►
From YouTube: IETF-UUIDREV-20230119-1900
Description
UUIDREV meeting session at IETF
2023/01/19 1900
https://datatracker.ietf.org/meeting//proceedings/
B
C
Yeah,
so
I
thought
we
would
just
all
go
to
the
I'll
share
the
the
notes
for
now
and
yes
window.
C
So,
let's
see
let's
go
to
this
one
here
and
why
isn't
that
awake?
Oh
well,
hello!
Everyone!
So
I
wanted
to
suggest.
C
First
of
all,
since
we're
a
very
small
group
and
the
latency
to
unmute
yourself
with
the
upper
left,
mic
button
is
high
because
it
builds
and
tears
down
the
audio
Channel
each
time
that
everyone
just
go
and
click
on
the
mute
audio
on
the
mic
button
above
in
the
left,
and
then
please
mute
yourself
in
the
with
the
microphone,
the
local
mute
on
the
lower
right,
the
very
bottom.
That
way
you
can
just
unmute
yourself
really
quickly.
If
you
need
to
say
something
as
opposed
to
waiting
like
XML.
C
You
know
RPC
time
before
you
can
do
anything
so
welcome.
So
this
is
the
first
of
two
virtual
interims.
The
second
one
is
on
the
16th
of
February
and
the
hope
is
that
we'll
be
done
at
that
point
and
we'll
be
in
working
group
last
call
and
we
won't
need
to
have
a
meeting
at
ietf
116
in
Yokohama.
C
If
that
proves
not
to
be
the
case,
then
we
could
talk
about
that.
I,
don't
know
I,
don't
think
I'll
be
physically
present
and
I.
Don't
know
about
the
rest
of
you.
It
may
be
a
waste
of
our
time
to
try
to
meet
physically
anyway,
but
that's
an
open
question.
So
so,
just
to
remind
you
about
the
note.
Well,
we
have
a
bunch
of
of
processes
and
policies
relating
to
your
contributions.
C
We
have
an
IPR
policy
and
there's
a
whole
list
of
different
documents
that
apply
to
this
and
in
particular,
if
you
know
of
any
patents
or
IPR
relating
to
the
document,
then
you
are
asked.
Please
disclose
it.
Even
if
you
it's
not
your
IPR
code
of
conduct,
basically
BCP
54,
please
be
nice
to
each
other.
Is
the
simple
summary
and
I'm
gonna
go
on
to
on
from
that.
C
So
today
we're
going
to
be
dealing
with
the
with
this
4122
this
document
and
we're
going
to
be
going
through
most
of
the
issues
and
pull
requests.
So
guess
I
could
turn
on
my
video
while
I'm
talking
and
with
that.
Are
there
any
changes
to
the
agenda
that
are.
We
should
talk
about.
C
Go
so
do
you
wanna,
do
you
want
me
to
share
the
GitHub
as
we
go,
or
do
you
want.
B
To
do
that,
no
that's
completely
fine.
What
I
want
to
do
is
I
want
to
start
bottom
up
because
specifically
Jim
put
some
notes.
You
can
see
in
number,
35
and-
and
those
are
already
addressed
here
so.
B
So
specifically
tagged
everything,
so
you
can
nicely
see
what
we're
doing
and
if
you
know,
I'm,
just
going
to
look
at
35
on
my
other
screen,
but
they're
32.
Sorry,
pull
request,
32
kind
of
hits
a
big
grouping
of
things
so
there's
some
spelling
errors
that
are
caught
under
number
18.
and
number
18.
If
we,
you
know,
if
you
click
the
commit
on
the
right
side,
we
can
see
what
that
is
or
yeah
exactly.
You
can
hover
it,
but
it's
a
it's
a
it's.
Basically,
a
quick
little
spelling
error.
It's
mixed
case
mixed.
B
One
to
say
yes,
this
is
probably
okay
to
fix.
Yeah
I
tried
to
do
all
of
my
commits,
like
you
said
last
time
and
one
at
a
time.
So
we
can
look
at
this
right.
This
one's
just
changing
a
mixed
case,
so
I
don't
really
see
any
problem
with
number
18
going
into
draft
one,
but
I
just
wanted
to
talk
about
it.
So
that's
that's
the
first
one
19.
we
can
just
go
down
like
that.
