►
From YouTube: YUI Open RoundTable July 18, 2013
Description
YUI Open RoundTable July 18, 2013
A
A
A
A
B
A
Yeah
and
in
addition
to
that
tony
has
finally
landed
paginator,
which
came
about
through
his
desire
to
work
on
data
table.
So
basically,
upstream
from
data
table,
was
one
of
the
big
features
for
peginator.
So
he
wrote
it
in
such
a
way
that
it'll
work
for
modules
other
than
data
table,
but
he
also
landed
the
page
in
a
day
table
page
reader
this
to
release.
B
All
of
these
trickle
down
with
base
being
such
a
core
low
level
infrastructure
component
within
yui
everything
else
that
uses
base
also
gets
performance
boost
by
anything
going
into
any
improvements
to
base
so
yeah
I'll
rerun
the
numbers
across
the
whole
library
and
kind
of
see
if
we
come
up
with
any
interesting
numbers
but
yeah
just
some
of
the
other
performance
tests.
We
have
out
there
for
like
app
and
scroll
view
anytime.
B
A
C
A
A
B
A
B
B
D
A
B
A
The
next
release
well
we'll
cover
that
after
this
okay,
because
that
is
that's
up
for
grabs
as
well
so
ryan
proposed
back
on
july
12th.
He
was
unhappy
with
the
fact
that
we
had
to
wait
for
you
know
several
weeks
for
anyone
to
you
know,
issue
poor
requests
and
there's
there's
a
lot
of
validity
of
that.
Because
imagine
if
you've
got
this
pull
request,
you've
got
all
finished,
got
to
the
finish
line.
A
You
can't
check
it
in
then
you
wait
a
week
and
a
half
and
when
the
tree
opens
again
you're
like
what
was
that
I
was
working
right.
What
was
that
thing?
So
this
is
great
for
that,
because
all
we
have
to
do
is
cut
a
short-lived
branch
that
will
have
the
what
we
believe
is
the
candidate
for
this
release.
Any
bug
fixes
or
changes
that
we
need
to
do
revealed
during
testing
for
that
week
of
testing
that
will
go
into
that
and
then
from
there.
A
A
A
C
E
A
B
A
B
Yeah
and
one
of
the
things
that
is
kind
of
neat
about
that
too,
I
was
talking
to
reed
with
or
read
about
was
so
if
we,
if
we
break
off
the
released
candidate
branch
and
just
give
it
a
generic
name
like
yui
three
release,
then
we
set
up
a
ci
task
to
listen
for
any
changes
on
that.
The
only
time
you'll
actually
get
a
bunch
of
new
code
introduced
is
when
we
go
when
we
start
approaching
the
release
process.
B
C
B
A
So
yeah
it's
another
thing
that
happened
behind
the
scenes
for
this
release.
The
point
of
release
was
huge
improvements
on
the
ci
system
yeah.
We
also
include
now
the
automated
functional
tests,
it's
all
the
cell,
like
tests
that
we
are
doing
so
that
was
a
fun
one
yeah
that
was
like
solving
a
problem
driving
a
truck
through
the
brick
wall
as
fast
as
you.
C
B
No,
so
I
mean
that
that
kind
of
yeah
that
recaps
the
at
least
the
discussion
that
was
going
on
about
freezing
stuff.
There
was
also
we
had
a
discussion
internally
about
using
github's
release
pages.
So
if
you
go
to
github.com
yui,
slash
yoy3,
slash
releases
you'll
now
see
a
list
of
all
yui
releases.
So
that
was
one
thing
that
was
mentioned
that
we
should
probably
discuss
today.
A
Yeah-
and
that's
that's,
I
think
the
main
reason
why
eric's
here
so
we
will
bring
that
up
now.
I
know
that
there
was
a
maybe
we
could
recap
a
little
bit
of
the
discussion
we've
had
via
email
or
we
just
basically
just
started
from
scratch,
and
the
idea
is
that
now
that
we
have
this
feature
from
github
for
releases,
how
would
best
make
use
of
it?
A
And
I
think
the
the
thing
that
you've
done
for
the
pre-release
for
the
pr1
was
nice
and
that
you
have
basically
you
link
off
to
the
resources
that
we
have
yeah.
