►
From YouTube: Jupyter/IPython Dev Meeting, December 20, 2016
Description
Meeting of the Jupyter/IPython development team, December 20, 2016
Meeting Notes: https://jupyter.hackpad.com/December-2016-weekly-meetings-EqtMljhwq6X
A
Hello,
everyone
welcome
to
the
Tuesday,
December
twentieth,
Jupiter
online
team
meeting,
and
we
realize
that
a
lot
of
people
will
be
leaving
for
holidays
soon
and
we'll
probably
talk
about
the
holiday
schedule.
14
meetings
at
some
point
here,
but
let's
go
ahead
and
get
started
and
Jamie
do
you
want
to
go
first
with
a
project
management
related
stuff
sure.
B
It
looks
like
Ryan
has
submitted
materials
to
a
trademark
lawyer,
so
we're
continuing
to
pursue
the
trademark
in
the
US
first
off
and
then
we'll
look
at
exploring
trademark
stuff
internationally,
and
it
looks
like
Brian
you're
working
with
Fernando
to
create
a
process
for
writing.
The
Joss
paper
for
your
different
repos
I've
been
working
on
still
trying
to
catch
up
from
that
tub
meeting
and
being
out
of
the
office
for
four
days,
but
that's
happening
slowly,
but
surely
so,
probably
probably
will
see
something
from
me
after
the
new
year
based
on
vacation
schedules,
etc.
B
The
talk
at
Bloomberg
went
really
well.
Last
week
we
did
a
talk
and
then
we
didn't
open
a
demo
and
hands
on
Pete
portion
and
carol
was
there
and
Fernanda
was
there
in
jest
hamrick
and
Mike
pacer
over
there
as
well?
And
we
got
some
good
feedback
from
the
attendees
and
it
was
a
really
fun
night
and
Paul
I
had
an
office
there
as
well.
B
A
I
could
make
just
one
comment
on
the
jaws
articles
where
this
came
from
is
I
recently
participated
with
a
symp
I
project
in
authoring,
a
peer
J
article
that
we
submitted
for
sin
PI
and
there
the
process
they
use
for
that
paper
is
just
spectacular.
They
were
able
to
get
really
high
participation
across
different
people
in
terms
of
all
aspects
of
the
paper,
including
the
actual
writing
of
the
article.
The
whole
thing
went
really
fast,
and
so
we
been
talking
to
Fernando
about
that
and
we
feel,
like
so
testing
out
some
of
those
process.
A
Ideas
on
something
simple,
like
the
Joss
articles,
where
the
writing
is
pretty
trivial,
would
be
a
good
idea
and
then
we
would
like
to
take
the
ideas
or
that
sort
of
the
process
ideas.
We
learn
from
that
and
start
to
apply
them
to
the
other
papers,
including
the
one
that
Thomas
has
been
working
on,
that
he
has
some
notes
on
further
down,
and
so
it
we're
just
trying
to
get
the
figure
out
a
way
for
us
to
write
papers,
sort
of
more
efficiently
and
more
often.
B
D
C
Foreign
up
43,
we
I
did
auto-generated
token,
if
you
don't
have
a
password
or
security
reason
to
avoid
cross-site
request
and
any
website
on
the
internet
to
take
over
your
computer.
It's
not
as
bad
but
just
to
be
safe
to
be
safe.
So
when
you
with
43
and
above
when
you
type
to
paternal
book
on
your
terminal,
it
should
auto
open
the
browser
and
pass
the
authentication
token
to
the
your
idol
should
automatically
yo
you
in.
C
The
other
problem
is
that
you
can
see
the
password
the
terminal
if
you
think,
to
look
at
the
terminal,
but
if
a
lot
of
activity
has
gone
by
then
your
password
is
way.
Patro
token
is
way
up
in
the
terminal
and
you
might
not
even
be
able
to
access
it
so
mean
me,
the
pr
remember
correctly
that
try
to
make
it
clearer
on
the
password
page,
if
you're
not
automatically
logged
in
what
the
steps
are
to
actually
get
sisto,
chem
and
login
to
the
server.
C
It's
still
not
perfect,
especially
if
you
have
a
remote
notebook
server
which
listening
on
a
public
IP,
because
you
might
not
have
access
to
the
terminal
and
you
might
have
to
find
a
way
to
the
Gimli
luckily,
but
for
now
we
believe
that
the
443
one
we
will
be
soon
should
be
better
from
a
user
experience
on
to
Laguna
still
not
perfect,
but
it's
a
good
compromise.
If
we
want
to
have
still
tight
security,
we
discuss
with
run-on.
Do
I
want
to
develop
their.
They
are
just
in
person.
I
might
I
can
do
a
writeup.
C
F
Oh
sorry,
I'm,
you
will
be
nice
after
day.
The
point
releases.
We
can't
have
additionally
a
blog
post
about
this
because
I
know
there
was
some
documentation
update
on
that
matter,
but
I
think
that
if
we
have
a
good
exposition
about
the
issue
and
how
we
solve
it,
people
will
be
innovative
in
the
space.
