►
From YouTube: Jupyter/IPython Dev Meeting, April 18, 2017
Description
Meeting of the Jupyter/IPython development team, April 18, 2017
Meeting Notes: https://paper.dropbox.com/doc/April-2017-Jupyter-Weekly-Meetings-kML3Uu32Y4HNAlpJkHRXX
Meeting Archive: https://jupyter.hackpad.com/ep/group/zCqMrBELzrv
A
Hi
everyone
it's
april
eighteenth.
This
is
the
weekly
Jupiter
deb
meeting
and
we
have
our
notes
meeting
notes
over
in
dropbox
paper.
The
link
is
in
the
comment
section
beneath
the
video,
if
you're
watching
this
later
and
hi
to
the
team
online
right
now,
so
we
usually
start
off
the
meeting
by
going
over
action
items
from
last
week.
A
So
you
might
see
notes
from
us
notifications
from
us
asking
if
we
can
close
issues.
I
think
I
signed
a
couple
out
yesterday
or
last
week
and
we'll
probably
see
more
of
those
from
us
over
the
next
and
next
week,
or
do
also
grant
shared
with
me
last
week
that
when
we
assign
issues
to
the
backlog
milestone-
and
this
is
true
for
all
issues
that
are
assigned
to
the
milestone-
you
can
actually
drag
and
drop
the
issues
within
the
milestone
view
to
start
prioritizing
the
work.
A
So
if
we
were
folks
aren't
using
that
feature
right
now,
you
can
you
can
check
that
out
in
any
of
the
words.
By
going
to
your
milestone
and
starting
to
arrange
your
issues
by
you
know
any
order
that
you,
like
I,
think
in
the
notebook,
the
notebook
repo
we're
organizing
them
by
priority
and
then
knocking
them
out
one
at
a
time
kind
of
combine
style
slowly
from
the
top.
A
So
that
is
thanks
for
filling
that
in
Fernando
target
is
about
270
open
a
shoot
by
the
end
of
april
right
thanks
for
making
that
more
explicit
great.
So
we
also
have
some
other
action
items
from
April
or
a
pre
as
it
is.
This
week
we
have
our
Jupiter
lab
extension
handling
is
going
undergoing
a
major
change
and,
let's
see
do
we
have
Steve
on
the
call
yep
great.
A
So
the
github
issue
is
in
the
Jupiter
lab
org
under
the
Chi
burlap
repo,
it's
issue
2011,
and
it
says
here
once
the
plans
we
would
like
others
to
keep
the
tires
on
it
by
writing
extensions,
so
feel
free
to
start.
Brainstorming
extensions
you
like
to
add
to
test
out
extensions,
okay
and
you've,
your
lab
and
then
finally,
Matthias
is
still
looking
for
feedback
for
a
sink
ruffle
implementation.
Oc
Matthias,
are
you
on
nope,
so
we'll
need
to
figure
out.
Matthias
usually
sends
out
the
meeting
notes
at
the
end
of
the
meeting.
A
So
if
that's
something
that
you
know
we
want
to
continue,
we
don't
don't
have
a
point
person
to
pick
up
when
he's
not
here.
So
if
we
could
we'll
work
on
the
Berkeley
and
this
round
to
figure
that
out,
but
hopefully
in
the
future,
when
Matthias
is
not
able
to
attend,
we
can
identify
different
folks
to
pick
up
that
task.
A
Alright,
that's
it
for
the
action
items.
Next,
we
have
our
weekly
news
by
organizations
section,
so
team
leaves
feel
free
to
share
the
information
and
in
each
of
these,
these
orgs
a
Python,
the
ipython
Jupiter
Jupiter
lab
Jupiter
widgets
in
Jupiter,
hub
or-
and
it
looks
like
we
have
a
new
note
here.
Weekly
news
is
available
via
the
dropbox
paper
files
or
stored
in
this
news
repo.
A
So
I'm
guessing
carol
added
that
which
is
super
useful
thanks
Carol
next
in
the
project
management
section
like
I,
said
earlier
today,
I
attended
a
talk
by
Facebook
last
week
and
I
would
love
to
pursue
their
buck
dashboard
where
they're
running
some
mouth,
NLP
algorithms,
under
its
the
corpus,
the
text
corpus,
but
there
there
are
bugs.
So
we
would
love
to
incorporate
that
into
our
management
of
our
our
issues
as
possible.
All
right
ipython,
if
you
Celyn
here,
will
move
into
notebook
and
that's
grant
and
Thomas
you
guys
here
today.
