►
From YouTube: CALEXT WG Interim Meeting, 2020-04-21
Description
CALEXT WG Interim Meeting, 2020-04-21
A
B
A
A
Well,
that
gives
us
more
space
to
see
each
other's
faces.
I
guess,
as
we
chat
in
terms
of
running
this
meeting
I,
don't
expect
that
we're
going
to
have
to
do
an
official
queue
as
such
I
think
just
talk,
and
if
we
discover
that
it's
getting
too
difficult
to
manage,
then
we'll
do
a
speaker
queue,
but
otherwise
we're
a
small
enough
group
that
it
should
just
be
fine
to
chat
amongst
ourselves
heading
over
to
page
2
on
the
slides.
It's
just
details
of
how
to
participate.
A
We're
all
here
so
clearly,
everyone's
worked
that
out
page
3
on
the
slides
is
to
note
well.
This
is
this
is
an
IETF
meeting,
so
we
are
operating
under
the
note.
Well,
all
the
standard,
IPR
etcetera
rules
apply.
This
is
public,
it
is
being
recorded,
notes
will
be
published,
etc.
So
don't
say
anything
you
don't
want
the
world
to
know
slide.
A
Four
to
the
minutes
of
the
previous
meeting
that
we
had
I.
Think
if
106
in
Singapore
throughout
this
slides,
I've
put
notes
of
the
action
points
that
we
decided
for
the
various
documents.
There
should
remind
us
what
we're
supposed
to
have
done
by
now
and
hopefully
get
the
discussion
started.
Is
there
any
other
agenda
batch?
Anything
people
want
to
add
to
our
agenda
I.
C
Either
sink
I
think
on
how
far
Institute
Germany
and
auto
for
CDL,
so
I
I'm
here
basically
for
the
Jas
calendar
thing,
be
the
Jake
dress
card,
sir.
So
but
unfortunately,
I
am
only
available
on
to
the
next
top
of
the
hour.
So
I
assume
this
fits
nicely,
but
if
it
is,
does
not
could
you
squeeze
the
Jas
calendar
stuff
in
before
the
yeah?
It
should.
E
A
B
A
F
A
Okay,
sorry,
this
is
the
one
that
I
sent
out
with
the
link
Daniel.
Did
you
create
a
different
etherpad
I
created
a
different
one,
I
think
so.
You've
got
notes.
Atf
interim!
Oh,
that's
the
official
one
and
I've
created
one:
oh
yeah,
okay,
all
right!
Apologies
that
must
have
been.
That
must
be
the
official
one
and
I
created
one
separately,
so
I
messed
this
up,
alright
put
them
wherever
you
want
and
and
let's
let's
do
the
official
one,
because
that
will
be
the
one
that
they
archive
for
it.
Where's.
A
D
D
D
Yeah
I'll
it
turns
out.
Yes,
it
I
think
I,
look
at
this
and
decide
I
got
I
just
got
bogged
down
with
it.
It
sort
of
means
having
two
things:
doesn't
it
you
could
have
the
value
and
a
URI
which
is
not
not
the
way
things
have
been
typically
and,
and
it
is
that
I
mean
every
time
we
got
a
URI
we're
gonna
do
that
it
just
doesn't
seem
still
seemed
altogether
right
to
me
in
the
end,
yeah.
A
D
This
kind
of
thing,
as
always,
has
come
up
in
the
past
and
and
the
general
agreement
that
that
I
recollect
with
people
and
not
just
the
idea
in
various,
is
that,
well,
you
got
a
URI
and
you
don't
want
to
be
hanging
around
downloading
this
thing.
There's
always
the
possibility
of
caching
it.
So
these
we
can
deal
with
the
offline
mode
in
that
way.
Yeah
there
is
the
issue
of
whether
you
want
to
download
it
at
all
anyway
yeah
and
that's
just
a
general
just
a
general
listing.
D
Should
you
just
go
and
download
a
URI,
it
doesn't
know
whether
you
go
down
download
it
to
make
it
work
for
offline
work
or
they
go
download
it.
So
people
look
at
it.
The
answer
is
probably
you
want
to
be
asking
them
whether
they
want
to
download
this
thing,
but
then
I
get
into
thinking.
Well,
what's
the
difference
in
that
and
say
a
webpage,
that's
full
of
links
and
things
that
you
don't
ask
people
generally
I
mean
there's
some
places
say:
oh
you're
gonna
go
to
another
site.
D
D
Images
and-
and
you
know
things
that
I'll
do
without
in
the
in
the
in
the
current
sort
of
mail
client
world
is,
you
know,
do
you
want
to
view
the
HTML
you
want
to
just
have
some
sort
of
raw
text
representation
of
it,
and
then
you
say:
yeah
I
want
to
see
the
site
and
you
and
it
displays
the
full
content
in
it.
Sort
of
generic
answer
to
the
profile
deal
with
URLs
rather
than.
A
D
D
D
In
in
in
five
five
four
five,
the
contact
is
just
a
text
field
where,
if
you've
actually
got
a
a
real
card
of
some
kind,
that's
obviously
a
better
thing
to
be
to
be
referring
to,
because
it
more
likely
to
be
up
to
date.
If
changes
have
got
made
like
the
phone
number
or
whatever,
the
fallback
is
to
just
ingest
countless
just
turn
into
a
data
URI
and
shove
it
in
there,
which
I
don't
know,
is
necessarily
any
better.