One
thing
I
won't
change
about
that
is
that
we
can
in
fact
upload
the
zip
files
to
that
release
page,
so
it
could
be
hosted
from
there.
So
that
might
be
one
possible
thing
too.
Yeah.
B
No,
I
can't
remember
what
the
exact
url
is,
but
anyways
you
can
download
a
zip
file
or
tower
of
of
any
releases
on
yui,
so
I
mean
do
we
have
to
continue
using
our
own
cdn
to
host
the
zip
files,
or
can
we
just
use
github?
I
don't
know
that
we
have
to
make
that
determination
right
here,
but
I
mean
that's.
A
B
B
F
We
did
we
had
it
hosted
on
an
updates
section
of
the
website
and
then
the
the
releases
feature
was
added
to
github.
So
I
tried
out
moving
or
copying
one
of
our
our
you
know
essentially
updates
posts
over
to
it
and
it
worked
out.
It
worked
out
fine
and
I
was
able
to
add
the
zip
file
that
our
grunt
stuff
generates
when
you
run
grunt
release
and
pure
and
that's
the
same
file
we
upload
to
the
cdn.
F
F
F
So
that
was
the
main
thing
we
wanted
to
maintain
that
we
didn't
break
any
old
urls
by
doing
that,
and
we
were
able
to
do
that
with
this
and
yeah.
You
know
basically
in
lieu
of
having
a
blog
for
pure.
We
decided
to
use
this,
because
this
is
the
only
things
we
wanted
to
announce
so
far
like
we
haven't
had
any
other
need
for
announcing
other
stuff.
Yet
so
it
was
just
been
the
two.
A
So
I
guess
that
one
that
begs
the
question
I
guess
so
github
has
changed
their
position
on
hosting
files
before
it
used
to
be.
You
know
they
stopped
doing
the
downloads
page
because
they
said
we
didn't
want
to
do
hosting,
but
now
it
looks
like
with
the
releases
they
do
they
sort
of
reinstated
that
right.
You
know.
F
C
F
A
B
F
B
F
A
A
F
Yeah,
so
this
is
something
I
was
trying
to
push
the
bauer
maintainers
to
to
implement
and
use
bauer
works
well
over
get
and
github,
and
I
was
trying
to
push
them
to
leverage
this
releases
feature
such
that
they
could
then
download
or
prefer
to
download
the
binary
binaries
instead
of
the
source
if
they're
available
right,
because
that
was
one
thing
for
pure.
We
don't
check
in
the
build
files,
but
it's
only
useful
if
you
use
the
the
css
that's
generated
in
the
build
directory.
B
Yeah-
and
I
feel
like
having,
though,
because
our
release
posts
are
really
long
or
not
very
detailed,
at
least
so
I
don't
know
I
mean
putting
all
of
those
on
on
github
releases
might
be
a
little
bit
too
much
I
feel
like
it
probably
is
better
served
to
be
on
the
blog
or
elsewhere.
I
don't
know
at
least
this
just
kind
of
stores,
a
little
bit
of
a
snapshot
of
anybody
who
actually
is
curious
to
look
at
release
history.
They.
A
B
F
Then
edit
the
release
and
drag
and
drop
it
onto
the
little
drop
area
for
files,
but
yeah
the
other.
The
other
thing
that
that
we
want
to
make
sure
to
maintain
is-
and
I
think
that
we
could-
we
should
abstract
out
or
provide
some
sort
of
service
for-
is
around
the
feeds.
So,
like
people
getting
notifications
like
rss
feeds
around
this
stuff.
So
I
was
suggesting
that
we
should
probably
create
set
up
something
very
simple
for
like
feeds.yui
library.com
and
what
it
could
do
is
have
a
releases
endpoint.
F
That
would
then
redirect
over
to
the
github
releases,
atom
feed,
and
we
could
have
a
you
know,
basically
a
way
where
somebody
can
follow
just
the
releases
that
are
happening
and
that
way,
if
we
ever
switch
away
from
using
github
releases-
and
you
know
we
can
maintain
that
url.
Similarly,
like
for
security,
we
probably
want
to
have
a
special
security
feed
which
are
all
like
blog
posts
that
are
categorized
or
tagged
with
the
security
tag
or
category.
A
C
F
A
A
Next,
well,
there's
two
things
that
are
going
on
there.