A
B
Just
a
minor
logistical
thing,
but
some
people
may
have
noticed
tests
have
been
failing
on
on
the
notebook
repo
starting
yesterday.
Maybe-
and
that's
because
of
a
release
of
tornado,
it
doesn't
appear
to
affect
any
actual
code,
but
we
were
had
some
stuff
to
run
an
old
version
of
web
sockets
on
the
server
in
order
to
work
with
phantom
jf.
Fortunately,
that
was
to
work
with
an
old
version
of
phantom
jazz
and
we've
updated
so
that
we
actually
don't
need
that
bit
anymore.
So
it
was
fixing.
B
It
was
just
a
matter
turning
off
some
special
casing
for
our
test
environment.
So
now
testing
is
actually
more
like
the
real
world,
so
everything
is
better,
but
if
people
are
seeing
test
failures
over
the
last
few
days
sets
that's
probably
a
10-6,
the
master
now
so
a
few.
There
are
a
few
local
PRS
that
are
probably
fine,
but
that
has
red
marks
next
domes
that
just
need
to
get
a
rebase
tour
or
something
to
figure
filled
with
the
new
version.
A
C
So
last
week,
I
made
a
PR
that
sort
of
solidifies
the
work
that
I
was
doing
on
getting
custom
configuration
files
so
that
you
can
have
in
browser
like
properly
can
durable
NB
convert
actions
being
triggered
so
that
exists
now,
there's
a
lot
of
you
guys
stuff
that
still
needs
to
happen,
but
that
I
mean
if
we
get
this
in
there.
It
dramatically
improves
our
story
for
easy
usable
customization
of
output,
closer
to
the
lines
of
markdown,
notebooks
and
whatnot.
C
Additionally,
because
of
the
way
that
I
had
set
it
up,
it
sort
of
obviates
the
need
for
the
download
as
menu,
because
you
have
to
choose
a
file
format
and
if
you
just
don't
kind
of
same
a
custom
to
a
custom
configuration
file.
But
you
could
lend
your
running
the
download
as
and
the
way
that
I'm
doing
it
now
won't
trigger
a
download
in
another
window
and
I,
don't
know
about
everyone
else.
But
that
is
a
much
nicer
use
case
like
I.
C
Don't
if
I'm
downloading
I
mean
I
know,
it's
come
up
more
for
me,
because
when
I'm
debugging
stuff,
opening
a
bunch
of
windows
is
not
fun,
but
just
in
terms
of
life
usability
having
something
appear,
you
download
it
and
you're
still
back
right
in
the
environment
that
you
already
wearing
sort
of
ideal
and
yeah.
So
I'm,
looking
forward
to
feedback
on
that,
I've
already
got
quite
a
few
comments
and
I'm
looking
forward
to
discussing
them
and
yeah.
That's
where
a
lot
of
my
otherwise
a
lot
of
find
yourself.
Questions
like.
D
E
So
for
I
think
the
only
thing
I
report
there's
for
Colonel
gateway,
we're
considering
making
a
pseudo
release
I
opened
an
issue
with
a
few
items
for
discussion.
Rowland
Weber,
who
is
an
IV
ever,
has
been
contributing
more
there.
So
somehow,
if
you're
doing
forward
left
so
that
each
time
did,
you
know
we're
kind
of
run
into
growl.
We
think
we
need
to
do
to
get
it
out
before,
but
the
link
is
in
papers
if
anyone
else
wants
to
take
a
peek
feel
free.
E
D
It
yeah
I
was
saying
that
that's
a
pretty
important
goal.
I
was
wondering
what
do
you
know?
Who
else
is
using
Colonel
gateway
and
how
we
can
kind
of
it's
a
very
nice
abstraction
layer
for
a
lot
of
things
and
up,
but
I
suspect
a
lot
of
people
may
not
be
using
it
because
they
don't
even
know
about
it.
So
I'm
wondering
what
we.
E
I
think
one
of
the
open
issues
that
someone
reported
about
a
month
ago,
like
he's
getting
right
to
the
heart
of
that
which
is
it,
was
a
documentation
request
saying
you
know
there
were
these
great
blog
posts
on
the
Jupiter
blog
on
the
IBM
et
blog,
that
more
clearly
explained
or
proposed
to
explain
likely
the
way
to
say
what
the
colonel
gateway
is
and
what
it
does
and
it's
great
to
get
that
documentation.
It
could
simply
be
that
you
know
people
are
coming
to
is
seeing
and
going.