A
Her
data
URI
yeah,
often
with
things
like
context,
that
it's
messy
you
kind
of,
want
to
know
what
the
value
was
at
the
time
as
well
as
being
able
to
get
the
updated
one
cuz.
If
they,
if
the
contacts
missing,
then
you
don't
have
the
information
anymore
about
what
was
there
and
certainly
for
archival
purposes.
You
need
to
know
what
it
was
at
a
point
in
time,
but
for
publishing
for
an
event
which
is
now,
and
maybe
you
don't
care
about
all
that
stuff.
D
Yeah
and
I
guess
the
contact
thing
it
maybe
made
me
sort
of
rethink
it.
I
mean
maybe
the
best
approach
is
to
is
to
it
is
to
have
a
the
names.
Keep
getting
used
up.
I
was
think
of
having
a
source
URI,
but
then
the
name
is
used
in
other
places
for
there's
something
slightly
different,
but
maybe
have
something
which
always
indicates
the
source
of
what
you've
got
and
then
have
the
the
thing
itself
encoded
in
some
form.
D
D
F
Mean
again,
it
depends
on
the
context
accessing
it.
You
know
a
fast
mail,
we
properly
everything
that's
connecting
to
an
external
server,
so
whoever
it
is
only
gets
knows
that
someone
from
fast
mail
connect
that
they
don't
know
which
user
they
don't
know
where
the
user
is
dependent
on
the
yeah
service,
to
be
able
to
provide
that
it's
not
inherent
yeah.
D
And
my
response
to
the
to
the
thing
I
think
we
never
really
I,
think
we
sort
of
came
to
some
agreement
on
that
was
that
this
is
such
a
general
problem
in
in
the
stuff
we
deal
with.
It
turns
up
your
eyes
turn
up
everywhere
and
and
the
the
the
issue
of
handling
things
like
this
really
belongs
in
almost
a
separate.
You
know,
RFC,
informational
or
otherwise.
A
D
It
is
it
just
seemed
not
say,
I
feel
that
this
particular
draft
was
being
picked
on,
but
how
many
drafts
we
had
and
specs
that
we
had
that
have
exactly
this
in
them.
They
have
for
a
URI
of
some
kind,
pointing
at
something
and
yeah
event.
Pop
happened
to
be
the
one
that
popped
up.
That
where's
me
thought
about
it,
but
I'm
not
sure
it's
the
best
place
to
actually
resolve
the
issue.
D
A
H
Calendar
entry
to
to
have
people
clicking
on
the
website
or
because,
if
you
don't
download
it,
then
it
means
you're,
clearly
tracking
the
person
so
I
think
we
it's
as
you.
As
you
mentioned,
it's
an
it's
something
that
has
been
probably
picked
up
and
we
don't
want
to
have
the
general
debate
into
that
document.
So
we
should
try
to
to
see
what
makes
the
ad
happy
on
that
for
that
draft
and
eventually
committing
in
handling
this
more
general
thing
in
another
document.
But
I
think
it's
going
to
be
hard
to
find
a
solution.
E
Otherwise,
this
is
very
similar
to
web
pages.
Actually,
I
think
it
will
probably
will
help
people
yeah
at
least
to
sort
of
cover
mental
model
of
how
they
think
about
this
or
to
emphasize.
This
is
not
particularly
different
and
I
think
also
talking
about
phishing
and
traction,
give
girls
being
automatically
referenced
difference.
That's
also
probably
good
texture.
D
E
D
And
if
we
had
a
separate
document
that
covered
these
things,
there's
information
so
that
we
could
update
I
guess
at
least
we
could,
when
these
things
popped
up,
we
have
them
to
point
people
at
say
a
third
of
this
sure.
Okay,
all
right
I'll
I'll
deal
with
that
I
think
that's
the
only
thing.
That's
really
outstanding
on
the
only
event
pub
thing
unless,
except
that
I've
started
to
get
bothered
about
it
in
relation
to
J's
calendar,
so.
A
D
A
All
right
next,
let's
talk
about
Jay
s
calendar
so
that
we're
skipping
ahead
a
slide
now
back
on
the
slides,
we're
going
all
the
way
down
to
slide
seven.
D
Well,
I
guess
guess:
since
I
suddenly
became
active
on
this
I,
better
say
where
I
got
and
why
and
how
so
when
I
have
been
sort
of
engaged
on
other
things
and
I
started
on
them,
building
myself
a
library
to
do
so.
They
chrome
to
iCal
for
Jay.
But
you
know
it
did
Jase
calendar,
but
not
quite
like
I,
know
anyway,
sort
of
a
library
and
I
built
all
that
part
with
the
intention
of
moving
on
to
getting
a
J
map
implementation
going.
D
Yes,
I
think
it'll
be
a
really
good
thing
for
the
people
who
take
feeds
out
of
our
system
and
but
the
next
step
I
realized.
We
I
had
to
convert
this
stuff
from
my
internal
representation,
which
sort
of
follows
the
iCalendar
model
but
isn't
and
convert
that
into
into
J's
calendar
and
the
first
two
properties
I
came
across
just
immediately
pose
difficulties,
and
it's
been
the
case
ever
since.
Every
every
property
I've
looked
at
there
have
been
its
raised
questions
and
if
it's
raising
questions
for
me,
it's
gonna
raise
questions
for
anybody.
D
It
it's
and
that's.
