation,
I
kind
of
like
that,
but
at
the
same
time
like
it
was
probably
put
in
for
a
reason.
D
G
Yeah,
so
in
terms
of
accessibility
right,
the
all
the
actions
are
currently
available
as
shortcuts
right.
This
is
just
about
connecting
a
keyboard
action
to
visual
navigation,
to
the
visual
elements
right.
E
E
I
am
not
confident
that
keyboard
shortcuts
constitute
accessibility,
because,
even
though
you
can
access
the
keyboard
like
the
shortcuts
through
your
keyboard,
which
is
some
concern
that
we
have
it
requires
you
to
be
aware
of
these
shortcuts,
which
may
not
be
surfaced
through
a
number
of
ways
like
I.
I
right.
E
G
Yeah,
it
will
work
for
some
users,
but
not
others.
I
see
your
point
so
then
the
question
is:
how
would
this
does
this
work?
Well,
is
this
inside
the
code
mirror
element
the.
D
In
terms
of
like
in
well
it's
inside
the
cell,
it
it
reacts.
The
logic
reacts
to
changes
to
the
code
in
your
entity,
but
like
it's
not
inside
the
code,
mirror
editor.
Does
that
make
sense
right.
A
D
Sorry
to
to
clarify
the
the
duplicate,
the
duplicate
cell
below
is
a
new
command
that
I
actually
don't
think
is
in
the
command
palette
right
now.
So
that's
probably
the
only
novel
functionality-
that's
in
the
the
cell
toolbar,
but
it's
not
exactly
a
potential
functionality.
A
Well,
you
could
add
that
to
the
command
palette
without
any
yeah.
Okay,
so
it
seems
as
though
there
are
there
are,
on
the
one
hand,
there's
definitely
a
conversation
to
be
had
about
how
to
make
the
cell
toolbar
the
most
accessible
it
possibly
can
be,
but
there's
also
the
how
to
mitigate
any
shortcomings
by
making
sure
there's
multiple
ways
to
access
all
the
functionality
there.
A
I
would
suggest
now
that
you've
got
this
to
a
fairly
polished,
demobile
state.
Bringing
this
to
the
next
accessibility
call,
so
that
you
can
spend
a
good
deal
of
time
on
it
or,
as
isabella
says
you
guys
can
schedule
a
private
call
if
you
wish,
whatever
whatever
you
want,
but
it
seems
like
it
seems
like
it's
at
that
point
in
its
life
cycle,
where
it
would
be
valuable
to
do
that.
E
D
E
D
Okay,
let's
continue
the
conversation
in
there
then
I'll
I'll.
I've
got
a
couple
other
things
on
my
calendar,
but
I
can
move
those.
A
Okay,
we
have
jeremy
next.
A
I
love
how
understated
that
is
and
how
big
it
is
at
the
same
time,
this
is
amazing,
nice
work
and
thank
you
for
everyone
who
looked
at
this
to
try
to
validate
before
it
went
out.
Thank
you
for
all
the
work
you've
done
on
this
jeremy
and
yeah
pip
install
pre-notebook
everyone.
A
Okay,
jason
grow.
C
On
releases
I'll
just
say
that
was
months,
slash
years
of
work,
so
congratulations
again,
jeremy
and
all
the
others
involved.
In
that
I
was
planning
on
doing
a
3.3.1
release
today
to
exercise
the
new
you
know,
learn
more
sharpen
the
skill
on
the
new
release
system
was
the
procedure
for
3.3.0
recorded.
I
It
was
not
recorded,
but
this
one
should
be
simpler
in
that,
in
the
case
that
yeah
we
just
used
the
next
version,
spec
in
the
user,
okay,
so
yeah
and
the
still
has
a
video
around
and
I
think,
can
be
used
as
a
documentation.
C
C
There's
been
lots
of
back
and
forth
on
that
discussion,
but
it
led
to
3.3.1
very
clearly
signifying
that
the
collaborative
flag
that
turns
on
rtc
in
jupiter
lab
3..