E
D
D
My
reflexes
becomes
go
to
the
reason
top
stage
and
figure
out
what
it
is,
and
so,
if,
if
some
of
that
blog
material
was
cause
well
kind
of
thing
to
the
right
diction
is
putting
into
the
docs
will
be
already
a
big
help
and
once
that's
done,
then,
maybe
a
little
blast
them
in
with
the
hate
people.
This
is
not
well
documented,
please
let
us
know
if
it's
useful
might
have
the
right
the
right
effect.
Yes,.
A
Great
all
right,
thank
you.
Okay,
great
we're
getting
some
more
text
added
to
the
service
sixer
vez
of
section
Jupiter
lab
see
Darien
and
Chris.
F
That's
gone
pretty
well.
Last
week
we
merchant
to
master
the
single
document
mode
that
dropped
in
phosphor
couple
months
ago,
and
so
now
you
can
toggle
between
having
the
back
panel
having
you
know,
all
your
tabs
or
just
showing
one
document
at
a
time
and
then
the
sidebar
there's
a
in
the
left,
sidebar
there's
a
tab
widget
that
shows
you
all
the
open
all
of
the
open
documents
and
other
things
that
are
sitting
in
your
dock
down
that
you
can
switch
between.
F
So
if
you
try
it
out
and
have
some
feedback,
there's
a
feedback
issue,
that's
open
and
we
would
appreciate
any
comments
you
have.
The
extension
work
is
coming
along.
It's
dramatically
simpler
than
the
original
proposal
for
how
we
would
do.
We
want
to
use
just
off-the-shelf
components,
and
the
support
for
the
base
use
case
will
will
will
require,
having
noted
an
MPN
on
your
machine
and
also
an
internet
connection,
but
we're
thinking
through
how
to
deal
with
the
air-gap
scenario.
G
Just
just
briefly
we're
trying
to
define
an
abstraction
for
document
models
that
different
backends
can
implement
that
maybe
in
memory
or
real-time
enabled
so
there's
a
PR
open
on
that.
Welcome
comments
that
the
basic
idea
is
to
have
the
model
database
be
able
to
provide
various
observable
objects
like
strings
and
lists
and
objects
that
can
be
implemented
by
various
backends
cool.
F
And
one
last
thing:
that's
not
in
in
the
document
but
may
be
worth
mentioning
is
chris
is
currently
working
on
the
grid,
widget
and
phosphor,
and
the
original
the
initial
release
is
going
to
be
a
read-only
mode.
So
one
thing
that
we
could
use
that
for
even
even
in
just
read
only
mode
would
be
the
way
we
render
tables
and
notebooks
and
it
would
solve
it
solve
the
problem
of.
F
What
do
you
do
when
you
have
a
gigantic
table,
because
right
now
we're
just
rendering
HTML
and
that
creates
a
mess
problem,
but
the
the
phosphor
grid
will
be
able
to
handle
both
they
would
do
virtual
scrolling
and
it
would
just
be
more
efficient
and
the
resources
going
to
require
concept.
So
that's
something
that's
coming
down
the
pike
and
you
should
keep
an
eye
out
for
and
that's
it
do.
F
Any
any
static,
rendering
that
is
going
to
be
able
to
handle,
say
static
versions
of
other
widgets.
It's
going
to
be
able
to
handle
this
as
well.
It's
not
it's
just
the
way
cells
get
rendered
for,
like
you
like
the
way
I
would
just
let
you
have
that
interactivity
without
actually
talking
to
the
server
whatever
mechanism
that
allows
for
that
should
allow
for
table
runners
as
well,
but
I
don't
have
an
immediate
answer
because
we're
just
not
there.
F
H
So
the
I
can
comment
on
that,
and
this
is
moving
towards
having
both
a
rep
for
HTML
and
a
json
table
schema
and
both
of
those
would
be
published
by
pandas.
And
yes
for
not
a
week
like,
we
could
render
the
table
schema
using
the
the
grid
component
and
then
still
have
HTML
output
available
for
static
renderings.
Since.
I
D
Great
Chris,
I
I
just
found
the
pipe
I
package
for
the
pandas.
The
published
schema
I'll
added
to
the
notes
right
now.
Great
thanks,
well,
I
also
wanted
to
say
to
darien,
see
you
and
everyone
else,
but
I
in
particular,
Stephen
dairy
and
have
really
toiled
super
heart
of
the
from
the
expansion
stuff
that
work.
D
We
know
that
that's
a
it's
a
nasty,
complicated
messy
problem
kind
of
in
the
its
high
effort,
low
reward
and
defense
I
mean
for
the
developers
in
the
sense
that
is
you're,
doing
infrastructure
plumbing
without
any
kind
of
visible
output
quickly,
and
we
really
appreciated
it.