B
Exactly
this
one
adds
a
forward
reference
to
unidentifiable
section
to
effectively
allow
unique
identifier
and
and
kind
of
just
or
allow
us
to
kind
of
have
a
reference.
So
it's
just
a
reference
Edition
which
shouldn't
be
any
problem.
There
number
20
moves
some
text
and
this
text
specifically
when
I
was
moving
content
from
RFC
4122
over
I
I
put
this
into
a
section
and
when
I
was
going
through,
the
document
I
realized
that
we
have
a
section
on
distributed.
B
Node
applications
right
and
this
text
on
distributed
applications
was
way
past
that
section,
so
it
makes
sense
in
the
grand
scheme
to
kind
of
just
marry
this
up.
Put
it
back
up
there.
The
conversation
here
would
be:
do
we
need
this
text
at
all,
right
and
I
think
the
text
still
needs
to
to
to
be
there.
B
It
makes
a
lot
of
sense,
but
if
we
do
keep
it,
we
move
it
up
in
a
distributed
section
I
mean
that
that's
the
first
of
like
okay,
this
isn't
a
draft
change,
or
this
is
like
just
a
quick
little
change.
So
no,
no
major
technical
changes
that
need
to
go
through
and
if,
at
any
point,
somebody
has
something
like
I'm
willing
to
discuss
any
of
these
25.
B
Is,
oh
sorry,
24
is
exactly
sorry,
go
to
24.
First
I'm
mixed
I
missed
up.
You
want
25
yeah,
we'll
do
25,
that's
fine!
I
I
did
them
in
in
alphabetical
order,
but
after
I
did
that
all
right.
So
this
is
a.
B
A
b
and
F
reference
and
specifically
I
had
the
wrong
number.
This
was
brought
up
in
the
last
interim
meeting
or
the
last
meeting
we
had
I
just
completely
put
the
wrong
RFC,
so
it's
just
a
typo,
so
we're
just
pointing
a
b
and
F
to
the
correct,
like
actual
forward
reference
in
all
applicable
sections.
B
So
it's
again
a
nothing,
nothing
change
20,
let's
just
go
to
24,
because
it's
on
the
same
topic
of
this
I,
don't
know
why
they're
out
of
order
I
committed
them
in
the
wrong
order.
It's
the
very
bottom!
This
is
the
one
Jim
was
talking
about.
This
changes,
the
the
a
b
and
F
hex
formatting
to
actually
follow
RFC
40
5234-
and
this
is
what
you
were
talking
about-
Jim
in
the
review
of
the
other
of
Bradley's
pull
request.
B
So
I've
already
got
these
so
they'll
get
merged
down,
so
it
should
be
covered
good
and
it
looks
like
based
on
just
my
checks
of
the
two.
It
looks
like
it's
solid
and
you
know
the
specific
thing
that
you
were
talking
about
was
just
adding
the
star
in
there
and
that
that
is
correct.
So
you're
you're,
right
and
I
did
change
that
just
want
to
let
you
know
that
I
I
did
address
it,
so
we
don't
have
to
address
it
in
that
pull
request.
B
E
C
So
I
think
that
there's
two
things
happening:
I
think
that
the
Syntax
for
hex
octet
is
not
a
b
and
F
correct
like
this
is
not
yet
it
is
okay,
okay,
so
the
the
change,
though,
to
say,
there's
four
hexotex
I,
think
Carson
you're
saying
mandates
that
they're
before
and
there
couldn't
be
three.
So
if
you
had
a
leading
zero
or
something
that
you
deleted,
that
would
wouldn't
fit
into
this
pattern.
E
Now,
what
I'm
saying
is
that
this
change
says
you
can
have
more
than
four
hex
other
kids
before
the
first
Dash
and
then
more
than
two
before
the
second
dish,
because
the
star
really
means
from
Two,
and
the
from
here
is
four
and
the
two
is
infinity,
because
you
didn't
put
it
in
a
two.
It's.
B
On
that,
in
the
first
place,
a
b
and
F.
B
B
Like
it
does
a
lot
that
was,
my
question
was
because
the
very
first
RFC
actually
Carson
did
have
it
as
four
hex
octet,
two
hex
OCTA,
two
hex
yeah
I
just
copied
from
there
and
then
I
got
a
bunch
of
recommendations
that
it
needed
to
be
for
Star,
so
I
I,
reviewed
it.