It
will
be
noted
in
gibraltar,
3.3.1
as
an
experimental
feature
and
and
just
in
general
that
that
the
collaborative
feature
is
considered.
Experimental
apologies
to
people
that
have
been
using
it
and
haven't
see
haven't
been
told
that
it
was
experimental
still.
A
I
linked
that
issue
in
your
bullet
point
jason.
So
this
is
one
of
those
cases
where
having
that
experimental
signifier
is
helpful
to
people,
but
it's
a
really
bad
experience
if
work
disappears
and
people
may
not
use
that
feature
for
a
very
long
time
to
come,
I
don't
know
what
it's
going
to
take
to
fix
it
because
it
seems
hard
to
reproduce
and
it
seems
like
it's
somewhat
dependent
on
a
networked
file
system
and
there
seems
to
be
a
lot
of
circumstances
that
make
it
hard
to
reproduce
this
bug.
A
But
if
you
have
the
capacity
to
work
on
this
issue,
this
is
one
of
those
things
where
it
is.
It
could
really
turn
off
the
user
to
have
lost
something
they
worked
on.
They
might
just
shy
away
from
this
whole
area
for
quite
some
time.
So
please
take
a
look
if,
if
this
is
in
your
wheelhouse,
because
I
think
it's
pretty
important,
okay,
cool
martin,
you
are
up
next
where's,
the
smart
town
which
martin
is
this.
There's
more
time?
Hey?
Yes,
it's
me
so.
J
I
wanted
to
discuss
that
pr.
I
will
add
the
gif
to
to
the
the
chat
so
this
api.
It
was
it's
about
rtc,
so
I
I
took
some
work
that
carlos
started
some
months
ago
about
adding
a
left
panel
for
collaborative
editing
and
this
pr
takes
some
pieces
of
it,
and
so
what
we
did.
A
J
It's
I
I
have
the
tendency
to
break
jupiter,
yeah
right
now,
so
yeah,
it's
not
working
for
me.
Right
now,.
A
Okay,
vidar,
you
are
up
next.
G
Yes,
sorry,
I
lost
track
of
the
notes,
because
I
was
looking
at
a
gif
yeah.
The
I've
been
looking
recently
at
the
debugger
about
improving
support
for
stepping
into
library
code
and
things
like
this.
So
I
don't.
I
don't
know
if
this
is
the
right
call,
because
it
lives
in
ipad
kernel.
But
I
don't
know
if
there's
a
good
relevant
call-
and
I
know
most
of
the
people
who
worked
under
the
debugger
might
show
up
here.
G
And
the
only
filtering
support
that
debug
pi
supports
out
of
the
box
is
by
module
name
or
path
which
isn't
really
what
we
want,
because
you
would
have
to
add
async,
io
and
set
mq
and
traitlets
and
other
packages
to
that.
To
avoid
stepping
into
the
kernel
loop
logic
and
if
they
make
changes,
you
would
have
to
manually
curate
that
list
and
it
would
prevent
you
from
from
stepping
into
similar
code
in
your
own
notebook.
G
So
what
hypothenused
to
do
for
the
pdb
extension
was
to
use
a
special
magical
variable
called
traceback
height
and
set
it
equals
to
ipython
bottom
and
use
that
as
a
cutoff
for
the
stacks.
So
I
was
wondering
if
there
was.
A
If
you
open
an
issue
and
tag
johan
on
it,
I
bet
he
knows.
The
answer
to
this,
and
one
question
I
have
is,
is
the
experience
of
seeing
all
that
messaging
back
and
forth
the
same
in
ipython.
Sorry
ipi,
kernel
as
it
is
in
zeus,
python
do
both
end
up
stepping
through
all
that
code.
If
the
answer
is
yes,
can
we
fix
this
in
two
spots?
If
the
answer
is
no,
I
wonder
how
zeus
python
prevents
the
user
from
seeing
that.
J
The
I
guess
you
would
step
into
a
python
code
yeah,
but
you,
you
wouldn't
see
anything
related
to
the
kernel,
but
you
you
would
step
into
a
python
code.
I
guess.
