►
From YouTube: WebPerfWG call - May 29th 2019
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
For
the
next
call,
we
can
probably
set
that
at
the
face-to-face
because,
like
I'm
guessing
like
it's
gonna,
be
a
couple
of
weeks
after
the
face-to-face,
so
we
can
set
it
at
that
point.
That
makes
sense
okay,
so
for
the
face
to
face
a
few
things,
if
you're
planning
to
attend
either
in
person
or
remotely,
and
you
haven't
signed
up
in
the
linked
form
from
the
agenda,
please
do
so.
A
A
A
C
A
A
A
A
D
A
Short
presentation,
followed
by
questions
and
discussions,
and
then
we
thought
of
you
know
in
the
afternoon
do
a
bunch
of
updates
related
to
new
platform
features
talk
whole
bit
about
metrics
for
single
page
apps.
That
is
a
subject
that
keeps
coming
back
as
a
pain
point,
or
you
know
an
unaddressed
use
case,
so
we'd
love
to
discuss
that
a
bit
and
understand
like
get
a
better
understanding
of.
A
If
we
can
tackle
that,
and
then
memory
talk
about
JavaScript
and
frameworks
needs
where
we
also
like
I'm,
not
sure
if
this
part
should
be
part
of
the
morning
or
part
of
the
afternoon,
but
we
need
to
to
see
like
see
what
scheduling
needs
we
have
for
those
frameworks
with
other
needs
we
have
for
those
frameworks.
It
can
also
tie
into
the
single
page
app
metrics
discussion.
A
A
A
C
A
Either
or
if
people
have
used
cases
that
can
be
addressed
by
extending
existing
specifications,
that's
great
if
they
have
use
cases
that
fall
like
that,
don't
fall
under
any
one
of
the
current
specifications,
any
one
of
the
current
features,
but
still
seems
like
something
that
falls
under
the
scope
of
the
working
group.
That's
also
great,
and
it's
like.
We
don't
expect
people
to
come
with
solutions,
but
just
present
their
problems,
and
then
we
can
discuss
them.
C
B
I
think
that's
consistent
with
what
we're
aiming
for
here.
So,
for
example,
I
can
speak
some
of
the
topics
that
Marcel
was
interested
to
talk
about
with
his
experience
of
implementing
some
of
the
performance,
primitives
and
tracking
across
Google
properties.
So
the
idea
there
is
to
presents
the
kinds
of
things
that
they've
done
the
learnings
along
the
way,
and
perhaps
that
leads
to
some
gaps
that
we
have
in
a
platform
or
not,
but
just
kind
of
help
inform
how
the
experts
are
actually
using
these
tools
at
a
while.
C
Okay,
okay,
great
yeah,
I
think
it'll
be
really
good
for
how
people
using
these
thing
tools,
a
lot
has
walked.
Polling
right,
I
know
some
like
much
about
what
might
be
lacking
in
books.
Sure
just
easiest
thing
ideas
would
be
good,
I.
Think
in
general,
the
folks
that
I've
spoken
with
my
coaching
to
them
has
been.
D
A
Yeah,
you
think
it
might
discuss
that.
So.
For
me,
that
is
more
of
a
discussion
for
the
issue
related
hackathon,
because
it's
not
something
that
we
need
to
figure
out
as
a
group
where
we
want
to
go
it's
just
something
like
work
needs
to
be
done,
and
then
we
can
figure
out
what
work
and
how
that
work
will
be
divided
on
the
second
day
as
part
of
the
issue,
discussions.
C
A
A
B
A
A
A
A
Sure
I
was
thinking
well,
maybe
after
the
face-to-face
cuz,
with
there's
still
like
we're
still
a
few
months
out,
but
generally
like
deep,
broad
strokes.
Is
that
we're
meeting
up
on
Monday
and
Tuesday
I'm,
pretty
sure,
yeah
and
but
we
haven't
yet
started,
dividing
it?
A
We
are
likely
to
do
like,
like
we
did
in
the
past
one
day
talking
about
new
designs
and
one
day
talking
about
issue
trash
and
if
there's
a
preference
for
which
day
is
which,
in
terms
of
conflicts
with
other
groups,
please
let
us
know,
but
otherwise,
for
the
detailed
split
between
those
days.
We
haven't
yet
started
working
on
that.
So.
A
E
Sure
so
well,
these
are
more
like
open
questions
rather
than
design
presentations.
So
the
first
one
is
element,
timing,
interaction,
way,
shadow
Dom,
so
I
I
just
put
an
issue
there
in
the
agenda.
It's
basically
asking
how
we
can
expose
information
from
performance
information
from
shadow
elements
to
the
basically
the
window
of
that
that
contains
those
that
contains
a
shadow
tree.
So.
E
But
there
is
some
conflict
as
well
with
the
encapsulation
principles
of
shadow
DOM.
So
after
some
discussion
he
can't
act
which
is
like
someone
from
Google
suggested
two
possible
solutions.
One
of
them
is
having
performance
observers
coped
to
shadow
route
instead
of
scoped
to
a
window,
which
would
mean
that
web
component
author
would
be
able
to
get
the
performance
information
from
such
such
scoped.
