►
From YouTube: IETF106-CALEXT-20191118-1330
Description
CALEXT meeting session at IETF106
2019/11/18 1330
https://datatracker.ietf.org/meeting/106/proceedings/
A
B
D
C
D
C
D
C
D
E
D
People
know
there
on
the
front
on
the
front
hall.
Don't
trust
don't
trust,
my
co-chair,
it
doesn't
know
what
he's
talking
about.
D
Alright,
it
was
definitely
an
employment
with
it.
E
D
What
is
going
on
here,
I
already
put
your
name
on
push
it.
Alright,
still,
no
sound
for
Mike
I.
B
E
D
B
Ok,
this
is
Barry
liebe
I,
don't
know
if
we
have
critical
mass
here
and
was
really
hoping
that
this
would
have
gotten
discussed
at
Cal
connect
where
the
right
people
are
really
there
for
it.
But
the
the
gist
of
my
discuss
on
this
is
that
there
there's
a
lot
of
malware
and
spam
and
other
crap
coming
through
calendar
things
that
calendar
clients
often
automatically
resolve
dereference.
You
are
eyes
and
there
have
been
generally
limited.
Use
of.
You
are
eyes
in
in
the
calendar
standards
to
this
point.
This
draft
add
the
whole
boatload
of
new.
B
You
are
eyes
into
this
I'm
really
concerned
about
what
will
happen
if
calendar
clients
start
dereferencing
them
left
and
right
and
and
they're.
The
places
that
the
URIs
are
in
are
our
places
that
I
think
would
be
likely
to
have
clients
automatically.
Do
you
reference
them?
So
we
need
to
say
something
about
that.
My
my
first
preference
would
be
to
change
the
draft
to
not
have
URIs
in
places
where
it's
not
really
necessary.
B
If
we
can't
do
that,
if
that's
not
practical
or
don't
meet,
doesn't
meet
the
needs
of
the
users
or
whatever,
then
the
next
best
thing
would
be
to
have
a
very
strong,
very
strong,
advise
against
dereferencing
these
things,
without
user
interaction
and
warnings
to
the
user
that
this
could
cause
bad,
which
is
a
technical
term.
So
I'll
stop
talking
and
you
can
let
Mike
talk.
F
B
F
Is
I
I
do
agree
with
your
points,
but
there's
URLs
that
could
be
automatic,
dereference
turnip
all
over
the
place
and
I
I.
Just
don't
feel
that
in
your
the
event
pub
draft
slash
spec
is,
is
the
is
the
best
place
to
deal
with
it?
We
talked
about
this
as
that.
Possibly
what
we
probably
need
to
do
is
come
up
with
a
say:
I
know
some
some
know
best
practices
thing
that
is
standing.
F
F
F
Some
kind
that
that
deals
with
what
what
a
clients
do
when
they
get
links
to
things,
and
you
know
they
can
have
an
email.
They
get
open,
link
headers
there.
They
can
turn
up
in
all
sorts
of
places
and
and
yeah.
I
I
I
guess
I
guess
the
event
pub
draft
does
add
some.
I
don't
think
it
I,
don't
think
it
there's
an
explosion
there
there
in
your.
F
If
you've
got
a
location,
that's
based
on
a
vCard,
it's
a
reference
to
the
V
card,
and
but
we've
had
them
all
the
time
in
in
in
encountering,
we
had
the
the
river,
it's
called
the
URI
property,
which
you
know,
clients
could
have
followed
blindly,
but
there's
no
guidance
for
that
and
I
I
I
get
your
point,
but
I
just
don't
think
it's
the
best
place
to
deal
with
it.
So.
B
F
B
F
F
H
I
wanna
said
the
the
main
problem.
One
is
the
style
description,
I
think
it's
called
in
here,
which
is
basically
lets
you
put
in
an
HTML
version
of
the
message
body
or
calendar
description,
which
a
bunch
of
people
are
doing
anyway,
just
in
the
regular
description
property,
even
though
it's
meant
to
be
plain
text
what's
unclear
to
me
is
why
that
is
I
think
that
one
will
be
better
not
being
your
I
both
because,
because
that's
the
one
you
have
to
actually
fetch
to
display
the
event.
H
When
you
open
it,
everything
else,
I,
don't
think
you
necessarily
do,
but
that
one
you
do
and
also
it
means
it
can't
just
be
arbitrarily
changed
without
actually
updating
the
event
data
later,
which
is
I,
think
a
good
thing.
I,
don't
think
you
want.