G
Yeah,
because
the
point
is
that
for
ipython
right
it's,
it
breaks
down
your
cell
content
to
a
bunch
of
ast
statements
and
you're
stepping
back
into
ipython
code
between
each
statement,
and
it's
it's
also
async.
So
suddenly
you
might
step
into
some
other
async
task
if
you're,
not
careful.
So
it's
it's
not
a
pleasant
experience
out
of
the
box.
G
A
Okay,
so
yeah,
I
think
then
your
next
avenue
is
to
hit
up
johan.
He
will
probably
have
insightful
things
to
say
about
this,
but
I
think
you're
right,
like
I've,
noticed
this
myself
and
it
isn't
it's
not
helping
me
in
debugging
the
code.
That's
actually
broken
right.
It's
just
showing
me
a
lot
of
the
back
and
forth.
That
has
to
happen
to
get
my
results
on
the
screen,
which
isn't
exactly
the
same
as
what
I
want
to
see.
So
I
think
this
is
a
good
avenue
to
improve
the
user
experience
of
debugging.
G
All
right
I'll
head
up
johanna
thanks.
H
Yeah,
so
this
is
a
more
of
an
open
question
and
a
bit
of
a
callback
to
a
previous
discussion.
I
forget
how
many
months
ago
it
was
now,
but
at
one
point
it
was
raised
that
we
want
to
completely
overhaul
the
launcher,
make
it
more
extensible
and
kind
of
make
it
behave
like
widgets
do
in
like
ios
or
android
being
able
to
do
lots
of
different
things,
and
that
was
really
cool
discussion.
H
It
was
kind
of
a
future
task
and
I,
for
the
life
of
me,
can't
find
the
github
tracking
issue
for
it
or
any
of
our
previous
discussion
to
follow
up
on
whether
or
not
anybody
started
that
because
if
nobody
started
it
I
would
love
to,
but
I
also
don't
really
yeah.
I
don't
know
where
that
conversation
came
from
and
where
any
documentation
on
it
is,
and,
despite
all
my
googling,
I
can't
find
it
so.
I
was
kind
of
that
was
an
open
question.
A
This
conversation
hasn't
come
up
recently
in
the
calls.
As
far
as
I
know,
what
I
would
suggest
is
since
you're
motivated
to
work
on
it
and
no
one
who
is
working
on
it
has
come
out
of
the
woodwork
recently
to
talk
about
it.
What
you
might
do
is
open
a
feature
issue
on
what
you
plan
to
work
on
that
way.
A
F
That
launcher,
if
you
know
now's
a
good
time,
we're
almost
at
the
end
of
the
call
if
you're
looking
at
working
on
it,
just
share
your
screen,
real
quick.
So
the
other
day
I
was
oh,
you
know
for
most
people
I'll
go
light,
just
they
kind
of
get
a
feel
for
the
real
deal.
But
I
was
asking
myself:
how
do
I
open
the
command
line
and
I
I
was
like
how
do
I
get
to
that
launcher
thing
and
I'm
thinking
to
myself?
F
Okay,
you
know
here's
my
nav
over
here
and
it
turns
out
you
have
to
open.
You
know
the
file
browser
and
then
you
know
the
plus
button
right
to
get
to
the
command
line,
and
I
was
like
that.
Doesn't
you
know
there's
all
kinds
of
cool
nifty
apps
in
here
I
feel
like
the
launcher
should
be
like
a
top
level
thing
on
the
left
side
of
the
menu,
because
it
has
nothing
to
do
with
my
files.
F
H
I
feel
like
that's
actually
pretty
comparatively
low
hanging
fruit
to
redoing
the
whole
launcher
yeah.
If
you
want
to
just
like
open
an
issue
on
that,
and
then
you
can
tag
me,
I'm
willing
to
actually
tackle
that
as
a
little
bit
of
preliminary
work
towards
a
better
launcher.
A
K
A
Links
an
interesting
extension
built
by
fred
for
an
enhanced
launcher.
What
is
the
story
with
this
extension?
