►
From YouTube: IETF-CBOR-20220309-1500
Description
CBOR meeting session at IETF
2022/03/09 1500
https://datatracker.ietf.org/meeting//proceedings/
B
B
Okay,
let's
get
started
because
I
I'm
not
sure
well,
there
will
be
too
many
more
and
if
so,
we'll
notice.
After
having
gone
through
the
agenda
speaking
of
agenda,
there
is
not
much
on
the
plate
today.
For
today,
file
magic
is
in
the
pipeline
and
I've
just
finished
and
uploaded
the
the
overhauled
shepherd
write
up,
and
I
don't
think
there
is
much
more
to
say
unless
someone
wants
to
say
something
to
that
right
now,.
B
Yeah,
no,
not
ask
for
comments
on
the
shepherd
write
up
more
on
on
the
topic
of
kind
of.
Is
there
and
is
there
anything
on
the
on
the
document
as
a
whole
that
would
like
to
report
michael
and
you
I
uploaded
to
the
the
two
latest
versions
right
before
the
cut
off.
So
I
think
people
have
seen
that
hello,
michael.
C
C
So
I'm
trying
to
avoid
the
term
sibo
objects.
So
I
usually
say
sibo
data
items,
but
that's
maybe
more
important
for
the
actual
documents
than
for
the
shepherd
write-up.
D
C
Well,
actually
I
tell
my
students,
they
have
to
write.
Well,
that's
about
the
introduction,
not
really
the
abstract,
but
the
same
thing
kind
of
applies.
They
have
to
write
it
twice.
They
have
to
write
it
at
the
beginning,
so
they
know
what
they
are
setting
out
to
do
and
then
they
have
to
rewrite
it
at
the
end,
because
then
they
know
what
they
should
have
been
setting
out
to
do.
B
But
with
that,
I
think
the
plan
stands
to
submit
the
document
from
the
working
group
into
into
the
iesg
review,
with
a
with
a
shepherd
review
that
just
points
out
that
there
is
this
small
single
word
in
the
abstract
that
will
be
changing
and
something
about
the
ayana
considerations.
But
I'm
writing
I'm
just
about
to
send
that
mail,
so
it
doesn't
need
further
discussion.
I
think.
A
So
so
I
don't
know
how
I
can
promise
that
will
happen
before
the
itf
meeting.
That's
for
your
information.
B
I've
put
one
more
more
of
a
question
to
the
agenda
in
about
other
documents,
but
really
the
latter
question
is.
Is
there
anything
anyone
would
like
to
bring
up
today
that
we
should
that
we
should
also
talk
about.
C
Well,
I
would
like
to
know
whether
people
are
implementing
the
pact
specification,
because
I
have
been
a
bit
slow
in
finishing
that
implementation
and
of
course
there
are
many
ways
to
implement
it,
because
the
the
compressor
is
not
specified
only
the
decompressor
specified.
So
you
can
do
amazing
things
in
the
compressor.
Do
a
very
basic
thing
both
would
be
valid
implementations,
but
it
would
be
good
to
get
some
some
more
feedback
from
other
people
as
well.
C
C
Yeah,
so
just
to
remind
people,
an
implementation
of
pact
is
part
of
the
sibo
diagnostic
package.
So
if
you
have
zebra
diagnostic
installed,
you
have
packed
so
you
can
play
around
with
what
my
compressor
and
decompressor
are
doing
at
this
time.
C
B
B
B
So
so
picking
that
up
there
do
you
have
to
have
you,
do
you
have
any
plans
or
or
tried
things
around
a
diagnostic
notation
for
packed
things,
because
I've
written
them
down
a
lot
for
coral
and
readability
is,
let's
say,
lacking
because
pact
means
there
are
different
tags
around
that
are
all
numeric,
and
so
do
you
have
ideas
yet
how
that
could
be
made
more
readable.
C
Yeah,
one
obvious
way,
of
course,
would
be
to
have
a
special
notation
for
the
references,
so
you
don't
have
well.
It
doesn't
make
a
big
difference
with
simple,
but
for
for
the
tags,
it
would
probably
be
useful
to
have
some
special
notation
that
shows
the
actual
index.
C
Now
additional
function
of
a
diagnostic
tool
might
be
to
actually
include
more
more
information
when
you
actually
have
a
reference
more
information,
so
you
don't
have
to
to
look
up
the
the
item
by
yourself,
but
I
I
haven't
designed
that.
So
if
somebody
has
an
idea
how
this
might
look
like,
so
that
in
the
diagnostic
condition,
you
essentially
are
seeing
the
unpacked
form,
but
you
you
see
it
with
the
breadcrumbs
in
place,
so
you
can
see
where,
where
that
data
was
packed
to
or
gotten
from
in
the
table,.
C
So
because,
then
maybe
you
have
the
the
strongest
intuitions
about
what
would
help
you
when
working
with
these
items
in
diagnostic
notation.
B
Well,
the
thing:
well,
the
thing
that
I'm
that
I'd
like
to
have,
but
I
think
it's
spanning
too
many
layers
to
be
practical,
would
be.
That
would
be
that
they
work
a
bit
like
kuri's,
because
that's
what
I
would
eventually
use
them
for,
but
I
don't
quite
see
how
that
would
work
out
syntactically.
