►
From YouTube: Jupyter/IPython Dev Meeting, April 11, 2017
Description
Meeting of the Jupyter/IPython development team, April 11, 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
good
morning,
everyone
today's
april
eleventh,
this
is
the
weekly
Jupiter
dev
meeting.
We
keep
our
notes,
/
and
Dropbox
paper
and
I'm.
Looking
at
that
right
now,
we
start
off
by
looking
at
our
action
items
from
last
week
and
folks
will
check
off.
What's
done
as
we
as
I
go
through
them.
Usually
so
it
looks
like
things
are
getting
checked
off
right
away.
We
have
one
outstanding
item
which
is
Jupiter.
A
A
A
Let's
see
if
Matthias
is
on
the
call
nope,
so
Allah
Matthias,
to
just
make
a
note
for
next
week
to
let
us
know
what
other
kinds
of
feedback
he's
looking
for
been
on
for
a
couple
weeks.
So
I'd
love
to
know
what
else
is
needed
to
move
this
forward
and
y'all
can
also
comment
in
the
issue.
It's
an
eye
python
org
under
ipython,
and
it's
pull
request
10399
the
list.
A
A
So
we
don't
usually
go
through
these
in
the
meeting,
but
everybody's
welcome
to
click
on
these
and
check
out
what's
going
on
in
each
of
the
works
on
their
own.
So
on
the
project
management
side,
it's
been
pretty
quiet
the
last
couple
weeks
and
I'm
just
catching
my
breath
and
getting
caught
up
on
things.
So
you
can
look
for
a
dutiful
for
the
recent
for
the
Jupiter
widgets
tutorial
that
we
talked
about.
A
I
think
last
week
or
the
week
before,
I'm
going
to
look
to
schedule
at
for
the
end
of
april
or
beginning
of
May,
sylvan
and
Jason,
so
they'd
be
happy
to
give
an
overview
of
what's
in
the
new
new
widgets
release.
I
think
that
would
be
really
cool
alright.
So
next
we
have
ipython
looks
like
that
men
are
you
doing
a
ipython
I.
B
A
A
D
E
D
F
Hey
grant
fits
Carol,
we've
done
some
stuff
on
Jupiter
hubs
and
one
thing
only
to
talk
about
it
offline
as
well,
but
one
thing
we
did
do
is
some
of
the
issues
that
he
wanted
to
close,
but
didn't
necessarily
have
a
good
place
to
close
them.
We
just
close
them
and
said
we
were
marking
them
as
a
reference
because
it
made
people
feel
a
little
bit
better
and
then
there
was
some
standard
wording
that
we
kind
of
used
in
different
situations,
so
that
might
help
so.
F
G
D
F
C
F
C
C
H
F
H
The
integral
lab
we
have
a
label
for
needs
info
and
very
often,
especially
for
the
things
that
have
been
left
open.
We,
if
we
ask
a
question
and
somebody
hasn't-
responded
in
quite
some
time-
those
needs
info
ones
we
can
close,
because,
basically
we
were
not
hanging
like
you-
can't
really
take
action
on
those
anyway.
I
know
that
doesn't
cover
anything,
but
that
does
cover
a
subset
yeah.
I
D
C
This
is
starting
to
look
like
it
might
dramatically
change
the
UI
model
for
how
download
should
work,
because
as
soon
as
you're,
starting
to
upload
a
configuration
and
having
to
choose
the
file
format
in
case
it
wasn't
actually
specified
in
your
configuration
file.
Because
you
want
to
use
the
configuration
file
for
multiple
different
output
format,
then
you
also
need
to
have
that
choice
and
then
all
of
our
current
options
are
more
or
less
equivalent
to
having
no
custom
configuration
passed
in
and
just
choosing
that
particular
output
format.
C
Also
it
made
me
realize
that
post
processor
is
a
pretty
bad
term
for
an
analogy
to
preprocessor,
because
what
what
we're
doing
with
post
processor
is
not
doing
something
in
memory,
but
rather
manipulating
objects
on
the
file
system,
because
it
has
to
come
after
the
writer,
which
is
why
I'm
thinking
for
the
rather
than
creating
all
of
the
documents
and
then
zipping
mortaring
them
just
writing
them
to
raise
it
or
tar
director.