So
it's
a
super,
hard
nasty
problem.
I
know
you
guys
have
been
fighting
on
it
for
quite
a
while,
and
so
don't
worry
about
it.
No
one!
D
J
We've
been
cleaning
up
a
lot
of
bugs
and
and
again
pushing
hard.
We
hope,
through
this
a
beta
set
point
seven
point
out
pretty
soon
again.
J
And
so
yeah,
the
big
thing
for
this
last
week
is
that
we
are
widgets,
are
now
output
elements
instead
of
special
things
inside
the
widget
container
and
we're
working
hard
on
getting
output
areas,
output
widgets
to
work
well
in
the
embedded
context.
So
that
means
that
interacts
will
display
and
in
the
end,
be
viewer
and
the
safe
notebooks.
D
J
J
D
Okay,
the
biggest
is
just
to
have
that
information
in
the
kinder
dollar
and
the
release
out
should
I'm
not
saying
it's
a
problem
for
that
to
be
the
case
so
just
to
have
it
documented
so
that
at
least
people
know
now,
when
we
can
point
it
out,
it
says
that
look,
that's
the
version.
You
need,
that's
just
what
it
is.
Don't.
A
K
The
work
on
getting
some
of
the
data
eight
stuff
documented
into
a
generic
formats
ongoing
and
lots,
were
cut
down
on
that
last
week,
which
is
great.
Thank
you,
the
folks
from
bids
and
UV
and
others
that
commented
I
am
going
to
be
doing
some
work
on
getting
all
the
great
Bob
stuff
that
men
did
recently
in
Jupiter
have
documented
and
Pete
very
kindly
offered
to
review
stuff
as
well.
So
that
will
be
helpful
and
I
don't
have
a
whole
lot
other
than
that.
What
about
you?
Men,.
K
And
then
one
little
miscellaneous
thing:
Tony
D
from
Adafruit
and
just
released
a
really
nice
video
on
Friday
about
how
to
write
a
Jupiter
kernel
and
wrote
one
for
micro,
python
and
sort
of
he
and
I
met
gosh.
Now
it
was
back
in
I,
can't
remember
when
but
talked
about
the
notebooks
and
he
did
a
longer
video
last
year,
but
thinking
that
adafruit
may
see
notebooks
as
a
good
way
of
doing
tutorials
and
things
like
that.
So
it's
worth
checking
out
lots
of
positive
feedback
and
he
really
liked
the
rapper
colonel.
So.
D
K
He
rewrote
he
wrote
one
that
was
specific
to
micro
Python,
not
to
the
micro
bit.
So
I
don't.
I
don't
know
if
he
used
any
of
Thomas's
stuff.
It
didn't
look
like
it
from
looking
at
his
code
and
the
way
he
set
up
the
connections
to
the
board,
but
I
know
he's
they
really
like
Jupiter
over
at
Adafruit.
So
it's
a
good
thing
and
I'll
probably
see
them
in
over
the
next
couple
days
to
when
I'm
in
New
York.
So
so.
D
K
It
horse
/
dev,
it
is,
it
was
kind
of
his
act
way
of
doing
it.
So
there's
no
guarantees
it'll
work
on
anything,
but
it
was
sort
of
a
proof
of
concept
that
hey.
This
is
a
cool
way
to
interact
with
hardware,
and
you
guys
can
do
this
too
so
and
he'll
be
at
PyCon
as
well
at
the
education
summit,
so
I
suspect
we'll
hear
a
little
more
there
too.
Oh
cool.
D
I've
been
playing
with
Thomas's
on
on
the
micro
bit
and
I
like
it.
I
do
like
it
but
abate
and
in
fact,
I've
been
going
back
and
forth
between
the
micro,
great
editor
in
the
notebook
and
I
think
with
with
a
couple
of
tweaks
to
how
the
notebook
interacts
with
the
microbe
it,
it
would
be
way
better
than
the
standard
ID
there's
a
couple.
Things
like
very
minor
convenience
think
that
we
could
do,
and
it
would
be
a
really
really
nice
solution.
Yeah.
K
I
actually
wrote
one
for
circuit
python,
which
is
a
de
fruits
fork.
That
kind
of
follows:
micro,
pythons
development,
which
is
really
targeted
as
being
education
friendly,
even
more
so
than
micro,
Python,
but
first
and
foremost,
so
so
yeah
there's
good
stuff
happening
there.
