►
From YouTube: Jupyter/IPython Dev Meeting, November 22, 2016
Description
Meeting of the Jupyter/IPython development team, November 22, 2016
Meeting Notes: https://jupyter.hackpad.com/November-2016-weekly-meetings-ODhDM77jlsd
A
Good
morning
this
is
the
November
22nd
Jupiter
dev
meeting.
This
is
the
week
of
Thanksgiving,
so
we
have
a
hungry
group
here
we
have
a
slightly
different
format
on
our
hack
pad.
We've
moved
to
partitioning
the
work
summaries
by
as
sub-projects
or
feature,
so
you
can
see
that
over
at
the
Jupiter
hack
pad
com
from
the
project
management
side,
my
continue
to
move
things
forward
on
many
different
fronts,
including
continuing
to
work
through
the
we
have
a
opening
for
a
tech
writer,
so
we'll
have
a
request
for
proposals
going
up
on
the
Berkeley
website
shortly.
A
A
B
B
We
can
we'll
have
a
look
at
them
ahem
and
then
one
specific
thing
that
we
were.
We
were
working
on
a
pull
request
to
add
view
and
edit
buttons
for
a
variety
of
file
types
and
particularly
for
HTML,
because
it's
something
that
you
can
both
view
in
the
browser
and
edit
the
raw
code
in
the
text
editor
view.
B
And
we
realized
that
there
was
a
kind
of
security
concern
with
the
HTML
view
and
because
it
can
execute
arbitrary
JavaScript
and
talk
to
a
kernel
and
then
run
now
area
code
on
your
computer,
which
is
normally
what
you
expect.
When
you
view
a
file,
and
so
kyle
has
some
ideas
about
how
to
how
to
make
that
safe,
using
I
frames
and
but
we're
going
to
postpone
that
toy
later
release
to
give
ourselves
a
bit
more
time
to
work
it
out.
D
C
Is
partially
motivated
by
some
issues
from
NB
convert
I
realize
that
the
way
that
we're
doing
header
tags
is
somewhat
problematic
in
terms
of
some
of
the
IDS
that
we're
generating
and
by
tags.
Mn
IDs
are
not
valid
HTML
for
in
the
first
place,
but
then
also
could,
just
in
general
cause
some
issues
with
how
characters
are
interpreted,
especially
when
you
do
get
to
the
env
convert
level
and
not
all
of
the
characters,
especially
punctuation
characters
are
allowed
in
ids
like
for
tech
and
whatnot.
C
So
I'm
thinking
that
this
is
probably
not
something
that
we
want
to
fix
for
the
4.3
release,
because
this
is
going
to
break
notebooks,
because
people
are
currently
using
the
header,
like
ID
mechanism
to
have
live
links
in
their
notebooks,
and
so
some
of
those,
if
we
want
to
even
if
we
wanted
to
go
to
sort
of
the
minimal
set
of
changes,
are
going
to
break
some
notebooks
somewhere.
I
don't
know
how
many
I
don't
know
if
it's
worth
trying
to
give
that
out
ahead
of
time.
C
The
one
thing
that's
up
in
the
air
that
I
think
I
agree
with
thomas
on
this,
and
we
talked
about
it
somewhat
in
a
person
and
on
a
number
of
issues,
is
whether
we
want
to
try
to
take
the
full
pan
doc
approach
in
terms
of
automatically
generating
header
IDs,
which
involves
normalization
of
the
everything
to
be
lower
case
and
also
what
the
general
rule
out
for
this
should
be
kind
of
wanted.
Some
feedback
on
that.
A
E
D
A
E
Other
than
that
the
mime
handling
for
Jason
outputs
that
was
discussed.
The
dev
meeting
in
both
and
earlier
is
pretty
much
ready
to
go.
So
as
people
have
been
talking
about
using
those
and
interact
in
super
lab
and
places,
we
need
to
actually
ship
a
release
of
the
updated
and
before
Matt
spec,
to
avoid
people
getting
validation,
errors
and
things.
A
E
Pending
release,
just
like
notebook,
4.3,
Jupiter
hub,
which
we
put
the
bait
out
during
the
dev
meeting,
I'll
probably
release
by
the
end
of
the
week,
we've
got.
We
found
a
few
more
little
things
to
fix,
but
nothing
significant
here
at
simula
working
on
pushing
out
a
first
actual
release
of
envy
dime,
which
mainly
needs
some.
Some
work
on
making
sure
installation,
documentation
and
stuff
are
well
covered
and
that
you
know
and
make
sure
that
we
do
all.
E
Since
it's
one
of
those
python
packages
with
typescript
up
web
application,
bundled
in
nike
making
sure
that
all
the
estes
steps
and
everything
do
the
right
thing.
But
we're
working
on
that
should
probably
have
that
this
week
and
then
Carlos
Cordova's
we're
going
to
help
him
do
a
release
of
QT
console
as
well
yeah.
So
lots
of
things
ready
to
really
sick.
