►
From YouTube: IETF-CELLAR-20220628-1900
Description
CELLAR meeting session at IETF
2022/06/28 1900
https://datatracker.ietf.org/meeting//proceedings/
C
C
C
A
Yeah,
if
you
look
on
the
bottom
right,
there's
also
another
microphone
which
is
a
local,
mute
and
so
yeah.
E
D
A
All
right,
I
did
this
right
now.
I
can
pull
up
this
one.
A
So
let's
get
started,
I
guess.
A
A
D
Better
because
I
can
see
that
anybody
have
worse
eyes
than
I
do.
D
So
I
got
an
email
from
rato
saying
for
the
125th
day
in
a
row.
He
is
still
in
ukraine.
Oh
really,
yeah!
That's
what
I
said.
A
D
D
It
looks
like
his
home
home
base,
maybe
zurich
or
someplace.
He
may
have
gotten
halfway
home.
Oh
maybe.
D
So
I
wanted
to
be
sure
and
include
him
just
so
that
we
could
have
him
in
our
thoughts.
A
I
wanted
to
look
like
you
spencer,
so
I
shaved
my
head
on
the
weekend.
A
C
A
Yeah:
okay,
any
objections
to
the
minutes.
From
last
time,.
C
A
Do
we
have
the
status
of
the
drafts?
I
think
that's
probably
correct.
D
I
think
I
think
the
question
for
me
is:
do
we
need
to
do?
We
need
to
talk
about
that
today
to
make
a
little
progress
on
it,
because
if
we,
if
there's
anything,
we
need
to
talk
about
on
a
call?
That's
it's
not
going
to
happen
today.
If
we
don't
save
time
for
it
and
we
don't
in
july.
C
Well,
I
I
don't
know
what
the
importance
is
of
having
a
milestone
set
on
the
right
date,
but
this
is
obviously
the
last
meeting
before
the
set
milestone.
Yeah
yeah,
I
don't
know
whether
it's
important
really.
A
It's
it's
mostly
to
set
expectations.
My
feeling
is,
if
you
said
to
me,
hey.
I
think
that
we're
gonna
finish
the
document
by
the
end
of
the
summer,
then
I'd
be
like
wow.
That's
pretty
actually
close
to
the
milestone.
So
you
know
that's
totally.
That
makes
sense,
if
you
say
say
well,
it's
going
to
be
the
end
of
the
year
before
we
finish
it,
then
I
would
update
the
my
milestone
just
to
set
expectations
right.
C
D
We
can
write
that
down
and
and
then
we
will
not
bug,
then
we
will
not
bug
the
ads
about
trying
to
get
the
milestone,
update
done.
A
A
Yeah
martin,
do
you
do
you
need
I
mean
I'd,
be
totally
happy
to
spend
all
of
the
august
call
on
flack
issues
if
that's
required
or
important,
but
you
tell
us
right
it.
Maybe
you
don't
need
a
call,
maybe
you
don't
need
in
person
time
to
to
resolve
them,
but
if
that
you
know
if
the
day
before
you
say
that's
the
case,
then
we
can
totally
do
that.
C
And
the
thing
I'm
now
mostly
working
on
is
some
well
specifics.
Really,
so
I'm
working
out
some
details,
they're
in
the
last.
I
don't
know
whether
it's
last
one
of
the
pull
requests
on
the
specification.
A
Okay,
so
you
just
it's
just
a
question
of
of
pushing
it
over
the
line,
for
you,
I
think,
is
what
you're
saying.
D
And
having
a
call
does
acts
as
a
forcing
function
to
make
people
make
people
suddenly
realize
they
haven't,
read
it
yet,
and
they
need
to.
C
A
A
There's
a
ripe
meeting
in
belgrade
that
I
will
likely
go
to
part
or
most
of
that
week
and
then
there's
a
net
dev,
which
is
totally
unrelated
to
anything.