We
wanted
to
just
randomly
change
under
you
each
time
you
open
it
for
spam
scanning
all
sorts
of
things
like
that.
If
nothing
else
and
in
jeaious
calendar
we
let
you
set
the
content
type
of
the
description,
so
you
can
put
an
HTML
description
in,
but
it's
not
your
I
reference.
I
F
Well,
here's
a
long
history
behind
all
all
this,
but
the
the
the
idea
was
to
try
to
make
locations
and
and
participants,
and
all
this
it'd
be
based
on
information
that
came
out
of
things
like
vCard
right.
That's
where
the
information
is
is
in
directories,
and
you
know
they're
they're
promised
at
you.
If
you
copy
information
into
something
like
it
like
a
a
counter
event,
it
neva
tably
gets
out
a
date.
You
don't
have
any
idea
of
where
you
go
to
get
the
most
up-to-date
version
of
that
information.
B
F
B
H
Attachments
are
generally
just
a
URL
to
where
I
can
download.
This
attachment
from
that
difference
to
me
is
that,
for
privacy
anyway,
is
that
that
is
explicit,
user
action
to
download
that
or
in
it
almost
everywhere,
there's
no
just
I
open
any
of
engineer
shows
yeah
so
and
I
think
there's
a
big
difference
being
between
being
able
to
refresh
something
by
giving
the
UI
of
where
I
can
fetch
an
updated
thing
at
some
point
versus
I
have
to
fetch
this
just
to
show
it
in
the
first
place.
H
B
H
H
I
think
this
is
this
is
diverged
a
bit
from,
what's
in
desk
calendar
for
a
similar
kind
of
location,
object
and
stuff,
which
I
expect
it
to
kind
of
be
a
mirror
in
that
that
tends
to
embed
all
the
information
that
you
can
can
always
add
links
to
external
places.
If
you
want
to,
if
the
user
wants
to
fetch,
more
or
refresh
or
other
stuff,
but
is
never
expect
it
there,
that's
gonna
be
fetched
automatically
just
to
show
the
event,
because
the
information
is
already
there.
So.
B
That
would
make
me
much
happier
having
the
information
that
you
expect
to
retrieve
from
the
URI
the
current
version
of
that
information
in
there
have
the
URI
available
to
refresh
it
and
have
a
warning
text
in
the
document
about
what
can
happen
both
from
a
security
and
a
privacy
point
of
view.
If
you
start
going
and
doing
that,
I
I
could
live
with
that
that
that
would
be
good
I.
B
B
D
H
We
actually
need
ten
minutes.
It's
been
in
last
call
since
the
previous
meeting.
Basically
what's
what's
happened
since
then,
is
mainly
just
creating
registries
for
everything,
as
discussed
in
Montreal,
so
that
it's
easily
extensible,
which
is
one
of
the
goals
of
it
so
and
I've,
had
some
good
feedback
in
the
last
call's
have
been
a
few
minor
tweaks
based
on
that,
but
I
think
I
should
publish
the
new
version
with
just
the
final
tweaks
that
have
been
made,
and
hopefully
we
can
push
that
out.
I
Yeah
so
Daniel
yeah
regarding
the
registry
I
sent
those
to
the
IANA
for
an
early
review,
I'm
realizing
either
mr.
email
or
haven't
received
any
feedback.
So
I
have
to
check
that
through
the
IANA
yeah,
so
yeah
so
I
hope
we
made
the
right
thing,
I
think
yeah
most
likely
we
even
do
more
because
I
don't
think
it
was
pretty
required
to
have
all
this
registries
at
the
I
knob.
I
So
I
think
we
I
mean
we're
good
students
at
that
point,
but
I'm
waiting
for
confirmation,
and
once
we
have
the
confirmation,
as
well
as
the
IPR
from
Robert
I,
wouldn't
get
a
cent.
But
that's
another
problem.
It's
just
it's
resolved.
In
a
few
weeks.
That's
a
yeah
yeah.
My
my
console
is
most
with
the
IANA
and
it's
one
of
the
document.
I
think
we
got
the
most
review
on
so
I.
Think
it's
a
it's
a
good
example
cool.
J
H
It
just
because
it's
on
the
slide,
there's
actually
a
second
document
being
discussed
here,
which
is
the
informational
one
about
translating
between
J's
calendar
and
I
calendar.
There's
been
no
work
on
that
since
Montreal,
but
yeah
I
think
Robert
will
finish
that
off
when
he
gets
back
in
a
month
or
so,
and
it
doesn't
need
to
hold
this
up.
So
just
short
update
on
that.
