►
From YouTube: JupyterLab Team Meeting - April 6, 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
april
6th
jupiter
lab
weekly
call.
If
you
would
like
to
speak
about
something
in
today's
call,
please
add
an
item
to
the
agenda.
A
If
you
have
something
that
will
require
more
conversation,
please
add
it
to
the
additional
discussions.
So
if
you
think
you'll
need
more
than
say
five
minutes,
please
add
it
to
the
other
section
and
let's
get
started.
My
name
is
first.
I
just
wanted
to
announce
to
everyone
that
I
have
a
new
job.
A
A
Hopefully,
we'll
be
able
to
do
a
bit
more
work
on
jupiter
than
I
could
before,
but
but
there
shouldn't
be
that
much
difference
in
your
interaction
with
me,
vis-a-vis
jupiter,
because
that's
what
I'm
there
to
do.
I
want
to
take
the
chance
to
thank
two
sigma,
my
former
employer
of
four
years,
for
allowing
me
the
latitude
to
spend
so
much
of
my
time
on
jupiter
and
things
that
I
care
about
and
yeah
nothing
but
good
things
to
say
about
them
and
that's
it.
B
B
B
And
thank
you
for
for
on
yours
from
me
for
about
you
for
reviewing
some
pr's.
So
this
is
the
least
out
of
the
ones
that
are
still
open.
There
are
the
two
that
have
been
reviewed
by
alex
this
one.
The
bot
code
mirror
go
to
line
and
this
one
about
search
and
replace
commands.
B
So
they
are
good
to
go
so
they
would
be
merged
soonish,
let's
say,
and
I
think
the
two
that
are
pending
no,
there
are
maybe
three
that
are
pending
work,
there's
the
one
of
json
on
the
side
toolbar.
I
don't
know
how
much
you
imagine.
You
still
need
to
work
on
except
feedbacks.
E
I
actually
listed
that
under
my
agenda.
I
got
some
great
feedback
from
you
frederick.
Thank
you
so
much
for
that
incorporated
your
textual
and
code
suggestions.
So
I
am,
I
am
looking
for
when
I
say
I'm
looking
for
feedback,
really
I'm
looking
for
an
approval
on
it.
I
think
this
is
just
about
ready
to
go,
but
I
don't
want
to
tip
the
scales
too
much.
Please
take
a
look
at
that.
B
B
F
So
I'm
here
yeah
I
I
will
be
trying
to
get
this
ready
soon.
F
B
So,
but
maybe
what
we
can
do
is
we
we
could
start
list,
let's
say
next
wednesday
or
doing
a
beta
release
and
and
and
so
that
we
do
a
candidate
last
week
of
april.
Does
that
seems
possible
for
you
mike
because
it
will
be
the
only.
B
Okay,
so
I
I
propose
that,
as
a
plan
like,
I
think
everybody
every
every
pr
here
will
be
ready
except
them.
One
of
mike
so
plan
would
be
to
to
do
a
better
release
in
a
week
and
to
do
a
candidate
release.
E
B
The
idea
was
to
back
for
them,
but
there
are,
as
there
are,
no
three
dot
x
branch,
yet
neither
so,
but
I
could
I
could
decide
or
we
could
decide
of
using
them
for
another
meaning.
It's
it's
same
for
me,.
B
Nitpicking
so
okay,
so
if
everybody
agrees,
then
that
will
be
the
plan
and
I
will
take
care
of
them
on
of
creating
the
branch.
I
will
wait
for
for
creating
the
branch.
B
I
will
wait
a
week,
so
I
will
create
the
branch
just
before
starting
the
release
to
avoid
the
need
for
of
new
blackboard
to
3.3.x
and
3.4.x,
because
I
know
that
that's
the
point
of
eric
that
cannot
represent
it,
but
he
plans
to
do
some
batteries
for
3.3.x
tomorrow.
So
in
case
there
are
some
still
in
the
coming
days.