So
if
we
have
something
like.
B
Michael,
could
you
move
down
your
cursor
for
some
reason?
I
you
you're
precisely
blocking
it.
Thank
you.
So
this
is
obviously
not
exactly
not
exactly
a
cri,
but
it
looks
a
bit
like
it
might
look
like
one
and
that
would
be
compressed
to
sorry
that
would
expand
so
that's
another
all
without
without
packing.
Yet
and
if
packing
came
in,
it
would
certainly
be
nice
to
have
some
way
to
say
what
this
is
actually
short
for.
B
Okay,
that's
sort
of
that's
a
bad
example.
Are
you
trying
to
share
something?
No,
I'm
in
the
I'm
in
the
notepad
lines,
37.
B
Thing
is
any
any
compression
that
way
would
interleave,
with
kind
of
it
would
interfere,
also
with
the
with
the
dictionary
setup.
So
that's
probably
not
what
it
could
be.
What
I
think
it
could
rather
be
is
something
like
there
could
be
the
concrete
number.
This
is
compressed
too
say
entry
at
250,
and
then
I
don't
know
yet.
Maybe
maybe
a
comment
right
in
there
or
just
or
just
a
comment.
Yeah.
B
In
a
separate
position,
but
the
edn
is
probably
already
suited
for
for
doing
something
like
this.
It
would
be
just
another
edition
that
says:
50
p
p,
something
numeric
expands
to
the
tag
from
the
dictionary
with
the
number
50
and
then
the
additional
information.
B
C
B
C
B
Then
again,
the
diagnostic,
the
tool
for
the
diagnostic
notation
might
have
additional
information.
Even
about
that
dictionary,
depending
on
how
the
dictionary
is
set
up.
That
would
allow
it
to
find
a
shorter
way
to
express
that,
because
the
expanded
form
is
not
necessarily
more
useful
to
the
reader
than
the
compact
form.
B
So
if
the
dictionary
were
transferred
over
the
network,
there
might
be
a
kind
of
a
short
form,
that's
necessary
for
expanding
and
one
with
additional
data.
That
is,
that
is
obtained
by
diagnostic
tools.
That
has
such
this.
That
has
that
shorthands.
But
then
again,
this
is
all
only
shaped
by
what
I
would
what
I
could
use
in
coral
and
things
might
look
completely
different
for
say,
students.
C
So
this
this
is
what
my
built-in
algorithm
for
creating
compact
names
for
dictionary
entries
would
generate
from
the
first
line.
In
your
example.
B
B
A
B
B
D
C
I
mean
it's
just
showing
how
to
use
something
we
have
in
a
domain
where
it
hasn't
been
used.
Yet
so
it's
it's
not
really
even
a
specification,
although
it
would
become
one
when
when
people
would
use
that
for
defining
an
actual
csv
document,
but
I
think
it's
actually
useful
to
think
about
areas
of
implications
like
this,
because
we
we
do
have
a
lack
of
formal
description
techniques
for
various
things
we
are
doing,
and
people
pointed
me
to
that
problem
with
a
csv,
and
I
said.
B
D
C
D
I
I
think
the
open
data
community
is
probably
desperate
for
something
more
detailed.
It's
very
right
now
very
much.
Here's
a
word
document
that
badly
explains
what
yeah
entity
is
exporting
yeah,
it's
usually
an
excel
document.
But
yes,
oh
well,
that's
that's
more
advanced
because
in
many
cases
it's
it's
a
it's
a
word
document
where,
from
a
person
you
know
where
the
where
the
save
as
pdf
is
because
that's
beyond
their
capability.
D
Anyway,
I
I
I
don't
know
if
it's
relevant
to
this
working
group,
but
I
I
don't
object
to
you
doing
it,
but
I
just
not
sure
that
it's
the
right
place
to
get
review.
D
Now
sorry
I
just
I
asked
I
just
was
curious,
what
other
things
were,
keeping
this
working
group
up
at
night
and
I'm
not
hearing
that
there's
a
lot.
B
Well,
there's
the
whole
topic
of
cddl2
brewing,
but
that
is
does
not
have
a
corresponding
internet
draft.
Yet
so.
C
Well,
we
have
the
freezer
document,
which
has
some
of
that,
but
okay
yeah
in
particular
the
the
specific
way
in
which
multiple
cdl
specifications
are
combined
into
a
semantic
hole
that
that
is
still
wobbling
around
a
little
bit
at
the
moment.
C
B
So
one
thing
in
the
working
group
one
thing
around
and
one
thing
brewing
is
not
not
a
lot,
but
still
it's
not
a
lot,
but
still
keeping
us
busy.
D
B
C
Right
so
the
the
one
thing
that
came
up
was
the
the
language
tags
discussion.
That
was
relatively
urgent,
so
we
we
need
to
have
that.
B
There
appear
to
be
implementations
that
don't
check
for
the
length
of
the
tags,
but
I
will
have
to
check
the
precise
implementations
to
see
whether
it
can
make
sense
to
go
language
text
direction
or
whether
we
should
just
which
would
be
the
easier
way
and
also
something
that
peter
okay
suggested
just
pick.
Another
tag.
B
Last
name
I
I'd
say:
okay,
but
that's
that's
a
guess.