When
I
came
after
after
I
started
early
on
I,
really
we
have
to
have
a
very
explicit
mapping
with
lots
of
musts
and
as
to
how
you
get
from
I
Canada,
just
calendar
and
back
again,
because
if,
if
we
want
you
know,
if
we
want
people
to
migrate
to
using
J's
calendar,
I
cameras
not
going
away,
you
know
ten
years
down,
the
road
we'll
still
be
doing
high
counter
and
people
have
got
I
kind
of
embedded
deeply
in
their
system.
Many
people,
it's
the
data
model
they
use
internally.
D
It's
just
not
going
to
go
away.
It's
a
primary
part
of
what
we
do
so
the
so
somebody's
going
to
convert
this
stuff.
Send
you
Jake,
Andrew,
I,
counter
Jess
can
try
calendar
and
you're
going
to
have
to
convert
it
back,
so
you
have
to
know
exactly
how
they
did
it,
or
at
least
exactly
how
they
should
have
done
it.
D
So
that
was
that
was
the
motivation
behind
this
thing
that
we
can't
be
just
saying
implementation-dependent
anywhere
and,
and
it
does
affect
the
dress,
counter
draft
in
in
a
number
of
ways:
they're,
not
major
changes,
I
think,
but
they're
they're
changes
all
the
same.
You
know
the
the
issues
that
I'm
coming
across
are
yeah.
Well,
we
talked
about
anti.
We
need
to
look
at
some
my
year.
My
own
notes,
if
I
can
find
them
are
things
like.
D
D
We
can't
have
it
converted
one
way
and
then
come
back
again,
looking
different,
that's
all,
and
we
with
that
turns
up
in
a
number
of
places
that
there
are
places
where
we
have
accompanies
the
rel
type,
all
over
I
think
they
think
the
same
as
each
other,
with
contact
with
with
with
their
geo,
but
I
think
contact
was
another.
Wasn't
imaged
that
we
can't
do
that,
a
contact,
it's
absolutely
vital
in
in
an
event
publication.
Everybody
has
a
contact
and
yeah
down.
The
road
would
be
great
to
have
multiple
contacts,
but
we
can't
be.
I
D
D
Sorry
to
interrupt
NTS
I,
don't
understand
which
contact
you
are
talking
about
the
contact
property
in
I
counter
that
there
has
to
be
a
one-to-one
conversion
back
words
and
thoughts.
You
have
to
be
had
to
identify
exactly
where
it
went
and
how
to
get
it
back
and
there's
it's
true
of
a
lot
of
things
and
I
mean
how
do
we
do
with
it?
I
mean
not
that
many
people
use
the
geo,
but
just
because
not
many
people
use
it
when
you
scale
it
up.
It's
still
a
lot
of
people.
I
It
is
a
forty,
you
know,
yeah,
just
as
an
example
for
this
I
agree
with
very
most
what
you're
saying
for
geo.
As
we
already
discussed,
you
can
describe
it
in
a
location
as
it
is
defined
in
she
s
condo,
but
it's
again
a
good
example
of
those
were
I.
Canada
only
allows
one
of
these
properties
to
be
said,
whereas
cheese
can
multiple
instances
of
this.
This
is
indeed
an
issue
that
we
counter
at
various
places
in
the
mapping.
I
D
I
I'm
saying
but
I
agree
with
you
that
at
least
at
the
moment,
for
example,
in
this
iris
I'm
at
implementation,
we
are
using
heuristics
to
identify
which
of
these
changes.
Calendar
locations,
for
example,
is
the
most
is
the
one
that
most
properly
fit.
Saudi
I
carry
the
location
and
which
locations
are
these
that
we
put
into
an
X
location,
property
or
whatever,
and
this
is
indeed
showing
up
at
different
places
in
the
mapping.
So
it
needs
to
should
be
a
good
or
a
hard
rule
which
your
first
set
of
candidates
for
the
respective
I
clean.
D
And
it's
yeah
it's
sort
of
different
when
you're,
just
just
generating
I
counter
out
of
bureau,
if
he's
starting
off
with
a
J's
calendar
model-
and
you
need
to
export
this,
but
when
it
when
it
comes
back
at
you
after
having
perhaps
been
modified,
being
converted
back
to
Jay
s,
calendar
modified
and
maybe
somebody
added
another
another
location
and
then
sends
it
all
the
way
back
to
you
again.
How
do
you
know
which
one
was
the
geo
to
star
with
the
way
you
do
it
is?
D
It's
that
we're
not
talk
about
something
that's
difficult,
though
we
just
have
to
know,
and
and
and
these
that
turns
up,
that
kind
of
she
turns
up
all
over
the
place.
You
know
the
content
we
win.
Any
contest
are
not
even
allowed
fall
in
in
the
current
spec,
they
have
to
be
a
participant.
But
how
do
you
know
participant
is
that's
that's
where
the
tie
back
to
event
table
turns
up
in
Ischl
ii
was
I?
D
D
So
that's
that's
why
I
say:
there's
a
there's,
a
there's
that
there's
a
two-way
thing
going
on
here.
Is
that
when
I
stand
like
that,
well,
it
looked
like
I
need.
I
can't
really
changes.
I
mean
I,
realize
I
need
to
make
a
change
to
event
pop
to
to
to
make
this
work,
but
I
think
we
need
to
use
the
participant
type,
which
is
an
existing
in
Jeff's
calendar.
The
assumption
is
they're
all
attendees.
I
I
Translation
guidelines
with
icalendar
since
I
mean
I
know
it's
a
I,
know
the
real
world
pragmatic
thing
why
we
wouldn't
wanna,
put
some
guidelines
here
or
some
some
tools,
but
in
the
other
hand,
I
really
wouldn't
want
to
mix
these
formats
so
closely,
because,
basically
it
would
mean
it
chase
handled.
