►
From YouTube: Jupyter/IPython Dev Meeting, May 9, 2017
Description
Meeting of the Jupyter/IPython development team, May 9, 2017
Meeting Notes: https://paper.dropbox.com/doc/May-2017-Jupyter-Weekly-Meetings-NaidawTLtrI06iFJzIdwZ
A
A
Next,
we
have
some
older
action
items.
So
those
of
you
who
are
on
the
call
in
the
meeting
this
week
and
haven't
attended
in
the
last
couple
weeks,
it
would
be
great
to
check
some
of
these
office
if
these
are
complete.
So
first,
if
you
have
thoughts
on
user
stories
for
a
configurable
exports
from
the
notebook
interface
take
Mike
next
I
PI
widgets
need
to
check,
needs
to
check
as
minimal
notebook
dependency
tests
and
document
it.
A
So
that's
something
that
we'll
do
as
soon
as
Jason
and
Sylvan
and
team
hits
the
I
PI
widget
7.0
the
beta.
So
we'll
look
back
in
on
Lestat's
release,
then
grant
and
I
have
been
working
through
the
Jupiter
notebook
issue
log.
We
close
out
a
hundred
issues
last
month
in
April,
so
we're
going
to
keep
hitting
that
keep
working
toward
that
goal
and
that's
about
270
open
issues
by
the
end
of
May.
A
We
seem
to
be
adding
about
50
to
60
new
issues
a
month.
So
if
we
can
at
least
get
a
hundred
done
a
month
will
eventually
get
ahead
of
ourselves.
But
I
don't
know
if
you
guys
are
familiar
with
the
Red
Queen
hypothesis.
It's
basically
run
as
fast
as
you
can
to
stay
in
the
same
place,
it's
starting
to
feel
like
that
that
model
of
evolution
currently
so
hopefully,
we
can
check
that
off
soon.
Next
SuperLab
extension
handling
underwent
a
major
taint
change
in
point
to
zero.
A
C
A
Okay,
I
gotcha:
are
you
able
to
capture
action
items
or
should
we
move
okay?
Cool?
Thank
you
so
much
wonderful!
Alright!
Next
we
have
our
weekly
news
buy
organization,
so
folks
who
folks
from
the
community
who
are
interested
in
these
different
project
areas.
I
python
Jupiter
Jupiter
lab
Jupiter
widgets
in
Jupiter
hubs.
You
can
click
on
these
links
to
see
what
we've
been
up
to
over
the
last
week
and
of
course,
team
leads
you're.
A
Welcome
to
use
that
information
to
report
on
your
work
today
and
then
moving
into
the
different
project
areas,
the
project
management
side.
Things
have
been
winding
down
a
little
bit
at
Berklee
this
last
week,
so
it's
in
fairly
quiet
I've
been
working
with
Fernando,
Brian
and
Anna
on
the
team
meeting
agenda,
plus
some
other
things
that
we've
had
it
in
the
works
for
a
long
time
trying
to
wrap
those
things
up
and
much
all
I
have
for
this
week
under
project
management.
So
let's
go
to
notebook
grants
and
Thomas
madly
editing.
You
guys
ready.
A
Wait
that
sounds
great
Mike.
We
got
a
lot
of
head
nod.
Silent
nod
sounds
good,
good
response.
Okay,
the
best
time
yes,
no
finger,
wags.
All
right
great!
Did
you
want
to
move
into?
We.
E
Talked
with
a
good
number
of
people
last
week
about
about
how
to
solve
the
problem
of
dropping
through
tech
straw
from
within
mark
themselves
without
having
to
be
in
a
row
and
be
convert,
sell
plenty
of
discussions
about
the
dangers
of
extending
markdown
and
all
of
that,
but
nonetheless
we're
at
least
for
the
site.
Hi
proceedings.
It
looks
like
it's
going
to
be
needed
in
order
to
or
something
like
it
will
be
needed
in
order
to
get
the
proceedings
up
and
running
in
a
reasonable
amount
of
time,
and
it
may
be
useful
to
precedent.
E
E
That
was
another
use
context
for
hypothesis
who
sort
of
from
the
publication
point
of
view.
So
sensitive
nd
convert
that's
closest
thing.
We've
got
so
for
one
of
the
problems
with
notebooks
on
github
because
they
are
rendered.
You
can't
do
line
comments
on
them,
because
there
are
no
lines
and
one
of
the
solutions,
so
that
might
be
using
something
like
hypothesis
annotations,
to
highlight
the
appropriate
amount
of
the
text
and
then
have
that
linked
into
some
kind
of
github
automation,
tool.
E
F
And
I'll
quote:
if
I
might
throw
a
quick
thing
on
the
Monday
convert
angle,
my
initial
experiments
with
using
headless
Chrome
PDF
generation
are
okay.