E
For
from
this
observer,
because
it'll
be
able
to
register
it
for
its
own
web
component
and
then
potentially,
if
it
wants,
it,
can
fire
events
to
the
frame
to
expose
its
performance
information
to
the
website
that
is
using
the
web
component.
But
maybe
they
won't
necessarily
do
that
and
the
other
alternative
that
was
suggested
was
having
boolean
in
the
edge
that
shadow
root
method
or
that
shadow
yeah.
That
shadow.
E
That
is
a
boolean
saying
that
the
performance
information
for
this
shadow
tree
should
be
exposed
to
the
performance
observer
of
the
window,
which
is
basically
saying
that
the
author
is
okay,
with
its
shadow
element,
timing,
information
being
exposed
to
regular
performance
observer.
So
those
are
the
two
proposals
that
were
suggested.
So
I
was
wondering
what
folks
think
about
it
and
if
they
have
other
ideas
in
mind
for
solving
this
problem,.
A
C
A
So
the
solution
is
not
like.
The
question
is
perhaps
not
how
we
want
to
resolve
it,
particularly
because,
yes,
this
is
something
that
will
have
to
be
decided
talking
to
the
web
components
folks,
but
in
principle
the
the
pieces
where
we
need
to
decide
is.
Do
we
think
that
exposing
performance
entries
to
a
shadow
root
is
the
right
thing
and
there
will
be
some
discrepancies
here
because,
for
example,
if
a
custom
element
is
triggering
long
tasks
or
triggering
resource
downloads,
those
will
be
exposed
on
the
document
not
on
the
shadow
root.
F
I
think
it
would
be
useful
to
have
more
information
on
the
cases
where
one
might
not
want
their
shadow
root
timing
information
exposed
to
the
global
collection
of
timing
information
because
you
know
naively
thinking.
Oh,
you
know,
website
developer
wants
to
find
out
what's
slow
and
if
their
shadow
reviews
is
slow,
they
want
to
know
about
it
and
not
have
a
million
places
to
look
for
it,
but
that's
just
because
I
haven't
thought
through.
C
Youtube,
it's
pretty
shirtless,
isn't
it
I
mean
there's
lots
of
broad
usage,
but
there's
like
a
few
big
ones.
Yeah
right
here
is
rather
than
thinking
about
this.
Theoretically.
Would
it
make
more
sense
to
just
ask
a
couple
of
the
big
folks
using
this
I
guess
I'm
wondering
if
we
need
to
solve
this
problem
right
now
or
if
it's
possible
to
come
up
with
a
flip,
a
coin
answer
for
now
and
wait
till
we
hear
back
the
other
way
well,.
E
G
Stephen,
are
you
gonna
call
yeah
yeah
there
I'm
there
that's
correct.
So
what
really
say
is
exactly
right.
So
me,
what
I
can
do
is
to
go
back
to
the
team
with
designing
the
you
know,
a
framework
based
on
web
components
and
talk
to
them
and
report
in
a
you
know.
Maybe
either
talk
to
me,
co
or
pipobag
next
goal.
A
A
A
C
Like
in
a
selection
of
current
proposal
is
to
have
see
photos
for
an
unfamiliar
section,
API
the
way
you
get
the
section
right
now
is
you
call
the
window
they'll
get
session
and
then
that
object
expose
much
information,
namely
range
get
flush
and
don't
get
rained
out.
The
0
will
give
you
the
dome
range,
dotted
that
spans
the
represents
that
whether
user
has
selected
the
problem
in
approach.
It
obviously
is
that
the
range
Channel
expose
things
inside
a
shadow
root.
So
the
way
we
want
it
is
the
we
learn.
C
So
that's
another
prototypes
and
all
the
functions
like
performance
observer
or
like
he
get
entries
on
a
performance,
timeline,
halter
or
couldn't
take
a
list
of
shadow
routes.
It
has
asked
the
Nicola
has
access
to
and
that
could
hit
entries
that
you
know
we
don't
like
even
cherrywood.
That's
a
whole
packet
design,
I'm,
not
saying
that
that's
necessarily
better
design,
but
I'm
saying
that
is
those
are
like
a
toggle.
You
know
that's
an
option
and
I'm
sure
there
are
few
options
to
get.
E
C
C
E
C
A
Yeah
I
think
we
have
a
rough
idea
and
yeah
folks
that
or
yeah
either
have
opinions
because
they're
using
web
components
or
opinions,
because
they're
part
of
the
specifications
of
web
components
yeah
feel
free
to
chime.
In
on
the
issue
and
yeah,
we
can
figure
that
out
later
I
think
that
yeah,
the
first
step
would
be
to
like
not
necessarily
exposed
element
timing
on
this.
Until
this
is
decided
right,
yeah.
E
A
C
Mean
I
think
the
legislation
model
what
my
my
concern
I
was
just
raising
the
concern
the
forest
mentioned
right.
I
was
a-going
is
like.
If
we
have
a
central
I
mean
we
have
a
dome
teachers,
support
speech,
your
function,
which
basically
didn't
do
anything
useful,
because
everyone
said
that
there
were
this
moment
of
each
of
even
that,
even
if
it
not
so
that's
the
risk.
If
you
have
a
centralized
list
of
features,
somebody
would
just
start
you
know
naively
coming.
They
just
got
all
of
the
things
right.
Obviously,.