You,
like
from
the
get-go
using
a
writer,
is
more
appropriate
approach,
but
I
am
still
figuring
out
how
to
do
that.
C
One
of
the
things
that
is
a
little
the
guy
Lang
but
I
think
I
figured
it
out
is
that
tar,
the
tar
file
library
in
Python
expects
to
see
only
files
being
tested
using
its
standard
like
ad
interface.
It
doesn't
have
the
equivalent
of
what
the
file
has
with
right
string
and
I'm
figuring
out
what
to
do
around
that,
because
I
think
I
just
need
to
create
file
like
objects
and
then
pass
those
around.
C
But
beyond
that,
oh
and
as
men
hopefully
made
a
PR
about
mathjax
is
shutting
down
their
CDN,
and
so
we
need
to
update
hours
that
updates
what
we
are
pointing
to
to
make
sure
that
it
doesn't
end.
Effet
updating,
I
mean
they're,
closing
it
so
yeah
and
anyone
elses
projects
that
are
relying
on
mathjax,
particularly
downloading
it
from
their
CDM.
You
should
probably
switch.
B
C
B
J
B
Yeah,
maybe
try
to
even
pull
reduce
the
configurability
of
that
to
the
very
specific
case
of
serving
serving
files
and
not
make
and
don't
make
post
process
or
some
extension
kind
of
pull
the
pull
back
on
that
as
an
extension
point,
because
it
it's
rather
not
the
right
time
to
do
much
of
anything.
It's.
C
Probably
good,
then,
that
we
didn't
make
post
processor
entry
points,
which
was
one
of
the
old
issues
I'm.
Actually,
if
this
is
going
to
be
our
strategy
going
forward,
I
think
there
are
actually
a
couple
of
issues
that
I
can
either
close
or
leave
a
message
on
over
future
closing,
because
this
was
little
when
you
say
getting
rid
of
it.
Do
you
mean
like
actually
getting
rid
of
the
post-processor
like
class
out
of
d
convert
or
because
then
external
okra.
C
F
B
B
C
A
All
right
thanks
Mike
next
we've
got
the
dashboard
and
related
extensions.
Pete,
unfortunately
can't
make
it
today,
but
let
me
just
share
what
he
provided
in
the
our
notes.
He
said
we
cut
releases
various
extensions
last
week
for
compatibility
with
notebook
five-point.
Oh
and
I
skipped
one
above
that
we
have
Colonel
gateway,
dr
stocks
for
services.
Men
did
you
want
to
provide
any
updates?
I
notice
that
peeps,
the
one
that
entered
those
bullet
items,
so
he
hides
so
there's.
B
What
there's
one
pull
request
left
for?
Colonel
gateway
to
use
the
activity
monitoring
stuff
from
a
notebook
like
10
and
as
of
a
minute
or
two
before
the
meeting
today?
That
one
is
working
and
ready
to
go.
So
that
should
be
relationship
and
I
stare.
To
thank
great
thanks,
and
then
we
relate
to
that.
I
I
need
little
patches
to
Jupiter
clients
and
notebook
and
Colonel
gateway
for
do
to
look
a
little
source
of
a
trouble
that
was
causing
the
test
to
fail,
and
I
carly
rae
gotcha.
B
I
This
week
more
lessons
learned
in
iterations
on
the
extensions
front,
we're
learning
more
and
more
than
we
need
to
use
the
standard
tools
and
workflows.
If
possible,
you
can
see
the
the
latest
iteration
is
as
close
as
possible
to
a
standard
JavaScript
developer
as
we
can
get.
So
it's
very
much
like
NPM
install
this
thing
and
web
pack
this
thing
and
then
met
your
application,
so
we're
starting
there
with
the
grants,
give
me
a
thumbs
up,
I'm
starting
there
with
the
you
know.
I
So
there
are
ways
forward
that
are
for
the
more
advanced
cases,
but
for
the
simplest
case,
using
the
standard
tools
for
JavaScript
development
is
the
least
brittle
of
the
painful
pat
forward
right
now.
So
that's
where
we're
headed
with
that
and
there's
link
to
the
discussion
on
going
on
that
issue
for
plan.
For
extension,
Hanlon
Ian's
been
doing
work
on
the
real-time
model
database
I'll.