It
was
like
stars,
looks
relevant
so
that
added
logic
that
the
stars
is
not
present
and
not
correct.
I
I.
E
Yeah,
you
could
also
say
four
star
four,
which
is
kind
of
redundant,
so
why
not
just
use
the
the
existing
for
hexocket
a.
C
B
B
A
B
A
Main
comment
there
was
that
you
have
a
a
duplicate
of
this
text,
basically
over
in
the
registration
section,
and
that
the
two
should
be
the
same,
whatever
whatever
it
is.
I'm
yeah.
F
B
C
C
B
It's
in
the
urine
namespaces
with
Diana
and
that's
the
only
place.
I
know
if
I've
checked
that,
specifically,
if
you
look
up
the
urine
namespaces
it
has
it
registered
there
and.
A
C
B
Okay,
so
do
I
can
fix
that
good
action
item
appreciate
the
feedback
there.
I
think
we're
good
to
go
to
the
next
item.
Then,
on
that
list,
which
was
we
stopped
at,
we,
we
deviated
it
24,
so
27.
F
C
All
right
boom
and.
B
27.
27
yep,
so
this
one
is
very
simple:
you
know
it's
implementation
specific.
We
put
a
should
and
it
needs
to
be
a
must
there's.
This
goes
into
our
larger
should
and
most
and
verbiage
review
that
I'll
probably
end
up
doing
a
draft
two
or
draft
three.
But
this
is
pretty
simple
like.
If
we
look
at
the
document,
you
know
we,
it
can
be
anything
and
you
you
do
version
eight.
So
a
must
make
sense,
like
don't
assume
what
you
know
is
in
there,
because
we
don't
know
everything's
very
specific
to
an
implementation.
C
B
B
29
is
a
typo.
It's
a
really
old
typo
back
when
uuid
version
8
was
specific
to
alternate
time
based
and
we
relaxed
it
to
be
just
any
old
uuid
you
cook
up
for
experimental
or
very
specific
purposes.
We
just
yank
that
out,
because
it's
now
used
for
anything
right.
So
you
know
time-based
doesn't
need
to
be
there.
So
it's
just
leftover
garbage
more
or
less
30
is.
B
Is
a
descriptor
that
effectively
changes
a
few
things
in
in
the
uid
version,
seven
in
space,
and
it
just
adds
some
clarifications
as
to
what's
going
on
specifically
for
some
feedback
from
testers
out
in
the
field
and
implementation,
folks,
where
they
may
have
been
thinking
about
old
version,
seven
and
and
some
of
the
different
levels
of
bit
trickery.
We
were
doing
and
just
being
able
to
call
out
that
it
is
milliseconds
and
then
calling
out
where
the
counters
go
if
they
want
to
do
some
counters
in
Randomness
inside
the
randomness.
B
B
And
that
should
have
been
the
last
one
for
that
pull
request,
so
that
basically
hits
all
of
those
and
obviously
I
can
change
the
a
b
and
f
as
an
option,
and
then
we
can
get
that
into
draft
to
this
draft,
specifically
for
draft
one
to
push
and
my
goal
for
this
is,
if
you
guys,
are
okay
with
these.
B
Well,
you
know
we
submit
draft
by,
like
the
end
of
the
week
like
Friday
or
something.
If
there's
not
a
ton
of
changes,
we
can
get
drafted
ones,
buttoned
up
stamped
and
out,
and
then
we
can
start
working
on
draft
two
stuff.
If
that's
that
sounds
awesome,
okay,
so
let's,
let's
go
to
the
next
pull
request,
which
is
I
believe
the
next
two
are
Brad's.
A
B
D
C
So
so
not
wait
to
the
next
interim,
because
then
we
won't
get
the
cycle
review.
I'm
happy
with
you
posting
a
document,
the
end
of
Friday
yeah
great.
If,
if
you
know
you
have
a
a
stealth
snowstorm
and
it
winds
up,
Monday,
whatever
okay,
but
I,
think
that
num
integers
are
cheap
and
being
able
to
see
a
succession
of
diffs
and
an
energy
helps
people
review
things
well,.