It
does
still
seg
both
all
over
the
place,
but
it
is
Chrome
Canary.
So
that's
a
very
fun
but
I'm
very
bullish
on
our
ability
to
make
good
use
of
that
for
a
non
check
workflow
for
generated.
Yes,.
E
F
A
Gastral
great
thanks
thanks
Mike
right
services
like
Colonel
gateway
doctor
stacks
MB
viewer,
none
today,
so
we
might.
We
have
something
up
in
action.
Items
with
them
be
viewer,
but
there's
no
work
to
report
today,
dashboard
and
related
extensions-
that's
usually
piqued
he's
not
here
today
either.
So
we
will
continue
down
the
list
to
Jupiter
lab
team,
Darien
and
Chris.
If
you're,
here,
okay,.
G
This
past
week,
I
worked
on
another
iteration
of
the
extension
system,
hopefully
moving
closer
to
the
final
implementation
based
on
some
feedback
from
Sylvan,
who
is
our
first
person
trying
to
make
an
extension
that
uses
another
extension.
So
thanks
for
that
feedback
that
will
be
released
today
as
super
lab
21.
G
The
other
things
that
the
other
activities
that
moon
feed
into
that
is
brian,
has
been
moving
forward
with
notebook
cell
design
and
parts
of
that
will
be
in
21
as
well
past.
That
nick
has
been
working
on
using
yarn
as
a
replacement
for
NPM,
because
yarn
allows
us
to
do
is
completely
offline
installs,
so
it's
offline
mode
actually
doesn't
try
to
talk
to
registry.
If
you've
got
all
the
files,
whereas
NPM
will
so
we
think
we
should
be
able
to
get
that
to
work,
but
he's
still
working
through
some
of
the
issues.
G
Placement
of
file
equals
installing
of
of
Cooper
lab
extensions,
so
that
a
comma
package
or
a
wheel
can
just
put
a
file
in
place.
We
pick
that
up
and
rebuild
the
app
as
necessary,
and
then
the
user
can
just
refresh
the
page
with
that
installed
app
without
having
to
do
anything
else
on
their
own,
so
making
network
flow
work.
The
offline
install
work,
I
think
that
shows
a
lot
of
promise,
so
that
makes
is
forging
ahead
with
that
part.
G
Darian
is
working
on
the
Settings
system,
implementation,
with
the
initial
being
just
using
some
private
files
in
the
directory,
where
the
notebook
server
is
launched
for
the
settings.
Eventually,
that
being
moved
to
is
more
like
what
the
current
NV
config
settings
system
looks
like,
where
it's
actually
ready
to
no
known
other
locations
on
disk
and
I.
Think
that's
it
for
people
that
anybody
else,
wanna
chime
in
I
got.
G
From
what
was
there
now,
it's
just
a
slight
suit
should
be
just
slight
CLI
changes
in
terms
of
how
you
launch
it
so
I've
updated
the
contributing
guide
and
the
other
tutorial
guides
to
reflect
the
changes
in
workflow.
There's
NPM
is
no
longer
part
of
our
PIP
command.
So
that's
those
separated
now
and
there's
the
the
way
the
fallback
core
Jupiter
lab
is
handled
is
slightly
different
than
it
was
before
the
change
that
accommodated
what
Sylva
needed
was
that
you
can
now
link
in
a
package.
That
is
not
an
extension.
G
D
D
G
F
Our
yarn
right
and
nobody
else
can
mess
with
it
and
then
it's
an
implementation
details
the
outsider,
so
you
basically
want
to
end
up
where
you
have
a
valid
NPM
package.
However,
you
get
there,
we
don't
care
right,
but
then
once
it
crosses
over
yarn
takes
care
of
it.
It
once
you
put
it
into
the
system,
so
yep.
G
F
Now
and
then
again
with
it
yeah
with
the
as
we're
thinking
about
this,
we'll
probably
end
up
with
a
Jupiter
lab
dev
sub-command
tree,
you
know
the
one
that
actually
requires
all
the
yarn
junk
and
with
that,
the
way
that
I've
got
it
set
up,
live
editing
of
your
extension
if,
for
example,
you're
just
editing
straight
yes,
yes,
but
renders
in
under
two
seconds
as
though
you're
into
that
loop
of
packing
I'm,
writing
and
packing
and
writing.
So
it
has
a
lot
of
potential.
We.
F
So
dirty
I
don't
know
how
to
solve
it
for
pip,
but
you
know
we
can
look
at
time
an
eraser
or
whatever
the
right
hack
is
to
get
node
on
on
a
pure
pip
install
or
maybe
we
can't
support
it.