So
I
think
I
see
Tony,
maybe
doing
something
in
the
future.
D
Excellent
answer:
thanks
for
letting
us
know,
maybe
maybe
sometime
later
we
could
happen
to
a
little
teen
tutorial
about
that
stuff.
I
think
I
notice.
His
notion
of
having
interactivity
with
hardware
is
a
very
fun
fun
and
actually
potentially
valuable
a
vein
that
we
could.
We
could
mine
a
little
bit
more
yeah.
A
B
L
A
And
then
and
see,
could
you
add
your
releases
to
the
sections
below
so
I
did
something
so
much
much
fun.
I
did
for
Thomas's
request
I
added
in
another
small
section
on
the
release
police's,
because
it
was
a
little
ambiguous
as
to
which
what
releases
were
supposed
to
be
listed.
So
I
have
a
section
for
things
that
we
released
last
week,
but
we
didn't
announce,
announce
them
and
then
anything
that
we're
planning
to
release
this
week
so
today
going
forward
and
then
releases
soon.
A
So
those
are
the
more
the
ones
that
we're
planning
to
release
them,
but
we
don't
really
have
a
date
for
them
yet,
alright,
so
dumb
ants
on
the
call
as
well
hi
Dom
ian,
so
glad
that
you
could
join
us
today.
We
have
a
really
good
group.
Today
we
have
about
14
people
here
Nick's
here,
so
good,
hey
Nick,
let's
see
so.
The
next
section
is
the
conference's
outreach,
so
quoc
khanh
is
May
second
to
fit.
A
So
that's
going
to
be
here
in
Oakland,
and
then
we've
already
announced
this
big
long
list
of
other
conferences,
Carol's
going
to
head
over
to
New
York
City.
In
a
couple
days,
probably
tomorrow,
actually
for
colon
with
Sofia,
so
I
should
be
really
interesting
from
watch
the
videos
later
I'm
guessing
we'll
be
live
stream.
K
A
Perfect
and
then
we've
got
craft
2017
in
Budapest,
April
26
to
28
and
my
glutes
excellent,
giving
a
tutorial
on
I
by
widgets
and
then
more
hiding
in
London
darienne's,
going
to
be
there
and
Jupiter
day,
silly
May
nineteenth.
So
if
you
go
over
to
our
event,
calendar
in
github
project
you'll
see
that
it
may
is
it
busy
month
for
our
team,
so
firm,
buddy
here.
So
it's
fun,
icon
coming
up
as
well
rely.
D
A
We've
had
a
couple
meetings
where
we
walk
through
it,
but
that's
a
good
idea
to
put
the
link
right
there
in
the
notes.
I'll
do
that
after
the
meeting
since
I'm,
going
off
to
different
devices
right
now,
great
Jupiter,
con
soft
launch
of
registration
started
this
week.
There's
a
link
and
Fernando
says
they're
finalizing
the
talk
section.
Do
you
have
anything
you
want
to
elaborate
on
their
Fernando?
No.
D
No
is
just
kind
of
internal
internal
finishing
that
up
and
we
should
be
done
in
a
couple
of
days
and
we're
making
good
progress.
The
o'reilly
team
has
been
has
been
on
top
of
things,
and
that
was
over.
The
one
task
that
I
could
not
completely
ignore,
while
I
was
in
family
leave,
was
was
kind
of
put,
pushing
pushing
that
through
and
Andrew
auto
on
and
Brian
were
huge
health
and
very
flexible
with
me.
So
appreciate
that
ok.
A
Great
so
great
I
mean
and
Carolyn
check
into
action
items,
so
we've
got
a
request
from
a
contributor
on
get
her
to
have
a
landing
page
for
dropbox
paper
weekly
for
the
weekly
meeting
notes
that
the
link
is
consistent
weeks
a
week,
I
think
that's
a
great
idea
and
I
will
work
to
get
that
done
by
the
next
meeting,
starting
to
give
it
them
by
today.
If
you're
watching.
A
So
then,
this
week
we
have
Jupiter
hub
NVR
session
proxy
released
Monday
for
17
and
Jupiter
lab
zero
point
20
with
single
document
mode,
then
releases
that
will
be
doing
soon.
I
python,
that
so
that
hasn't
been
updated
in
awhile
and
a
really
candidate
was
released,
says
last
week,
but
I
believe
that
was
a
couple
weeks
ago.
At
this
point
and
then
an
NB
dime
0.3
get
rest
support
will
be
released
soon
as
well
as
been
mentioned
all
right.
Well,
that's
the
end
of
our
note.
Tears
is
anybody
else.