They
should
be
really
to
date
of
different
data
formats
and
necessarily
I.
D
Don't
think
what
I
don't
think
what
I
don't
think
I
pointed
have
to
be
able
to
move
on,
but
we
also
have
to
be
able
to
coexist
for
the
next
few
years.
No,
it's
a
decade
or
so
we're
gonna
have
to
coexist,
and
anything
that
is
done
in
iCalendar
from
now
on
has
to
has
to
it's,
it's
the
it
flips
it
around.
It
has
to
be
compatible
with
what
we're
doing
in
Jai's
calendar,
but
we
have
to
get
to
a.
D
We
have
to
be
at
a
point
now
where
we
can
do
a
a
reversible
mapping
as
near
as
possible
between
the
two.
So
if
you
start
off
with
icalendar,
you
can
turn
a
jeaious
calendar.
You
can
send
that
to
somebody
else.
They
can
turn
it
back
into
AI
calendar
and
get
pretty
close
to
what
you
actually
starred
with
and
and
and
it's
not
a
matter
of
guesswork
or
or
heuristics
or
anything
else.
D
It's
got
to
be
in
the
data
and
that
probably
means
if
they
absolutely
means
changes
to
the
current
J's
calendar,
spec,
to
reserve
certain
rel
types
to
add
some
flags
to
add
types,
the
participant
and
so
on
and
so
forth.
Then
I
don't
think
any
of
them
are
huge
changes,
but
I
don't
think
we
can.
We
can
just
push
the
mapping
up
to
one
side
and
say:
I
will
do
that
later.
It's
part
of
what
we
have
to
do
otherwise
I'm
gonna
end
up
like
we
started
off
in
can't
connect
yet
over
a
decade
ago.
D
The
whole
reason
for
having
I
can
calculate
was
the
terrible
interoperability
problems.
We
had,
you
simply
couldn't
send
an
event
to
somebody
and
have
it
having
to
understand
what
it
meant
they
would.
There
were
problems
all
over
the
place.
We
want
to
do
that
again
and,
and
that's
the
best
that
the
the
danger
of
not
specifying
this
properly
from
the
outset,
I.
A
Think
reserving
specific
case
sounds
totally
reasonable.
Gee
did
you
get
to
a
list
of
what
you
think
needs
to
be
done.
D
D
The
a
there's
another
issue
is
that
we've
buried
all
there's
a
with
Boyd
everything
inside
link
and
and
I
just
spent
an
entire
day
trying
to
implement.
How
would
I
deal
with
somebody
who
has
an
override
where
they've
just
removed
an
attachment,
and
now
I've
got
a
disentangle.
Let's
say:
we've
got
a
bunch
of
images
and
attachments
and
all
these
other
links
and
things
now
we're
going
to
disentangle
these
things
from
that's
that
collection
and
build
an
override
which
looks
right
for
what
we're
doing,
and
these
are
unrelated
properties
that
we've
buried
inside
link.
D
You
know,
there's
images,
there's
those
whatever
it
is
then,
and
maybe
we'll
have
to
put
some
other
stuff
in
there.
So
there
are
some
places
over
that.
I
think
that
we'd
be
a
lot
better
off
having
an
attachments
property
even
in
Jay's
calendar,
it's
a
different
usage
and
then,
when
you
get
a
bit
that
this
once
I
started,
doing
with
the
overrides
the
the
complexities
of
it's
difficult,
I
mean
it's
not
not
sort
of
difficult.
It's
just.
You
could
spend
entire
day
implementing.
D
I
I
D
Simply
because
their
usage
is
different,
I
mean
that
their
likelihood
is
that
if
you've
got
say
a
meeting
or
whatever
and
old,
it's
an
event
going
on
over
time
and
you've
got
different.
The
chances
are,
you
might
have
different
agenda
every
meeting.
The
other
thing
they
look
pretty
much
the
same
from
from
meeting
to
meeting.
But
you
change
the
agenda.
You
go
hospitals.
F
D
I
D
F
I
think
I
think
we're
all
in
agreement
that
documents
important
and
possibly
when
we
make
that
standards
track
rather
than
information
as
we
were
originally
planning
I
think
the
disagreement
is
on
how
whether
or
how
much
changes
would
still
be
required
juris
calendar
and
whether
to
hold
up
the
jurist
can
expect
waiting
on
that
all
come.
We
push
ahead
with
this
well.
D
Here's
the
I
I,
absolutely
believe:
we've
got
Update
discount
of
some
extent
for
this
to
make
this
work
and
and
I
don't
think
we
can.
We
can
avoid
that
I'm
I'm
in
my
own,
my
I,
just
don't
think
matters
too
much
one
or
the
other.
Whether
thing
is
is
is
is
one
document
or
two
people
there.
You
know
there's
enough.
People
complain
that
they'd,
be
they
don't
like
following
multiple
documents:
they'd
like
it
in
one,
but
then
there's
people
the
other
way
around.
D
But
it's
fine
I,
don't
think
that
that's
a
big
deal
but
I
think
the
two
are
inextricably
linked
that
that's
the
problem.
We
could
we
got
decades
of
eight
decades
or
years
of
having
to
deal
with
this
stuff.
The
likelihood
is
that
almost
anybody
who
wants
to
implement
Jai's
Kerala
is
going
to
have
to
do
that.