So
maybe
a
blog
post
is
something
reasonable
to
do
about
that.
After
the
poor,
Fugees
I.
B
Okay,
great
thanks
guys
looks
like
the
next
two
things
on
the
list
or
Jupiter
lab
and
I
pie,
widgets
and
those
are
within
daring
and
Jason's
realm,
but
they're
out
this
week.
So
we'll
move
on
to
Jupiter
hub
min.
Take
it
away.
E
Bronson
was
in
the
front,
oh
my
only
thing:
I'm
jupiter
has
been
just
ironing
out
a
couple
issues
that
have
got
now
that
we've
rolled
out
0.7
there's
a
particular
configuration
of
goober
hub
and
docker
spawner.
That
needs
a
little
help.
Carol's
been
helping
me
on
those
be
ours
and
should
be
able
to
get
a
bug-fix
release
to
those
two
packages
shortly
and
then
Carol's
been
doing
a
lot
of
work
on
some
of
the
deployments
and
documentation
of
the
Mickey,
making
the
teaching
deployment
easier
to
understand
and
follow.
G
It's
been
just
a
lot
of
just
responding
to
people,
and
you
know
we're
going
to
start
updating
some
things
for
Python
3,
dot,
6,
just
primarily
Travis
with
the
tests
and
for
some
reason,
Jupiter
hub
docks
are
now
failing.
Probably
for
the
similar
reason
that
and
be
convert
was
so
we'll
take
care
of
those
as
well.
E
G
H
E
So
the
main
reason
I
didn't
do
too
much
on
Jupiter
have
was
because
I
was
helping
out
here,
getting
NBD
mounted,
or
so
we
have
our
first
they're
gonna
have
been
doing
a
lot
of
different
things.
I
haven't
made
a
lot
of
progress
on.
E
One
over
the
last
week
or
two,
but
this
will
yeah
I,
guess
I'll
go
so
every
dime
as
a
release.
So
if
you
want
to
see
how
definitely
merging
their
books
can
or
can
get
feedback
on
that
stuff,
it's
in.
I
think
it's
it's
fairly
usable.
The
difficulty
is
in
pretty
good
shape.
There's
a
lot
of
room
on
the
merging,
but
it
works.
E
Next
is
also
me
yep
and
be
you
so
we've
been
having
some
issues
with
NB
viewer
that
seem
related
to
the
fact
that
it
was
running
on
kareena,
tough
at
Rackspace,
and
I
moved
it
I
just
so
we
have
our
viewer
deployment
scripts
on
on
github
and
I.
Just
ran
them
on
I
do
better
have
instance.
Them
pointed
the
domains
to
that
and
are
pointed
fastly
at
that,
and
the
error
rates
have
gone
from
relatively
high
to
effectively
04
and
stayed
there
for
a
week.
So
that
seems
seems
to
have
been
the
issue.
B
H
Right
so
last
week
we
actually
released
and
became
her
twice
so
the
first
one
was
actually
because
Silva
had
asked
specifically
for
and
be
viewer
to
be
able
to
have
live
widgets,
so
I
just
wanted
to
push
something
out
and
back
ported
it
and
then
sent
it
out.
So
that's
what
43
does
and
then
50
is
the
one
that
we've
been
working
on
for
most
of
year
and
yeah.
So
and
then
we
already
broke
something
and
then
Thomas
fixed.
It
though,
and
now
that's
just
waiting
an
emergen,
then
I'll
send
out
a
patch
release.
H
E
A
So
so
have
we
pulled
and
the
reason
I'm
asking
is
the
in
notebook
for
dot
3.
We
updated
the
styling
of
HTML
tables
to
match
what
we're
doing
in
Jupiter
lab
and
I'm
one.
It's
a
significant
improvement
and
I'm
wondering
if,
if
that's
going
to
end
up
the
one
of
those
versions
of
and
be
convert
and
subsequently.
H
H
H
A
H
Oh
and
there
were
some
a
bunch
of
stuff
about
docks-
that's
been
temporarily
taken
care
of
by
the
release
process,
but
we
still
don't
have
a
good
way
of
handling
the
thing
where,
if
you
are
dependent
upon
a
project
that
is
dependent
upon
you
and
not
building
on
master
or
a
dev
version,
but
rather
whatever
the
most
stable
releases
and
you
change
the
API,
it
overwrites
the
editable
like
dev
install,
and
so
you
run
into
errors
on
read
the
docs
furtive
inevitably
and
the
solution
that
we
thought
we
had
had
for
that
didn't
seem
to
continue
working
so
still
trying
to
figure
that
out
and
if
anyone
has
figured
out
how
to
get
read
the
docs
locally
working
on
OS,
10.
E
H
E
H
H
Good,
since
it's
working
now,
it's
not
as
urgent
as
it
was
before,
so
it's
only
going
to
become
really
urgent
if
we
change
the
API
again,
but
it's
also
something
that
any
other
project
that
has
that
sort
of
cyclic
like
package
dependency.
You
should
keep
an
eye
out
for
that
because
it
will
cause
you
problems
or
if
it
doesn't
tell
me
how
it
doesn't
cause
you
problems,
because
I
want
to
know.