I
don't
know,
but
Wakanda
we
build
slug
for
everything,
including
web
pack,
Co
CSS,
all
that
noise,
no
typescript
is
17
megabytes.
So
it's
17
megabytes,
tar
balls
that
basically
doesn't
compact
any
further,
which
I
think
is
not
unreasonable
and
then
the
individual
extension
packages
have
no
over
it
at
all.
H
F
F
Sure
yeah,
just
what
can
we
push
the
build
time
right?
What
can
we,
what
can
we
actually
have
accomplished
statically?
So
you
know
if
all
of
these
second
party
Jovian
widgets,
if
you
want
an
environment
with
all
those
that
should
be
no
op
or
not
in
a
lot,
but
it
should
be
download
this
one,
five
megabyte
file
or
ten
or
whatever,
however
big
it
is,
but
yeah
works
for
developer
first
for
sure,
so.
H
One
thing
that
happened
to
us
when
we
were
trying
to
base
everything
for
a
demo
was
that
some
similar
dependency
had
a
breaking
change.
We
need
Union
minor
version,
so
they
were
birthing
center,
and
so
this
isn't
see
that
we
noticed
as
install
time
instead
of
their
time.
So
with
your
way,
were
we
ship
a
bundle,
only
the
packager,
actually
producing
good
bands
or
would
have
noticed
the
breaking
change
in
the
second
of
the
dependency
were
here,
it's
anyone
who
is
closes.
So
it's
more
free
fun.
I
So
in
this
case
it
would
have
been
fine
if
you
were
the
only
one
shipping
the
dependency
and
you
were
on
the
old
non
broken
version
as
soon
as
somebody
else
to
use
the
better
broken
version
that
violated
some
very
would
have
broken
for
the
runtime
user
as
well.
So
it
was
only
one
sitting
package,
it
so
yeah,
but
it's.
J
H
When
we're
all
over
to
San
Francisco
at
last
week
and
we've
had
a
demo
with
might
at
the
data
center
and
then
they
talk
about
Ida
widgets
at
threaten
which
one
really,
though,
and
that's
when
we
worked
on
updating
everything
to
the
latest.
If
that
happens,
that's
why
we
had
this
feedback,
also
yeah
with
just
a
little
educational
environment.
H
C
Man
today,
I'm
in
is
a
traveling
and
everybody
is
gearing
up
for
Pike
on.
It
will
be
an
amazing
showcase
of
all
things.
Jupiter
we've
got
a
lot
of
our
developers
going
to
be
there.
It
kicks
off
in
Portland
on
Wednesday
with
tutorials,
and
there
are
quite
a
few
Stata
science,
scientific
programming
ones
as
well
as
Jupiter
and
ipython
specific
ones,
and
talks
at
the
education
summit.
C
So,
if
you're
at
PyCon
check
them
out,
if
not
the
videos
should
go
up
pretty
quickly,
so
you
can
watch
from
home
and
last
year
we
did
some
open
spaces
with
that
Allen,
Downey
and
Luciano,
and
under
is
Lanka
just
really
informal
for
an
hour
answering
questions
that
hopefully
will
do
that
as
well
and
feel
free
to
tweet.
At
me,
if
you're
there
are
looking
for
Jupiter
books
and
as
far
as
Jupiter
hub
we're
just
moving
right
along
Thanks,
great.
B
A
A
A
Thomas
has
disappeared;
well,
I'll
read
it
so
I,
don't
I'm,
not
sure
how
much
you
want
to
get
into
this
Matthias,
but
it
says
we're
getting
various
issues
reported
with
Jedi
completions,
principally
around
performance.
Do
we
have
a
reasonable
expectation
of
getting
this
properly
solved?
Is
it
a
question
of
wacking
performance
moles,
one
by
one
I've.
B
Been
working
on
that
and
I
have
some
time
today,
master
is
already
much
faster,
isn't
stable
and
people
can
tweak.
We
have
option
I
by
count
two
to
change
how
much
time
we
give
to
Jedi
to
do
completion,
but
it's
also
a
question
of
some
upstream
libraries
are
doing
weird
stuff
that
you
technically
should
not
be
doing
Python,
but
you
know
doing
stuff
that
you
shouldn't
be
doing
is
not
something
away
with
that.
B
A
Right
thanks
I'm
channeling
Thomas.
Today
he
is
only
listening,
so
Thomas,
you
have
further
questions,
feel
free
to
post
them
in
the
chat
and
we
can.
We
can
verbalize
them
here
in
the
meeting
right.
You're
welcome.
Okay.
Next
we
have
another
section
with
plumbing
and
packages
etc,
and
usually
the
one
who
fills
out
that
section
so
he's
not
in
the
call
today
or
on
the
meeting
in
the
meeting
today,
so
we'll
skip
through
that
governance.