Is
it
something
you
envision
as
something
that
could
go
into
core?
Is
it
experimental?
What's
the
status
on
it.
B
We
could
make
it
into
core,
so
the
starting
point
was
a
pr.
I
added
in
the
note
about
some
some
thinking
to
refactor
the
current
launcher,
mainly
by
merging
the
two
section
kernel
and
console,
but
I
think,
as
alex
mentioned,
the
there
is
an
opportunity
to
make
it
even
more
extensible,
because
I
think
what
people
would
love
would
be
to
to
be
able
to
plug
like
kind
of
panel
within
the
launcher,
with
custom
things.
What's
this
even
not
possible
in
that
that
extension,
jeremy
is
thinking.
A
Oh
so
alex,
I
would
suggest
then
making
that
one
of
your
ports
of
call
for
where
you
want
to
start
and
what
functionality
you're
thinking
about
and
yeah.
I
think
that
the
launcher
like
where
I
work,
we,
we
overrode
the
launcher
because
we
needed
custom
functionality
and
I've
heard
this
story
from
other
people
as
well,
so
the
extensibility
of
it,
I
think,
is
probably
going
to
be
welcomed
by
lots
of
people
in
the
community
and
as
prime
real
estate.
A
H
What
would
be
a
good?
What
would
be
a
good
avenue
for
getting
some?
I
guess
what
would
be
the
term
some
like
design
ideas
like
because,
as
you
said,
there's
lots
of
people
who
have
overridden
it.
What
would
be?
Would
there
be
a
good
way
for
me
to
find
out
what
some
of
the
other
ideas
people
have
had
when
they've
overridden
it
before
kind
of
like
as
a
basis,
because
it's
really
hard
to
start
from
scratch?
Saying
oh
yeah?
H
We
want
it
to
be
extendable
and
overrideable
and
be
able
to
do
all
these
things,
but
when
you
start
like
actually
getting
to
it
like
that's
a
lot,
which
is
probably
why
this
current
old
pr
took
what
we
currently
had
as
a
base
and
it
improved
it.
But
it's
still
from
what
I
can
tell
from
the
screenshots
and
and
gifs
it's
it's
still
stuck
with
like
categories
and
and
and
panels.
L
I
just
wanted
to
say
like
as
an
experiment:
jupiter
lab
starters
is
a
super
cool
project.
Nick
wrote.
We
wanted
a
way
to
customize
the
launcher
for
classes.
So
like
open
me,
this
notebook
for
this
class
and
there
are
some
nice,
affordances
and
design
features
in
there.
So
it's
a
good
place
to
start.
Perhaps.
H
Okay,
yeah,
I
want
to
say
nick,
was
the
other
loud
voice
in
the
room
when
we
previously
had
the
discussion.
I
was
hoping
he'd
be
on
the
call
today,
but
I
didn't
see
him.
K
Rick,
what
were
you
thinking?
Well,
I
was
just
thinking
that
the
that
I
had
a
revelation
some
months
ago
that
if
I
looked
at
the
launcher,
it
was
essentially
a
desktop
much
like
a
mac,
os
desktop
or
a
windows
desktop,
and
if
you
think
about
how,
if
you
think
about
how
that's
organized
it's
simply
a
a
file
in
the
or
simply
a
directory
under
the
user
directory
desktop
on
mac
os,
for
example,
and
you
just
and
when
you
double
click
on
something
you
just
open
it.
K
K
A
G
G
I
think
I'm
assuming.
If
you
had
a
similar
kind
of
collection
of
favorites,
you
somebody
you
would
say
then,
okay,
I
write
the
five
favorites
extension.
I
will
want
my
favorites
to
show
up
on
the
launcher
so
have
some
way
to
like
make
your
own
space
for
an
extension
on
the
launcher
right.
A
So
if
you
do
launch
your
terminal
all
the
time,
you
drag
the
terminal
button
onto
your
sidebar
and
if
you
have
a
specific
kernel
like
that's
the
one
you
always
want
to
launch,
and
then
you
drag
that
there
or
something
like
that.