B
All
right,
so
it
looks
like
we
had
the
github
automation
that
Matthias
is
working
on.
He
shared
shared
some
good
information
with
the
Berkeley
team
yesterday,
but
he's
not
spilling
mom,
so
he's
not
going
to
share
today.
Maybe
when
we
comes
back
from
this
break
next
in
line,
we
have
a
design
and
that's
Brian.
A
Yeah,
so
the
students
at
Cal
Poly
are
making
another
pass
on
the
brand
guidelines
to
address
the
feedback
from
the
dead.
The
G
meeting.
We
do
have
a
curved
version
of
the
brand
guidelines
up
on
the
main
readme
of
the
design,
repo,
so
feel
free
to
use
that
pass
at
around
and
then
Cameron,
ferrika
and
charm
free
are
working
on
a
stink,
Jupiter
sinks
team.
That's
based
on
the
the
alabaster
theme
and
I
wanted
to
mention.
A
So
it
would
allow
us
to
have
the
documentation
on
a
separate
to
continue
to
host
it
on
read
the
docs,
but
from
the
users
perspective,
have
a
uniform
navigation
between
the
documentation
and
the
main
website,
eventually
we'd
like
to
also
add
that
same
site
nav
and
be
viewer.
So,
basically,
people
can
sort
of
navigate
between
these
different
sites
in
a
very
consistent
uniform
way
and
so
they're
they're
working
on
that
and
then
another
thing.
A
We
are
starting
to
use
material
design,
icons
and
but
it's
also
quite
simple-
to
create
new
icons
as
well,
basically
and
just
drop
an
SVG
file
in
and
create
a
CSS
class
for
that,
rather
than
having
to
go
through
the
whole
difficult
process
of
you
know
creating
an
icon
font,
which
is
a
lot
more
difficult
to
do
so
at
least
wanted
to
mention
that
and
we'll
we'll.
Let
people
know
how
that
goes.
A
Is
people
try
it
out,
though
the
only
possible
disadvantage
is
for
if
you
had
a
web
page
with
many
many
many
icons
to
the
SVG
based,
you
could
run
into
performance
problems,
but
I
think
are
feeling
right.
Now
is
that
in
general,
we're
probably
going
to
end
up
with
fewer
icons
on
the
page
then
github
has
they
have
quite
a
few
SVG
based
icons
on
your
page
and
I.
Don't
think
anyone
that's
been
having
performance
problems
with
github.
So
what
will
keep
everyone
posted
on
that.
A
So
Thomas
at
some
point,
I
think
minimally,
UI
and
fernado
should
talk
about
the
paper
process.
I
think
we've
learned
a
lot
and
I
think
we
can
sort
of
make
some
small
adjustments
to
the
process
and
I
think
it
really
helped
us
to
get
get
this
paper
and
other
papers
out.
So
we
should
probably
get
to
a
phone
call
or
chat
at
some
point.
B
Way
to
go,
let's
see
I
have
a
blog
on
here.
There's
no
name.
Next
to
it.
Matthias
it
looks
like
you
may
have
added
it.
It
says
how
to
update
a
costar
version
is
almost
two
years
old.
B
Soft
piece
of
software,
that's
actually
was
created
by
a
person
who
graduated
from
Berkeley,
so
he's
really
excited
to
work
with
us
work
with
Berkeley
in
general
to
deploy
across
the
different
projects
that
are
going
on.
So
this
was
set
up
by
a
research,
IT
group
and
data
carpentry.
So
we
had
about
a
I
think
was
about
a
30
to
45
minute
demo
from
them,
and
it's
a
beautiful
piece
of
software.
B
B
But
everything
needs
to
go
through
aha,
first
to
create
github
records,
so
I
thought
that
was
pretty
much
a
deal-breaker
and
so
we're
going
to
continue
to
look
at
tools
on
our
end.
I
know
different
folks
have
been
creating
useful
tools
in
github
and
I'm
working
on
scheduling
that
tutorial
Sophie,
so
people
can
share
what
they've
been
working
on
separately,
Brian
and
Jason,
and
the
TS
me
and
I
think
some
other
folks.
B
Sorry
if
I
left
off
your
name,
so
that's
the
status
on
that
today
is
the
last
day
to
submit
any
names
for
our
tech
writer
candidate.
We
need
to
get
the
names
into
our
business
office,
so
they
can
send
the
candidates
a
request
for
proposal
for
the
contract.
Currently
we
have
three
names
and
it
sounds
like
Fernando
didn't
receive
any
via
email.
B
So
if
anybody
has
knows
of
anyone,
who's
looking
for
a
tech,
writer
position
or
job
send
them
our
way
before
the
end
of
today,
ok
and
I'll
be
working
in
github
over
the
next
couple
weeks.
Creating
some
schedules
that
are
easily
accessible
in
github,
so
I'll
be
using
the
github
projects
tool
to
create
some
just
some
basic
roadmap
like
a
calendar,
style
roadmap
and
start
slotting.