Mapping
there'll
be
a
few
people
who
have
the
luxury
of
being
able
to
look
at
just
Jai's
calendar
there'll,
be
the
people
who
maybe
have
a
web
page
they're
trying
to
display
this
stuff
on.
That's
it.
D
Everyone
else
is
gonna,
be
doing
the
mapping
and
a
partner
if
the
two.
If,
if
we're
for
reserving
stuff
in
Jeff's
calendar
for
the
mapping,
it's
a
little
hard,
not
okay,
my
only
my
only
concern
about
as
I
said
about
not
having
it
as
a
as
a
embedded
in
the
document
is
it
makes
it
easier
for
people
to
ignore
it
and
then
just
do
their
own
mapping
that
doesn't
work.
That's
that's!
That's
my
feeling
on
it.
D
I'm
prepared
to
I
mean
I,
I'm,
gonna,
insist
it
be
one
document
I
think
we
better
is
one
document,
that's
my
opinion
and
but
I
don't
think
we
can
weaken
them
until
we
have
substantially
completed
the
mapping.
I,
don't
think
we
can
say.
J's
count
is
done.
That's
my
feeling,
and-
and
you
know
we
I
don't
think
this
is
what
amount
years
it
depends
how
fast
people
respond
to
any
other
questions.
I
put
on
the
list,
I've
known
a
single
response.
D
Yet
so
you
need
to
get
on
there
and
start
responding
to
some
of
these
things.
I
mean
yeah.
I
could
be
wrong
on
some
of
the
issues,
but
some
I
know
I'm
writing
some
of
them.
So
we
need
to
resolve
it,
and,
and
and
I
can
do
this
I'm
trying
to
work
through
this
thing
to
some
extent
in
a
piecemeal
basis
and
I
started
trying
to
implement
the
stuff
one
at
a
time
and
then
thought
this
is
gonna,
be
too
slow.
D
At
least
we're
gonna
make
a
good
pass
through
every
property
and
what
I've
been
finding
is
the
way
there.
Any
we
figure
out
where
the
issues
are
is
at
least
take
five,
five,
four
five
and
work
through
every
damn
property
in
there
and
try
converting
the
examples,
because
I
keep
running
into
problems
as
I.
Do
that
and
I
don't
believe
all
my
problems
that
are
illusory
or
anything
else,
I
think
they're
real.
H
H
I
had
the
impression
that
we
stated
that
I'm
gonna
do
two
to
document
so
I'd
like
us
to
keep
that
full
on.
With
that
decision
we
made
and
to
keep
I
mean
the
update
is
frigid
I
mean
we
may
improve
gs
calendaring,
but
like
then,
the
update
should
be
mostly
related
to
the
model.
For
example,
we
we're
missing
something
more,
so
maybe
the
conversion
that
can
help
into
that.
But
I'm
I
would
like
to
keep
those
translation
and
the
model
as
two
separate
document.
D
Find
that
I'm
not
gonna
push
that
and
but
I
think
yet
there
are,
there
are
tree.
I
I,
think
you're
substantially
J's
counter
is,
is
gonna
work
if
there
is
missing
stuff
like
say
a
participant
type,
and
then
there's
there's
things
like
reserving
values
for
the
mapping
which
I
think
has
to
be
in
the
in
the
J's
counter.
Spec
yeah
I
substantiate.
H
H
H
D
H
D
H
D
How
you,
how
you
increment,
but
I,
we
shouldn't,
fall
in
the
trap
of
doom,
what
what
I
can
I
did
of
having
a
version
and
never
changing
it?
We
I
realized
a
week
or
so
ago,
when
I
said
we
should
have
a
version
number
we
should
use.
It
was
I.
We
we
must
spend
half
our
time
when
we
try
and
add
another
account
of
property
or
a
change
I
counter
going
on
about
backward
compatibility,
and
it's
just
ridiculous.
We
end
up
okay,
you
you
supposed
to
ignore
anything,
you
don't
understand.
D
C
A
F
J
It's
probably
the
best
we
can
do.
We
absolutely
should
not
wait
for
a
PCP
on
so
don't
take
that
off
the
table.
That
doesn't
make
sense.
We
probably
should
work
on
such
a
document,
though,
but
yeah
I
guess
that
suggestion
is
about
the
best.
We
could
do
I'm
a
little
worried
about
an
option
to
to
tell
the
client
to
download
all
URLs
or
your
eyes.
I
would
rather
have.
J
Maybe
it
may
be
at
the
should
level,
that's
the
best
we
should
do.
There
might
be,
you
might
add,
a
little
more
text
to
explain
the
issue
with
the
the
security
issue
with
downloading
you
our
eyes
and
and
recommend
at
a
non
normative
level
that
the
user
be
asked
for
each
URI
user
be
warned
about
each
URI.
You
know
that
this
could.
J
Of
whitelist
right,
yes,
whitelist
would
be
good,
so
a
little
more
text
than
that
would
be
helpful,
but
yeah
I
think
that's
all.
That's
really.
All
that
makes
sense
to
do
the
problem.
I'm
looking
at
is
people
embedding
URIs
that
distribute
malware
people
embed
in
URIs
that
have
that
have
one-time
use
issues
where,
if
the
client
downloads
them
without
explicit
action,
it
could
use
them
up.
I'm
worried
about
URIs
that
cause
side-effects
like
or
just
purchase,
one-click
purchases
or
unsubscribes,
or
things
like
that.
D
Okay,
so
let
me
say
so:
I
said:
if
I
come
up
with
some,
if
I
come
up
with
some
text
and
I
I'll
circulate
on
the
list
and
make
me
we
can,
maybe
can
be
a
little
more
than
just
you
know
a
brief
mention.