But
there's
a
lot
of,
I
think
space
here
to
improve
the
user
experience
of
it
and
to
make
it
well
to
take
advantage
of
the
the
environment
that
the
you're
sitting
in
I
think
alex.
A
One
thing
you
could
do
is
in
addition
to
opening
a
github
issue
linking
that
to
a
discourse
thread
to
get
a
wider
audience
to
talk
about
this
and
to
solicit
some
ideas
for
how
to
improve
it,
because
you
know
like
off
the
top
of
our
heads.
I'm
sure
we
can
come
up
with
what
seems
really
cool
to
have,
but
it'd
be
good
to
have
some
validation
and
some
some
sort
of
numbers
from
you
know.
A
wider
community.
H
Okay,
then,
later
after
this
meeting
sometime
today,
I'll
open
up
an
issue
and
I'll
essentially
add
all
the
discussion
points
from
the
hackmd
and
the
various
links
everybody
gave
me
in
the
chat
and
we'll
use
that
as
a
launching
point
to
find
out
what
work's
been
done.
What
people
are
thinking
and
give
it
a
good
amount
of
time
to
percolate
before
I
start
diving
directly
into
code,
which
is
great
for
lane's
idea,
which
will
allow
me
to
work
on
something
else
related
to
this.
H
That
will
get
my
get
my
toes
wet,
while
I'm
waiting
for
that
to
percolate.
A
Thanks
for
putting
some
thought
into
this
okay,
next,
we
have
andrew.
M
Hey
everyone,
my
name
is
andre
yaroshenko.
You
might
remember
me
from
the
summer
I'm
a
cal
poly
jupiter
intern.
I
was
a
part
of
summer
cohort
and
we
worked
on
notifications
extension
with
darian
and
cameron
and
also
with
support
from
brian
and
from
kevin
and
from
isabella
and
from
many
other
people,
even
including
elaine.
Here
right
and
many
other
people
who
I
wouldn't
mention
so
yeah
I
mean
you,
you
might
know
that
I'm
from
ukraine
and
there
is
a
war
going
there.
M
So
I
I
thankfully
was
able
to
my
family
was
able
to
get
out
like
couple
of
days
before
the
war
started
and
before
the
skies
were
closed
down.
M
But
I
still
have
like
extended
family
and
my
friends
there,
and
in
particular
he
was
getting
shelled
like
every
day,
and
there
are,
you
know
real
trauma
going
on
physical
trauma
and
my
good
friend
I
was
like
a
best
friend
at
his
wedding,
he's
a
he's,
a
surgeon
and
he's
currently
in
field
teams
that
work
with
like
people
who
are
getting
their
their
stuff
dramatized
so
to
speak,
and
he
reached
out
to
me
and
asked
if
I
can
maybe
try
to
organize
helping
like
a
fundraise
for
like
merchants
and
trauma
kids.
M
So
that's
what
I'm
here
to
do
and
I
have
created
a
amazon
wishlist.
It's
amazon,
deutschland
amazon,
germany,
wishlist
and
I
have
added
like.
Is
this
drama
kids
there
and
we
have
a
family
member
of
his
in
vienna,
who
can
like
accept
it
shipment
from
amazon,
germany
and
then
like
deliver
it
to
kiev.
It's
still
possible
like
it's,
not
surrounded.
M
So
if
you
would
like
to
participate
or
help
me
spread
the
call,
I
would
really
really
appreciate
it.
So
there
is
also
like
alternative
ways
to
help.
I
have
also
edited
to
the
description.
I
will
also
post
this.
All
in
chat,
I'm
really
thankful
and
sorry
for
the
off
topic.
I
have
never
done
this
before
and
you
know,
but
it's
an
emergency
situation
and
I
can't
not
do
it.
M
I
would
be
also
very
like
thankful
for
feedback
about
this
message
about
like
a
better
ways
to
do
it,
because
it's
the
first
time
I'm
attempting
getting
into
kind
of
fundraiser
there
were
like
ideas
floating
around
that
something
like
a
gofundme
or
kickstarter
would
be
a
better
way,
but
I
thought
that
amazon
is
a
good
way,
because
you
see
what
what
you're
like
helping
and
where
it's
going
and
you
can
like
track
it.
