►
From YouTube: JupyterLab Team Meeting - November 30, 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/152
A
Thank
you,
Darian
welcome
everybody
to
the
weekly
Jupiter
up
call.
We
don't
have
so
much
point
to
the
agenda,
but
I
will
skip
myself
and
let
Gabriella
start
if
it's
fine
for
her.
C
To
and
they
cover
usability
issues
in
the
current
Jupiter
lab
and
also
do
some
research
on
new
features
that
don't
exist
yet,
but
that
we
all
know
are
coming
soon
and
that
we
would
want
to
know
like
how
to
implement
them
and
where
to
put
the
buttons
Etc
I,
think
I
won't
go
through
the
whole
process
of
it.
C
But
I'll
just
share
a
quick
presentation
of
the
results,
maybe
just
one
or
two
results
with
you,
and
so
that
you
can
see
what
kind
of
things
I
was
testing,
and
so
all
of
all
of
the
results
that
they're
sort
of
easily
fixed
the
issues
that
are
easily
fixed,
I've
created
issues
on
GitHub,
so
that
anyone
can
go
and
have
a
look
and
create
either
discussion
or
actually
tackle
the
suggestion.
I
provided
following
the
issue.
C
C
With
other
sharing
a
notebook
with
somebody
else
being
NC
tags
being
able
to
see
where
other
people
are
currently
getting
the
notebook,
so
those
are
the
things
we
have
in
mind,
but
but
haven't
been
implemented
yet
and
also
in
this
prototype,
I
recreated
things
that
currently
exist
in
Jupiter
lab
that
I
wanted
to
test.
So
I
created
them
exactly
identical
to
how
they
behave
and
look
like
today.
C
So
I'm
gonna
consisted
of
so
discovering
their
usability
issues
in
the
current
features
of
Jupiter
lab
in
a
very
basic
workflow
to
test
the
new,
so
I
will
never
features
so
see,
get
a
few
insights
about
the
toolbar.
What
actions
are
used
or
Not
by
certain
types
of
users?
C
How
do
users
use
their
tag
management
feature
and
how
do
they
add
cells
when
they
need
to
add
a
cell
and
also
to
test
new
sharing
features
so
we're
seeing
who
is
editing
what
so?
It's
happened
with.
C
Yesterday,
in
the
community
call
also,
you
can
go
and
have
a
look
at
the
details
of
how
the
study
plan
was
created,
but
the
results
and
the
bonus
that
I
did
the
test
with.
So
there
were
beginner
users
and
experienced
users,
the
beginner
users,
so
they
remain
in
between
they
were
between
18
and
25.
They
have
a
frequent
use
of
Jupiter
lab,
but
they've
only
been
using
it
for
one
to
two
months:
maximum,
whereas
the
experienced
users
were
over
40
and
had
been
using
it
for
approximately
10
years
or
more.
C
The
I
had
seven
participants
that
represented
the
beginning
user,
so
that
was
largely
enough
to
uncover
most
of
the
usability
issues,
but
for
the
experience
user
I
only
had
two.
So
most
of
the
results
that
follow
mainly
concern
the
beginning.
C
So
results
Family,
Guy,
two
columns
inverse
of
files,
so
that
doesn't
exist
today
to
see
I
was
wondering
where
how
can
we
inform
users
with
a
document
or
a
notebook
has
been
shared
with
them,
and
how
can
we
diff
make
a
difference
with
the
other
files?
So
the
first
idea
was
to
create
two
tabs
and
a
user.
That's
nose,
they're
gonna,
open
a
shed
file,
go
to
the
web
shared
with
me,
so
this
one
example
doesn't
work,
and
so
at
least
now
I
can
iterate
and
it
works.
C
This
for
was
never
understood
by
the
users.
For
various
reasons,
then
there's
were
results
about
being
able
to
see
where
people
are
in
the
different
cells.
