►
From YouTube: IPython/Jupyter Dev Meeting, May 3, 2016
Description
Meeting of the IPython/Jupyter development team.
A
B
B
So
now
I'm,
working
with
the
last
on
writing
up
a
statement
that
we'd
like
other
projects
to
join
and
say
that
they
will
be
dropping
by
10
2.
I've
passed,
attested
that
the
link
in
the
hot
pad
you
can
go
there
and
read,
see
how
it
looked
like
and
contribute.
If
you
want-
and
I
would
also
like
to
push
ipython
50
frozen
in
the
next
two
months-
I've
been
a
lot
of
change.
B
So
I
would
appreciate
that
if
you
can
triage
issues
and
polish
everything
that
we
can
start
to
freeze
features
and
have
something
which
is
developed
because
it
will
be
the
long-term
support
version.
So
I
think
it
should
be
well
well
polished
and
ready
to
ship
long
before
she
just
not
to
have
to
fix
crazy
break
after
a
while
and
otherwise.
I
would
like
some
experiment
view
on
my
chocolate,
editor
and
switching
to
javascript
six
personal
look
at
that
has
been
ready
for
a
few
days
now
and
would
a
tuna
for.
A
A
So
I
would
like
to
share
that
with
the
group
and
then
y'all
can
make
comments
on
where
the
dates
fall.
One
thing
that
that
Fernando
and
I
discussed
was
breaking
the
work
into
two
month,
increments,
so
put
the
next
into
May
and
June
X
plus
1
into
July
and
August,
and
then
the
future,
and
it
just
goes
into
the
future.
But
we
would
pick
up
with
that
two
month,
increments,
starting
with
whatever
work
is
in
that
future
category.
E
G
Ok,
so
I
am
the
next
week
time
to
get
the
next
version
of
every
greater
released.
It's
been
really
close
for
a
while
I've
been
working
on
fixing
some
issues
with
windows
which
are
fixed,
but
the
tests
are
still
not
passing
and
Jenkin,
so
it's
still
figuring
that
out,
but
but
hopefully,
I'll
have
that
released
sometime
this
week,
I've
also
been
I.
You
know
gathering
responses
for
the
Jupiter
education
survey,
which
you
all
have
probably
seen
at
this
point.
We
have
over
100
responses
now,
which
is
awesome.
It's
going
to
close
on
friday.
G
If
you
know
anyone
else
who
know
should
fill
it
out.
Who
hasn't?
Please
bug
them
for
me
and
then,
when
the
survey
closes
I.
My
plan
is
that
I'm
going
to
do
a
little
bit
of
initial
analysis
on
it
myself
and
then
post
the
data
and
the
analysis
that
I
have
and
then
anyone
who's
interested
in
it
can
do
further
analyses
and
contribute
stuff,
and
there
would
be
awesome
if
we
could
put
together
on
something
like
a
dashboard
of
like
there
was
for
the
UI
survey.
G
A
G
Yeah,
so
we
had
a
meeting
Matthias,
Jamie
and
I
with
some
people
from
Wharton,
which
is
in
Lavinia,
and
they
have
they
have
this
website,
where
they
share
a
bunch
of
data
that
they
have
with
business
schools
all
across
the
country
and
in
the
world,
and
they
have
something
like
forty
thousand
users
of
this
and
they're
interested
in
setting
up
a
Jupiter
hub
deployment
for
their
users,
who
may
currently
just
ssh
in
and
work
through,
ssh
or
they
download
the
data
through
the
website,
but
they're
hoping
to
set
up
a
trigger,
have
deployment
to
make
it
easier
for
other
users
to
access
the
data.
G
H
Hey
josue
yep
Jess.
You
should
probably
at
some
point
Lupin
Jacob
Kaplan
Moss,
because
the
Django
folks
do
a
lot
of
stuff
with
Wharton,
so
something
to
keep
in
mind
for
the
future.
Okay,.
G
H
H
Read
the
docs
just
updated,
not
only
their
suffix
from
read
the
docs
org
to
read
the
docs
I
owe
all
the
docs
still
work
because
they're
redirecting,
but
they
also
just
released
common
mark
zero
point
six
point
three.
So
hopefully
that
will
go
smoothly
and
maybe
will
help
with
the
rendering
of
some
of
the
notebooks
that
are
converted
and
then,
if
folks
want
to
start
taking
a
look
at
the
installation,
QuickStart
section
of
138.
That
would
be
great
I'm.
H
A
I
More
than
what's
written
there
couple
releases,
we
did
actually
do
her
first
mpm
release
at
the
dashboard
server.
So
that's
out
there
in
the
enhancement
proposal
you
could
barely
be
called
in
it.
An
sment
proposal
is
a
sitting
there
waiting
for
comments,
not
quite
clear
on
how
to
proceed
on
that
yet,
but
it
did
beliefs
encouraged
us
to
write
down
metadata
format
for
the
dashboard
extension.