So
I
don't
know.
Thank
you,
everyone,
so
yeah.
Thank
you.
A
Thanks
angie,
I
don't
know
if
this
is
a
subject
that
anyone
wants
to
comment
on
I'll.
Just
leave
some
space
for
that
for
a
second
before
jumping
to
the
additional
discussion
section.
A
Okay,
thanks
for
the
links,
please
follow
them.
Folks,
if
you
are.
A
Interested
in
helping
out
and
let
us
move
on
to
the
additional
discussion
section,
I
have
one
question,
though:
do
we
record
additional
discussion
or
do
we
stop
the
recording
here?
I
believe.
Last
week
we
didn't
have
additional
discussion
and
we
got
all
the
way
to
the
end
of
the
call.
So
we
didn't
have
to
sort
that
out.
A
All
right
well,
this
looks
like
any
other
bullet
point
to
me
and
it
could
have
been
a
regular
one.
It's
just
that
it
might
be
unbound.
So
I
propose
we
talk
about
this
until
55
minutes
past
the
hour,
and
then
we
stop
recording
just
in
case
there's
any
other
subject
with.
That
being
said,
I
assumed
this
was
added
by
jeremy,
but
who
added
this
plan
and
who
wants
to
talk
about
it.
E
I
Yeah,
I
was
just
a
little
another
thing.
Another
update
from
last
week
is
due
to
the
the
funding
issues
that
mybind.org
is
experiencing
right
now.
It
means
that
there
is
a
lot
less
capacity
available
on
binder
and
one
of
the
solutions
we
investigated
for
a
while
was
to
switch
some
of
the
demo
environments
to
use
jupiter
light.
I
I
A
One
immediate
piece
of
feedback-
and
I
don't
know
how
to
address
this-
is
if
you
click
on
the
try,
jupiter
lab
and
it
launches
jupiter
light.
The
redirect
is
the
sort
that
eats
up
your
back
button
and
you
can
no
longer
use
your
back
button
if
there's
a
way
to
resolve
that.
That
would
be
really
nice
because
that
could
be
really
frustrating.
A
But
I
think
this
is
really
smart,
because
we
don't
need
compute
power
for
people
to
see
jupiter
lab
in
action
and
to
use
python
and
to
get
it
working
on
their
computer.
So
if
it
turns
out
that
we
were
burning
a
lot
of
money
on
this-
and
this
solves
that
that's
amazing,
that
sounds
really
cool
to
me.
I
Yeah,
so
the
issue
is
tracked
on
the
live
repo,
it's
a
known
issue
and
regarding
the
resources
yeah.
If
you
go
to
the
grafana
dashboards
on
mygrad.org,
it's
pretty
visible
that
I
think
the
number
of
buds
has
been
cut
by
half,
so
you
could
assume
that
the
cost
follow
the
same
trend
and
yeah.
That's
already
again.
Of
course,
there
are
a
couple
of
things
that
don't
work.
Terminals
don't
work,
but
maybe
maybe
there
is
still
room
to
improve
that.
A
Do
you
think
that
jupiter
light
could
have
a
terminal
based
on
the
js
linux
library
by
fabry
spell
art,
because
oh
man
that'd
be
kind
of
amazing
right?
You
have
a
whole
linux
kernel
running
in
javascript
inside
jupiter
lab
running
in
javascript,
with
a
python
kernel
running
in
javascript.
It
would
be
kind
of
crazy,
so.
G
Yeah
I
had
another
question:
is
how
large
does
this
make
the
the
resources
per
hit
to
to
jupiter,
because
if
you're
changing
compute
for
bandwidth,
I
don't?
I
don't
know
how
the
calculation
of
those
resources
compare
like
how
how
large
is
troop
delight
compared
to
just
regular
troops
lab?
I
have
no
idea,
but
I
would
assume
it
was
quite
a
bit
larger.