There
may
be
a
point
release
coming
up
because
there's
a
number
of
small
issues
that
we've
discovered
in
the
course
of
you
know
people
testing
things
out,
so
there
there's
a.
What
do
you
think
is
the
probability
of
that
probably
fairly
high
low.
You
think
it's
low
now,
because
it's
not
as
not
as
critical
as
we
personally
thought-
yeah.
A
A
Yeah
I'd
like
to
find
out
from
him,
but
what's
on
next
on
his
roadmap,
because
you
know
there's
a
lot
of
pent
up
now
that
he's
this
was
this
paginator
stuff
was
supposed
to
open
the
door
for
him
to
get
going
on
data
table.
Yes,
this
was
kind
of
in
the
way.
So
I
think
now
that
this
is
kind
of
behind
him.
Now
he
can
start
building
up
more
data
table
features
so
I'll
be
curious
to
find
out.
What's
next
on
this
list,.
G
F
Correa
yeah
demarco,
no,
that's
his
first
name,
derosa
he's
yeah.
He
he
was
like
tons
of
help
like
so
tilo,
and
I
were
both
on
vacation
for
a
couple
like
basically
three
weeks
in
a
row,
and
so
while
we
were
gone,
he
he
was,
you
know
on
on
all
the
issues
and
making
sure
stuff
was
was
getting
taken,
care
of
and
and
then
opened
a
bunch
of
pull
requests
for
them.
B
F
C
A
Yeah,
while
you're
doing
that,
I
just
wanted
to
mention
that
I
really
feel
like
this
is
a
good
confirmation
of
the
original
idea
of
like
open
sourcing,
this
stuff,
where
there's
a
period
of
time
where
any
given
person
may
not
have
time
to
work
on
something.
But
the
community
can
really
chip
in
and
either
go
deep
on
a
specific
feature
or
they
can
like
maintain
things
as
we
go.
So
I
think
that's
a
really
great
confirmation
of
that
idea.
C
Yeah,
I
think
I
think
we've
done
really
well
with
peer.
Is
we
like
for
any
issue
that
comes
in
we're
really
good
with
at
labeling
them
as
what
type
of
issue
that
they
are?
I
think
every
issue
has
at
least
one
label.
You
know
like
discussion,
bug
enhancement
and
we
change
the
labels
like
if
we
discuss
something
we
decide
to
do
something.
We
change
it
to
like
an
enhancement,
and
every
issue
generally
has
a
milestone.
So
we
know
this
is
what's
going
to
go
into
the
next
release
so
that
helps
the
community
also
figure
out.
C
A
A
Meanwhile,
we're
waiting
on
that
tony
said
that
the
now
column
is
the
next
thing
he's
going
to
be
working
on,
as
well
as
a
batch
renderer
pulled
in
to
help
offset
the
rendering
of
the
row
similar
to
yy2
and
he's
also
looking
at
a
way
of
a
smaller
footprint
for
data
as
model
views
are
concerned.
So
that's
just
a
quick
recap
of
what
he's
working
on
next.
What's
the
now
column,
does
it
see
what
that
means.
D
C
Okay,
so
I'm
ready
here
so
basically
the
pure
menu
that
we
have
right
now,
like
the
menu
css
module
for
pure,
is
pretty
closely
tied
with
the
y
dot
menu
javascript
module,
that
ryan
wrote
and
this
new
version
of
pure
menu
kind
of
like
so
the
current
css
is
very
heavily
leveraged
from
the
work
that
ryan
did
this
new
version
kind
of
assumes.
I
just
kind
of
started
from
scratch
and
I
just
said:
what's
the
simplest,
dom
structure
that
we
can
have
for
a
menu?
C
What
would
the
css
for
that
look
like
and
trying
to
make
it
have
less
class
names
and
easier
to
customize?
So
right
now,
one
of
the
things
we,
which
is
a
priority
for
us
in
pure,
is
to
make
things
really
easy
to
customize
right.
So
if
you
have
a
button-
and
you
want
to
change
the
color
in
the
in
the
background
with
it,
you
should
be
able
to
say
class
pure
button
and
then
pure
and
then
you
can
have
like
my
button
and
then
my
button
could
have
you
know
color
and
background
rules.