C
B
Do
that
I
was
saying
32-37
all
just
get
well
at
the
end
of
this
Friday
I
fix
whatever
we
say
do
here.
All
of
this.
That's
in
the
open
PR
as
it
gets
bundled
up.
That's
draft
one.
Then
we
start
the
pull,
requests
and
and
changes
for
draft
two
for
all
the
stuff
that
kind
of
was
going
on
on
the
email
thread.
As
of
yesterday
that
we
haven't
had
a
chance
to
really
digest,
or
even
this
morning.
B
C
B
The
34
here,
which
is
issue
21,
Brad,
updated.
If
we,
if
you
click
just,
do
files
changed
or
even
click
on
the
the
commit
with
either
one.
This
specifically
is
changing.
B
G
G
G
I
think
there
were
a
bunch
of
I
think
what
it
is.
Yeah
I
think
I
think
what
it
is
is
that
each
time
there
was
there
were
some
exact
rules
in
the
earlier
one.
That's
what
this
retained
the
old
mock
address.
Behavior
is,
but
that's
talking
about
the
node
I
apologize
I,
trying
to
remember
the
context
of
where
this
came
up.
Is
there
a
way
to
see
in
here
what
problem
this
was
solving?
Does
it
have
a
link
to
like
where
what
we
were
fixing
with
this.
G
B
We
could
you
could
see
lyos's
a
specific
comment
about
what
what
he's,
what
we're
trying
to
fix.
Specifically,
he
was
the
one
who
recommended
right
he's
not
not
here,
but
lyos
has
been
a
very
good
resource
throughout
the
previous
drafts
for
feedback,
so
we
definitely
value
his
input.
G
Right,
okay
got
it
so
there's
two
things
here:
yeah,
so
the
node.
What
they're
saying
is
that
in
the
in
the
MAC
address
version,
the
node
was
like
a
static
value.
So
he's
saying,
if
you're
going
to
tell
people
to
regenerate
the
node
with
a
random
number,
you
should
be
very
clear
that
this
is
every
single
new
uuid.
So
that's
that's
one
specific
thing
that
we're
fixing
and
then
I
think
on
the
clock
sequence.
G
That's
that's
a
that's
optional!
That's
the
point
of
that.
C
So
let
me
ask
a
question
about
the
word:
should.
A
D
A
Is
Legacy
behavior
from
previous
implementations
right.
A
B
G
Yeah
just
for
a
con
sorry,
just
to
clarify
one
more
thing:
what
happens
is
uid
V6
is
specifically
just
a
bit
swap
from
the
the
uuid,
the
other
uuid
one
I.
Think
so
because
of
that
direct
translation,
we're
saying
there's
a
lot
of
people
are
going
to
implement
this
by
taking
a
uidb1
and
swapping
some
bits
around
and
we're
trying
to
communicate
that
it's
totally
okay
to
just
keep
whatever
the
prior
implementation
was
doing.
C
C
D
B
Oh
I
I
clean,
I,
clean
them
up
later
guys,
don't
don't
fret
on
it
yeah
you
can.
You
can
knock
labels
on
the
side
or
you
can
there's.
C
B
Cool
okay,
so
the
the
35
is,
you
know
we
had
a
meeting
last
time
about
moving
this
template
and
then
making
sure
it
it
fills
out
and
is
correct.
So
we
ain't
got
the
to
do.
We
put
it
in
the
right
spot
that
the
actual
to
do
here
specifically,
is
around
the
ABN
F.
So
all
this
is
is
updated.
Namespace
registration
template
with
all
the
right
things
pointing
to
the
right
rfcs,
all
that
fun
stuff
and
it's
it's
90
the
same
from
what
I
can
see.
B
B
Yeah
and
that's
because
I
think
we
the
way
we
you
know
have
it
merged
up,
they
all
came
from
the
same
dock
because
they
had
opened
but
yeah.
We
can
fix
that
and
we
can
fix
the
ABN
f.
A
So
should
I
just
it
sounds
like
my
suggestions
are
all
wrong.
There
should
I
just
delete
that.
B
Yeah,
you
can
yeah
you.
Can
you
can
delete
those.
A
C
There
really
no
way
that
we
can
have
the
Regis
registry
point
to
this
document,
rather
than
repeating
the
ABN
f.