So
not
what
else.
A
report.
I
It
may
one
day
do
dashboards
and
has
Zito
deeds.
Xyz,
for
that
use
case
and
dashboards
are
very
different.
Beast
and
rise
is
going
to
be
another
one
and
a
Indian
and
try
to
come
up
with
one
specification
for
laying
out
any
possible
layout.
You
can
imagine
for
different
use
cases
just
eludes
me
at
least
at
the
moment.
You
know
how
to
write
that
kind
of
spec
feels
like
trying
to
reinvent
the
browser.
So
I'm
not
quite
sure.
I
A
I
There
I
think
Sofia
wrote
a
bot
that
actually
posts
the
mailing
list
as
soon
as
the
enhancement
proposal
goes
out,
so
it
was
on
the
list
week
ago.
We
could
have
to
go
so.
C
I
A
A
I
J
Hey
so
hi
folks
yeah.
So
besides
the
release
of
the
claritin
widgets,
that's
in
the
hack,
pat,
something
that
we're
just
debugging
here,
while
on
the
call
just
wanted
to
bring
it
up,
there
seems
to
be
something
happening
in
Jupiter
4.2
and
it
might
be
coming
from
I
pie,
widgets,
so
Sylvan.
This
might
be
interesting
to
you
we're
seeing
on
a
safe
of
the
notebook.
J
The
cells
are
getting
re-render,
the
output
cells
are
getting
me
rendered
and
we
see
are
our
declarative
widgets,
getting
reinstated
on
just
pressing
the
Save
button,
something
we
didn't
see
in
the
past
and
I'm.
Sorry
bugging
it
to
delete
snapshots,
and
that's
why
I
say:
I
pie
widgets,
but
I
I'll.
Just
we
can
have
a
conversation
offline
on
on
on
Gator,
which
version
of
a
pie
widgets
are
using
just
as
53
x,
105
5
11,
whatever
the
last
one
you
released,
I
think,
is
5
11.
J
Further,
is
it
the
widget
snapshots
table,
that's
yeah,
it's
it's!
It's
when
I
see
when
I
look
at
my
stack,
trace,
consistently
I'm
sitting
deletes
snapshots
in
the
stack
trace
coming
in
and
then
basically
causing
all
our
declarative
widgets
to
go
through
the
life
cycling
of
recreating.
It's
like
the
Dom
gets
deleted
and
we
regenerated
check
check.
E
And
I'll
just
add
been
working
out
with
the
continuum
guys
on
the
Jupiter
lab
and
in
particular,
Steve
and
I
have
been
hammering
through
multiple
iterations
of
an
architecture
for
a
no
quick
execution,
and
we
have
a
very
rough
outline
of
what
we
think
is
one
of
the
best
ones.
We've
come
up
so
far
and
we'd
like
to
run
it
by
Brian
or
Matthias
or
men,
or
you
know,
Chris
and
maybe
right
after
this
meeting
or
maybe
145
or
something
like
that,
we're
sorry
in
like
30
minutes
or
so
yeah
for
rough
comments.
E
F
A
F
A
C
E
G
E
So
I
was
at
I,
gave
a
talk
at
the
philly
emerging
tech
conference
about
a
month
ago
and
I
talked
to
the
tech,
lead
important
data
services,
and
he
was
like
this
is
great.
If
we
would
have
known
about
this
two
or
three
years
ago,
we
could
have
just
totally
done
it
with
this,
instead
of
with
whatever
proprietary
system
we're
using.
How
do
we
get
in?
What
do
we
do
so
I'm.
A
F
I
bet
even
good
yeah
yeah.
This
is
this
is
great,
that
it
does
put
some
emphasis
for
me
on
the
group
at
Berkeley.
That's
working,
there's
some
students
at
Berkeley
and
UV
at
Wikimedia
is
helping
work
on
some
kind
of
scaling
testing
performance
testing
so
that
we
can
have
some
actual
quantification
of
what
you
know
when
it's
gonna
fall
down.
What
are
reasonable
numbers
before
you
start
needing
to
do
more
exotic
deployment
choices
and
right
now
we
just
have
no
idea.
We
just
know
that
a
couple
hundred
is
fine,
but
we'd.
E
F
But
I
mean
every
yeah
every
piece
that
would
fall
down.
There
is
a
way
to
deal
with
that.
We
just
don't
have
the
information
for
which
pieces
are
going
to
fall
down
when
in
order
to
do
all
those
scaling
choices
and
so
I
I'll
I'll.
Try
to
talk
to
the
berkeley
folks
to
see
if
we
can
push
that
that
quantified
scaling
stuff
faster,
great.
F
A
D
K
D
A
I
E
E
A
Who
are
more
good
to
go
here
and
get
back
to
work
and
okay
snack
on
what
they
were
doing
before
meeting
used
to
do
anything
else
wanted
ad,
no.