Perhaps
we
could
put
his
ace
I
know
as
a
subsection
somewhere
on
handling
your
eyes,
so
I
think
if
we
can
put
something
in
there,
I
mean.
Perhaps
you
know
that
verifying
the
you
are
eyes
to
some
extent
before
you
download
them.
J
Warning
of
the
the
issues
that
can
arise
a
recommendation
to
to
have
a
white
list
and
to
check
with
the
user
on
things
that
are
not
on
the
white
list:
yeah
I'm,
not
looking
for
pages
I'm.
Looking
for
a
couple
paragraphs
but
yeah
I,
think
that
would
do
it.
Please
CT
me
explicitly
so
I,
don't
miss
it
when
you
put
it
on
the
list.
Okay,.
J
D
C
And
I
saw
the
emerging
draft
that
is
Jay,
scarred
or
Jewish
contact
assume
it's
called
better
than
the
cut
part
of
the
world,
and
so,
as
I
was
always
one
background,
I'm
from
from
Cydia
the
concise
definition,
the
division,
language
and
I'm
from
from
doing
basic.
These
ITF
versions
of
the
things
that
some
people
call
schemas
for
Jason
and
and
binary
equivalent
supercenters
the
Seaboard
and
an
a
big
fan
of
iCal
and
I
card.
C
So
I'm
natively
moved
here
basically,
because
I
am
interested
in
this
work
in
general
personally
and
then
I've
saw
oh
you're,
doing
Jason
definitely
was
trying
to
read
and
understand
them
and
everything
that's
just
written
in
English
text
I
was
like
okay.
Maybe
you
would
be
interested
in
having
this
in
a
data
format,
data
definition,
language
type
of
way
that
would
accommodate
both
Jay
s,
:
affairs
Carol,
as
is
chase
card,
and
because
I
assumed
that
you
would
have
better
readability
and
very
fine
structure
that
you
can
validate
and
create
by
so
the
construction
part.
C
The
messages
that
are
basically
age
is
cotton.
J
is
Cal
the
moment
you
have
very
complex
and
sometimes
a
little
bit
ambiguous.
Sorry
only
my
and
my
high,
but
the
120-day
singer,
CDI
for
Carly
and
I,
try
to
to
make
sense
of
it,
so
the
ambiguous
wording
there,
and
so
my
question
to
this
group
here
before
I
write
anything
just
out
of
the
blue
to
the
list
and
by
the
way
hi
Alex
I
would
like
to
ask.
I
I
D
This
gets
back
to
my
point
I
made
earlier
on.
It
would
be
great
if
the
ITF
didn't
treat
these
documents
as
as
immutable
when
they
really
aren't.
We
could
just
bring
one
out
down
the
road
and
say
this
is
the
document
and-
and
we
could
add
things
like
that
as
an
appendix
and
people
would
always
get
pointed
the
latest
thing,
maybe
maybe
we
could
help
with
calculate
and
package
all
this
stuff
together
in
a
way
and
pointer
these
things,
it's.
C
And
to
add
stuff
to
things,
the
the
data
definition
could
include
extension
points.
For
example,
I
saw
that
by
definition,
some
of
the
major
items
I
think
metadata
items
was,
and
the
text
highlight
these
maps
of
these
objects.
I'm
sorry
objects
are
too
intended
to
be
extensible
and
have
more
keys
in
them
later
on,
and
this
can
be
accommodated
today.
We
are
the
formal
definition,
so
there
can
be
additional
RFC's
to
just
create
extensions
to
this,
or
if
you
have
a
fundamental
change,
you
can
change
the
basic
structure.
The.
D
History
of
J's
calendar
is
gonna,
be
exactly
the
same
as
a
history
of
AI
counter.
In
that
respect.
In
them
we
have
been
constantly
adding
stuff,
and
this
is
why
I
referred
to
earlier
with
versioning
is
that
we
add
properties
to
iCalendar
or
we
add
parameters
to
I
count
of
properties
or
we
change
how
we
try
and
add
new
value
types,
which
is
the
worst
thing
you
can
possibly
do
in
iCal,
but
because
there's
no
versioning
and
and
everything
we
do.
We.
D
We
have
to
worry
about
breaking
the
older
clients,
which
will
just
swallow
the
stuff
and
then
explode.
So
we
got
to
assume
that
J's
calendar
will
be
extended
in
in
very
many
ways
as
as
time
goes
on
and
any
any
specification
has
to
allow
for
that
and
that's
where
versioning
would
would
help.
If
you
had
a
version
that
you
knew
explicitly
what
was
allowed
in
that
version
by
the
spec,
then
you
could
just
pause
it
with
the
appropriate
version
and
anything
you
didn't
understand
will
just
get
shuffled
off
to
the
side
as
being
stuff.
C
You
really
need
versions
for
that
due
to
the,
if
you
have
fine,
grained,
extensibility
and
and
using
objects,
objects,
I
think
as
some
are
in
the
lane
of
you
might
not
need
versions
here.
So
you
can
just
omit
items,
you
don't
know
and
the
items
you
know
or
basically
do
two
extensions
that
you
know
you
can.
You
can
have
provide
a
procedure
for
so
yes,
why
I
assume
versions,
look
like
a
good
idea,
always
I
assume
that
never
be
a
single
version
being
deployed
and
therefore
it
creates
a
rift
a
fraction
or
something
like
that.