Then
we
don't
need
to
back
for
them
two
minutes
too
much.
B
Okay,
thanks
people,
then
I
have
a
couple
of
other
points.
This
jeremy,
there,
I
don't
know.
I
think
I
have
a.
We-
have
some
error
on
the
ci
by
time
to
time,
so
I
have
pusher
pr
for
being
able
to
to
customize
the
toolbar
of
the
file
browser
mainly
to
to
be
able
to
reuse
that
features
on
the
classical
notebook
so
on
the
notebook
before
the
trouble
with
that
is.
B
The
plugin
in
which
the
toolbar
is
defined
is
not
the
same
as
the
one
in
which
the
commands
of
the
file
browser
are
defined,
and
so
we
are
facing
a
risk
condition
and
if
the
the
the
five
browser
is
added
to
the
shell
before
the
comments
are
defined,
then
it's
failing
to
find
the
icon
link
to
a
link
to
the
comment,
and
so
the
buttons
are
displayed
with
a
text
and
not
with
the
icon.
B
I
have
seen
that
only
on
the
ci,
but
I
presume
it
may
happen
also
on
the
for
the
user,
and
I
don't
really
know.
A
B
A
A
A
B
Okay,
then
I
will,
I
will
propose
a
pr
on
the
with
that
solution
and
the
last
one
forget,
oh
yeah.
I
think
it's
it
was
more
for
not
forgetting
to
bring
the
subject,
because
I
think
there
hasn't
been
much
discussion
on
the
issue.
B
The
question
was
a
good
place
where
to
to
move
that
that
exploratory
repository,
knowing
that
brian
commented
later,
that
there
are
some
restrictions
coming
from
amazon
on
which
kind
of
organization
is
needed,
and
so
I
think
it
would
be
nice
if,
like
some
some
other
people
want
to
comment
of.
B
What's
the
what's
the
status,
because
I
know
in
the
past
we
have
with,
we
have
had
a
couple
of
extensions
directly
in
jupiter
lab,
but
the
organization,
but
I
don't
know
if
we
had
some
early
like
experimental
repository,
although
there
are
some
kind
of
things
like
monaco
editor
thinking
out
loud
sorry,
so
yeah
don't
hesitate
to
comment.
The
the
idea
really
of
I
think
and
the
question
from
push
and
brian
is,
is
jupiter
organization
the
good
place
for
for
creating
that
new
repository,
and
they
just
want
to
agreement
on
that.
H
Yeah,
so
I
I
did
see
brian's
comment.
I
actually
was
not
aware
that
aws
has
this
restriction,
so
brian
probably
knows
more
about
this,
but
yeah.
If,
if
anybody
had
any
feedback
on
that,
how
do
we
move
forward
for
this?
You
know
feel
free
to
comment
on
this
issue.
A
Will
it
continue
to
exist
after
data
registry,
or
is
this
really
just
the
next
version
like?
Could
we
simply
rename
data
except
yeah.
H
This
is
the
data
registry
yeah.
This
is
the
next
version,
and
I
think
that
was
the
idea,
because
there's
there's
really
not
much
from
the
earlier
code
base
that
we're
gonna
incorporate
in
this
version.
So
if
renaming
is
the
solution,
then
we
can.
We
can
definitely
work
with
that.
A
D
No,
I
was
just
wondering
what's
stopping
us
from
just
making
another
repo
with
the
new
name
with
the
new
content
and
just
letting
the
old
one
exist
where
it
is
because
it's
this
it
because
it
sounds
like
this
is
a
spiritual
successor
more
than
like
a
direct
code
successor.
So
is
there
a
reason
it
can't
just
be
in
its
own
repo.
A
D
Yeah,
I
just
maybe
it's
the
the
get
ocd
in
me,
but
the
idea
of
completely
replacing
a
code
base
with
new
code
as
part
of
the
same
repo
that
just
I
guess
it
irks
me
in
the
wrong
way
from
a
get
cleanliness
perspective.