You
might
have
heard
of
it's
a
linux
networking
stack
meeting
in
lisbon
that
they
scheduled
that
week,
but
my
plan
actually
was
to
show
off
on
friday
in
amsterdam.
E
We'd
probably
do
our
best
to
accommodate
you.
If
you're
going
to
come
on
on
the
friday,
the
venue
is
going
to
be
in
the
in
the
hague,
which
is
I
I
guess
the
airport
is
sort
of
on
the
hague
side
of
amsterdam.
So
it's
after
you
land,
it's
maybe
yeah.
A
A
So
my
intention
is
to
be
there
anyway
on
the
friday,
which
is
the
28th
and-
and
maybe
that
means
I'm
arriving
on
thursday
night
or
maybe
it's
friday
morning
early,
but
I
don't
know
yet.
I
haven't
made
any
I
haven't
made
any
I'd
rather
take
the
train
from
belgrade,
but
it's
a
two-day
event
to
do
that
so
and
they're
not
running
they're,
not
crossing
the
borders,
yet
they
don't
know
how
to
cross
borders
with
kovid
yet
again
yeah.
A
A
A
Yeah,
the
the
net
dev
was
usually
supposed
to
be
the
week
before
ietf
and
then
they
moved
it
two
weeks
before
ietf
and
it
was
supposed,
then
I
don't
know,
I
reminded
them
of
this
fact,
and
they
said,
oh,
I
think
we
screwed
up
but
anyway,
whatever
there,
you
go
all
right.
So,
let's
get
on
to
the
substantive
issues.
So
the
status
was
that
I'm
the
document
shepherd
and
what
I
did
is.
A
I
did
a
top
to
bottom
read
most
of
most
of
the
words
except
for
that
very
long
legacy
appendix
which
I
decided
to
skip,
and
I
I
don't
think
I
have
any.
A
I
don't
think
I
have
any
problems
with
the
document
as
a
technical
specification
for
the
protocol
right
and
that
wasn't
my
goal
to
object
on
that.
All
I'm
really
trying
to
get
at
in
this
review
was
to
identify
things
that
are
going
to
be
problems
for
the
other
reviewers
and
the
isg
in
their
reading
of
it,
and
so
I
wrote
those
things
down
steve
I
think,
has
done
80
of
it
and
then
you
sent
an
email
a
couple
days
ago
with
some
of
the
issues
and
I'd
like
to
walk
through.
F
B
F
Over
three
weekends,
I
went
through
all
the
issues
you
mentioned
and
my
replies
and
I
applied
basically
what
I
replied
to
you
in
the
emails
yeah.
I
made
them
into
a
pull
request.
Some
are
very
short,
some
are
bigger,
so
basically,
I.
F
A
So
so
these
these
things,
these
comments
here,
you've
already
handled
this
or
you
think
that
we
don't.
F
A
A
F
A
Let's
come
to
encryption
in
a
minute
just
I
I
kind
of
want
to
walk
down
this,
this
email
just
a
little
bit,
I
could
have
replied
to
it.
I
didn't
have
a
chance
to
to
really
grok
at
all.
Yet
I'll
just
say
recommended
is
not
is
exactly
equivalent
to
should
in
terms
of.
A
14
term,
both
of
them
are,
they
mean
the
same
thing.
They
mean
the
same
thing
and,
and
you
can
use
recommended
or
should
it
really
doesn't
matter
there,
but
what
I'm
saying
is
they
don't
have
any
there's?
No,
we
wouldn't
argue
whether
it
should
be
recommended,
or
should
they
mean
the
same
thing,
and
so
what
I'm
trying
to
say
is
that
if
you
write
should
or
recommended,
then
that
means
that
there's
some
scenario
in
which
maybe
there's
an
exception
right,
yeah.
A
All
the
time
too,
where
I
want
to
make
my
document
kind
of
future
proof
against
someone
deciding
something
something
new,
and
so
I
want
to
write,
should
all
the
time
and
I'm
almost
pretty
sure
that
it's
almost
all
the
time
that
that,
unless
you
can
document
an
exception,
you
should
you
must
write.