C
C
It's
just
really
big
and
the
reason
is
because
you
have
to
go
in
and
you
have
to
have
this
double
class
selector
to
give
to
give
these
rules
precedence,
and
then
you
have
to
like
dig
deep
into
the
inspector
to
figure
out
where
exactly
I
have
to
tweak
my
colors
and
stuff
and
it's
really
ugly.
So
that
was
one
of
the
main
things
that
I
want
to
focus
on.
So
this
is
the
new
pure
menu.
It
looks
the
same,
but
it's
half
the
size
of
the
current
version.
C
It
has
some
responsive
stuff
built
into
it
and
it's
super
easy
to
customize.
So
I'm
going
to
give
you
some
examples
I'll
skip
this.
Well,
I
I
can
kind
of
go
into
this
one,
but
this
this
example
here
talks
about
how
you
can
leverage
the
new
grid
system,
which
I'm
which
I've
been
working
on
to
work
with
the
existing
menu.
So
here
I
have
a
vertical
menu
and
it
has
the
classes
so
you're
here
I
know
it's
kind
of
small.
C
C
If
I
make
my
windows
small,
it
goes
to
100
because
of
the
mobile
one,
and
then,
if
I
make
my
window
really
large
and
kind
of
zoom
out,
it
becomes
25
because
of
the
desktop
one-fourth,
so
your
menu
can
be
kind
of
the
grids.
One
thing
that
I'm
really
happy
about
is
we
now
have
responsive
menus.
So
if
you
look
at
the
yellow
menu
here
and
I
kind
of
shrink,
my
viewport
a
little
bit-
it
actually
comes
in
gives
you
this
like
open
thingy.
C
So
if
you,
if
you
bring
it
in
it,
you
open
it
and
then
you
like,
take
it
out
wide
it'll
stay
there,
and
then
you
can
just
so
that's
pretty
cool
and,
lastly
yeah
you
can
edit
the
menus
and
style
them
in
two
lines
of
css.
Basically,
so
I
can
do
something
like
pure
menu,
one
with
the
green
background,
and
I
want
the
links
in
pure
menu
one
to
have.
I
think
this
is
like
dark
green
and
then
I
want
the
heading
to
be
white,
and
that's
all
you
that
that
is
all
the
css.
F
C
F
C
Effectively
you're
talking
a
class
yeah.
The
only
thing
I
do
right
now
is:
I
do
some
trickery
where
so
this
changes
from
an
open
to
a
close.
I
use
some
javascript
for
that.
Oh
yeah
and
arguably
I
could
use
some
javascript
to
change
this
arrow
to
point
the
other
way
sure
but
yeah
apart
from
that
is
just
one
class.
It's
there's
an
there's,
a
pyramid,
pure
menu,
active
class
that
gets
toggled,
okay,
but
yeah.
C
So
and
if
you
notice
at
the
top
here,
I
have
this
fixed
menu
and
that
I
also
made
that
one
responsive
so
so
yeah.
I'm
really
pleased
with
how
this
how
this
it
seems
like
it's,
the
win
on
all
counts
because
it's
it's
lighter.
It
has
more
it's
easier
to
work
with
and
it
probably
has
a
few
more
features.
So
it's
cool.
G
C
Yeah
and
I've
tested
it
pretty
thoroughly
and
it
seems
to
work
everywhere.
So
if
it
wasn't
for
that
grid
dependency,
which
is
not
really
dependency,
but
just
that
example
we
can.
I
can
send
a
pull
request
for
this
like
right
now.
The
only
thing
we
lose
out
on
with
this
is
the
current,
so
the
current
pure
menu
has
some
code
in
there
to
help
with
drop
down
menus
and
specifically
with
why
doll
menu.
You
know,
specifically
with
how
light
on
menu
does
drop
downs.
C
I
took
those
out
which
is
part
of
the
reason
why
it's
so
much
smaller,
I
think
I'll,
add
some
drop
down
menu
code
in,
but
what
separated
out
from
any
sort
of
javascript
module
out
there
and
I'll
look
I'll
look
into
how
I
can
how
I
can
do
that,
while
still
maintaining
you
know,
ie7
plus
compatibility,
so
it
won't
be
a
css
only
drop
down
menu,
but
it'll
be
something
where
you
can
add
a
like
a
sprinkling
of
javascript
to
make
it
drop
down.
Possibly.