When
you
essentially
just
like,
say:
oh
yeah,
it's
the
same
repo
but
we're
completely
replacing
the
code,
which
is
understandable
that
we
need
to
completely
replace
the
code
with
this
new
version.
But
that
feels
like
it
shouldn't
necessarily
be
a
continuation
of
the
current
git
refill
from
a
git
perspective.
A
That's
interesting
because
I
I
think
of
git
almost
exactly
the
opposite,
as
I
want
the
whole
history
of
a
thing.
Even
if
one
day
you
decided
to
rewrite
the
extension
of
fortran,
I
still
want
to
see
the
original
javascript
version
before
the
new
fortran
version.
So
anyway,
that's
a
philosophical
debate.
We
don't
know.
D
That's
a
that's
a
good
counterpoint
to
that
so
yeah.
As
I
said,
no
strong
opinions
just
open
question
on
why
we
were
specifically
tunneling
on
just
doing
a
full
code
complete
like
white.
D
Oh
the
issues,
the
issue
tracking;
no,
that's
actually
vidar
just
posted
in
the
chat
about
it's
also
for
issue
tracking.
That's
actually
a
really
good
reason
to
just
keep
the
repo
and
potentially
rename
it
so
not
thinking
of
it
as
git,
but
thinking
of
it
as
github.
A
H
Yeah,
I'm
not
blocked
at
this
point,
but
I
think
to
to
move
forward
with
this
question.
I
think
I
think
people
are
suggesting
this
rename
right.
That's
what
you're
suggesting
so
is
that
a
rename
is
there
a
button
somewhere
in
the
github
that
I
should
go
and
just
screening
or
how?
How
does
that.
A
H
H
A
A
D
Yes,
I
have
a
quick
demo
slash
pitch.
This
is
a
little
bit
of
code
that
I
wrote
in
my
work
with
alyra
and
I
wanted
to
kind
of
share
with
everybody
and
see
how
people
feel
about
this
feature
and
see
if
it's
something
that
I
might
want
to
move
into
core,
it's
a
standalone
extension
right
now,
I've
nicknamed
code
viewer.
D
We
have
released
it
as
part
of
the
lyra
as
of
last
week,
but
if
we
moved
it
into
core,
it
would
probably
just
be
a
command
that
gets
bundled
into
something
like
the
file
editor.
So
right
now
you
can't
open
the
editor
for
files
that
don't
exist.
So,
for
instance,
you
have
a
chunk
of
code
and
you
want
to
display
it
to
the
user.
You
would
have
to
persist
it
to
open
it
in
the
file
editor,
so
I
created
a
way
to
open
an
editor.
That's
in
this
case
read
only
that
is
transient.
D
It's
not
persisted
anywhere
on
the
system.
So
in
this
case
this
is
a
definition
yaml
file
for
that
node
in
our
pipeline.
The
file
gets
returned
by
the
server,
but
we
didn't
want
to
persist
it.
So
we
want
to
be
able
to
display
it.
This
is
done
by
using
a
command
that
that
we
register
and
call
that
just
opens
up
the
code
viewer.
The
code
viewer
wraps
our
code
editor
instead
of
the
file
editor,
since
the
file
editor
itself
requires
a
file
in
the
file
system.
D
So
it's
a
cool
way.
It's
also
transient.
It's
extremely
transient,
because
if
you
open
it
again,
it
opens
a
second
copy,
like
it
doesn't,
have
a
persisted
id
of
any
sort.
I
if
I
remember
my
implementation
correctly,
I
actually
randomized
the
id
for
the
widget
and
it's
transient
like
when
you
go
when
it
goes
away.
It's
gone,
so
I
was
wanted
to
show
this
to
everybody
and
see
if
people
thought
that
this
was
a
feature
that
would
be
worth
me
putting
in
core
before
I
do.
