►
From YouTube: CBOR WG Interim Meeting, 2019-08-28
Description
CBOR WG Interim Meeting, 2019-08-28
A
A
Yes,
so
first
item
on
the
agenda,
something
that
I
forgot
to
put
on
was
about
the
conflict
for
the
Singapore
meeting
for
sea
boar.
So
me
and
Jim
has
started
talking
about
this.
We
have
a
couple
of
working
groups
that
we
have
added
so
far
but
like
at
least
our
you
know,
editors
and
a
more
active
participants
to
tell
us
what
their
conflict
is.
If
they're
not
already
covered.
B
A
A
A
A
B
A
B
A
B
B
A
C
A
A
The
vertical
black
hole
ended
today,
and
hopefully
we
with
that
and
during
last
interim
that
today
we
will
discuss
any
issue
that
the
working
group
last
call
would
create.
So
I
saw
there
is
a
couple
of
reviews
and
if
you
want
to
discuss
those
fast
and
if
there's
anything
that
you
need
input
or
you
will
too.
B
B
It's
not
a
pull
request,
I'm,
sorry,
so
there
is
a
branch
that
has
who
is
that
maybe.
B
B
In
master,
sorry
about
that,
so
that
did
her
I'll
never
commit
you
that
are
addressing
the
various
comment
and
I
think
they
are
all
straightforward,
Mowbray
my
accepts,
and
so
we
can
discuss
whether
the
actual
wording
there
that
I
use
is
the
right
one,
but
I
think
there.
There's
no
discussion
is
about
the
actual
problems.
Okay,.
B
B
B
B
C
C
A
C
I
will
probably
sit
down
and
and
review
stuff
private
today
or
tomorrow.
Okay,.
B
A
B
A
B
Sadly,
Jesus
for
the
same
and
that's
reflected
in
the
ini
registry
and
the
IP
Rd
ipfs
people
have
actually
with
na
player
playing
administration.
They
had
done
that
two
years
ago,
but
for
some
reason
that's
like,
and
the
nice
example
that
you
should
make
sure
that
your
registration
still
gets
back,
because
somebody
step
on
your
feet.
B
Nice
so
syllabus
hold
submitted
national
seven,
which
is
basically
just
a
checkpoint
because
they
are,
there
was
a
ton
of
editorial
changes
and
when
we
now
start
making
taking
your
changes
again,
it's
very
confusing
to
have
all
these
editorial
changes
in
a
disc.
So
that's
the
main
reason
why
there
is
additional
seven
out
there
and
when
you
review
that
that
the
point
really
was
to
get
the
tank
toilet,
terminology
right
and
I
have
some
more
site.
For
that
later,
we
have
one
outstanding
pull
request.
C
That
tends
to
be
the
way
I've
read
goto,
simply
because
is
a
lot
easier
to
a
to
a
it's
a
lot
easier
to
fix
mistakes
and
B.
If
you
find
you
missed
a
corner
case,
I,
don't
really
want
to
update
the
standard
to
deal
with
that,
but
PPA
people
think
that
that
is
a
comprehensive
suite.
Then
they
won't
see
it
yep.
Let's
see
that
updated.
B
And
actually
wrote
a
test
for
the
half-length
floating
point
that
just
includes
all
the
numbers
they
are
because
you
can
do
that
for
four
16-bit
states
and
we
wouldn't
do
that
in
RFC,
but
we
can
easily
do
that
on
the
test
suite
in
the
repository,
so
I
think
it
would
be
a
good
idea
to
really
do
this.
Nestle,
okay,
and
this
means
that
we
actually
have
to
be
task
number
16
that
that's
in
the
issue
exists
now.
B
Okay,
so
so
much
about
the
purpose
on
the
issues
and
I
said
that
the
main
item
was
the
tag
technology
update
for
all
seven,
that
issue
number
85
I
haven't
closed
dead
yet
because
I
think
there
this
still
needs
review.
Whether
we
correctly
did
this.
So
basically,
the
idea
is
to
make
it
more
clear
that
we
have
the
kinds
of
containers
in
seawall.
We
have
arrays
and
maps
obviously,
and
we
have
tags
which
is
a
very
special
container
which
just
combined
the
tag
number
and
then
those
item.
B
So
when
we
talk
about
arrays,
we
have
a
word
for
the
things
that
go
into
raised.
These
are
elements
and
when
we
talk
about
maps
we
can't
a
member
pday,
you
care,
and
we
didn't.
We
have
a
great
term
for
the
thing
that
goes
into
a
tag.
Why
we
said
tagged
item
but
then
also
get
the
whole
thing.
Is
we
take
item
so
this
is?
This
can
be
very
confusing,
so
we
now
say
enclosed
item
and
I
think
this
requires
one
one
more
round
or
correctly
the
text
and
making
sure
we
hit
everything.
B
B
B
What
we
haven't
really
done
very
well
is
talk
about
validity
and
how
this
is
both
a
civil
concept
and
an
application
concept
and
I
think
what
what
we
really
need
is
some,
some
better
terminology
that
that
separates
the
is
the
kind
of
validity
yet.
So
there
is
some
basics
in
the
validity,
but
there
are
also
application
aspects
that
enter
basic
theology.
So
if
you
are
talking
about
duplicate
map
t's,
there
is
an
application
view
on
what
actually
is
public
--it
that
influences
receiver
validity?
B
Okay,
but
which
is
a
bit
weird
so
yeah
the
terminology:
is
it's
not
really
there
and
then,
of
course,
there
is
the
set
of
expected
inputs
that
application,
what
they
could
be
for
this
valid
or
expected.