C
D
A
C
B
Yeah,
so
I
posted
a
a
blog
post
of
kind
of
recapping,
the
the
team's
experience
at
velocity
in
2013.
So
it's
a
pretty
detailed
post,
there's
a
lot
of
really
cool
content
in
there
links
to
slides
and
videos
and
presentations,
and
I
gave
just
kind
of
a
high
level
summary
of
what
some
of
what
each
of
the
talks
was.
But
this
is
really
only
about
a
fifth
of
all
the
talks
that
were
actually
there.
B
B
If
you
want
to
one
I
did
want
to
highlight
was
andrew
betts
of
the
financial
times
from
financial
times
labs.
He
did
a
really
cool
talk
about
escaping
the
uncanny
valley.
Basically,
the
the
point
of
the
talk
was
to
relay
all
of
his
findings
and
learnings
from
building
the
financial
times
mobile
application
that
got
rave
reviews
and
a
lot
of
people
kind
of
mistook
it
for
a
native
application.
B
Even
though
it's
a
it's
a
web
application,
so
in
these
slides
is
having
worked
on
scroll
view,
there
was
a
lot
of
really
useful
tidbits
of
information
in
here
just
about
how
to
properly
build
mobile
interfaces
on
these
low
powered,
slow
connection
devices.
So
let
me
see
if
there's
anything
anything
too
notable
in
here.
Aside
from
just
going
going
through
the
slides
on
your
own,
I
don't
know
yeah
anyways
you
can
you
can
kind
of
go
through
the
slides.
B
B
That
equates
to
60
frames
per
second
much
faster
than
that
and
people
will
notice
much
slower
than
that
and
that's
when
you
start
to
get
into
what
people
commonly
refer
to
as
jank
the
the
choppiness
when
things
are
scrolling
and
animating
so
yeah,
just
to
ensure
all
of
your.
Your
frame
paint
times
are
sub
16,
milliseconds
and
just
kind
of
with
any
user
interaction
design.
B
It's
always
been
a
long
known
rule
to
keep
any
responses
to
under
100
milliseconds,
because
that's
the
threshold,
where
again
users
of
your
of
your
application
or
interface,
are
going
to
determine
whether
or
no
will
notice
if
it
seems
slow
but
much
faster
than
100
milliseconds
the
threshold
yeah
yeah
they
don't.
They
won't
notice
much
faster
than
100
milliseconds,
that's
at
least
the
threshold
where
they
expect
it
to
happen.
Yeah
just
some
other
css
tips
like
not
using
flexbox
and
disabling
other
effects.
B
There's
more
details
on
those
in
the
slides.
One
of
the
cool
things
I
thought
was
their
approach
to
text
re-encoding
to
optimize,
basically
how
much
you
can
actually
store
in
local
storage.
So,
basically
inside
of
a
web
browser,
you
have
a
set
amount
of
space
that
you
can
drop
whatever
in
whatever
text
strings
you
want
to
in
there
and
in
this
blog
post
they
he
outlines.
B
B
Yeah
yeah
and
but
what
you
can
actually
do,
there
is
get
quite
a
bit
more
storage
capacity
out
of
it.
If
you
look
at
it
inside
of
devtools
or
something
you're
not
going
to
be
able
to
make
sense
of
it,
but
yeah
so
anyways,
some
interesting
stuff.
There
also
yeah
they
developed
a
couple.
Libraries
like
there's
a
fast
click
and
fast
scroll.
B
I
think
it
was
called
fast
scroll.
I
don't
remember
exactly
but
yeah
some
cool
stuff.
There,
then
another
one
that
I
thought
was
pretty
cool
was
ilia,
grigorik
he's
one
of
the
performance
gurus
at
google
and
he
announced
that
they
had
moved
http
archive
information
into
bigquery.
So,
basically,
you
have
all
of
the
http
archive
data
available
inside
of
bigquery
that
you
can
write
sql
around
and
actually
query
into
it.
B
Now
what
the
http
archive
is
so
every
month
or
I
think,
twice
a
month,
they
do
a
crawl
of
like
the
top
million
websites.
I
believe
google
does
and
they
catalog
all
of
the
http
data
that
they're
collecting
from
the
initial
page
load
and
then
storing
that
just
into
mysql
database,
so
you
can
down
in
the
past.