I
So
that's
now
in
Darien
is
currently
working
on
the
single
document
mode
for
Jupiter
lab
and
the
way
we're
going
to
initially
manifest,
that
is
that
be
a
command
and
the
command
palette
to
switch
toggle
between
the
modes.
And
so,
if
you
go
in
a
similar
document
mode,
the
tabs
go
away.
The
doc
kind
of
goes
away
to
have
single
widget
at
a
time,
they're
looking
a
single
item
and
then
we're
adding
a
left
side
panel
item
that
allows
you
to
talk.
I
Toggle
switch
between
the
open
items
which
can
return
those
consoles
files
and
it
looked
very
much
like
the
tab
would
have,
but
as
a
list
of
open
items,
so
that
more
like
a
single
document
view
that
you
would
guide
with
the
classic
notebooks
we're
now
offering
for
all
file
types.
So
that's
a
work
in
progress,
Darren's
working
on
and
he
and
I'll
turn
it
over
to
you
have
anything
you
want
to
add
about
the
model
database,
work,
I'm,
sure.
K
The
back
end
is
more
or
less
of
the
running
at
the
moment
is
the
Google
Drive
real-time
back
end,
but
looking
at
targeting
other
things
like
like,
possibly
the
Easter
bjs
back
end
and
other
possible
real-time
frameworks
that
exist
out
there.
So
that's
that
the
basic
structure
that
I'm
working
with
rich.
So
if
anybody
wants
to,
let
me
take
a
look
at
that
I'll
throw
wake
up
and
eat
it.
A
A
And
Ian:
alright,
thanks
guys
next
I
by
widgets,
you
guys
here
today,
Jason
yes,.
L
I
had
doesn't
eat
myself,
yeah
lots
and
lots
of
work
towards
I
pie.
7.
You
can
see
the
pulse
report
for
the
whatever
10
15
20
full
request
stories
closed.
One
thing
of
note,
though,
for
everyone
we're
moving
right
now
at
pull
request
is
in
flight
to
move
the
widgets
in
the
classic
notebook
from
a
single
widget
area
to
just
being
rendered
in
in
line
with
any
other
output.
L
Unless
there
are
huge
objections,
it
is
a
kind
of
a
big
change,
workflow
wise
with
widgets,
but
I
think
it
brings
more
benefits
and
then
the
downside
of
backwards
and
compatibility
and
the
backwards
in
that
ability
is
really
just
understanding
how
widgets
work
that
it's
in
line
with
all
the
outputs,
not
so
much
any
backwards
and
compatibility
with
people
actually
using
widgets
or
programming
woods
thoughts
+1
when
it's
when
I'm
on
the
phone.
So
I
can't
see
what
you
guys
are
doing.
L
Good,
okay,
we
will
proceed
with
that.
The
eigen
I
think
awesome,
we'll
proceed
with
that.
Vi
were
fixing
out
some
of
the
some
of
the
lingering
things,
but
it's
allowing
us
to
believe
a
ton
of
code
as
well
the
plan
of
code
to
manage
the
widget
area
and
everything
and
and
I,
don't
think
we
need
to
delete
very
much
good
in
the
notebook
because
I
think
we've
stripped
it
out.
But
anyway,
it's
nice
to
clean
things
up.
A
B
Yeah,
so
the
web
stuff
has
landed
to
know.
Jupiter
hub
is
internally
using
OAuth
to
authenticate
every
could
handle
the
negotiation
of
authentication
inside
and
there's
been
a
lot
of
work
recently
on
deploying
Jupiter
have
whisker
burn
a
nice,
that's
UV
and
and
grindin
and
Caroline
can
folks
at
Berkeley
Hanson
working
on
getting
some
print
dealer
documentation
for
doing
wrong
deployments
and
I.
You
piece
is
working
a
lot
on
getting
the
taking
the
generic
parts
of
the
Berkeley
deployments
and
making
something
reusable.
A
Thank
you
alright,
so
we
have
a
new
section,
it's
other
so
for
those
who
are
watching
most
of
the
groupings-
and
these
notes
are
based
on
our
team
structure.
Now
we
finalized
the
team
structure,
wait
2016,
but
some
work
falls
outside
of
those
those
team
groupings.
