►
From YouTube: JupyterLab Team Meeting - 21 June 2023
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/170
A
A
Weekly
call
Solstice
Edition.
Today
we
have
16
people,
nope
17,
a
17th
person
just
joined.
We
have
17
people
on
the
call
and,
if
you're
just
joining,
please
look
in
the
chat.
The
meeting
minutes
are
linked
there
and,
if
you'd
like
to
sign
in
please
do
and
if
there's
something
you
would
like
to
talk
about.
Please
add
an
agenda
item.
A
C
Yeah
Frederick
just
just
joined,
oh.
D
E
F
Okay,
then,
let's
go,
let's
start
by
the
latest.
This
is
the
easiest
Just
for
information.
I
pushed
appear
on
the
Jupiter
Road
laser
to
add,
support
for
pipei,
crested,
publisher
and
and
PM
proven
nose.
F
So
for
that's,
one
of
the
new
security
features
that's
available
on
pipei
and
on
npm,
and
basically
it
gives
you
the
opportunity
to
to
link
what
was
done
for
publishing
the
a
release
so
like
you're
gonna
on
npm
you're
gonna
have
a
badge
saying:
okay,
this
release,
like
was
done
on
from
that
GitHub
organization
that
repository
at
that
commit,
and
on
Taipei
the
what's
these
things
are.
F
Those
is
that
on
the
releaser,
you
don't
need
to
set
up
a
bypi
token,
any
longer,
basically
use
you
need
to
set
up
on
pipe
icing.
One
of
the
way
of
publishing
a
package
is
through
a
GitHub
action
from
that
repository
using
that
workflow,
and
so
it
allows
to
to
avoid
to
set
up
one
of
those
tokens
for
the
releaser,
so
yeah,
and
you
should
expect
me
to
push
some
pure
on
Jupiter
related
project
for
switching
to
to
those
to
that
new
system.
F
As
a
it's,
it's
good
for
security
yeah,
that's
that
was
the
last
the
info.
One.
B
And
do
we
have
it
already
for
collab
we're
not
right
not
yet
or.
F
I
haven't
set
up
anything
yet
the
under
the
the
one
I
tested
hit
on
is
a
Jupiter
lab
kit,
the
extension
for
managing
git
repository
the
what
needs
to
be
done
is
is
to
set
up
on
pipei
that
no
URL
the
publication
from
a
specific
GitHub,
so
that
you
need
to
do
some
action
on
pipei
and
then
that's
about
it
like
you
need
to
change.
Also,
the
the
workflow,
of
course,
on
the
repository
yeah.
F
F
Token
yeah,
they
decided
to
like
add
some
signatures
and
some
like
improve.
F
F
So
back
to
the
highlights
and
the
question:
I
have
there's
been
a
a
big
issue
reported,
but
like
a
bug
in
the
way
we
this
realize
the
kernel
messages.
Basically,
we
are
introducing
empty
buffer
like
it's
a
bit
technical
for
people
that
are
not
familiar
with
the
kernel
protocol
and
there's
been
David
was
very
reactive,
I,
don't
know!
F
If
he's
on
the
call,
no,
it
doesn't
look
like
it
and
basically
I
would
like
people
that
are
that
I
have
a
better
history
background
and
I
think
Json
not
to
point
the
finger
at
you,
sorry,
but
you
were
pinged
on
the
pr,
so
you
can
leave
if
some
of
the
older
developers
of
Jupiter
that
can
never
look
and
give
their
opinion
on
it
would
be
great
yeah
and
then
the
second
point
so
Eric
chart
languages
and
for
different
thumbs
up
average
charts
as
a
pusher
PR.
F
That's
improving
the
the
way
we
display
the
kernel
available
to
to
be
more
pertinent
for
the
user,
so
this
has
been
merged
for
in
the
four
Branch.
The
idea
is
to
back
Port
it
in
36x
and
because
it
has
somehow
an
impact
on
the
interface
I,
I
told
Eric
I
will
bring
the
the
point
today
at
the
meeting.
I
think
it's
okay,
because
basically
it's
just
improving
the
order
of
the
kernel
list.
F
When
again,
when
the
people
is
in
the
select
kernel
dialogue,
when
the
people
just
ask
okay,
what
are
the
corners-
and
please
pick
one
but
yeah
I-
wanted
to
hear
like
to
highlight
it
here,
so
that
if
people
wants
to
comment
right
now
or
on
the
pr,
they
can
do
it
if
they
have
opinions
that
we
should
not
back
Port
it
on
36x.