Must
you
should
write?
Must
so
that's
all
I'm
trying
to
say
this
is
one
of
those
things
where
the
reviewers
will
go
through
and
they
will
ask
you
know
what
are
the
exceptions
to
the
shoulds,
and
so
that's
it.
A
So
so
anyway,
in
this
case,
we're
talking
about
meta,
seek,
it
cannot
be
a
must
because
it's
not
usable
in
a
live
stream.
So
I
think
that's
all
right
I
mean:
if
medicine
is
you
should
use
or
medica
is
recommended
unless
it's
a
live
stream
stream,
and
I
think
that's
just
fine-
you
can
also
have
conditional
musts
where
you
say
the
same
thing
right,
but.
A
And
and
usually
it
it's
usually
that
it
works
just
fine
that
way.
If
you
include,
if,
if
you
are
going
to
drive
a
car,
you
must
have
a
driver's
license.
That's
a
completely
reasonable
thing
to
say
right,
because
it's
not
optional.
There's
no
case
in
which
you
can't
you
can
drive
a
car
and
not
have
a
driver's
license.
It's
that
the!
If
you
don't
need
to
drive
a
car,
then
you
don't
need
a
driver's
license
right.
So
there's
a
there.
A
On
the
other
hand,
you
know
some
countries
would
say
you
should
have
a
driver's
license.
The
exception
is
that
you
are,
you
know,
driving
your
wife
to
the
hospital,
and
you
know
it's
an
emergency,
but
most
of
them
don't
have
that
exception
anyway,
yeah,
okay,
so,
for
example,
for
the
metac.
F
The
problem
is
that
we
also
supposed
to
describe
what
exists
now.
I'm
not
sure
the
library
is
actually
write
one
because
it's
I
mean
you
can
use
a
file
without
one,
it's
just
a
nice
thing
to
have.
So,
basically,
if
we
say
must
we
it's
like
making
all
these
files
invalid,
even
though
they're
quite
valid.
A
So
what
we,
what
we
can
do
is
we
can
say,
is
that
encoders
that
are
compliant
to
this
specification
must
use
a
meta
seek
unless
the
stream
is
live.
D
Okay,
that
means
that
that
would
be
a
well-formed
that
would
be
a
well-formed
bcp-14
comment,
yeah,
and
then
we
can
say
that
parsers.
A
F
A
A
F
You
have
some
files
with
just
chapters
and
or
maybe
tags
so
you
don't
you
usually
make
tastic
is
useful
when
you
have
large
files
when
you
need
to
know
exactly
where
at
the
end
of
the
file
you
need
to
seek.
But
if
your
file
is
small,
it's
not
that
much
too
sick
to
actually
find
it
right,
especially
you
have
just
chapters
and
tags,
it's
very
easy
to
seek.
F
F
D
Here
it
seems
like
to
me
is
that
if
you
say
yo
so
michael
was
saying
if
you,
if
you
receive
this
stuff
and
you
need
to
be,
you
need
to
be
tolerant
of
of
implementations
that
don't
do
the
must
basically.
D
The
isg
ought
to
hold
you
up
to
say
what
does
it
mean
to
be
tolerant?
You
know
yeah
what
what
what
what's
the
recommended?
What's
the
recommended
action
that
you
would
take
if
you
realize
that
you're
looking
at
a
if
you're
looking
at
something
that
is
not
conforming
to
the
specifications
that
we've
got,
you
know.
A
So
maybe
maybe
dave
you
could
answer
the
question
is:
would
would
an
archivist
an
archiving
organization
want
to.
A
Essentially,
be
less
tolerant
of
content
or
reprocess
it,
or
something
like
this.
E
If
it
was
content
made
produced
by
the
archive
as
the
result
of
reformatting,
then
then
maybe
true,
but
if
it
was
like,
if
we
were
trying
to
receive
material
like,
for
instance,
if
we're
trying
to
read
a
live,
matraska
stream
to
store
it
and
it
doesn't
have
medic,
because
that's
one
reason
why
it's
not
you
know
it
doesn't
need
to
be
used.