B
C
B
C
B
I
think
we
can
move
to
the
next
full
request,
which
I
believe
is
from
Corbin
on
the
meeting
real
quick
one,
nice
nice
catch
Corbin.
A
C
Yes,
so
we're
gonna
Kaiser
is
going
to
fix
the
ref,
the
remove
the
duplicate
abnf
and
make
it
a
reference
and
then
we're
going
to
merge.
It.
B
B
So
Corbin
specifically
was
fixing
a
typo
in
the
md5
low,
where
I
it
was
the
Shah
one,
and
you
know
I
copy
and
pasted
from
in
from
the
uuid
version
three
into
five
and
I
didn't
change
one
of
the
descriptors
for
the
field,
so
we're
fixing
the
descriptor
here.
There's
one
conversation
on
there.
B
That's
new
six
minutes
ago
from
from
from
a
gentleman,
if
we
could
look
at
that,
real
quick
that'll
be
another
change
which
you
probably
need
to
hit
itches,
also
and
I
believe
in
the
same
section,
I
pointed
to
the
mb5
security
security.
E
Well,
actually,
yeah,
someone
is
probably
better
I
mean
Shaw,
one
is
kind
of
dead
as
well,
but
let's
ignore
that.
Furthermore,.
A
So
I
I
had,
in
my
pull
request,
made
a
made
a
different
change
that
you
probably
shouldn't
accept,
then,
which
is
that
I
I
had
thought
that
this
text
was
just
misplaced
and
moved
that
sentence
back
to
the
mb5
section,
because
there
wasn't
a
similar
statement.
There
I
didn't
actually
check
with
the
RFC
referred
to,
but
because
I
assumed
it
was
md5
considerations.
So
you
know
whatever.
F
The
right
thing,
there's
one
sorry,
there's
one
more
shallow,
that's
right
above
that
we
may
want
to
change
a
shot
one
low
for
consistency.
B
B
B
A
And
and
I
used
a
completely
different
style.
I
took
a
lot
of
things
that
I
thought
were
you
know,
rather
than
create
a
an
issue
for
every
typo
which
really
doesn't
scale.
I
I
opted
to
create
a
pull
request
that
basically
just
describes.
A
You
know
what
the
kind
of
the
specific
changes
that
I
thought
that
I
had,
and
some
of
them
are
wrong,
and
some
of
them
are
right,
but
I
mean
that,
rather
than
rather
than
trying
and
create
a
lot
of
individual
work
for
each
for
each
individual
change,
this
seemed
like
a
a
better
approach.
C
B
A
B
I
think
the
other
gentleman's
going
to
catch
that
in
his.
So
you
can,
you
can
just
leave
it
be
for
the
moment.
B
A
C
B
Would
I
I'll
merge
them
in
a
very,
very
specific
way
the
stakehold
merger
so
I'll
handle
the
merge.
We're
we're
good
there.
B
A
Yeah
that
that's
fine
I
mean
these,
were
these
were
the
things
that
I
wasn't
as
clear
on
what
the
what
the
appropriate
resolution
on
these
is
many
of
them,
some,
like
the
some
of
them,
are
straightforward.
Like
the
the
the
binary
representation
I,
think
you
just
put
that
in
two
lines,
just
put
a
courage,
return
in
the
middle
of
it
and
you'll
be
fine.
B
Yeah
it
looked
like
they
wanted
a
slash
at
the
end
kind
of
like
you
do
in
like
a
unique
shell
to
indicate
that
there's
like
you
know
it
doesn't
stop
here,
but
I'll
do
it
as
per
whatever
the
RFC
doc
says,
and
it
looked
like
it
was
just
a
trailing
slash
to
indicate
that
one
thing
I
did
want
to
comment
is
brufa
Robert,
specifically
on
the
call
has
a
typo
called
out
in
section
33.
B
I
have
to
double
check
if,
if
that
was
caught
in
the
review
that
you
did
Jim,
if
so
we
could
probably
close
33,
but
if
not
I'm
gonna
I'm
gonna
sneak
that
into
draft
zero
one,
because
it's
just
a
typo
fix.
But
you
know
I'll
I'll
merge
that
in
there
so
30
33
specifically
is
probably
going
to
get
in
Draft
zero
one.