G
Thank
you
Frederick,
so
maybe
I'm
going
to
bring
a
bit
more
context
because
we're
looking
at
the
kernel
selector
you
you
can
see
some
some
inquiry
runs
or
something
which
is
not,
which
can
be
better,
especially
when
you
define
when
you
launch
your
server
a
default
kernel
and
when
you
upload
a
notebook
without
any
assign
kernel
or
if
the
assign
kernel
is
no
more
present.
G
That's
a
default
kernel
is
pre-selected
and
at
the
top
of
the
group,
when
the
notebook
has
a
kernel
assigned
which
is
not
existing
or
has
no
assigned
kernel,
there
is
another
action
which
is
open
and
which
I
linked
at
the
bottom
of
the
agenda,
to
rethink
a
bit
more.
The
naming
and
the
grouping
of
those
kernels
and-
and
this
may
have
more
impact
on
the
UI.
G
G
F
F
F
So
please
hesitate
to
comment
on
the
pr
if,
if
you
have
arguments
against
I
guess
from
the
silence
that
people
agree
on
that
otherwise
and
then
the
last
PRM
linking
here
is
the
huge
one
by
Nick,
it's
very
good
Improvement.
So
the
idea
is
to
to
have
the
same
features
as
in
GitHub
to
be
able
to
display
mermaid
GS
syntax
as
a
graph
in
the
markdown
and
the
main
question.
F
So
the
this
feature
is
targeting
4.1.0
that
doesn't
have
any
T
yet,
but
it's
going
to
be
in
the
next
feature
release.
But
the
question
on
the
pr
that's
left
right
now
is
should
so
it's
this
is
opening.
F
To
delegate
more
out
of
more
GS,
so
the
library
we
are
using
to
to
enroll
stuff
ourselves,
so
should
we
create
an
API
to
make
that
pluggable
and
in
the
Jupiter
lab
sense,
or
should
we
keep
it
closed
in
the
sense
that
we
want
to
give
like
to
ever
a
better
control
of
what
should
Mark
down
look
like
in
Jupiter
tooling?
F
So
basically,
Nick
also
opened
a
PR
with
that
to
to
add
the
support
in
and
we
convert
and
if
we
start
to
say,
okay,
people
can
change
the
way
those
graphs
are
displayed
or
stuff
like
that.
Then
it's
going
to
be
tougher
for
us
to
ensure
consistency
and
I.
Think
people
like
William
I,
don't
know
if
William
integrated
in
cocalc,
it
will
be
probably
more
happy
to
know.
Okay,
they
included
that
version
of
mermaid
so
to
have
a
bad.
F
H
Think
you're,
taking
the
best
feature
from
missed
one
of
the
best
features
for
mist
and
putting
it
in
the
core
and
Mr
Eddie,
handles
all
these
directives
right
so
like.
If
you
wanted
to
really
extend
the
language,
go
use
Mist,
but
like
this,
one
is
just
more
that
this
identifier
has
become
canonical
like
in
boot
camps.
At
this
point
you
know
so,
like
I
think
mermaid's
way
more
special
than
just
like
extending
it
personally.
A
A
F
Thanks
and
the
last
point-
but
maybe
I
don't
want
to
discuss
here-
maybe
we
can
schedule
the
discussion
for
next
week
because
basically
I
went
through
all
the
repository
in
the
Jupiter
lab
organization,
and
there
are
a
couple
of
repository
that
are
officially
not
archived
but
that
they
haven't
seen
any
activities
for
quite
some
time
and
I
think
it
will
be
good
if
we
discuss
what
should
be
the
status
of
those
repository
and
I
propose
to
maybe
schedule
the
discussion
for
next
week,
so
that
people
can
have
a
look
at
those
and
make
a
better
opinion
for
themselves.
E
F
B
Frederick,
maybe
you
could
add
a
list
of
those
in
the
agenda
for
next
week,
so
we
can
prior
look
into
it.
C
Might
be
good
to
do
like
a
team
Compass
issue
as
well.
Just
so
we
don't
lose
sight
of
it
and
we
can
continue
to
discuss
it
because
I
like
the
idea
of
having
a
plan
to
kind
of
declare
a
repository
as
end
of
life
and
to
you
know,
redirect
people
to
other
newer,
more
supported,
repos.