C
D
Not
it's
not
the
extensibility,
adding
adding
new
stuff
has
always
been
relatively
easy,
then
add
a
new
property,
you
can
add
a
new
component
or
whatever,
though
the
component
blew
up
most
the
libraries
that
they
were
around
they
it's
when
you,
when
you
make
changes
to
existing
properties
and
ingest
counter
changes
too
existing
objects
that
if
we
we
have
something,
that's
incompatible,
they
just
some
extent
or
an
assumption
changes.
We
realize
we've
got
a
problem
and
need
to
fix
it.
We
could
have.
D
Well,
I
did
that
and
it's
a
pretty
imperfect
way
of
dealing
with
it,
but
yeah
there
are
ways
around
it,
but
they're
always
euros
working
around
an
issue
that
would
have
been
much
better
dealt
with.
By
simply
saying
this
is
version,
you
know
X
and-
and
this
is
how
you
interpret
it-
I
mean
that
the
parsing,
the
thing
is
one
thing:
we're
not
changing
the
structure
which
it
say
it
doesn't
that's,
not
a
problem.
You
can
pass
it
all
out
and
to
into
the
appropriate
objects
and
things.
A
A
Issue
is,
if
you
have
something
at
nose
head
of
a
past
version,
3.2
and
version
3.3
comes
along
and
it
changes
the
semantics.
You
don't
know
how
much
has
changed
the
semantics
until
you
go
into
semantic
versioning
horrors,
and
so
you
don't
know
how
much
for
you
can
still
understand.
Can
you
understand
any
of
it
or
none
of
it
until
you
get
upgraded
to
three
partha?
Well
change.
C
F
D
F
With
icalendar
was
the
fact
that
you
wasn't
an
object,
it
wasn't
a
component
in
I
kind
of
the
format
for
these
things
and
say
you
had
cram
just
flying
over
the
place,
but
adding
that's
not
from
the
jessman,
because
lot
of
things
are
which
objects
already.
So
you
a
participant
is
a
rich
object
and,
yes,
you
can
easily
add
extra
properties
to
that.
That
is
the
whole
point
of
the
registry
that
is
created
a
great
length.
The
end
of
the
dog.
C
Didn't
want
to,
despite
a
extension
version,
discussion,
I'm,
sorry
I
just
wanted
to
provide
more
instructions
way
to
to
basically
describe
what
you
are
defining
right
now.
I
think
that
the
extension
inversion
and
whatever
you
want
to
do
here,
will
be
expressed
in
that
definition.
In
any
case,
so
now
there's
none.
Maybe
there
will
be
some
indication
of
a
I,
don't
know
qualified
name
and
maybe
even
a
version
of
the
thing.
You
are
the
revision
here
and
then
that
is
fine.
C
C
J's
contact
thing,
so
I
can
paste
and
it
took
me
like
three
hours
to
create
a
specific
specific
definition
and
I
can
paste
the
URI
in
the
chat
window
right
now
you
go
so
this
I
brought
in
the
last
two
and
a
half
hours
literally.
So
by
reading
the
draft
and
creating
the
the
data
definitions,
it's
the
link
is
a
candy
I
hope
that
everybody's
fine,
with
that
it's
a
cloud
web
2.0
thingy,
and
so
it
shows
basically
the
the
how
I
would.
C
C
But
you
have
the
definition
of
chemicals,
so
it's
related
uppercase
and
so
all
these
names
that
are
actually
going
over
the
wire
sorry,
the
bottom
of
the
document
and
at
the
very
top
of
its
basic
structure.
As
you
can
see,
it's
a
very
big
document
and
it's
taking
basically
a
screen
size
input
here.
That
is
the
whole
definition
of
the
thing,
leaving
out
all
the
labels
that
are
basing
the
frosting
up
there.
So
I
think
it's
very
comprehensive
and
and
yeah
ssame
name
implies
it's
concise
and
it
hates
to
talk
about
things
and
and
I.
C
C
Yeah,
you
are
using
different
ways
to
group
things.
For
example,
the
preferred
contact
language
is
an
iteration
of
raised
by
diffusion
and,
for
example,
in
4,
is
the
iteration
of
maps
by
definition.
So
you
could
your
mom
organize
this.
You
don't
need
to,
but
you
can
see
this.
Basically,
if
you
look
at
this
ad
hoc
also.
A
C
From
the
top
of
my
head,
the
only
RC
that
is
using
this
was
not
an
RFC.
It
was
cozy
the
concise
object.
Signing
and
encryption
thing
is
basically
a
complement
to
Jose,
so
cozy
uses
CDJ
notation,
but
only
information,
because
at
the
time
of
RFC
in
cozy
CDT
I
was
not
the
RFC
yet
so
there
are
other
things
in
the
queue.
I
don't
know.
E
C
D
C
To
drop
out
so,
but
you
have
an
example
again:
I've
whipped
this
up
in
less
than
three
hours
and
so
I
assume
that
even
Jake
al
will
have
an
estimated
effort
of
like,
as
in
two
days
and
one
day
of
integrity
checks
because,
as
you
said,
there
might
be
consistencies
between
English
prose
and
that's
this
one
here.
But
but
my
assumption
is,
it
is
not
that
hard.
Oh.
C
C
I
A
I
D
A
A
D
D
Yeah
I
mean
I,
basically
wrote
the
whole
thing
in
one
one
wrong
sitting
in
a
hotel
and
presented
and
there's
been
no
real,
substantial
objections
to
it.
But
I
think
Neil
came
up
with
something
at
some
point.