So
I
can
close
that
out.
B
30.
you
actually
clean
up
90,
it's
the
third
from
the
top.
C
Sorry
talking
about
issue
90.
B
C
B
C
Is
your
reply
to
Jim's
email
for
the
rest
of
the
group
if
you're
looking
for
it.
B
C
Haven't
we
have
20
minutes
scheduled,
of
course,
if
we're
done,
then
let's
not
beat
the
dead
horse.
So.
A
Well,
let
me
let
me
kind
of
hit
a
couple
of
the
the
ones
that
I
think
are
are
more
important.
One
of
them
has
to
do
with
the.
What
is
it.
A
Trying
to
find
the
the
document
that
the
where
the
the
time
stamp
and
version
number
are,
are
merged
there
and
and.
A
Was
kind
of
the
way
it
was
represented
in
in
version
one
time,
high
inversion
and.
B
A
Sir,
so
yeah
I
realized.
That
was
the
way
it
was
done
in
version
one.
It's
very
confusing
the
way
it
is
I
mean
you're
you're,
basically
not
basically
deciding
to
to
merge
two
things
into
a
field.
I
would
recommend
that
you
know
like
for
version.
One
you
put
in
a
version
field
call
the
the
other
field
time
High
actually
shorten
the
whatever
it
is:
clock
sequence
high
res.
A
That
also
includes
the
the
variant
bets,
and
you
know
and
explain
in
the
text
that
this
was
you
know
this
concatenated
with
the
version
field
was
referred
to
as
high
time
high
inversion
in
previous
versions
of
the
document.
I
think
you
know
you
know,
for
people
that
that
are
familiar
with
the
previous
version,
that
that
provides
the
the
continuity.
But
this
you
know
having
having
fields
that
contain
two
completely
different.
Things
is
very
confusing
to
somebody.
That's
just
reading
this.
For
the
first
time.
C
B
Yeah
but
yes,.
F
I
agree
with
you:
Jim
yeah
I
do
generally
agree
I
kind
of
got
into
this
view,
ID
specification
stuff
because
of
research
on
uuids,
because
partially
because
of
this
confusion
around
mixing
of
fields.
F
B
Yeah
we
we
do
visual
because
I
hated
it
too,
just
my
two
cents,
anything
that
was
created
by
me
and
Brad
net
new
did
not
Multiplex
Fields
like
that
and
I
agree.
We
can
remove
it.
That
will
have
some
Rippling
changes
throughout
the
document
that
will
need
to
be
vetted
and
double
checked,
to
make
sure
that
they're,
they're,
good
and
they're
correct
right,
technical
things
will
have
to
be.
You
know,
really
put
put
a
fine
microscope
too,
but
I'm
on
board
with
changing
it
too.
So
that's
that's
my.
G
Vote
yeah:
we
might
need
to
come
up
with
a
good
way
to
communicate.
Some
of
that
stuff
like
what's
currently
says
octet
number
might
need
to
be
like
octet
number
five.
You
know
three
most
significant
bits
or
something
like
that
went
to
figure
out
how
to
fit
it
into
where
we
have
those
slots
right
now,
but
assuming
we
can
solve
that
problem,
I
think
it's
a
good
idea
to
fix
it.
A
B
B
Yes,
well,
Max
knew
nil
was
was
in
the
previous
and
there
are
from
I
believe
lyos
or
maybe
Sergey
looked
I.
Don't
remember
the
exact
gentleman
who
looked
at
it,
but
there
was
a
person
who
reviewed
some
libraries
and
they
saw
these
carved
out
and
the
idea
was
well
if
they're
carving
out
the
minimum,
we
should
carve
out
the
maximum
as
well
just
so
that
people
know
that
these
are
special
key
special
use
case
scenarios.
B
B
Do
we
call
that
out
and
the
issue
tracker
kind
of
goes
into
the
weeds
a
little
bit
and
I
kind
of
mauled
on
it
molded
over
I
think
if
we
just
put
a
one-liner,
something
in
the
variant
section
to
say
hey
by
the
way.
This
spec
also
defines
these
two
uuids
in
these
spaces,
which
should
be
okay,
and
really
it's
just
one
more
we're
just
giving
it
better
clarification.