B
A
A
Thanks
for
it
so
yeah,
if
you
could
please
look
at
these
repo
stuff,
that's
listed
sometime
between
now
and
next
week
or
right
now,
I
see
people
writing
comments
in
them.
A
Just
weighing
in
that
would
be
helpful
cool
okay.
So,
let's
carry
on
the
next
person
on
the
agenda.
Today
is
Jason
w.
C
Hey
everyone,
a
couple
quick
notes:
one
is
a
plug
for
the
weekly
issue.
Triaging
call
we
do
this
at
9
00
a.m,
Pacific
on
Tuesdays,
if
you
haven't
been-
or
you
haven't,
been
in
a
while
we'd
love
to
have
you
come
along
the
more
people,
we
have
to
look
at
new
issues,
the
more
likely
it
is
that
we'll
be
able
to
make
a
firm
decision.
C
You
don't
have
to
be
an
expert
or
particularly
experienced
about
it
about
the
projects,
but
we
can
take
a
look
at
new
incoming
issues
for
Jupiter,
lab
Jupiter,
notebook
and
Jupiter
lab
desktop
and
determine.
Are
these
new
issues?
Are
these
not
duplicated
by
existing
stuff?
Do
we
need
more
information
from
the
requester
so
particularly
for
new
contributors
to
Jupiter?
This
is
a
great
way
to
make
them
feel
heard
and
welcomed
and
provide
feedback
on
their
issues.
C
So,
if
you'd
like
to
help
out
with
that
we'd
love
to
see
you
Tuesdays
nine
o'clock
on
the
same
Zoom
Channel
one
other
question
that
I
got
from
a
co-worker
just
earlier
this
morning.
My
time
is
this
person's
curious
about
opening
a
file
within
Jupiter
lab,
using
a
link
inside
a
notebook
so
effectively
having
a
link
that
when
you
click
it,
it
opens
a
notebook
in
a
new
notebook
tab
inside
the
same
Jupiter
lab
session.
C
C
F
F
C
F
After
the
this,
let's
always
subtle
is
that
it
depends
like
if
you
just
spin
your
server
in
different
directory,
like
in
different
route
directory
than
the
relative
path,
is
going
to
change,
and
so
yeah.
D
C
Yeah
I
think
he's
using
HTML
instead
of
markdown,
for
it
so
I'm
wondering
if
using
markdown
for
that
link
might
be
more
explicit,
but
I'll
follow
up
with
with
my
colleague
and
if,
if
it's
either
a
bug
or
a
possible
enhancement
request,
maybe
we'll
file
that
and
we
can
keep
the
discussion
going
on
that
I
just
figured
because
of
the
because
of
the
coincidence
in
timing.
It
would
be
good
to
ask
the
group
that,
thanks
all
that's
all
I
have
so
Tony
is
next
on
the
agenda.
Go
ahead,
Tony.
E
Hey
hi,
everyone
great
number,
two
in
a
row
of
triaging
we're
just
taking
over
here
so.
H
H
If
anybody
ever
sees
men
from
Jupiter,
making
issues
they're
very
serious
and
worth
our
time,
and
specifically
this
one.
Currently,
it's
split
into
15
tasks,
I'm
not
done
triaging
it
yet.
But
what
makes
me
really
hopeful
in
this
that
I
wanted
to
communicate
to
you
all
is.
Let
me
see
I'm
trying
to
find
this
reference.
H
Folks,
this
the
suggest
this
okay,
here
it
is
related
to
Jupiter
lab
9399,
which
is
the
accessibility
audit
that
we
got
on
Jupiter
lab
notebook,
7
version
7
is
obviously
much
better
and
what
should
be
recommended
to
persons
with
disabilities
and
just
needs
a
few
small
fixes
to
be
fully
supporting
of
low
vision,
colorblind
and
ambulatory
users.
So
I
personally
see.
This
is
super
encouraging
that
we
can
make
a
big
dent
in
some
accessibility.
H
So
if
anybody
has
any
time,
there's
some
like
low-hanging
fruit,
perhaps
in
these
re
in
these
issues
that
I've
triaged,
if
anybody
has
any
extra
information
that
they
can
provide
in
any
issues,
please
do
the
the
more
information
that
we
have,
the
easier
we'll
be
able
it'll
be
to
implement
these
things,
and
if
anybody
has
any
time
or
has
any
knowledge
about
these
other
tasks,
it
would
help
me
in
triaging
these
two
or
also
so.