D
The
legwork
actually
put
it
in
core
and
if
people
think
it
should
go
into
core.
Where
should
I
put
it
into
core?
Because
I
have
the
widget
file
and
then
I
also
have
the
command
that
has
to
be
added
in
an
index
files
somewhere.
A
D
Sure
it
will
go
away.
Let
me
just
double
check
that.
So
at
one
point
I
actually
called
it
transient
viewer
specifically,
because
the
idea
is
it's
transient.
D
It's
it's
meant
to
be
used
to
display
something,
so
you
don't
have
to
display
it
in
like
a
pop-up
with
some
leg
work
we
could
add
it
to
the
restorer
and
like
do
the
work,
to
get
it
to
get
a
like
a
persistent
widget
id,
but
that
would
most
likely
require
some
sort
of
hashing
or
possibly
similar
to
launcher
a
way
of
just
an
incremental
id.
D
So
if
restore
is
something
that
people
think
would
be
something
we
want,
if
we
added
this
decor,
I
could
do
the
legwork
to
add
that,
but
it
was
not
something
that
we
needed
in
our
initial
implementation,
so
it
doesn't
currently
do
it.
J
D
They
actually
work
completely
separate
to
each
other,
but
we
could
add
some
sort
of
menu
bar
option
or
menu
option.
That
says
then
save
as
and
it
could
save
the
file.
But
as
of
now
it's
completely
separate
from
files
like.
I
essentially
looked
at
our
how
the
file
editors
implemented-
and
I
just
went
up
the
class
extension
change
until
it
no
longer
required
a
file
being
persisted.
D
Oh,
like
refreshing
screen,
I
did
refresh
it
goes
away.
A
Yeah
isabella
had
a
question
in
the
chat.
Could
I
ask
alex
to
play
the
other
side
and
say,
mr
bella,
you
can
ask
your
question
to
say
why
we
wouldn't
want
this.
Go
ahead.
C
Yeah,
I
was
just
gonna
say
it.
It
may
be
because,
where
I'm
coming
from,
but
like
a
lot
of
the
time,
I'm
like
okay,
what
are
the
cons
of
having
a
new
feature
like
I'm?
Not
really
sure
you
know.
Is
there
a
maintenance.
D
In
my
mind,
are
almost
always
bloat
is
this
something
that
needs
to
be
in
core,
or
should
it
be
a
standalone
third-party
extension
that
somebody
can
install
which
it
is
currently,
even
though
I
said
this
is
part
of
a
lyra,
we
do
publish
this
as
standalone.
You
could
just
do
a
pip
install
a
lyra
dash
code,
dash
viewer
and
it
would
pip
install
this,
and
the
command
would
be
available
to
you.
C
A
D
I
don't
believe
so.
I
I'd
call
this
a
100
additive.
I
don't
think
it
would.
I
don't
think
it
would
make
sense
to
replace
anything
we
currently
have,
though,
if
we
decide
to
extend
this
with
some
of
those
functionalities
that
people
pitched
those
functionalities
may
end
up
being
implemented.
D
That
way
like,
for
instance,
being
able
to
like
save
it,
save
ads
into
the
file
system,
how
we
handle
that
and
how
we
handle
if
they
do
a
save,
as
it
would
probably
need
to
be
reopened
as
a
file
editor,
because,
right
now
you
can't
actually
edit
this
I'm
typing
on
my
keyboard.
D
So
and
I've
had
to
disable
the
right-click,
the
built-in
right-click,
because
the
right-click
functionality
of
the
editor
is
tied
to
the
file
system.
So,
for
instance,
if
you
had
a
markdown
editor
open
in
the
background,
it
would
give
you
the
option
to
open
up
the
preview
for
that
markdown
editor,
because
again
this
is
transient,
and
so
it
doesn't
have
a
file
persisted
and
we've
actually
tied
in
the
context
menu
of
file,
editor
of
safari
of
the
editor
directly
into
the
directory
structure.