So
we'd
want
to
store
the
data,
as
is
so
like
I
don't
know,
there
are
a
lot
of
instances
where
we
wouldn't
intentionally
want
to
manipulate
a
file.
A
I'm
just
thinking
about
right
are
there,
would
you
want
to
market
the
stream
as
not
completely
compliant
in
some
way,
so
that
later
on?
You
would
know
that
it's
going
to
be
harder
to
work
with
or.
A
No,
I
mean
like-
let's
let's
say
I
I'm
thinking
about
you
know
I
don't
know,
I'm
thinking
about
archive.org,
which
I
know
more
about
than
anything
else,
but
somebody
brings
a
a
recording
of
something
and
it
was
recorded
lot
from
a
live
stream
or
from
a
legacy
recording
device,
or
you
know
converted
from
betamax
right
and
they
say
here
archive,
please
you
know,
keep
it
and
it's
not.
You
know
it's
not
it's
not
new.
A
It's
not
a
new
recording,
it's
an
older
encoding
and
would
would
would
it
make
sense
for
an
archival,
archival
organization
to
say,
oh,
look,
I
am
going
to
reprocess
the
stream
and
put
the
metaseq
headers
in
having
searched
through
the
thing,
because
it's
gigabytes-
and
I
don't
want
to
do
it
again-.
E
Yeah
I
mean
it
depends
on
context
like
some
archives
might
be
in
a
situation
where
they
have
to
show
that
they're,
preserving
a
bit
exact
copy
of
the
original.
That
there's
been
no
manipulation
to
what
they
received,
but
yeah.
Some
might
want
the
additional
features
and
flexibility
by
re-wrapping.
It.
E
A
You
might
actually
have
to
pay
the
cost
of
having
two
copies,
one
that
you
know
you
digitally
sign
forever
and
the
other
one
which
you
you
upgrade
to
the
latest
standard,
so
that
people
can
can
see
it
at
any
time.
And
you
know
whatever
I
don't
know,
I'm
just
trying
to
ask.
A
Two
copies:
I
guess
I'm
trying
to
I'm
trying
to
think
about.
Do
we
have
to
have
a
definition
of
what
is
what
was
the
word
I
used
before
not
legacy,
but
the
opposite
one
spencer
conformance
conformant.
Do
we
have
to
have
a
word
that
says,
there's
some
conformance
level
or
something
or
doesn't
matter
or
maybe
this
is
something
we
would
come
back
to
and
say
this
is
the
experience
we
have
later
on.
E
Yeah,
I
think
steve
has
some
experience
in
that
already
with
the
tool
called
mk
validator,
because
it's
a
it's
like
a
conformance
recommendation
checker
for
matriscophiles.
I
can't
remember
the
language
it
uses,
but
it'll
either
kind
of
complain
about
a
matroska
file
or
say
it's
good.
A
Okay,
well,
I
don't
know
what
exactly
to
say,
but
what
I'm
I'm
trying
to
say
is
that
it,
the
language
probably
needs
to
be
writers,
must
do
x,
okay,
and
but
that
that
readers
must
tolerate.
A
You
know
absence
of
x
right
because
it
doesn't
always
exist
or
there
are
some
contexts
in
which
it
won't
won't
be
there
and
that's
all
I'm
trying.
I
can't
think
I
can
actually
say
which
ones
should
be,
which
which
ones
must
be
which,
but
all
that's
all
I'm
trying
to
say
is
that
that's
the
kind
of
language.
I
think
that
we
need
to
make
sure
that
we
have
that
will
make
our
reviewers
happier
about
understanding
it,
because
they're
going
to
be
in
the
same
situation.
A
All
right,
so,
let's
continue
on
here.
So
that
was
should
messages.
And
then
I
asked
about
the
cluster
and
element
was
that
a
good
re-parsing
of
that
steve.