That
is
my
first
point.
D
Cool
I
wanted
to
mention
that
this
is
relevant
to
the
Jupiter
lab,
because
some
of
those
issues
at
least
ones
that
I
already
identified
charging
them
and
I,
would
expect
more.
They
have
to
fix,
be
fixed
on
the
Jupiter
lab
side,
so
at
least
identifying.
That
would
be
very
helpful
but
yeah
thanks
Tony
for
creating
that
them
and
thanks
to
anyone
who
would
be
willing
to
participate
and
contribute
to
them.
H
Yeah,
thanks
for
your
support
in
getting
that
those
issues,
these
issues
together,
Andre
I,
appreciate
appreciate
it
and
I.
The
next
thing
I
wanted
to
talk
about
is
an
overlapping
issue,
so
a
few
of
the
ones
that
I've
triage
so
far
are
fairly
small,
but
this
one
is
significant
and
would
affect
I,
don't
know
if
it's
Jupiter
lab
and
notebook
or
whatever,
but
right
now.
H
The
current
implementation
of
the
file
browser
is
a
bunch
of
list
items,
but
it
appears
as
a
table
and
it
would
have
a
lot
better
screen.
Reader
affordances
if
it
was
marked
up
as
such,
so
I
was
curious
about.
If
anybody
had
any
initial
reactions
to
changing
the
file
browser
to
a
table
like
layout,
perhaps
versus
what
it
is
now.
H
Or
just
comment
in
the
issue
later
on
I
feel
like
this
is
a
big
one,
but
being
able
to
navigate
to
files
is
super.
H
Cool
all
right,
the
next
one,
the
in
in
accessibility
land.
There
is
a
unit
of
documentation
called
the
Aria
authoring
practice
guides,
and
these
Arya
authoring
practice
guides
give
have
recommendations
for
patterns
when
you
implement
user
interface
components,
so
menu
bars,
landmarks
tables
forms
all
kinds
of
stuff.
Basically,
these
are
the
reference
implementations
for
how
the
interactivity
should
be
so
right
now.
One
of
the
comments
that
we
have
accessibility,
wise
is
that
it
takes
20.
H
Tab
stops
to
get
to
a
notebook
cell
in
notebook
7..
Now
there
are
patterns.
One
of
them
is
the
toolbar
pattern
so
where,
basically,
you
have
to
tab
through
all
the
buttons
in
the
toolbar
restart
run
all
so
and
so
forth.
Right
that
toolbar
pattern,
the
suggestion
is
to
have
one
the
the
first
element
receive
focus
and
then
remember
the
last
element
and
be
able
to
use
arrow
keys
to
navigate
there.
H
That
feature
is
something
that
we
actually
have
in
the
menu
bar,
so
we
abide
the
menu
bar
keyboard
pattern
with
left
right
up
down
sort
of
navigation,
but
some
of
the
other
components
like
tabs
and
menu
bar
or
and
toolbars
don't
abide
these
APG
practices
would
folks
could
folks
imagine
it
being
possible
to
update
these
interactive
implementations
are,
do
you?
Would
it
cause
too
much
of
a
stir,
changing
how
the
tab
stops
work
because
right
now
it
definitely
is
not
assistive.
H
A
D
A
Like
a
polite
thing
to
ask,
but
we
have
it
as
part
of
our
mission
for
this
application
to
be
accessible
at
a
bare
minimum,
for
legal
reasons,
but
really
for
moral
reasons
and
for
Community
reasons,
and
if
it
is
the
case
that
what
people
who
use
screen
readers
expect
is
to
be
able
to
tap
through
toolbars
that
are
tied
to
the
most
recently
used
text
editor
and
that
we
can't
currently
do
that.
We
should
find
out
how
we
have
to
do
it.
I.
Imagine
nothing
that
we
do
in
this.
A
Space
is
Backward
Compatible
api-wise,
so
everything
here
could
either
be
pitched
as
a
feature
and
do
a
DOT.
You
know
minor
release
or
it
could
be
considered
even
a
a
patch
of
bug
fix
and
it
could
be
a
passwords,
but
yeah
I
mean
I
think
what
might
be
useful
is
picking
components
one
by
one.