Again.
D
A
D
A
A
So
it
isn't
exactly
analogous
to
the
situation
that
the
code
viewer
solves
it's
a
little
bit
different,
but
it
was,
I
mean
I
thought
of
it
as
well.
It
does
seem
it's
in
the
same
universe,
but
it
is
slightly
different
in
that
they
need
to
have
a
file.
However,
those
files
are
not
always
accessible
via
the
file
contents.
Api
they're
coming
from
somewhere
else.
D
Yeah
then
that
yeah,
it
sounds
like
that
was
a
essentially
a
different
solution
to
the
same
problem.
In
fact,
we
could
have
probably
done
that
to
solve
our
problem
in
a
lyra
rather
than
writing
this,
but
I
do
believe
they
do.
They
can
fit
different
use
cases,
even
though
our
ally
problem
actually
can't
be
involved
with
that
solution.
Instead,.
D
That
was
actually
one
of
my
questions
is
whether
or
not
it
should
be
brought
in
as
a
package
or
if
this
should
just
be
in
included
an
extension
we
already
have,
because
it
is
just
a
widget
definition
and
a
command
and
and
then
defining
a
command
in
an
index.
So
in
theory
this
doesn't
have
to
actually
be
its
own.
Like
its
own
package
in
court,
it
could
just
be
put
into
a
package
that
already
exists
as
a
command.
D
Yeah,
the
fi,
the
file
editor
package
was
my
best
guess,
but
I
just
didn't
know
because
the
the
actual
widget
itself,
I
think,
belongs
in
the
code,
the
code
editor
package,
but
that
doesn't
have
an
associated
widget.
So
I
was
unsure
of
which
extension
index
file
we'd
want
to
add
the
creation
of
the
command
to
if
it,
if
that
belongs
in
the
file
editor.
Given
it
it's
loosely
related.
D
A
So
if
you
took
jason's
approach
and
added
another
exported
class
to
the
file
editor
package,
would
this
be
simply
a
regular
pr
to
core,
or
would
it
be
something
like
you're
bringing
along
if
yeah
just
be
a
regular
in
that
case?
In
that
case,
that's
not
a
team
compass
issue.
That's
more
like
just
create
a
jupiter
lab
issue,
saying
hey!
I
want
to
create
this
kind
of
thing.
What
do
you
guys
think
and
if
you
get
enough
thumbs
ups
on
it,
then
open
the
pr
for
it
cool.
D
Then
that's
what
I'll
do
and
I'll
retroactively
put
that
in
the
agenda
for
people?
That's
fine!
Thanks
for
the
feedback.
A
Alex
jason
w.
E
Okay,
well,
the
topics
that
I
put
in
the
agenda
it
seems
like
have
already
been
addressed.
3.3.3
and
3.4.0
release
date
expectations.
I
kind
of
summarized
what
was
said
earlier.
So
please
correct
me
if
I'm
wrong
that
3.3.3
is
expected
tomorrow,
the
7th
3.4.0
early
may
and
then
on
the
cell
toolbar
stuff.
I
got
some
great
feedback
from
from
frederick
and
some
some
code
samples
that
I
incorporated
with
some
other
changes.
E
So
there
are
a
couple
of
automated
tests
that
are
failing.
I
am
not
100
sure
whether
my
code
is
related
to
that.
So
that's
like
the
only
other
area
of
uncertainty,
but
if
anyone
else
could
take
a
look
at
that
pull
request,
I'd
appreciate
it.
C
Cool
I
similarly
just
have
two
kind
of
announcy
things.
One
a
couple
weeks
ago
I
brought
up
that
I
I
was
trying
to
solve
an
issue
for
jupiter
lab
citations
and
we
were
like
we
don't
have
a
process
open
something
on
the
jupiter
lab
team
compass.
So
I
did.
There
is
now
an
issue
on
the
jupiter
lab
team
compass.