She
was
working
being
able
to
display
tags,
but
in
a
better
way
than
it
is
done
today
being
able
to
add
new
cells,
because
I'm
noticed
that
in
the
usability
is
when
I
asked,
if
participants
to
add
a
new
cell
or
they
had
to
do
it
within
another
task,
they
never
ever
use
the
toolbar
and
I.
C
Think
one
of
the
reasons
that
came
out
is
because
most
of
the
code
is
normally
on
the
left
and
it's
actually
very
cumbersome
to
always
be
going
to
the
far
right
of
the
cell,
especially
when
you're
on
a
big
screen
and
another
result
is,
was
that
the
RAW
button
is
useless
and
nobody
uses
it.
So
these
are
some
examples
of
results.
I
got
from
the
usability
tests,
I
tested
the
tags.
C
So,
basically
from
here,
there
are
two
kinds
of
actions
I
took.
One
is
listing
some
really
easy
usability
issues
that
exist
now
in
Jupiter
lab
and
can
be
like
easily
corrected,
I.
Think,
for
example,
these
two
hours
to
move
a
cell
up
or
down
they
don't
have
a
disabled
state.
So
when
the
cell
is
all
at
the
bottom
or
all
at
the
top.
Well,
it
looks
like
it's
clickable
when
it
actually
isn't,
and
it's
not
it.
C
It
can
seem
obvious
because,
of
course
when
the
cell
is
at
the
bottom,
you
can't
push
it
further
to
the
bottom,
but
it's
still
a
cognitive,
something
that
takes
a
tiny
bit
of
your
cognitive
space.
Just
of
the
add
a
disabled
space
to
disabled
state
to
the
button,
and
so
those
are
very
little
things
that
I
tried
to
cut
down
to
the
bare
minimum
and
I
created
issues
for
each
of
them
and
then
for
the
others.
Some
that
are
big,
bigger
pieces.
C
A
B
C
To
be
honest,
I
don't
know,
but
in
my
prototype
I
mean
if
you
I,
didn't
use,
I
didn't
use
that
any
version
of
Jupiter
lab.
If
that's
what
your
question,
it
was
a
prototype
that
I
created
with
only
the
features
that
I
wanted
to
test.
C
C
D
A
A
B
C
Oh,
do
you
mean
like
on
the
real
Jupiter
lab
I?
Think
that's
what
request
mean
means
like
do
I
plan
to
well
yes
for
the
current
for
the
existing
features?
Yes,
but
for
those
that
don't
exist
well,
it's
not
possible,
but
for
maybe
I
can
show
you
quickly
what
it
looks
like
like
the
product
that
really
because
I'm
testing,
quite
a
basic
workflow
I,
asked
people
to
open
them,
notebooks
it
with
someone.
C
Imagine
they
are
working
in
a
team
and
they
would
like
to
see
who
are
they
I?
Don't
need
to
have
all
the
detailed
workflows
that
Jupiter
lab
provides
and.
B
C
C
B
C
C
And
yeah
getting
statistics
from
Jupiter
lab
would
be
great,
but
we
would
have
to
like.
We
could
think
about
what
we
want
to
get
statistics
about
and
measure
them,
but
it
would
be.
It
could
be
really
really
useful
to
have
this,
for
example
on
the
use
of
the
toolbar
and
if
there
are
features
that
are
actually
never
used
or
very
poorly
used
and
maybe
try
and
understand
why,
and
do
some
research
around
that.
C
Yeah
well
the
thing
about
the
keyboard
shortcuts
some
well
I
actually
can
reproduce
that
also
in
the
project
site.
But
what
happens
a
lot
is
it's
experienced
uses
it's
very
verbal.
The
the
usability
test,
so
at
the
beginning
of
the
test,
I
asked
the
participants
to
try
and
think
out
aloud
as
much
as
possible
and
so
a
lot
everyone
sort
of
forgets
that
it's
not
really
Jupiter
lab.
They
look
they're
working
with,
but
the
Prototype
so
very
quickly.
They
go
back
to
their
mechanism
and
we'll.
C
If
I
ask,
for
example,
to
add
a
cell,
a
lot
will
intuitively
use
their
keyboard
straight
away.