So,
starting
with
say,
the
toolbar
make
the
toolbar
tabable
and
then
figure
out
later
how
to
make
sure
you
get
from
tabbing
out
of
a
cell
text
editor
into
its
toolbar
or
something,
but
like
the
toolbar
problem.
Self-Contained
right.
A
F
Yeah
I
just
want
to
say,
I,
agree
fully
and
I
think
we
should
definitely
consider
that
as
bag,
because
It
Anyway
is
inconsistent,
as
you
said
like
what
is
done
for
accessibility,
it
basically
applies
to
all
softwares.
Even
if
people
use
keyboard
navigation
and
don't
need
it
for
accessibility
reason,
and
so
it
will
be
it.
We
could
consider
it's
a
bug
for
now
and
from
the
API
change
the
potential
APA
change
point
of
view.
F
I
think
we
can
do
what
Mac
has
proposed
when
we
change
the
dialog
tag,
meaning
that
we
can
bump
internal
package
of
Jupiter
lab
to
say:
okay,
it's
a
mineral
release,
but
it's
a
patch
release
at
the
application.
Level
I
think
it's
perfectly
fine,
because
it's
consistent
from
the
from
somewhere
point
of
view.
H
It
is
awesome,
y'all.
Thank
you
because
yeah
Darion
said
you
know,
there's
legal
reasons
for
this,
but
the
stuff
that
we're
actually
talking
about
is
assistive
and
like
making
the
best
experience.
These
aren't
things
you
can
check
for.
So
it's
really
good
that
you
all
are
supporting
it
in
this
way.
So
thank
you
and
thanks
for
the
comment
and
the
issue
Fred
too
cool.
F
H
E
F
It's
it's
one
of
the
pure
I
want
to
do
since
quite
sometimes,
but
because
there
is
a
an
easy
way
to
do
it
by
using
a
web
component
and
that
support
it
internally
and
will
be
transparent.
But
whatever
it's
the
implementation
detail.
Okay,
cool.
H
Cool
well
I'll,
open
up
I'll,
open
up
an
issue
or
I'll,
expand
on
this
issue
and
see
what
it'll
take
to
fix
it.
Yeah.
Thank
you.
A
H
Oh
yeah,
so
the
it's
a
today
we
are
focusing
on.
Hopefully,
when
folks
show
up
we're
gonna
work
together
on
Community
proposals.
It
was
acknowledged
last
week
that
we
have
not
been
with
all
of
the
development
Focus
and
things
going
on.
H
We
haven't
been
so
focused
on
the
newcomer
experience
and
that's
always
been
an
Avenue
that
accessibility
sort
of
shined
in
so
our
hope
is
to
figure
out
way,
write
some
proposals
to
figure
out
some
ways
that
we
can
help
onboarding
new
people
and
kind
of
yeah
get
back
to
that
mission.
A
Great
thanks
next
person
on
the
agenda
today
is
Eric
C,
S,
Young,
Nicole.
G
Yeah
I
I
think
we
already
covered
part
of
the
of
those
points
with
the
kernel
selector
just
wanted
to
to
link
again
to
the
ux
design
issue.
G
Well
because,
apart
from
the
default
kernel
selector,
you
have
the
notion
of
preferred
kernel
when
you
assign
a
kernel,
there
is
like
a
language
also
like
so
you
can
say
Python,
3
or
Pi
spark,
but
the
language
is
python
and
the
selector
shows
you
stuff
around
well.
This
is
all
the
kernels
with
the
Python
language,
and
this
is
all
the
sessions
with
the
Prof
with
the
python.
What
is
this?
G
This
is
all
the
active
sessions
with
python
language
as
kernel
and
I
think
there
is
some
terminology,
and
maybe
some
some
brainstorming
to
do
to
to
make
it
more
readable,
because
at
some
point
preferred
is
not
default
and
aside
is
not
preferred
and
so
on.
So
this
is.
This
issue
is
meant
to
maybe
do
do
something
around
those
terminology
and
yeah.
We
discussed
about
the
planning
for
3.6
so
that
that
was
my
points.
Thank
you.
A
Awesome
thanks,
sir
okay
great
so
now
we
have
Rosia.
I
Yeah
I
forgot
to
bring
this
up
in
a
notebook
meeting,
but
since
it's
some
tied
to
Jupiter
lab
and
there's
still
a
few
notebooks
I
mean
people
or
a
number
of
people.