I
So
for
light,
this
is
like
the
main
cost
comes
from
the
the
the
size
of
the
static
assets,
so
mostly
javascript,
and
then
the
pyrodine
assets
are
coming
from
the
pi
diet
cdn
right
now,
so
this
is
not
a
cost.
That
is
not
something
that
caused
through
the
department.
We
have.
I
One
thing
with
the
binder
setup
was
that
they
had
already
put
a
maximum
number
of
cochlear
users
on
these
two
gimbals
and
that
I
think
it
was
set
to
100.
So
and
very
often
this
one
was
maxed
out,
so
you
will
have
200
pods
200
users
at
a
given
time.
Kind
of
you
know
keeping
the
resources
busy.
I
So
this
this
is
not
the
case
anymore,
so
this
is
free
for
other
binders
to
run
so
for
jupiter
light.
I
think
the
the
main
main
issue
is
to
is
the
bandwidth
for
sure.
So
that's
something
we
discussed
with
chris
graph
on
the
on
the
pull
request
is
that
maybe
we
will
end
up
having
a
cdn
in
front
of
it
and
that
should
already
help
a
lot.
G
That's
I'll
talk
through
the
only
issue
I
could
see,
then,
would
be
if
for
some
reason,
it's
so
large
that
the
users
over
low
speed
internet
can't
access
it
in
a
reasonable
amount
of
time.
So
as
long
as
it's
within
bounds
of
reasons,
it
should
be
sounds
like
the
perfect
solution.
A
If
it
turned
out
that
bandwidth
not
bandwidth,
but
that
just
the
package
size
was
an
issue,
I
wonder
if,
for
example,
an
updated
launcher
that
did
some
sort
of
lazy
loading
or
a
js
kernel
only
version
or
something
like
that
would
be
one
thing
we
could
look
at,
but
I
imagine
jupiter
lab
itself
is
pretty
large
and
adding
the
python
runtime
itself.
A
I
wonder
I
wonder
if
it
like
basically
changes
the
order
of
magnitude
or,
if
it's
roughly
the
same
barrier
for
loading
as
jupiter
lab
itself
was
because
jupiter
lab
is
a
non-trivial
application,
as
it
is.
I
Yeah,
so
loading
paradigm
definitely
adds
a
couple
more
megabytes,
but
they
only
come
when
you
start
to
crowd.
So
when
you
open
jupyter
lab
and
if
you
are
only
interested
in
the
interface
which
might
not
be
the
case
for
the
trial
jupiter
but
still
like,
if
you
only
open
the
page,
then
you
will
only
pull
the
static
javascript
assets
that
are
almost
in
the
same
order
of
magnitude
as.
A
What
are
your
thoughts
on
updating
the
notebook
version
to
be
also
running
the
former
retro
lab
new
pre-release
at
some
point,
probably
not
until
we
do
the
full
release,
but
basically
soon
thereafter,
so
that
that
also
is
running
jupiter
light.
I
I
Maybe
in
a
couple
of
weeks
I
would
say
that,
then
we
need
to
decide
whether
we
want
that
to
also
be
on
the
try
jupiter
we
post
do.
We
want
users.
A
It's
running
retro
lab
right
now,
when
I
click
the
try
notebook
link.
Is
that
right?
Yep?
Oh
okay!
I
see
interesting
so
we're
gonna
change.
We
have
to
change
that
tagline
there,
the
original
web
application
for
creative-
oh
no.
Actually,
if
we're
calling
a
notebook
and
it's
notebook,
maybe
although
I
would
remove
the
python
icon
there,
because
it's
not
really
what
you're
running
the
icons
here,
all
don't
make
sense.
Actually,
because
the
jupiter
lab
one
says
jupiter
the
jupiter
notebook
one
says
python.
I
think
both
of
those
are
in
apt.
K
Well,
I
was
just
about
downloading
python
with
jupiter
lab
what
happens
when
people
are
running,
jupiter
lab
on
a
chromebook
or
other
device
that
doesn't
have
a
python
doesn't
have
really
ability
to
execute
python
right.
Typically,.