So
I
would
really
love
to
get
good
terms
wave
form.
The
vanilla
pods
are
so
on
from
XML,
which
were
the
which
they
were
or
invented
for,
and
nobody
usually
has
done
this
civilization
format
valid
versus
application
valid
thing
yet
so
we
never
invent
the
term
here.
All
these
be
some
automated
things
are
an
actual.
B
B
B
B
Okay
and
then,
of
course,
we
still
have
this
term
straight,
which
is
actually
used
to
talk
about
two
different
things.
One
is
validity
checking
and
the
other
one
is
enforcing
preferred
encoding,
and
we
have
to
separate
these
two
forms
of
straight,
which
I
called
speak
one
and
speak
to
on
this
slide.
It's
like
that
about
a
cousin,
get
better
terms
for
that
and
picture.
Probably
it's
not
a
very
good
term
anyway.
B
Yeah,
so
once
these
clowns
out
there
I
think
it's
pretty
straightforward
to
actually
doing
the
editorial
work,
I'm
sure
when
we
go
to
met
validity,
we
do
have
some
more
discussions
about
that
and
the
duplicate
keys
and
all
that
but
I.
Think
generally.
This
is
should
be
straightforward
to
do.
B
B
A
A
B
B
Okay,
so
these
are
two
major
technology
things
that
go
through
the
entire
text
and
change
it.
So
I
would
be
happy
to
do
these
in
in
one
piece
like
we
did
with
the
tank
terminology,
then
we
have
the
whole
issue
of
what
kind
of
advice
we
give
to
designers
of
seaboard
based
purpose,
and
there
are
a
number
of
issues
there,
some
of
which
are
really.
B
For
good
practice
and
I'm
not
so
sure
how
far
we
actually
want
to
go
there,
so
this
is
small
integers
as
not
to
use
that.
That's
certainly
something
that
should
be
said.
Oh,
but
maybe
we
actually
have
too
much
text
for
that
at
the
moment
and
then,
on
the
other
hand,
we
have
things
like
error,
handling
or
handling
of
unusual
cases
like
amount
X
and
then
simple
values
and
Laurence
brought
up
this
whole
thing
about
using
tags
in
unwrapped
way.
B
C
C
B
B
B
How
to
handle
unsupported
takes
that
that's
again
part
of
the
forcing
money
that
that
Johnny
could
go
into
the
obesity
as
well
and
then
the
whole
strictness
in
the
spectrum.
Sense
of
handling
tags,
the
structure,
the
semantic
validity
model,
I'm,
not
sure
this
can
be
entirely
on
the
DCP
side,
because
at
some
point
we
need
to
say
what
the
validity
of
tags
is
about.
B
So,
to
give
you
an
example
here,
JavaScript
recent
versions
of
JavaScript
now
have
begins
and
in
JavaScript
begins,
are
completely
separate
universe
from
other
numbers,
and
that
of
course
means
when,
when
you
are
serializing
JavaScript,
you
want
these
begins
to
be
different
from
from
everything
else,
and
one
idea,
one
might
have
is
say:
oMG
stakes
23
for
victims.
Why
don't
you
use
them
for
Bill?
It's
only,
and
the
problem,
of
course,
is
that
other
implementations
freely
converge
between
their
their
internal
number
formats
and
and
classic
integers
and
big
numbers
and
so
on.
B
B
B
Yeah
I
think
the
basic
idea
here
is
the
right
one.
We
don't
want
to
have
a
single
straight
load,
but
we
want
to
talk
about
some
ways
of
handling
validity
in
tags
and
to
be
able
to
do
that.
We
need
to
be
able
to.
First
of
all
talk
about
such
of
there
is
another
validity,
but
there
may
be
other
things
that
you
want
to
ask
you
a
generic
decoder
that
I
was
discussing
there.
B
B
B
B
B
B
One
issue
is
the
the
signaling
lens,
the
not
a
numbers
that
you
can
put
on
the
wire
and
which
are
really
little
booby
traps,
because
when
we
get
the
receiver
to
put
them
into
your
data,
so
it
will
be
a
data
structure
at
some
point
data
in
the
processing.
They
will
want
to
apply
for
different
operations
on
you
and
then
they
get
exceptions
there
and
that's
about
a
hundred
times
worse
than
the
clamped
integer
array.
Booby
trap
that
that
we
discussed
so
maybe
it's
worth
pointing
this
out
somewhere
and
I.
B
B
This
is
a
little
bit
conflicted
between
the
one
objective
to
be
able
to
use
all
of
actually
75
for
because
people
are
using
all
of
75
for
and
the
other
objective
to
keep
it
simple
enough
for
most
people,
people
who
just
want
to
voice
and
called
a
floating-point
number
without
becoming
a
floating-point
expert,
and
for
those
we
have
the
proposal
to
just
use
one
triage
man
and
there's
actually
text
in
there.
That
says
you
should
always
use
this
one
fight
them,
but
that
better
cause
nonsense.
B
B
Yeah
and
then
we
have
this
Jason
to
see
ball
issue
so
when
the
right
adjacent
receiver
converter,
when
do
you
stop
recognize
something
as
an
integer
and
I
think
we
want
to
go
2
to
the
power
2
to
the
power
53
as
the
default
pressure
diera
there
right
now,
it
says
32
or
64,
+
64
doesn't
make
a
lot
of
sense
and
33
is
kind
of
outdated.
At
this
point,.
C
C
C
B
C
B
B
Yeah,
so
do
we
have
a
name
for
the
Jason?
That
is
not
amazing.