I
wanted
to
kind
of
ask
a
question,
and-
and
that
is
I'm
wondering
about
the
development
like
process
like
when
there
are
features
that
are
requested
in
Notebook
7
or
something
along
those
lines
that
could
potentially
be
applied
to
Jupiter
lab
or
they
may
have
even
been
requested.
Bridge
repair
lab.
I
What
what
kind
of
questions
should
we
have
answered
before
like
implementing
a
such
a
future
notebook?
Is
it
like
a
matter
a
case-by-case
basis?
Where
depends
on
on
the
feature?
That's
being
sort
of
proposed
and
and
based
on
people's
willingness
to
to
have
that
included
in
in
Notebook,
7
or
even
Jupiter
lab?
Is
that
more
of
the
situation
for
a
reference?
I
I'm
kind
of
I
keep
I'm
having
mind
the
custom
CSS
issue,
where
that
was
that
functionality
was
added
into
notebook
7,
but
there
was
previously
some
issues
like
kind
of
requested
it
in
Jupiter
lab
so
now
like
in
this
case,
I
would
think
I
think
I'm.
We
would
go
with
removing
removing
that
functionality
from
notebook
7
to
Jupiter
lab,
potentially,
if,
if
people
are
okay
with
that,
but
yeah,
sorry
I'm
kind
of
rambling.
I
F
But
I
can
go
ahead,
then
and
I
think
there's
no
we're
using
Jupiter
lab
as
the
basis
for
most
of
the
code.
We
should
try
hard
to
do
everything
in
Jupiter
lab
because
it
will
be
beneficial
for
most
user,
so
yeah,
that's
my
take
and
I
think
I
don't
know
if
Jeremy
is
on
the
call
yet
still
but
I
think
it's
what
Jeremy
has
used.
Also
since
he
started
notebook
7
is
like
everything
that
I
can
backboard
to
Jupiter
or
doing
Jupiter
lab.
F
Let's
do
it
in
Jupiter
lab
and
then
consuming
in
Notebook
7..
They
may
be
like
right
now.
There
is
there's.
Maybe
that
there's
maybe
a
problem
of
of
timing
in
the
sense
that
know
that
we
are,
we
are
in
a
yeah.
We
have
released
Jupiter
app
for,
but
you
have
not
yet
released
notebook
7.
Maybe
we
are
a
bit
slower
in
on
the
Jupiter
website
to
do
new
release
of
the
packages,
but
I
think
it
shouldn't
be
a
problem
which,
like
if
it's
for
Notebook
7
and
you
need
a
release.
F
A
So
somewhat
surprisingly,
I
completely
agree
with
Brad
I
think
for
most
of
the
kind
of
features
that
we're
going
to
be
adding
to
notebook
they're
going
to
be
features
where
the
implementation
and
lab
will
be
the
way
it
actually
gets
delivered
anyway.
But
those
features
where
it
really
truly
is
like
a
notebook
application
feature.
You
know
we
can
visit
them
one
by
one
and
decide.
Does
that
belong
in
lab
as
well,
but
yeah
so
I
mean
if
you
run
into
a
feature
whose
implementation
is
not
going
to
be
inside
Jupiter
lab
core.
A
But
your
curious,
hey
should
this
feature
exist
in
both
then
bring
it
up,
but
in
general
I
think,
assuming
that
notebook
and
lab
or
sibling
applications
that
they
have
very
compatible
feature
sets
it's
just
that
the
notebook
user
experience
is
quite
different
than
the
lab.
One
then
I
think
that's
a
good
guiding
principle.
I
A
Thank
you,
okay.
Well,
if
we
reached
the
bottom
of
the
agenda,
is
there
anyone
who
has
been
holding
back
and
would
like
to
talk
about
something.
H
Do
you
think
anybody
would
care
to
see
this
idealized
semantic
notebook
that
I'm
talking
about
I
haven't
ever
shared
it
with
Jupiter
lab
before,
but
the
research
is
kind
of
at
an
end,
and
maybe,
if
there's
a
little
bit
of
time
and
we're
recording
it'll
be
cool
to
have
this.
A
Yeah
I
think
that's
a
great
idea
just
by
way
of
contextualizing
a
little
bit,
so
there
are
kind
of
one
and
a
half
or
two
accessibility
calls
that
regularly
happen,
and
one
of
them
Tony
has
presented
his
conclusion
to
researching
what
the
notebook
document
actually
is.