F
I
want
just
before,
but
subject
before
you
suggested,
that
we
have
a
general
recommendation:
how
to
create
files
somewhere
in
the
file,
and
I
already
created
one
merge
request
for
that
and
pretty
much
the
end
of
the
document.
If
you
want
to
create
useful
things
in
microscope,
you
should
do
this
or
that
maybe
the
metastic
could
be
one
of
them
as
well.
It
could
be
moved
there
explaining
why
you
should
do
it
and
like
it's
really
a
good
thing
to
do,
even
though.
A
Or
a
forward
reference
a
forward
reference
would
be
fine
to
that
part
to
say
you
know,
as
explained
in
section
blah
blah
blah
a
medici
must
be
present
unless
it's
live
and
then
that
that
or
that,
then
this
part
of
the
text
becomes
the
normative
part
and
the
other
part
becomes
the
recommendation.
A
F
F
A
Okay,
so
I
asked
whether
there
was
only
audio
and
video
components.
I
thought
there
were
other
things,
but
there's
never
been
a
need
for
that.
That's
fine!
Okay!
So
let's
go
on.
F
G
A
Yeah,
so
I'm
I'm
not
actually
trying
to
limit
it
to
block
group
or
simple
block
elements.
I
think
I
was
trying
to
parse
the
the
fact
that
the
block
group
and
the
simple
block
elements
can
are
are
each
there
may
be
zero
or
there
may
be
one
of
them
or
more
of
them
right
yeah.
That
was
all
I
was
trying
to
get
at
is.
Is
that
the?
I
think
that
the
numbers
that
a
cluster
element
that
contains
zero
block
groups
and
zero
simple
block
elements
is
not
valid.
A
F
G
A
That
I
was
complaining
about
the
should
probably.
A
I
would
say
something
like
the
crc
out:
32
element
must
be
first.
If
it
is
present,
the
I
would
say
the
timestamp
element
must
be
must
follow
right.
F
Yeah
we
already
had
some
similar
texts
in
other
places.
It's
just
that
that
piece
of
text
was
different,
so
I
just
copied
what
was
already
said
elsewhere
and
so
that's
merge
request,
a
pull
request,
608
and
basically
now
it
says
the
timestamp
element
should
be
the
first
element
in
the
cluster.
It
belongs
to
or
the
second.
If
that
cluster
contains
a
checksum
element.
A
All
right!
I
asked
the
question.
I
don't
think
there's
any
relevance
here
you
answered
it.
I
think
you're
telling
me
that
yes,
they
could
be
a
different
codex,
but
they
don't
get
played
at
the
same
time.
So
it
doesn't
really
matter.
A
So
then
you
try
to
change
the
element
from
there.
H
A
No
okay,
so
I
I'm
I'm.
I
just
think
that
we
could
make
it
a
little
more
concise
somehow,
but
you
know
what,
if
it's
really
a
hard
layout
problem,
then
let's
not
do
it
and
I.
F
Know
that
it's
all
generated
from
yeah
the
problem
is
that
it's
it's
the
markdown
to
xml
translation,
the
format
that
we're
using
allows
us
to
have
that
kind
of
formatting.
With
these
elements,
this
value,
that
kind
of
thing,
but
once
you
try
to
do
something
funky,
it's
not
possible.
Otherwise,
that's
to
be
a
table
and
then
it
would
look
bulkier.
I
think
so.
I
did
shrink
some
lines.
F
It
worked,
it's
still
readable,
but
I
mean
I
saved
the
20
pages.
Just
doing
that,
so
I
think
it's
useful,
but
it's
not.
I
mean
I
don't
really
like
the
formatting,
but
we
don't
have
a
lot
of
options
because
of
the
markdown
or
translation.
A
Yeah,
okay,
I
I
figured
we'd
run
into
something
like
that,
but
I
thought
I
would
ask,
but
you
said
you
saved
20
pages
already,
so
that's
pretty
significant,
already
yeah.