G
Hey
do
you
hear
me
I
have
like
my
computer
is
lagging
like
hell,
I,
don't
move
okay
and
so
I've
mostly
been
working
on
github
automation,
a
complete
release
and
api
out
of
you
to
create,
but
that
do
things
on
github
without
doing
that
as
a
specific
users
and
user
and
what
can
have
a
rest
rhetoric
permission
that
I
don't
have
come
it
right
on
things
like
that,
and
so
now.
G
I
have
something
on
an
org
mind
that
you
can
mention
and
can
do,
issue
migrations
for
you
and
I'm
going
to
have
to
improve
it
a
bit.
So
that's.
For
example,
we
can
give
users
some
users,
the
ability
to
tag
issue
or
closed
issue
without
the
exam
coming
right.
I've
been
mostly
working
on
that
this
week.
G
I
just
want
to
make
sure
there
is
no
no
critical
bug
before
before
making
that
a
little
bit
more
public,
and
I
also
want
at
some
point
to
add
the
ability
to
add
mention
this
and
ask
it
to
backport
ER,
because
right
now
bypassing
PR
is
quite
thankful.
You
need
to
do
things
on
your
own
computer,
which,
which
can
take
take
sometimes,
and
it
is,
in
my
opinion,
a
limiting
factor
for
why
some
things
have
not
me.
G
Feedback
poverty
and
we've
also
been
working
with
other
people
on
adding
a
better
book
sessions
section
on
I
python.org,
just
like
we
had
requests
to
add
books,
we
added
the
books,
and
now
there
is
a
page
that
lists
which
are
the
condition
for
us
to
ali
book.
My
back,
not
org,
so
it's
there
is
a
link
in
that
pad
and
that's
mostly
eat
for
for
me
this
week.
We
also
discuss
pleased
with
my
can
tell
us
about
that
bike
on
proposal,
and
I
will
expose.
C
A
B
C
Oh
and
more
generally,
the
Python
3
statement
any
of
the
packages
that
are
like
agreeing
to
that,
what
we're
putting
into
place
so
that
we
don't
break
things
for
Python
to
users
who,
for
example,
try
to
upgrade
and
there's
a
repo
in
the
Python
3
statements
that
describe
that
has
a
early
like
partial
draft
from
attias,
as
well
as
some
notes
that
from
the
meeting
that
we
took-
and
there
isn't
much
else
right
now-
that
I
think
is
concrete
about
that
thomas
and
smitty.
Estas
to
me.
H
Is
that
next,
actually
I'm
gonna
tutorial
submission
from
the
team
on
either
Jupiter
hub
or
Jupiter
lab,
or
something
because
Ruben
and
Alan
and
I
are
co-chairing
tutorials
and
it
would
be
nice
to
have
one
that
is
from
the
Jupiter
team,
because
I
don't
think
we
had
one
from
the
team
last
year.
I
think.
G
E
H
F
A
A
H
It's
just
a
kind
of
work
in
process
to
see
what
format
might
work
within
the
Jupiter
hub
team
and
then,
as
folks
figure
out,
what
is
going
to
be
the
way
that
we
do
the
whole
team
wide
communication.
The
thinking
is,
maybe
we
take
some
snippet
of
it,
not
the
whole
thing,
but
to
be
determined.
It
was
more
just
putting
a
stake
in
the
ground.
Oh
like
it.
A
H
A
So
you
know
this
actually
reminds
me
of
one
idea
that
I
that
I
had
one
request
that
we
had
before
the
project
meeting
was
some
tutorials
for
the
rest
of
the
team.
So
folks,
working
on,
like
Jupiter
lab,
could
get
cut
up
onto
the
hub
and
folks
working
on
Jupiter
hub
can
learn
more
about
Jupiter
lab,
so
I'm
thinking
mm-hmm.
Maybe
it
would
be
worthwhile
to
schedule
some.
A
You
know
periodic
tutorials
for
the
team
that
are
open
or
we
could
do
demos
and
just
do
walkthroughs
for
folks
to
just
kind
of
to
help
get
them
over
the
hump
of
like
installing,
whatever
it
is,
whatever
sub-project
we're
looking
at.
So
we
can
kind
of
do
some
of
some
internal
communication
and
education
as
well
to
help
bring
the
the
folks
on
the
team
who
are
interested
in
all
the
other
sub
projects
that
are
going
on
to
give
them.
You
know
a
little
bit
of
a
head
start
and
learning
about
that.
A
So,
if
that's
something
you'd
be
interested
in,
please
reach
out
to
me,
I've
reached
out
to
Jason
grout
and
grant
grant
Nestor
to
do
a
demo
on
some
of
the
github
project
management
tools
that
they've
been
using
to
know,
helps
for
other
ideas
on
the
team
for
how
to
manage
issues
and
work
through
some
of
that
stuff,
and
it
sounds
like
if
I
heard
I
heard,
maybe
every
four
or
five
words
from
a
TS.
But
if
that's
some
of
the
stuff
that
Matisse
is
working
on
as
well
Matthias.
A
I
So
I
just
got
back
from
spending
the
week
at
University
of
Washington
at
the
Geo
hack
week
event
where
I
was
the
Jupiter
ambassador
and
overall
helper
I
gave
a
little
tutorial,
so
not
a
whole
lot
to
report
other
than
that.