C
B
Talking
about
like
25
gig
of
data
per
crawl,
so
and
then,
of
course,
writing
queries
against
it
even
on
a
newer,
macbook
pro
still
really
slow.
Depending
on
how
complex
the
queries
are,
so
the
availability
of
this
data
inside
of
bigquery
is
really
cool,
because
now
you
can
actually
use
these
crazy
high-powered
machines
to
grab
all
the
information
for
you.
So
one
of
the
things
I
guess
I
can
actually
show
a
demo
of
what
you
can
actually
do
with
it.
So
working
on
a
library
that
lots
of
people
use
that's
hosted
on
a
cdn.
B
B
I
think
some
like
this,
this
one
that
I
could
show
it's
a
pretty
complex
query,
but
anyways
it
does
select
url
and
the
library
from
the
most
ranked
pages,
and
it
will
actually
what
we'll
see
here
is
the
results,
god
that
was
so
quick
too.
So
there's
200
000
rows
here,
but
we
can.
We
can
at
least
look
through
some
of
the
initial
ones.
So,
basically,
what
this
query
is
doing
is
I
can
bump
up
the
font
a
little
bit
or
yeah
now
that
actually
doesn't
help.
B
Yeah,
okay,
so
anyways
we
can
look
at
a
couple
of
them
so
yeah.
This
is
basically
getting
or
telling
us
what
libraries
all
the
most
popular
websites
on
the
internet
use.
So
yui
is
the
most
pop
it
I'm
sorry
to
rephrase
this
yui
is
the
library
that
of
all
the
websites
that
use
a
library,
yui
is
number
one
on
there,
because
google
and
one
of
the
other
ones,
probably
google,
microsoft
and
facebook,
they
all
have
their
own
custom.
Libraries
and
I'm
assuming
google.com,
doesn't
even
have
some
sort
of
a
library.
B
Product
that
they
that
they
open
source
them
so
yeah
you
can
look
at
some
of
the
other
ones.
Obviously,
a
lot
of
them
use
jquery.
So
like
here's
another
one,
it
will
actually
list
out
the
websites
that
use
multiple
libraries
too.
So
you
can
go
through
all
200
000
pages
here.
So
I
think
there
was
another
one
that
used
yui
so
like
yeah
flickr
is
the
number
or
91st
most
popular
website
and
it
uses
yui.
C
B
There's
a
ton
of
other
really
interesting
queries
you
can
throw
into
here.
I
don't
know,
I
don't
think
I
have
any
handy,
but
it
was
one
that
I
did
put
post
a
demo
image
of
was
sites
that
use
yui
by
version.
So
we
can
actually
query
that
information
out
of
there
and
figure
out
which
it
looks
like
yui.
Three,
four
one
is
the
most
popular
version.
A
B
Yeah
yeah
yeah.
I
guess
that
would
kind
of
make
sense
and
yeah
it's
interesting.
I
I
would
what
I
would
really
love
to
do
is
actually
gives
yeah
import.
All
the
data
sets
from
the
last
year
over
the
last
couple
years
and
actually
look
at
the
changes
of
yui
version
usage
over
time.
So
you
want
to
see
adoption
rates
yeah
and
then
we'd
be
able
to
see
the
upgrade
path
for
for
sites
you
use
yui.
I
think
that
would
be
really
cool
information.
B
Yeah,
I
did
another
career
that
actually
included
white
way
too,
and
the
usage
of
y2a2
is
still
astonishingly
high.
So,
but
it
was
one
of
the
few
popular
libraries
at
that
time.
So
there
was
quite
a
bit
of
adoption
and
those
sites
are
still
out
there.
They
haven't
pulled
them
down
or
anything
so
right.
B
Choosing
that
so
yeah
yeah
and
then
real
quickly.
I
think
there's
only
like
one
or
two
more
posts
here,
so
arya
hideout
did
did
a
really
cool
talk
about
just
overviewing
all
of
these
javascript
tools
that
both
he's
written
and
that
he
uses.
So
for
those
of
you
not
familiar
with
him,
he's
one
of
the
authors
of
phantom.js
and
esprima,
which
is
a
ecmascript
cursor
so
and
a
lot
of
these
tools
actually
use
esprima.