To
do
it,
and
even
if
it
doesn't
work
they'll
just
say:
oh
yeah,
I
well
I.
To
do
what
you
asked
me
to
do.
I
would
have
used
my
keyboard
and
I
know
because
I
just
saw
them
do
it,
and
so
there
are
some
limitations,
but
the
way
prototypes
are
created
now
is
actually
quite
powerful.
C
A
A
It
seems
it's
only
me,
so
it's
gonna
be
a
short
update.
So
last
week
we
discussed
the
introduction
of
the
enhancement
features
in
Jupiter
lab.
There
were
some
follow-up
tasks
that
were
decided
so
I
have
opened
an
issue
in
the
security
team
repository
for
asking
them,
but
having
their
opinion
and
review
of
the
features,
I
also
have
open
a
PR
I
did
not
go
into
the
path
of
using
an
iframe,
because
at
first
I
was
thinking
like
should
I
iframe
every
notification,
it
seems
a
bit
too
much
so
I
was
thinking.
A
B
A
So
for
now
what
I
decided
to
do
is
to
to
use
a
stricter
option
of
the
sanitizer,
so
the
the
pr
is
linked
in
the
note
and
so
for
for
people
interested
feel
free
to
to
comment
there
and
propose
a
succession
of
doing
it
differently
and
I.
Have
one
question
is
okay,
maybe
before
going
to
that
point,
maybe
Json
as
your
raise
your
hand.
G
Yeah
I
have
a
quick
thought
about
that.
I
I
like
how
you
really
tried
to
Pare
down
the
options.
G
I
would
suggest
we
even
go
further
and
just
say
like
no
formatting,
just
plain
text
messages
to
start
with,
and
a
single
URL
a
user
can
click
on
and
and
the
nice
thing
about,
having
just
plain
text
say
with
a
maximum
size
of
I,
don't
know
whatever
Twitter
or
message
on
or
something
you
know,
400
characters
or
something
is
it
forces
and
incur
it
encourages
by
forcing
messages
to
be
very
non-distractic
with
just
you
know,
a
short
sentence
or
two
and
a
single
read
more.
G
B
A
It's
because,
in
the
the
implementation
of
the
notification,
I
said
that
you
can
use
markdown.
If
you
want
mainly,
the
idea
was
to
be
able
to
easily
use,
add
links,
but
yeah
I,
don't
know
it's
probably
too
noisy
for
a
notification
to
have
a
too
much
formatting
in
the
in
the
notification
text.
G
I
would
suggest
we
start
simple
and
start
as
absolutely
simple
as
we
can
and
then
and
then
we
can
always
go
up
from
there.
But
it's
hard
to
like
pull
back
down
features
once
they're
launched.
B
A
H
What
I
was
going
to
say
when,
when
I
last
looked
into
some
of
these
things,
I
don't
know
I
got
into
using
open
graph
and
metadata
and
stuff
to
kind
of
make
it
richer
things
that
you
showed
to
convey
more
information
and
less
space?
You
know
like
a
thumbnail
and,
and
things
like
that,
but
yeah
I'm
not
sure.
H
So
you
want
to
store
their
state
client
side,
but
they
might
throw
out
client-side
State
sorry
server
side,
rather
that
they
might
throw
out
server-side
State
all
the
time
because
they're
using
like
condom
environments
or
they're
clearing
out
you
know
like
it,
it's
just
it's
it's
a
much
tougher
problem
than
the
problem
that,
like
vs
code,
has
because
vs
code
kind
of
owns
its
entire
space
and
those
users
never
going
to
clear
it.
It
knows
that
it
can
store
some
like
last
red
value
and
it's
always
correct.
A
Actually,
for
now,
I
use
a
trick
of
not
using
the
state
I'm
using
the
config
from
the
server
API
directly,
because
then
it's
single
placed
it's
independent.
On
the
on
the
workspace
and
anytime
you
re-run
the
server
or
whatever
from
another
environment.