Brian.
Do
you
have
any
government
governance
updates
that
you'd
like
to
share
with
the
group
canola?
A
K
Although
one
thing
I'm
going
to
one
here
in
New,
York
talk
with
Jason
the
other
Joker
lab
team
about
implementing
some
of
the
changes
to
be
labeling
system,
we're
using
some
of
that
based
on
Jupiter
labs
evolution
to
using
lerna,
and
now
we
have
a
bunch
of
different
endgame
packages.
So
what
I'm?
Imagining
will
probably
evolve
our
current
component
prefix
into
something
like
that
basically
map
it
on
to
the
packages
that
we
have.
K
K
The
milestone
and
I
think
that's
a
really
nice
more
cleansing
way
of
closing
issues
and
indicating
to
people
that
they
may
still
be
useful,
they're
just
not
currently
on
our
radar
and
so
I
I
think
I'm
going
to
propose
doing
that
in
Jupiter
lab
as
well,
and
it's
an
all
updates.
A
before
request.
I
have
describing
research,
the.
J
K
J
A
A
There's
quite
a
there's.
Quite
a
few
decision
points
in
what
we
want
to
do
with
the
issues
and
where,
if
you
store
them
after
we
take
action
on
them,
is
definitely
something
that
will
you
know.
Team
will
want
to
massage
out
over
time,
because
I
think
the
evidence
of
the
decision
doesn't
get
left
in
the
issue,
so
finding
it
finding
a
group
of
things
that
you
made
a
decision
on
it
can
be
a
little
bit
a
little
bit
challenging.
So
so
it
sounds
good.
A
Next,
that
was
governance
next
week,
conferences
and
outreach,
I'm
gonna
skip
to
the
team
meeting
the
spring
team
meeting,
so
Anna
is
listening
in
today,
she's
in
a
coffee
shop,
and
it's
too
noisy
so
I'm
going
to
read
the
information
that
she
wanted
to
share.
So
Brian
Fernando
on
and
I
are
putting
together
an
agenda
and
we're
at
the
point
where
we
can
share
that
document
with
the
team.
A
Also
Jupiter
Khan
registration
is
open
and
I'd
like
to
ask
the
team
members
here
in
the
meeting
to
go
ahead
and
follow
and
retweet
Jupiter
Khan
on
Twitter
and
that's
just
twitter.com.
Slash
do
and
in
any
events
that
you
go
to
over
the
next
couple
months,
go
ahead
and
mention
the
event
Jupiter
Khan
the
dates
and
the
website,
and
it
looks
like
on
I,
have
postcards
and
other
marketing
materials
available
and
let
her
know
if
you're
interested
in
distributing
those
should
she
can
ship
those
out
to
you.
A
Anna
says
O'reilly
is
tweeting
about
to
program
so
for
those
who
are
watching,
Riley
is
working
with
us
to
produce
our
first
super
Khan
and
they've
been
incredible
and
extremely
instrumental
in
moving
everything
forward
to
have
this
event
launched
in
August.
So
it's
been
great
okay,
so
for
other
conferences
coming
up,
we
have
pi
day
to
Paris.
A
A
So
as
as
sylvans
team
members
is
our
job
to
help
promote
pi
day
to
paris
and
help
support
someone
in
this
to
have.
H
A
Great
pack,
the
house
great
all
right,
Jupiter
day,
Phillies
coming
up
May
19th.
Of
course,
dr.
Khan
is
coming
up
starting
next
week.
Jupiter
team
meeting
is
also
at
the
end
of
the
month.
We've
got
pike
on
Taiwan
June
9
through
the
11th
hi
Paris
in
June
Europe
Python
in
July
Syfy
in
July
science
gateways,
community
institute
webinar
and
July
18th
or
20th
Carol's
presenting
and
then
in
October.
A
great
great
operation
and
sugar
will
be
participating
in
open
source
day.
All
right.
A
A
Okay
feel
free
to
interrupt
me.
If
you
do
actions
for
this
week,
we
just
have
one
remember
to
register
for
do.
Pecan
I'd
also
add
tweet
out
chip
pecan
any
announcements
already
promo
stuff.
We
could
add,
get
that
added
added
to
the
action
item.
That
would
be
great
releases
from
last
week:
main
release,
MB
dime
0.3
and
this
week
to
their
lab
zero
point,
two
one
or
point
21.
These
are
updates
to
Jupiter
lab
configuration,
handling
and
sounds
like
it
will
happen
after
the
Deaf
meeting
today
that
not
speech
cool,
alright
thanks
anything
else.
A
Folks,
all
right.
Let's
thank
the
community
if
you
could
stay
on
after
just
briefly
a
couple
more
announcements
method,
thanks.