H
D
F
F
D
Great
I
had
a
question
about
status
deleted
and
how
that
works
for
items
that
have
multiple
recurrent
so
days.
Do
you
always
have
to
send
all
the
recurrences
if
what
any
of
them
are
changed
and
if
you're
deleting
a
recurrence
to
send
the
status
to
latest
occurrence
ID,
or
are
you
just
sending
all
of
what's
left
for
that
recurrence?.
F
D
D
F
D
H
This
is
yeah
I
read
through
this
is
this
morning.
It
seems
mainly
find
there
were
a
few
question.
Does
somebody
else
send
through
to
the
list?
I
thought
the
water
spirit
was
kind
of
section.
7
is
where
it
says,
there's
a
number
of
different
ways.
You
can
get
stuff,
you
can
get
a
card
out
thing.
You
can
get
a
coward
eye
thing
with
auth
and
they
don't
have
almost
no
description
to
them
and
they
don't
actually
currently
have
references
to
the
specifications
which
they
definitely
need.
Yeah.
H
F
Get
was
the
enhanced
get
was
where
I
see
that
being
useful
is
is
in
relatively
primitive
clients.
Mostly
you
know
in
in
my
implementations
and
deployments.
We
have
a
lot
of
web-based
feeds
that
can
work
very
nicely
with
this,
that
they
they
just
look
for
updates
and
and
then
they
update
their
view.
So
our
lightweight
synchronization
process
works
works,
pretty
well,
you'd
have
to
implement
CalDAV.
We
told
wrapping
a
CalDAV
subset
for
a
while
and
then
decided
that
really
what
we
should
do
is
just
implement.
F
F
The
biggest
change
with
the
whole
thing
is
people
realizing
that
the
biggest
gain
for
this
is
in
mobile
clients.
We
sort
of
talking
from
server
perspective.
We
started
this
I
think
the
biggest
gain
is
for
mobile
clients
that
are
downloading
huge
these
times
and
upgrading
to
something
more
efficient,
probably
means
battery
savings
and
goodness,
as
Wells.
H
D
I
What
what
we
did
is
that
we
asked
for
the
GOP
people,
so
the
former
chairs
of
that
working
group,
and
he
actually,
he
told
us
he
well.
It
advised
us
to
to
send
our
comments
on
mist,
which
is
not
so
much
helpful
for
now,
because
I
think
the
the
real
problem
is
that
if
we
have
no
perspective
to
end
up
with
that
documents,
it's
going
to
be
hard
to
find
it
the
resource
and
the
energy
to
move
that
document
for
one.
I
So
back
to
the
comment
we
had
regarding
its
Joe
privacy
I
think:
well,
there
is
a
j.g
reprieve
framework,
so
we
should
not
ignore
that
one
and
I
think
the
previous
versions
of
the
draft
were
a
little
bit
ignoring
this
framework.
So
I
think
we
should
consider
that
framework,
but
I'm
regarding
I
mean,
if
you're
able
to
to
produce
the
localization
aspect,
I
think
it's
still
better
than
having
no
specifications.
I
If
we
do
not
think
while
full
GPS
information
is
going
to
be
provided,
so
it's
still
better
and
in
most
cases
the
people
you're
providing
you're
asking
for
this
time
zone
are
the
same
you're.
Providing
your
GPS
information
because
they
also
map
services,
so
I
think
well.
Privacy
is
important,
but
at
that
point
we
also
have
to
balance
with
what's
the
impact
of
not
doing
anything
and
what
would
be
the
impact
given
that
it's
not
a
single
the
only
application
that
use
geo
localization.
I
So
my
what
I
would
like
is
that
we
we
make
sure
that
this
draft
could
be
moved
forward
with
that
perspective,
and
so
I
mean
so
can
can
work
on
that
and
I
think
what
I'd
like
to
really
to
avoid
is
that
we
improve
the
draft
very
close
to,
and
then
we
we're
back
oh
yeah,
but
you
leak
some
information
yeah.
We
will
need
some,
so
we
should
work
on
leaking
the
smallest
amount
of
those.
J
Alexei
Melnikov
I
think
in
general
my
advice
would
be
state
your
assumptions.
You
know
if
you
are
saying
well,
if
you
are
leaking,
this
information
will
read
you
leaking
it
throughout
the
channels.
You
know,
at
least,
if
you
write
it
down,
people
might
agree
or
disagree
with
you
and
general
I.
Think
people
are
more
tolerable
to
you,
stating
that
there
are
certain
limitations
or
certain
risks,
as
opposed
to
saying
nothing.