C
I
moved
a
lot
of
the
notes
that
we
took
from
there
about
like
options
that
we
could
take,
I'm
not
sure
how
to
move
it
forward,
because
this
is
not
the
kind
of
thing
that
I
should
be
making
singular
decisions
on.
So
I'm
letting
you
all
know
it's
not
urgent,
but
it
would
be
nice
to
have
it.
So
we
can
close
the
existing
issues
and
any
future
issues,
because
I
know
this
has
come
up
a
few
times
like
people
post
this
issue
periodically,
so
it'd
be
nice
to
answer
them.
C
Yes,
so
that's
happening,
and
what
was
my
other
thing
in
here?
Oh,
we
have
our
accessibility
meeting
15
minutes
after
this
one
today
hooray
feel
free
to
join
in.
So
I
think,
that's
all
from
me.
Unless
anyone
has
questions.
A
Cool
thank
you.
We've
reached
the
end
of
the
agenda
before
we
jump
into
additional
discussions.
If
there
was
anybody
who
wanted
to
say
something
and
did
not
update
the
agenda,
now
is
your
moment
to
shine.
A
I
F
So,
at
each
release,
whenever
these
get
stacked,
there
is
a
hook
which
creates
a
new
identifier
for
each
release.
So
it's
it's
integrating
with
github
very
well,
but
I'm
not
currently
sure
how
the
autos
are
added.
F
C
Sometimes
people
like
to
do
either
additional
things
or
a
more
vague
like
something
I
see
a
lot
is
like
project
jupiter,
like
I
think
our
mit
license
has
like
project
or
like
jupiter
lab
contributors
as
like
a
title
instead
of
like
listing
people,
I
don't
know
if
that's
of
interest-
I
don't
know
if
that's
possible,
but
that's
something
to
consider.
I
can
go
write
that
on
the
issue
too,
I
didn't
get
to
it
yet.
F
C
Yeah,
I
agree,
my
main
thought
is:
if
we
were
to
do
it
manually,
my
guess
would
be.
We
would
use
the
same
language
in
the
license
that
just
says
generally
and
doesn't
list
specific
people.
That
would
be
my
instinct
for
from
a
maintenance
perspective
as
well.
I
I
like
the
idea
of
a
doi
per
release,
plus
an
overall
doi,
which
is,
I
think,
what's
provided
by
the
system,
because
then
you
can
cite
a
specific
release
that
you're
using
or
you
can
cite
the
project
that
large,
depending
on
sort
of
what
you're
trying
to
say
that
does
dilute
the
citation
rank.
So
yeah,
that's
a
disadvantage
of
that,
but
it
does
help
things
be
more
precise.
I
I
Just
I'm
can
I
just
call
for
comment
because
there's
several
people
here
that
have
been
professors
and
in
the
academic
community
in
the
academic
communities
who,
I
think
of
when
I
think
of
people
doing
citations.
What's
what's
what's
your
take
on
people
that
have
done
a
lot
of
citations
and
done
a
lot
of
papers.
K
Just
say
that
this
is
a
no-doubt
thing
with
the
pillow
example
looks
really
impressive
and
cool,
and
I'm
I
mean
yeah
for
the
reasons
you
stated:
it's
really
cool
that
it
shows
you
specific
authors
for
release
and
then
all
of
the
authors
and
has
the
ability
to
cite
either
a
specific
release
or
all
releases.
A
There
are
two
questions
here.
I
would
like
to
answer
them
backwards,
because
one
of
them
has
an
answer.
A
What
is
the
current
status
of
uploading
recorded
that
meetings
to
youtube
the
status
is
so
dumb
that
I
feel
embarrassed
to
say
it
on
a
recorded
call,
but
I
know
that
recorded
call
is
not
going
to
be
on
youtube,
so
don't
worry
about
it,
but
no.
It
will
eventually
be
on
youtube.