B
So
one
of
them
that
I
thought
was
pretty
cool
was
meta
js,
which
basically,
you
can
put
in
any
type
any
javascript
snippet
and
you
can
actually
watch
it,
go
through
and
step
through
the
code,
and
so
basically
you
can
watch
the
code
run
in
action
as
opposed
to
in
like
chrome
tools
or
something
sitting
there
hitting
play
play
play
play
play
over
and
over
again,
so
you
can
actually
just
see
yeah
the
animation
of
the
code
actually
executing.
So
that
could
be
pretty
useful.
B
Js
complexity
was
another
pretty
cool
one
that
you
could
find
a
bunch
more
detail
on
his
blog
that
I
was
linked
to
in
there.
So
js
complexity
basically
will
dissect
your
code,
your
javascript
code,
and
let
you
know
how
complex
it
is
and
give
you
a
maintainability
index
score
of
yeah
how
complex
or
how
simple
your
code
is
so
basically
to
show
this
real
quickly.
We
do
like,
if
true
and
then
just
do
an
empty
block
there
and
then
else.
Okay,
so
let's
click
analyze.
B
This
is
going
to
give
us
the
aggregate,
cyclomatic
complexity
score
of
two,
because
there
are
two
routes
that
this
code
can
actually
go
through.
So
if
we
do
another
one
inside
of
the
false,
if
we
do,
if
false
then
do
nothing
so
then
now
we
added
in
another
route
there.
So
that's
actually,
three
and
but
the
reason
the
route
was
there.
C
B
Because
so
alice
is
already
false,
so
if
we
do
if
false,
then
it's
going
to
go
to
the
same
place,
but
it
was
also
including
a
well.
There
is
also
an
else
branch
that
should
be
there,
which
would
add
in
another
route.
So
if
we
add
an
else
there,
you
can
see
the
score,
doesn't
change
at
all
or
the
complexity
score.
B
B
And
engage
me
and
unravel
some
of
that
right.
Yeah.
So
more
me,
look
at
the
maintainability
index
of
various
modules,
the
score
over
time.
Like
I
don't
care.
If
it
says
the
score
is
170..
That
doesn't
really
mean
a
whole
lot,
but
the
interesting
part
could
come
in
if
we're
actually
tracking
that
information
over
time,
and
we
see
that
it's
going
from
170
to
190
to
200
to
250
over
time,
and
that
means
it's
getting
more
and
more
and
more
complex.
B
And
then
we
should
it's
not
saying
that
the
complexity
isn't
justified,
but
it's
something
that
we
could
at
least
look
at
to
see
if
we
can
make
it
a
little
bit
more
simple
right,
because
you're
talking
you're,
correlating
complexity
with
maintainability
yeah,
so
there's
some
other
pretty
cool
tools
there,
including
istanbul,
which
we
use
on
yui
and
yeah.
Then
I
conclude
with
the
overall
takeaways
of
the
conference
and
you
can
kind
of
go
through
and
read
some
of
that
if
you
want,
but
overall,
a
really
really
cool
conference
velocity.
B
While
I
love
jsconf
and
a
few
other
conferences,
velocity
is
at
least
the
one
that
I've
been
to
in
the
past.
That
always
velocity
is
the
one
that
I've
been
to
in
the
past.
That
always
kind
of
is
the
most
educational
right,
because
it's
kind
of
like
it
doesn't
overlap
as
much
with
you
know,
it's
like
in
a
different
domain
that
still
touches
on
web
development
yeah.
It's
a
combination
of
devops
and
front
end,
mostly
around
javascript
performance.
B
B
You
get
a
little
bit.
You
get
some
some
of
the
same
information
it
when
it's
only
a
javascript
conference,
there's
not
a
whole
lot
of
new
stuff
if
you
get
exposed
to
a
chamber
exactly,
but
when
you're
exposed
to
all
of
this
other
really
cool
information
that
outside
of
your
realm
of
expertise,
I'm
not
in
I'm,
not
an
expert
in
devops,
but
there's
a
lot
of
really
interesting
stuff
in
there.
So
that
kind
of
gets
me
thinking
outside
of
the
box
cool,
so
anyways.
B
A
A
Using
the
word
awesome
today,
so
it's
very
hell
out
of
it.
So
for
me
there
are
a
couple
other
things,
my
bad,
a
couple
of
things
to
update.