So
we
have
another.
We
have
another
category
now
and
I
can
add
to
that.
I
wanted
to
just
give
you
guys
a
quick
update
on
binder,
so
we're
working
with
a
team
here
crystal
graph
is
leaving
a
great
effort
with
UV
again
and
Jeremy
Freeman
and
I.
A
Don't
want
to
leave
anybody
out,
but
yous
doing
a
lot
of
the
work
you
can
check
out
the
binder
room
on
git
er.
There's
a
lot
of
activities
they're
going
on
to
get
binder,
supported
in
a
more
sustainable
way,
and
that's
going
really
well
once
they're
finished,
we'll
probably
have
a
bigger
announcement,
but
I
just
wanted
to.
Let
you
guys
know
that
that
continuing
on
alright
so
exciting
development
yep!
That's
right!
Thank
you
so
conferences.
A
Outreach
project
Jupiter
will
participate
in
all
the
open-source
pay
at
great
popper
next
year,
so
the
organizers
reached
out
to
Carol
and
I
22
again
participate.
We
did
carol
led
an
open
hatch
session
on
the
Friday
of
grace
hopper,
where
we
worked
with
about
20
some
aspiring
developers
other
folks
for
had
been
working
in
industry
for
a
while.
So
it
was
a
really
great
range
of
folks
and
skill
sets
and
having
an
install
party
showing
them
how
to
use
Jupiter
and
then
taking
on
a
mini
project,
so
that
was
under
open
hatch
this
next
year.
A
We'll
do
we'll
do
our
session.
That
is,
is
for
project
Jupiter.
So
that's
what
the
name
of
our
our
group
will
be
titled,
as
so,
if
folks
want
to
go
to
great
offer
and
you
guys
want
to
participate
and
work
with
a
lot
of
really
fun,
energetic
developers
and
technical
women
come
on
over
it's
a
really
great
time.
So,
okay,
we
also
have
a
whole
bunch
of
other
events
that
have
already
been
announced.
A
So
we've
got
our
team
meeting,
come
at
the
end
of
May
code
land
in
New,
York,
City,
April,
21st
and
22nd
craft
2017,
budapest
april.
Twenty
fifth
of
28
slide
a
to
London
is
coming
up
next
month,
mais
fester,
the
seventh
there's
a
Jupiter
silly
day
looks
like
May
nineteenth
to
pike
on
taiwan
june,
nine
to
11,
pi
paris
in
june,
and
then
science
Gateway,
Community,
Institute
webinar
in
July,
eighteenth
or
20th
that
Carol
will
be
presenting
at.
We
also
have
pike
on
coming
up
in
may,
as
well,
so
I
think.
A
Every
week
I've
shown
what
our
schedule
looks
like
for.
May
we
all
know
we're
going
to
be
busy,
so
it
equals
in
last
month
that
will
have
peace
and
then
multiple
hits
hit,
may
Jupiter
con
we're
plugging
away
on
the
registration
site,
so
it'll
be
released
soon.
I,
don't
think
I'm
supposed
to
give
the
date
out,
but
that's
moving
along
Noah
based
on
the
governance
side
and
then
thanks
for
recording
your
action
item
for
this
week,
the
least
I
applied
Colonel.
A
Four
point:
six
point:
one:
so
could
you
guys
if
you
have
other
action
items
that
you
want
to
make
sure
get
on
the
the
notes?
Please
Adam
here
releases
this
week,
we've
got
moves
out.
Last
week
input
the
Seas
we
had
Jupiter
dashboard
of
these
are
compete
so
Jupiter
dashboards,
0.7,
Jupiter,
CMS,
0.7
g,
/,
dashboards,
bundlers,
0.9
and
assess
I.
Think
that's!
What's
so
funny!
A
Last
week,
Jupiter
lab
zero
point
19
and
associated
NPM
packages,
mostly
at
0.2,
so
we
don't
have
any
releases
soon,
but
I
think
I'll
probably
have
to
update
that,
because
I
I
know
that
there
was
the
ipython
released.
That
Matthias
was
working
on
and
a
couple
other
things
from
last
week.
I'll
have
to
check
back
in
March.
There's
that
deleted
so
great
anything
else
folks
have.