A
I
tried
to
sign
in
to
zoom
and
I'm
signed
into
zoom
right
now
as
project
jupiter,
but
the
videos
are
uploaded
to
the
cloud
zooms
cloud,
so
I
need
to
sign
into
the
zoom
web
ui.
The
zoom
web
ui
sent
a
two-factor
authentication
to
a
project,
jupiter
email
whose
credentials
I
don't
have
and
everyone
that
I've
asked
also
doesn't
have
it.
So
I'm
I
just
right
now
I'm
sending
a
message
to
the
steering
council
saying:
whoever
has
it
please
add
it
to
the
password
vault
and
then
this
problem
will
go
away.
A
B
So
as
talk
about
last
week,
I
think
the
best
was
to
open
some
discussion.
I
created
the
issue,
but
so
far
nobody
has
commented
on
that.
So
I
don't
know
either
we
can
do
some
kind
of
question
answer
session
about
that.
I
don't
know
what's
the
best
path
for
what,
because
I
think,
like
the
first
step
on
that
subject,
is
for
people
to
understand
to
get
their
own
understanding
on
the
technology
and
the
implication
for
the
project,
so
yeah.
A
A
I
I
think
the
main
thing
that
it
comes
down
to
is:
do
we
have
some
sort
of
consensus
on
whether
to
go
ahead
and
adopt
this
as
the
official
ui
toolkit,
we
want
users
extension
authors
to
use
basically,
and
if
that's
the
case,
do
we
want
to
start
grafting
it
into
the
current
ui,
like
bringing
those
things
forward,
swapping
components
that
kind
of
thing
it's
a
big
step.
It's
a
big
change
and
it's
a
new
dependency.
A
All
of
those
things
mean
we
need
to
have
discussion
about
it,
but
we
haven't
actually
discussed
that
much
despite
this
being
the
third
time
we're
talking
about
this.
There
hasn't
been
enough
back
and
forth.
I
feel
like
so
I
don't
know
people
you've
had
this
idea,
percolating
in
the
back
of
your
minds,
at
least
on
some
level
for
the
past
few
weeks.
J
A
A
Do
we
want
to
rely
on
these
specific
libraries
and
make
an
investment
of
time
and
effort
and
code
longevity
on
them
right,
because
pieces
of
code
that
we
write
using
these
components
might
be
there
five
years
from
now
might
be
there
longer
look
at
how
long
the
classic
notebook
ui
has
lasted,
for
example,
and
so
is
it
a
sound
investment
to
make.
K
G
B
So
there
is
a
jab
open
to
suggest
it
to
be,
if
adopted
at
the
jupiter
level,
to
be
used
more
broadly.
B
Personally,
I
will
love
it,
but
I
know
that
other
people
have
for
good
reason
with
a
further
opinion,
and
so
I
don't
think
it
should
be.
It
should
be
mandatory
for
for
for
especially
for
the
widget,
as
the
user
is
already
far
from
the
from
the
the
styling
and
the
and
the
dom
you
when
it's
using
the
widget.
B
I
I
think
like
if
it
gets
adopted
outside
of
jupiter
lab
the
best
other
candidate
is
probably
the
interface
of
jupiter
hub
for
trying
to
have
a
more
consistent
ecosystem
and
maybe
then
for
for
ipad
widget,
knowing
that
probably
a
good
path
will
be
first
to
have
it
as
a
an
extension
that
bring
widgets
that
are
built
on
top
of
that,
but
are
not
the
official
ones
and
then
maybe
if
people
think
that
okay,
that's
that's
good
to
make
them
official,
then
maybe
let's
merge
that
for
party
extension
to
to
become
the
official
standard.
I
I
think
it
it's
not
a
clear
given
considering
that
in
ipad
widgets,
we
worked
really
hard
to
stay
with
very
the
basic
html
form
controls
for
most
everything,
with
a
few
exceptions
and
the
idea
there
was
very
easy
cross
browser
compatibility,
accessibility,
typically
out
of
the
box,
because
we're
using
just
straightforward
html
form
controls
and
the
the
scope
of
ipad
widgets
is
deliberate.