C
C
A
A
A
Oh
yeah,
because
a
lot
of
them,
some
of
them
had
sections
had
usage
notes
and
some
of
them
had
rationale
in
the
area
in
the
in
the
definitions,
and
I
was
trying
to
understand
whether
we
were
putting
them
in
the
right
places
all
the
time.
I
think
that's
the
the
real
thing
in
some
cases,
maybe
we
weren't
just
a
sex
yeah.
F
Basically,
your
rational
is
just
an
explanation
of
why
we
have
this
element
because
in
the
definition,
sometimes
you
don't
want
to
put
too
much
text.
You
want
to
be
concise
and
really
explain
basic
and
then
irrational.
F
If
you
want
to
have
some
extra
explanation
on
on
the
elements,
whereas
users
use
h
notes,
it's
really
on
implementation,
what
you
have
to
do,
for
example,
which
we
were
saying
before
like
if
you
have
this
element,
you
must
have
that
one
that
kind
of
thing,
so
it
it's
more
usage
notes,
is
more
normative,
whereas
rational
is
just
extra
text,
hello,
it
is
we
can't
you,
but
we
couldn't
hear
you.
F
A
A
Okay,
and
can
you
tell
me,
is
this
this
new
thing:
what's
the
interaction
between
that
and
and
university
applications.
A
A
Sorry
did
I
get
the
mute
wrong
way.
My
my
son's
english
teacher
math
teacher,
just
called
so
I
had
to
take
that.
A
Sorry
about
that
yeah,
so
I
was
trying
to
yeah.
I
got
what
you're
saying
so,
I'm
just
trying
to
make
sure
that
my
problem
with
having
two
two
boxes
that
we
is,
is
that
we
don't
it's
making
sure
that
we
basically
just
always
know
that
we're
using.
Why
are
we
using
this
box
versus
that
box?
That's
all
and
yeah.
H
A
Today's
the
last
day
of
school,
so
that's
why
they're
telling
us
things?
Okay,
so
thank
you
for
changing
it
to
uuid.
That's
really!
A
good
change,
I
think,
will
make
everyone
happier
to
know.
A
So
that's
really
good,
I
think
that's
an
improvement
block
label.
You
did
this.
What
was
some
other?
Oh
yes,
I
think
this
is
a
good
change
from,
rather
than
calling
it
ihf
element,
let's
call
it
a
bcp-47,
which
is,
I
think,
the
right
reference,
and
otherwise
it
looks
kind
of
weird
in
the
ietf,
because
we
wouldn't
know
what
it
was.
A
F
A
F
A
Okay
got
the
references,
tell
me
which
let
me
see
here,
oh
yeah,
so
let's
talk
about
the
content
at
inc
alg,
so
this
is
going
to
get
tripped
up
by
some
security
guy
who's
going
to
read
this
so
the
question
first
of
all
is:
are
there
really
files
out
there
that
were
encrypted.
A
Okay,
so
I
would
say
that
I
would
move
it
to
a
legacy
to
a
legacy
space.
F
But
still
news
every
day,
if
you
go
on
youtube
or
things
like
that,
one,
maybe
not
youtube
or
google.
A
Okay,
so
so,
if
what
I'm
trying
to
s?
Okay,
so
somehow
we
have
to
say
something
that
that
this
is
you.
I
would
say
even
that
this
has
been
used
in
proprietary
video
systems
like
webm
youtube.
F
A
Okay,
that's
fine!
What
I'm
trying
to
say
is
that
would
you
want?
Would
you
want
somebody
new
coming
along
to
use
these
elements
to
do
their
encryption.
F
Well,
it
depends
by
use
these
elements
because
that's
exactly
what
webmd
they
didn't,
they
took
what
existed
and
used
it.
They
just
expanded
it
and
they
use
some
values
and
say
this
is
with
that
value.
You
do
that
someone
could
just
come
and
create
their
own
system
based
on
the
few
fields
we
have.