A
Semantically
and
Gabriel
and
I
have
volunteered
to
take
this
vision
and
try
to
implement
it
within
the
notebook
that
the
Jupiter
lab
notebook
package,
basically
so
that
the
rendered
markup
that
those
packages
generate
comes
out
in
a
way
that
is
accessible
to
assistive
devices,
so
yeah
take
it
away.
Tony
I'd
love
to
see.
E
Really
cool
thanks
for
that
intro.
That
was.
H
Super
concise,
so
I've
been
working
for
a
few
months
with
the
Space
Telescope
Science
Institute,
and
our
charge
has
been
to
make
the
rendered
version
of
The
Notebook
accessible,
so
this
doesn't
really
have
much
to
do
with
the
interactive
notebook
until
it
did
so.
This
is
my
rendering
here
and
it's.
H
This
is
the
Lorenz
notebook
and
it's
pretty
much
the
same
thing
that
we're
used
to
seeing
the
only
thing
that's
new
here
is:
we
actually
have
cell
numbers,
but
none
of
the
the
visual
part
isn't
really
what
I
wanted
to
get
at.
It's
that
when
we
go
and
inspect
the
Dom
there
aren't
any
divs
in
our
notebook.
H
H
So
a
feed
is
a
collection
of
containers
that
are
meant
to
be
read
so
they're
kind
of
like
non-interactive
widgets.
So
in
this
pattern
you
have
a
section
with
a
roll
feed.
There
are
16
cells
in
it
and
inside
of
there
every
cell
is
an
article.
H
H
H
So
we
do
have
a
form
and
what's
really
interesting
about
this-
is
that
now
every
cell
is
destructured
and
you
can
access
them
through
their
forms.
You
can
access
the
inputs
without
any
kind
of
like
Dom
Behavior,
so
they're
sort
of
independent
there,
each
cell
needed
a
label
and
it
needs
a
visible
label.
H
The
inputs
are
a
thing
called
a
field
set,
so
field
sets
are
used
to
group
interactive
widgets.
So
in
our
what
this
actually
winds
up
making
is
sort
of
a
group
area,
but
this
is
what
labels,
the
inside
of
the
cell,
are
the
input
of
the
cell.
We
have
a
toolbar
implementation
here,
that's
not
exposed
yet,
but
the
idea
is
is
that
from
this
like
accessible
resting
place,
we
should
be
able
to
make
an
accessible
interactive
experience
progressively.
H
We
have
another
field
set
for
the
outputs,
and
then
this
is
a
markdown
cell,
so
we
have
some
rendered
HTML.
This
is
what
a
code
cell
looks
like
so
on
so
forth.
So
in
accessibility
terms,
what
this
means
is
if
we
go
and
we
look
at
our
accessibility
properties.
This
is
actually
where
this
representation
shines
the
most.
H
H
If
we
were
to
go
and
look
in
here
right
now
to
my
oh,
it's
showing
up
in
another
place-
let's
not
worry
about
that
right
now,
but
nevertheless
what
we
can
actually
see
is
or
unders
learn
from
this
is
that
the
the
Dom
was
actually
designed
for
the
screen
reader,
and
you
can
start
to
see
that
there's
a
lot
richer
information
and
context
in
here
and
Screen
readers
read
camel
case
as
if
they're
separated
words,
so
this
mashing
of
things
together
is
okay
and
yeah.
H
This
has
been
tested
on
voice
over
chromevox
nvda.
It
has
not
been
tested
on
Jaws,
but
our
goal
is
to
take
these
things
like
to
each
of
these
things
in
The.
Annotation
object
model
here
relates
to
a
Dom
node,
and
what
we
can
see
is
that
by
using
certain
tags,
we
get
different
roles
in
here.
So
from
this
idealized
structure
that
I
showed
you
to
start
what
we
can
go
and
do
is
sort
of
reverse
engineer.
F
H
No,
that's
the
Hope
man
yeah.
F
Meaning
that
the
structure
of
the
document
is
is
mainly
driven
by
the
headings
yeah.
What
will
happen?
Basically
I,
really
like
the
idea
like
this
gives
a
very
that's
the
trouble.
Let's
say
of
notebook.
The
structure
you
propose
is
awesome
in
the
sense
that
it
really
represents.
What's
a
cell,
what
the
input?