The
core
controls
is
deliberately
limited
for
that
reason
to
make
it
simpler
to
maintain
to
make
it
cross-browser
compatible,
etc.
I
But
we
also
do
provide
a
number
of
tweaks
that
you
can
do
with
the
styling
that
may
mess
up
the
styling
of
of
these
widgets.
So
it's
not
a
it's,
not
a
given.
I
think
the
ipad
widgets
moves
to
these,
but
I
think
frederick,
that's
a
that's
a
good
plan
to
go
ahead.
I
We
can
make
it's
easy
to
make
a
third-party
ipad
widgets
library,
just
like
we
have
for
material
ui,
just
like
we
have
for
other
widget
frameworks
and
then
and
then
let's
see
and
have
the
conversation
about
whether
we
want
to
replace
the
core
controls
with
with
these.
L
So
another
question
is:
how
much
size
does
these
runtime
dependencies
are
added
to
the
bundle?
Do
you
know
roughly
the
order
of
magnitude
these
packages
add
to
the
compiled
bundle.
A
So
the
size
that
you're
showing
this,
what
is
it
the
size
of?
Is
it
the
size
like?
I
guess
a
question
I
have
is
suppose
you
have
an
extension
authored
with
these
components
and
that
extension
is
grafted
into
jupiter
lab
versus
an
extension
that
has
the
same
components,
but
it
implemented
you
know
with
html
or
with
lumino,
or
something
that
already
ships
for
free.
What's
that
size
difference
is
is.
Is
that
what
you're
showing
this
or
I
I
don't
really
know
what
these
numbers
indicate.
B
Yeah,
that
will
be
the
additional
size
that
will
add
to
jupiter
lab.
Let's
see,
okay,
knowing
that
we'll
start
losing
weight
when
we
start
ourselves
replacing
elements
from
the
core
with
them,
because
we
could
drop
some
codes.
We
could
drop
some
stars
so
not
to
say
that
we'll
at
the
end,
reach
zero.
L
B
I
In
the
past,
pete
has
advocated
for
web
components,
but
with
a
caution
that
web
components
have
this
global
name
space
in
the
browser,
and
so
you
can
have
two
different
libraries
that
package
you
know
they're
the
same
web
component
library
and
then
they
stomp
on
each
other
in
this
global
namespace
and
and
here
you're
repackaging,
even
a
library,
that's
already
out
there.
First
of
all
is
that
understanding,
correct
and
two?
Can
you
just
address
like
how
you're
thinking
about
these
things
and
how
you're
working
around
this
global
name,
space
issue.
B
B
So
what
happened
is
only
one
version
of
the
library,
because
so
you
lose
the
control
of
which
one
will
get
it,
but
with
the
webpack
federation,
only
one
extension
get
the
the
underlying
microsoft
library
loaded.
Actually,
there
are
two
library
from
microsoft
and
actually
in
the
demo,
I
didn't
got
into
the
detail,
but
one
part
came
from
one
extension
and
the
other
one
come
from
the
other
extension
and
it
worked
in
the
sense
that
it
was
not
major
version
change,
so
there
were
no
conflict.
B
What
happened
is
that
the
styling
of
the
the
element
was
not
the
latest,
because
the
fine,
the
one
that
make
it.
Finally
wasn't
the
latest
version.
But
no,
I
didn't
encounter
any
error
and
I
think
also
the
yeah
just
because
only
one
version
was
really
used
at
the
end.
B
It's
just
we
lose
control
of
which
version
what
what
won't
be
the
case,
if
it
it's
in
the
core,
because
if
it's
in
the
core
then
we
can
do
is
react,
we
can
make
it
make
sure
through
the
federation,
but
we
control
which
version
is
used.