I
may
have
incorporated
that
need
to
go
back
through
my
notes,
but
I
bring
out
another
draft.
So
it's
at
least
I
presume.
It's
expired,
I
thought
I,
don't
think
it's
worth
paying
time
discussing
for
the
month,
especially
as
were
running
later.
Yeah
subscription
upgrade.
A
D
I
did
that
again,
I
guess
I'll
go
says
we
can
discuss
it
on
the
list.
Oh
check
that
and
trying
I'll
go
through
this
and
the
next
few
days
and
I
can
I
can
push
that
out.
I
think
we
were
pretty
close
to
to
agreement
on
most
of
the
things.
So
so
it's
I'll
see
if
I
did
that
and
if
not
I'll
do
it
and
then
shove
it
out
cool.
B
H
H
I
think
the
main,
the
main
aspect
is
just
not
ignoring
those
privacy
issues
but
a
concerns,
but
there
is
nothing
more
we
can
do
about
it.
I
mean
it's
I.
Think
it's
I
mean
my
understanding
was
that
we
should
more
acknowledge
that
there
is
a
risk
so
that
people
using
it
or
implemented,
are
really
aware
than
trying
to
find
out
a
solution
of
a
very
complex
problem.
B
H
A
I
B
Removed
it
because
of
the
difficulty
in
how
a
Cal
deaf
client
notifies
a
server
that
it
wanted
to
remove
a
default
alarm.
Apple
does
some
wacky
stuff
by
basically
postdating
or
I.
Guess
we're
predating
the
alarm
so
uses
like
1976,
something
something
something
it
was
kind
of
a
hacky
way
to
get
around
it
and
a
while
ago
the
group
decided
is
to
take
it
out,
or
at
least
we've
decided
and
Cal
connect
to
take
it
out.
I
D
Wasn't
the
whole
problem
with
default
alarms?
Really
yet
another
backward
compatibility
issue
I
mean
it's
not
the
idea
of
default
alarms
is
fine,
I
think
and
I
think
you
could
implement
them
quite
nicely
if
you
didn't
have
the
I
count
of
baggage.
So
there's
two
issue:
it's
one
of
those
things
in
it.
We
got
two
issues
is
how
do
we?
A
D
Least,
I'm
just
kind
of
extensions
people
I
think
I
made
mostly
changed.
I,
don't
think
we've
iterated,
yet
I
I
think
the
big
change
to
V
Paul
was
to
it
was
to
use
participant
instead
of
the
voter
property,
and
it
was
a
lot
of
changes.
I
think
I've
made
most
the
changes
I
need
to
make
another
check,
and
then
ken
was
going
to
help
me
move
it
along
so
I
all
there
I
leaned
over
to
do
that
and
then
near
future.
D
D
D
I
think
I
think
that
I
think
maybe
in
the
end
we
should
just.
We
should
just
go
ahead
and-
and
we
didn't
have
a
plan-
I
mean
the
plan-
was
to
try
and
and
communicate
with
Apple
and
see
how
what
we
had
matched
up
with
what
they
had,
and
why
did
they
raise
it
in
the
first
place
as
if
they
didn't
have
it
yeah
and,
of
course,
with
circumstances?
We've
we've
not
managed
to
have
that
discussion
yeah.
B
D
I
well
I
mean
from
reading
a
mozilla
document,
which
was
it
no.
It
wasn't.
It
was
say
one
of
Emmet's
documents,
I
think,
is
how
to
interact
with
these
things,
like
I
thought
the
date
predated
the
now,
it's
quite
possible
that
what's-his-name
moans
gone
blank,
the
decline,
Apple,
client
guy
and
raised
it
it's
possible.
He
wasn't
even
aware
that
it
existed
as
a
service.
It
was
a
server-side
thing
which
is
pre
buried
deep
inside
the
client.
H
D
Raised
it
and
said
they
would
like
to
have,
it
was
the
Apple
client
guy
I
think
is
back
in
Tokyo,
said
it'd
be
great
to
have
server-side
subscriptions,
because
there
are
all
sorts
of
things
you
can
do.
When
you
have
server-side
subscriptions
like
even
cache
stuff,
you
only
need
to
download
it
wants
popular
things.
You
could
have
thousands
or
millions
of
subscriptions
the
same
thing.
H
D
D
Like
web
Cal
feeds
and
all
the
rest
of
it-
or
you
subscribe
to
this
thing-
and
it's
just
downloaded,
maybe
by
the
client
or
or
whatever,
but
when
people
are
or
subscribing
to
the
same
things,
it
curly
there's
a
opportunity
to
make
things
work
better
and
also
it
does
mean
that
it's
in
your
shared
it's
a
shared
thing
in
you
know
when
you've
got
multiple
devices,
you
can
you
subscribe
once
and
and-
and
it
turns
up
in
all
your
clients,
so
there
are.
There
are
many
benefits
to
doing
it.
Okay,.
D
A
D
D
D
D
A
D
Yeah,
sorry,
can
we
go
back?
Can
we
go
back
to
the
previous
one?
I
heard
suicide
subscription,
but
you
were
told
a
subscription
upgrade
one.
Yes,
sorry
yep,
oh
yeah,
there's
almost
nothing
to
do
with
that,
make
it
July.
So
not
everything
is
in
June,
I
I
heard
the
wrong
thing
and
I
server-side
subscriptions
I.
Think
that's
the
one
where
that
probably
is
October,
because
we
need
to
try
and
make
some
have
some
contact
with
the
album
and
the
wristlet.
Alright.