If
anybody
is
really
interested
in
geospatial
analysis
got
to
see
some
pretty
cool
tools
and
if
anybody
wants
to
do
massively
parallel,
geospatial
analysis
with
Google
Google
Earth
engine,
it's
a
pretty
neat
tool,
so
I
recommend
checking
that
out
nap
in
the.
A
C
Also
there's
some
more
general
issues
with
being
able
to
source
files
and
whatnot
and
I'm
trying
to
figure
out
how
to
do
that
and
I
might
have
come
across
a
sort
of
solution
that
more
generally
might
work
for
a
lot
of
the
different
exporters,
particularly
the
ones
that
need
to
rely
on
external
media
files
and
kit
that
can't
be
embedded
at
least
not
immediately
so,
and
then
there
are
some
other
things.
I
started
moving
away
from
using
the
ipython
genitals
for
the
temporary
directory.
C
A
code
I
also
realized
that,
in
general,
we're
supposed
to
be
moving
away
from
my
Python
genitals,
I,
don't
know
how
much
of
a
priority
that
is,
but
I'd
also
started
a
quick
sort
of
automated
project
management
type
thing
where
you
could
generate
a
list
of
all
the
a
nice
github,
flavored
markdown
like
checkpoint
list
for
all
of
the
links.
Thomas
wasn't
a
huge
fan
of
it,
so
I'm
probably
not
going
to
continue
moving
forward
on
it.
C
But
if
that
is
a
priority,
then
that
is
something
that
we
should
probably
revisit
at
a
future
meeting,
because
at
least
as
far
as
I
can
tell
there's
not
been
too
much
movement
in
that
regard.
C
E
Well,
I
just
a
little
background
on
the
the
genitals
stuff,
so
that
came
that,
let's
say,
fell
out
of
the
big
split
that
there
were
some
a
few
common
utilities
that
we
couldn't
quite
figure
out
where
they
should
live,
where
they
didn't
fit
in
the
dependency
graph
and
they're
kind
of
generic
things.
But
we
didn't.
We
didn't
want
to
be
responsible
for
maintaining
like
a
generic
utilities,
library.
E
But
it
exists,
it's
there,
it
works
and
unless
something
unforeseen
happens,
it
will
never
see
another
release.
So
there's
no
pressure,
but
there's
not
a
lot
of
pressure
to
remove
it
without
some
other
input.
But
it's
something
that's
you
know.
So,
if
you're
going
to
update
something,
we
would
like
to
not
increase
the
use
of
it
and
if
anything
in
there
doesn't
do
what
you
want,
we
shouldn't
fix
it
there.
We
should
use
something
else,
but
there's
not
a
lot
of
pressure
to
remove
it.
Okay
got
it
for,
for
its
own
sake,.
A
J
So
yeah
I
I
made
some
more
progress.
I
can't
get
to
my
terminal
right
now,
so
I
haven't
pushed
it
up,
but
I
don't
know
sometimes
sometimes
it
just
beach.
Balls
and
I
can't
do
anything.
But
I
can
talk
to
you
guys.
So
I've
got
a
pattern
in
place
for
building
the
different
excursions
of
installation
so
right
now,
it's
all
based
on
the
base,
notebook
or
minimal
notebook
er
which,
whichever
ones
small
and
it
tests
Python
or
a
test
ipython
as
installed
with
Conda.
J
It's
kind
of
hard
to
bake
installing
it
with
hip,
but
I'll
probably
have
a
more
vanilla
example
of
that
installing
it
on
different
platforms.
But
so
what
it
does
looks
through
all
the
metadata
finds
all
the
things
we
got
four
different
kernels
and
then
build
a
docker
file
for
each
of
them
and
so
I
think
it
by
the
end
of
it.
You
know
a
kernel
maintainer
should
only
have
to
create.
K
Yeah
just
little
odds
and
ends
since
coming
back
from
the
conference,
mostly
focusing
on
paid
work.
So
yeah
I
did
make
a
little
progress
on
major
bearhug
cameras
example
and
did
a
local
deployment
of
the
Jupiter
hub.
0.7
beta
and
you
know,
opened
a
couple
issues
like
men
mentioned,
but
otherwise
it
seems
to
be
working
well,
so
it
has
my
vote,
for
you
know,
release
other
than
that.
Just
answering
a
few
issues
here
and
there
nothing
major
gonna.
A
A
Okay,
great
thanks,
ok,
I!
Think
that's
everyone!
If
I
missed
you,
please
speak
up
pick
now
or
forever
hold
your
peace
or
the
sweet.
Peas,
I.
Think
with
the
Thanksgiving
coming
up,
I
have
the
worst
dad
jokes
ever
so
I
just
flow
out
my
tongue,
the
more
comfortable
I
am
with
a
group,
so
just
apologize
for
anything
any
bad
joke
I
make
from
here
on
out
great
okay
guys.
Well
thanks
so
much
for
connecting
with
us
today,
it's
great
to
see
you
all
and
have
great
thanksgiving.