D
C
F
F
Organizations
tend
to
produce
very
long-running
recurring
events
in
which
every
instance
is
modified,
but
in
reality
that's
sort
of
true
but
meetings
where
every
occurrence
every
currents
has
a
has
an
agenda
in
it.
So
the
the
you
know
the
the
recurrence
model
as
it
was
works
very
well
for
something
it
doesn't
change
much
like
birthdays,
it's
very
bad
for
meetings
and
for
the
kind
of
thing
that
libraries
and
places
do,
which
is
what
they
call
a
series.
F
F
H
D
F
H
Fine
good
so
yeah
III,
said
I
think
the
idea
is
good
if
we
need
it
means
various
things
for
start
to
make
sure
it's,
interoperable
and
I
think
we
want
to
be
clear
how
this
interacts
with
CalDAV
and
whether
a
service
should
be
generating
these
to
me.
It
has
to
be
a
server
generating
these
for
it
to
be
useful.
It's
it's
kind
of
it
wishy-washy
at
the
moment
that
a
client
or
server
may
generate
these
ahead
of
time.
H
There
were
few
other
things.
I
thought
were
a
bit
unclear
how
the
split
concept
I
did
not
understand
at
all.
It
didn't
see
what
was
going
on
there
and
there
was
a
discussion
that
the
series
master
is
not
a
member
of
the
series
which
so
it's
like
a
template,
I
presume,
which
makes
sense
that's
good,
but
that
means
that
you're
not
meant
to
show
that
as
an
event,
but
any
client
that
doesn't
understand
this
is
going
to
show
that
an
event
no
I
I.
F
F
Car
time
yeah,
let's
have
a
look
and
see
what's
going
on
it,
the
backwards
compatibility
thing
was
a
lot.
The
idea
was
to
try
and
and
build
this
so
that
clients
have
just
this
is
actually
the
missing
thing
from
Cal
dev
is:
is
the
idea
of
an
expansion
where
what
you
see
are
just
things
that
look
like
real
events
and
and
so
from
from
the
backwards
compatibility
point
of
view.
You
just
see
these
things,
look
like
real
events
and.
H
H
F
F
F
They're
very
lightweight,
if
nothing
changes,
but
if
every
instance
has
a
change
in
it
and
override,
then
that
turns
up
looking
like
a
real
event
and
recurs
recurrences
in
icon
or
not
a
change.
The
recurrence
override
is
actually
a
full
copy
of
what
the
event
looks
like.
So,
if
you,
if
you
change
every.
D
F
F
Work
on
that,
the
problem
we
have
is
is
is
largest
structure,
the
issue
of
whether
we
produce
a
single
document
that
is
V
fo
without
any
of
the
modes
and
then
and
then
have
basic
mode
and
a
few
other
modes
are
separate
documents
or
or
M,
merge
them
into
one.
It
so
I
think
that's
where
we
are
still
so
sweet
I
claims,
I'm,
shouting,
so
I
think
I.
D
D
B
D
D
B
D
B
D
Out
yep
so
for
the
documents
that
we've
discussed
today
on
the
list-
let's
just
zoom
back
to
the
start
here,
where
the
agenda
is
event
pubs
been
sent,
joyous
calendar
will
be
sent
subscription,
upgrade
I'd
say
we
should
aim
to
have
that
done
by
the
next
HF,
so
maybe
February
next
year.
Is
that
same
realistic
to
people
the
subscription
upgrade.
I
J
E
D
C
F
And
the
reason,
the
reason
why
that
was
done.
That
was
because
you,
you
know
user
applications
didn't
implement
it
and
I
I
get
that,
but
the
reason
that
I
always
disagreed
with
that
and
the
thing
that
the
reason
it
came
back
to
my
mind
was
I
had
to
work
on
a
bunch
of
data
for
some
site
that
I
support
and
discovered
they
had
seven
recurring
events
that
all
had
slightly
different
recurrences
and
the
only
reason
they
did.
That
is
because
they
could
not
express
that
with
Nas
blah
rule
out
of
interest.
F
F
H
F
H
F
C
F
H
F
H
F
H
D
F
D
Okay,
yes
you're
tuned
like
this,
is
when
it's
supposed
to
be
done
by
it
doesn't
mean
you
have
to
go:
oh
okay,
so
yeah,
but
it
just
lets
a
spaced
out
the
order
in
which
we're
expecting
things
to
happen
a
little
bit.
Okay,
yep
awesome,
alright
I
will
update
all
the
milestones.
Finally,
is
there
any
other
business.