What's
the
output
and
stuff
like
that,
and
unfortunately
we
have.
F
We
have
that
use
case,
that
people
structure
things
adding
headings
in
it
that
totally
are
uncoupled
with
actually
the
cell
structure
and
so
yeah,
and
that's
that's
always
the
the
question
like
how
do
we
reconciliate
I
I?
Think
there
is
no
answer
like
there
is
no
reconciliation.
It's
like
how
do
we
handle
the
two
point
of
view.
I.
Think
it's
more
like
the
question.
Probably
yeah.
H
So
so
what
my
experience
with
this
is
so
I'm
not
worried,
I'm,
not
working
on
the
heading
problem,
yet
I'm
trying
to
work
on
like
basically
it's
like.
If
we
can
there's
a
building
and
the
headings
are
inside
of
rooms.
I
need
the
hallways
to
be
accessible
before
I
worry
about
your
problem
right.
So
this
is
basically
a
structure
that
puts
landmarks
on
the
page.
So
landmarks
are
really
important
for
assistive
technology
because
they
offer
quick
ways
to
access
them
on
a
menu,
so
you
can
go
and
hit
like
Ctrl
shift.
H
M
and
that'll
show
you
all
the
landmarks
and
you
can
go
to
them
really
quickly.
Headings
are
also
controls
or
control
shift
H
right,
so
there
are
different
Navigation
structures.
Headings
are
one
navigation
structure
and
landmarks
are
another
navigation
structure.
H
One
of
the
this
is
a
I'm
going
to
jump
off
to
another
place,
because
I
feel
like
you've
vibed
with
that
one
of
the
other
things
that
I've
learned
here
in
this
process
is
that
when
we
talk
about
the
static
converted
notebook,
we
do
a
disservice
and
talking
about
it
statically
and
what
I
mean
by
that
is
like
tab
stops,
are
defined
by
interactive
widgets
right.
H
The
more
interactive
crap
you
put
on
the
page,
the
more
interactive
these
static
documents
are
right,
so
headings
as
a
whole
link
to
itself
huge
hit
Area
great
for
ambulatory
users,
but
it
also
adds
interactivity
there
right.
So
now
you
have
a
heading,
you
have
a
link
and
it's
inside
of
a
landmark,
and
you
have
multiple
ways
to
navigate
and
that's
what
we're
trying
to
do
here
is
kind
of
great
there's,
a
lot
of
different
ways
that
people
use
screen
readers.
It
might
be
tabbing.
It
might
be
using
these
special
menus,
so
on
so
forth.
H
But
what
we're
this
is
actually
opening
up.
Another
dimension
outside
of
the
headings
like
if
you
screwed
up
your
headings,
they're
just
out
of
order
when
you
go
and
access
them,
it's
actually
not
all
that
big
of
a
deal
but
like
if
you
go
and
you
click
on
that
heading
and
it
announces
this
heading
is
in
cell
five.
That's
a
helpful
thing
right,
because
now
you
can
go
and
access
just
cell
five,
so
we're
kind
of
creating
like
layers
of
redundancy
here,
because
it's
really
easy
to
get
lost
in
a
long.
H
A
Cool
thanks
for
showing
us
this
and
I'm
sure
we're
gonna
have
to
revisit
pieces
of
this
as
we
progress
with
trying
to
implement
it,
because
it
might
be
that
the
easiest
path
is
a
bunch
of
small
br's
or
it
might
be
that
the
most
reasonable
path
is
here's
a
PR
that
makes
the
notebook
work
with
screen
readers,
I,
don't
I'm,
not
sure.
Yet,
once
we
start
I
think
we'll
be
coming
back
to
this
call
to
show
people
things
and
getting
put
on
them
with
some
frequency.
H
I
after
what
you
said
earlier,
Darien
I
think
we
are
we're
going
to
focus
on
one
component
at
a
time
and
just
kind
of
become
experts
in
that
or
like
have
a
few
PRS
that
are
sort
of
expert.
Let's
focus
on
the
menu
bar,
let's
focus
on
the
main
landmark,
let's
focus
on
each
section
or
region,
and
that
that's
probably
going
to
make
the
most
focused
sense
but
we'll
find
out.
So
yes,
yes,
we
will
revisit.
A
Cool
great
is
there
anyone
else
who
would
like
to
bring
up
something
or
I
can
stop
the
recording
to
ask
if
again.