Then
it's
going
to
be
the
same
value.
That's
right
from
there.
A
A
And
I
I
plan
also
to
because
it's
related,
probably
to
other
options,
to
to
create
a
specific
section
of
four
binder
in
the
documentation
for
providing
like
a
good
default
settings
that,
like,
for
instance,
that
turn
off
the
the
notification
on
binder,
because
that's
probably
something
also
people
don't
want
to
see
like
every
time
they
they
launch
a
new
binder
for
testing,
something
they
don't
want
to
get
like.
First
time,
pop-ups.
F
Can
I
suggest
whatever
I
first
of
all,
I
tend
to
agree
with
Jason?
It's
always
better.
Just
start
simpler,
but
can
I
suggest
that
when
you,
then,
when
you
make
a
change
like
that,
you
also
put
a
new
notification
to
the
best
practices
for
the
extension
guide,
because,
ideally
they're
going
to
a
lot,
be
lots
of
people
writing.
A
B
A
E
Thanks
Fred,
so
I
just
want
to
give
a
heads
up
thanks,
Fred
and
Jeremy
for
handling
the
Fallout
from
this,
but
we
we
extracted
what
were
the
fixtures
in
Jupiter
server.
We
had
this
Pi
test,
plug-in
module
and
Jupiter
server,
where
you
could
use
it
by
installing
super
server
with
the
extra
test
and
then
get
some
nice
fixtures.
E
We
extracted
that
out
into
a
separate
package,
called
Pi
test,
Jupiter,
which
we
had
tried
to
make
like
a
year
ago
and
decided
to
back
out
of
it,
because
the
the
way
plug-ins
work
in
pi
test
is
like
they're
automatically
used.
If
you
use
entry
point
system
and
we're
like
clubbing
other
extensions.
So
now
it's
an
opt-in
I
try
to
try
to
overcome
some
of
that,
and
it
has
like
basically
three
tiers
like
if
you
just
installed
by
that's
Jupiter.
E
You
get
some
like
environment
overrides
to
make
sure
you're
running
things
in
a
clean
environment,
overriding
all
like
the
Jupiter
paths.
Things
like
that,
you
can
select
the
client
add-in
to
get
an
echo
kernel
and
a
fixture
to
be
able
to
run
Echo
kernels
and
then
the
server
plugin
has
everything
that
was
input
in
in
the
pythus
plugin
on
the
Jupiter
server.
E
So
another
recommendation
is,
if
you're
previously
relying
on
what
was
in
Jupiter
server,
podcast
plugins
to
switch
over
to
pythus
Jupiter
server,
and
if
you
work,
because
you
didn't
want
to
take
a
dependency
on
that
or
whatever
or
just
to
advertise
the
fact
that
those
environment
fixtures
are
now
available
in
a
self-contained
thing.
That's
not
tied
to
a
version
of
Jupiter
server.
You
don't
have
to
depend
on
the
server
for
your
tests
and
that
there's
a
condo
recipe
pending
on
that
swing
for
the
materials
to
merge.
E
That
and
that's
it
in
that
front.
I
I
Hi
I
have
a
very
brief
Point,
just
highlighting
the
pull
request,
which
originally
was
only
intended
to
fix
an
issue
with
code
mirror
6
migration
with
competition,
completion
of
tokens
not
working
but
then
had
to
be
had
to
have
the
scope
expanded
to
change
some
of
the
apis
and
that
most
notable
change
is
replacing
the
connector
proxy
with
a
new
kind
of
object,
which
is
reconciliator
so
with
the
new
completed
refactor
can
have
multiple
providers
of
completions,
so
we
can
have
a
kite
extension
and
a
tab,
9
extension,
an
LSP
server
and
whatever,
and
then
we
want
to
combine
the
completions
from
all
these
sensors.
I
K
A
And
so,
if
nobody
has
additional
points
on
record,
maybe
we
can
stop
the
record
or
if
somebody
has
some
discussion,
but
it
seems
not
so
so.
Darian
is
gonna,
get
the
recording.