One
of
the
things
I
was
working
on
last
quarter
was
updating
the
developer
documentation.
So
if
you
go
to
the
github
wiki
now
and
you
go
to,
for
instance,
developerworkflow
under
get
involved,
that
is
a
pretty
comprehensive
and
unified
document
that
can
take
you
through
the
mechanics
of
being
a
contributor
on
a
day-to-day
basis.
A
Another
item
that
I
updated
was
I
added
contributing.md
to
our
repo,
so
that
when
you
are,
whenever
you
issue
a
pull
request,
you
will
see
this
linked
off,
so
you
can
read
that
and
see
what
does
it
take
to
become
a
contributor
to
you?
So
it's
like
a
getting
started
guide
in
addition,
and
just
some
other
things
like
I've
updated
the
readme
to
refresh
the
links,
as
well
as
adding
the
branch
information
there.
A
So
a
lot
of
this
is
trying
to
look
at
all
the
documentation
that
we
have
on
the
site-
that's
scattered
everywhere
and
finding
ways
to
unify
it
and
to
bring
it
up
to
date.
So
they'll
be
continuing
efforts
on
that
report.
Basically,
when
I
give
a
call
if
anyone
who's
out,
there
runs
across
any
documentation,
they
find
say
on
the
github
wiki,
that's
either
inaccurate
or
that
some
omission
to
let
me
know
right
away
and
I'll
fix
it
right
away.
A
A
There's
gonna
be
a
lot
of
like
some
of
these
requests
getting
going
through.
So
what
I
want
to
do
is
take
this
week,
go
through
all
the
staple
requests
that
are
still
out
there
that
are
over
a
week
and
I'm
going
to
update
them
in
some
way
I'm
going
to
either
contact
the
owner,
or
I
will
contact
someone
who
can
be
an
honor
for
that
and
next
week,
a
report
back
to
you
how
we
can
get
that
reduced.
H
H
Keeping
the
file
name
and
to
do
that
it
was
pretty
interesting.
I
basically
had
to
use
the
html5
download
attribute
on
an
anchor
and
create
an
invisible
anchor
and
then
simulate
a
click
on
it
and
then,
by
doing
that
and
setting
the
href
to
the
file
that
I
was
trying
to
download,
I
was
able
to
keep
the
file
name
and
it
worked.
So
it's
a
lot
hackier
than
I
would
like
it
to
be.
But
that
seems
to
be
the
only
way
to
do.
E
H
E
H
E
H
H
It
at
the
time
there's
a
connection
limit
to
how
many
peer
connections
you
can
have
that's
set
by
chrome,
because
that's
what
I'm
using
right
now
and
that
is
fixed
in
an
upcoming
version,
but
not
currently
in
the
stable
version.
So
currently,
the
limit
is
that
you
can
only
have
10
peer
connections
at
a
time
in
one
browser
instance,
and
they
raised
that
to
256,
which
should
be
fine.
C
C
H
10
limit
so
when
they
do
raise
that,
I
have
a
feeling
that
the
only
limit
is
going
to
be
your
bandwidth
and
your
like
compute,
like
cpu
power
and
like
computing
power
and
stuff
like
that
and
ability
to
show
the
streams
of
the
video
streams.
If
that's
what
you're
dealing
with,
how
often
does
the
fan
kick
on
on
your
laptop.
H
B
C
H
Because
the
initial
connection
needs
a
server
to
do
the
handshake
to
tell
each
client
like
where
each
client
is
and
how
to
connect
to
it.
So
I've
been
I'm
probably
going
to
be
using
echo
echo
to
try
to
get
mock
server
going
and
get
the
responses.
I
need
to
continue
testing
outside
of
that,
I'm
working
on
chunking
files
when
you
send
files
over
webrtc
right
now.
The
limit
is
about
a
kilobyte
or
so
so.
H
In
order
to
send
a
file
over
a
data
channel,
you
need
to
chunk
it
up
into
pieces
and
then
send
it
over.
This
is
a
little
problematic
because
in
chrome
at
least
it
might
be
included
in
firefox,
I'm
not
sure,
but
the
reliable
transfer
isn't
implemented.
So
sometimes
I
lose
packets,
essentially,
which
is
annoying
because
then
you
get
like
a
corrupted
file
at
the
other
end.
So.