If
what
you
mean
is
you
have
ctr?
Yes,
cbs,
I
think,
and
if
I
have
this
value,
how
do
I
do
it
right
now?
We
only
describe
the
one
that
is
found
on
webm
there's
another
version
that
is
possible.
F
F
G
A
A
sentence
in
that
says
that
detailed,
a
detailed
description
of
how
to
do
this
for
modern
algorithms
is
the
subject
of
future
work:
okay,
okay
and-
and
I
hope
that
gets
rid
of
the
problem
right.
Otherwise
we're
going
to
have
the
security
people
basically
saying.
Oh
hang
on
here,
you
can't
use
static
keys.
You
can't
do
this.
You
can't
do
that.
Oh,
what's
the
key
agreement
protocol,
why
are
you
using
rsa
and
weddes
and
blah
blah
blah
blah
and
reality?
A
Is
we're
not
trying
to
come
up
with
a
a
video
encryption
system
here
at
all?
It's
not
our
goal,
we're
simply
documenting
that
these
containers
do
exist
and
if
you
see
them
you
know
they're,
it's
not
bogus,
but
it
is
part
of
essentially
a
legacy
system.
As
what
I'm
trying
to
say
and
legacy
is
anything
that's
not
really
well
documented.
It
doesn't
have
to
be
old
right,
so,
okay,
so,
okay,
so
then
we
had
there.
A
You
said
the
there's.
You
put
the
reference
in,
we
put
the
reference
in.
Do
we
need
a
given
that
discussion?
Do
we
need
to
have
anything
other
than
the
pointer
if
we
just
have
the
pointer
to
their
stuff
that
works
for
me.
A
I'm
just
saying
that
that
point
to
their
stuff,
if
they've
already
defined
it,
then
that's
great
okay,
we
don't
need
to
get
there.
F
F
Question
about
the
janai
registry
is
that
we
have
a
lot
of
fields,
integer
fields
that
are
actually
a
new
values
yeah
and
because
they
can
be
expanded
in
the
future.
You
could
want
to
add
a
value
for
anything.
A
So
there's
two
possibilities
here:
right:
if
we
don't
have
an
iana
registry
for
something,
then
the
implication
is
it's
specification
required
and
essentially
the
only
way
to
update
it
would
be
to
write
a
new
document
that
updates
this
document
and
that's
okay.
A
We
don't
have
to
imagine
every
single
extension
point
if
we
want
to
have
if
we
want
to
have
elements
that
can
be
extended
with
less
of
a
strong.
A
Thing
so,
in
other
words,
we
don't
need
another
ietf
rfc,
you
could,
you
know,
ask
for
it.
Some
other
way.
First
come
first
served
et
cetera
whatever.
Then
we
don't
need
to
have
an
eye
on
it.
Then
we
do.
Then
we
need
to
have
an
iana
registry
because
we
need
to
say
what
the
rules
are,
but
if
they're
mostly
just
like
elements
of
you
know,
you
say
fields
of
a
thing
and
if
we
did
come
along,
we
would
have
to
write
a
new
document.
Anyway,
then
that's
fine.
We
just
leave
them
that
way.
F
Yes,
well,
I
think,
for
most,
I
will
have
to
check
all
of
them,
but
basically
most
of
our
enums
are
I
mean
they
need
an
explanation
on
how
to
import
interpret
the
value
it's
not
just
and
how
it
works
compared
to
the
other
values,
but
basically,
for
example,
the
the
tag
do
not
detect
the
codec
id
is
not
needs
a
yeah
registry,
because
it
there
will
be
a
lot
of
values
added
over
time,
which
change
nothing
to
the
way
microscope
works
right,
so
so
that
needs
a
rest.
F
But
in
that
case
it's
also
not
an
enum,
and
I
think
basically,
we
use
a
menu
when
we
want
to
limit
the
amount
of
values
and
have
very
strict
meaning.
So
probably
not
a
lot
of
them
would
need
a
registry.
A
Okay,
so
I'm
just
trying
to
so
the
next
section.
The
next
comment
I
had
was
about.
C
A
C
D
Did
michael
did
we
need
to
talk
about
the
possibility
of
in
a
registry
for
this
draft.
A
If
someone
comes
along
and
and
and
for
many
of
these
spencer,
we
won't
it's
not
like,
you
can
allocate
a
new
color
like
if
you
think
about
some
of
these
are
red,
green,
blue
or
yuv,
or
whatever
yeah
yeah.
If.
D
A
Came
along
with
a
new
needed,
a
new
value
you'd
actually
have
to
go
and
define
a
whole
bunch
of
other
stuff.
So
actually
you
have
to
write
a
new
document
and
it
would
have
to
update
this
one
and
the
problems
of
adding
new
values
is
not
you
know
the
biggest
problem
right.
It
would
just
so
you
just
have
to
update
it,
and
you
know-
maybe
maybe
at
some
point
I
just
don't
think
I
I
I
as
I
was
saying.
A
A
A
A
A
We
had
a
long
conversation
with
zeke
element.
Okay,
never
mind:
do
we
do
this
page?
Am
I
repeating
this.
A
Part
two:
okay:
we
just
talked
about
crc
too.
That's
fine
must.
A
So
you're
gonna,
I
don't
add
a
hint
or
not,
I'm
just
concerned
about
questions
more
than
anything
else.
F
A
Yeah,
I
don't
know,
let's
let
them
decide
it's
a
bonus
of
the
chapters
before
the
data
and
eleno.
Does
that.
A
A
A
A
Is
the
I
isn't
the
eye
and
a
section
only
from
the
diana
registry
yeah,
but
also
it's
also
for
things
that
actions
activities
that
we
want
ayanna
to
do
in
other
registries,
so
the
iona
considerations
is
basically
ayanna.
Please
create
this
registry,
please
create
this
registry
and
then
we
also.
C
A
G
Okay,
let's
check,
let
me.
A
Now
we
don't
have
a
macintosh
file
type
code,
so
n
a
and
we
asked,
for
you,
know
this
media
type,
application,
blah
blah
blah.
A
Okay,
if
you
want
to
if
you
want
to
register
a
extension,
but
I
think
what
I
was
really
saying
is
that
I
think
we
needed.
I
thought
we
needed
to
have
a
media
type
extension
there.
So
what
did
I
do
with
the
copy
of
the.
A
Oh
yeah,
so
15
right
was
just
media
types.
Okay,
so
we
just
need
to
to
move
this
to
do
we
have
a
mime
type
already.
I
can't
remember.
F
Me
neither
which
we
have
a
few,
but
it
should
be
the
beginning
of
the
document.
B
A
Yeah,
so
so
yeah,
so
we
just
need
to
redo
this
as
a
media,
as
a
media.
Registration
is
what
I'm
trying
to
say.
A
A
So
there
is
a
I'll
send
you,
the
link,
regis
regis
ratio,
registration.
There
we
go.
That's
the
right
word
and
there's
a
there's,
a
template
from
another
document.
That
explains
how
to
do
that,
and
this
is
so
then
the
the
the
file,
the
file
name,
extensions-
would
go
here.
A
Okay,
so
we're
just
running
out
of
time
here
you
liked
this
reordering
the
hacks.
A
A
A
Okay,
yeah,
so
that's
fine
everything's
good.
So
why
don't
you
post?
Even
if
I
know
you're,
not
quite
finished?
Why
don't
you
post
the
document,
as
so
far
and
that'll,
make
it
a
little
bit
easier
to
see
what
else
is
left.
F
C
Yes,
I
was,
I
was
wondering
whether
jerome
could
review
to
pull
requests
from
the
flock
specification
repository.
A
A
Right,
thank
you.
Do
you
do
you
want
to
paste
those?
Do
you
want
to
paste
those
issues
into
here
or
tell
me
what
they
are
number
wise.