A
Okay,
okay,
understood
about
uuid
version,
5
being
in
the
previous
one
and
I
had
for
not
realized
that
that
was
the
case.
I
forgot,
where
the,
where
the
changes
happened,
oh
hold
on
something's.
B
Turning
yeah,
that's
the
problem
with
three
and
three
and
fives
they're
muddied
up
and
very
poorly
described
in
the
old
document.
So
they
get,
they
get
glossed
over
a
lot.
One
and
four
are
really
the
focus
of
that
document
and
three
and
five
just
kind
of
get
a
little
bit
of
text.
So
we
expanded
upon
those
nicely
I
believe
in
this
one.
A
What
do
we
think
about
adding
a
shot?
256
uuid
I
mean
I'm,
not
convinced
that
that
there's
a
a
particular
Pro.
The
way
that
we
use
sha1
here
is
a
is
a
particular
problem,
but
there
are
a
lot
of
people
that
look
at
anything
that
has
sha-1
in
it
just
like
they
look
at
anything
that
has
a
Unix
Epoch
time
stamp
on
it
and
they
say.
Oh,
that's,
that's
not
good.
So.
C
Yeah
we
we
need
to
either
say
it's
not
a
problem
or
because
of
how
we're
using
it
we're
not
using
it
for
its
cryptographic,
pre-image
resistance,
we're
using
it
as
a
you
know.
Essentially,
it's
a
hash
right,
a
non-cryptographic
hash.
C
And
we're
just
using
it
for
its
distribution,
so
we
just
need
to
say
something
that,
like
that
and
I
think
that's
fine,
but
you're
right.
Otherwise
we're
going
to
get
we'll
get
some
review
comments
later
on
about
this.
A
B
I
did
test
shot
256
when
I
was
reverse
engineering
and
kind
of
doing
some
independent
testing
on
three
and
five
and
looking
at
various
libraries
that
do
you
know
the
Shaw
and
the
md5
variants
and
kind
of
just
comparing
the
results
and
looking
at
the
code
and
just
kind
of
understanding
what
prior
art
is.
I
did
cook
up
with
shot,
256
and
thought
hey
this.
This
makes
sense,
but
I
don't
know
if
the
effort's
there,
because
those
two
uuid
variants,
first
or
versions,
aren't
the
most
ubiquitous
out
there.
In
the
first
place.
C
B
C
Me
ask
a
different
question:
someone's
coming
along
Shaw,
Shaw
writing
new
code
and
Shaw.
One
has
been
removed
from
their
libraries
and
they're
like
well.
I,
don't
want
to
stick
it
back
in
just
to
create
a
uuid.
C
A
D
D
Yeah
so
I'll
just
jump
in
with
a
little
bit.
So
if
I
remember
correctly,
I'll
get
to
that
question
in
just
a
moment.
The
the
research
we
did
showed
that
the
the
namespace
IDS,
both
B3
and
B5,
were
like
one
percent
of
use
cases,
so
they're
they're,
pretty
minimal
and
in
terms
of
the
real
world
usage
my
take
on
this
spec
and
Kaiser.
D
Please
correct
me
if
I'm
wrong,
but
the
this
issue
of
Shah
of
the
shaw
one
or
the
you
know
the
hash
to
use
for
the
namespace
IDS
was
pretty
far
out
on
the
edge
of
discussion
and
interest
in
all
of
the
the
discussion
that
led
up
to
this
I,
just
I,
don't
think
I've
I've
I
can
recall
any
conversation
where
people
were
concerned
with
the
quality
of
the
namespace
uuids
that
were
in
the
original
spec
other
than
you
know.
D
Md5
had
already
been
deprecated
at
that
point
to
your
specific
question
about
what
somebody
would
do
if
they
wanted
a
shot,
256
namespace
uuid.
D
My
answer
to
that
would
be
that's
actually
a
pretty
good
use
case
for
the
version
8,
which
is
the
experimental
or
the
the
vendor.
Specific
uuid,
okay.
B
Right
yeah,
we
could
put
verbiage
like,
like
Robert,
said,
to
steer
towards
version
8,
for
that
use
case
specifically
says
Hey
like
okay.
Do
that
in
this
right,
like
kind
of
as
a?
How
do
you
handle
that
scenario?
I
think
that
makes
sense
and
I
will
also
just
comment
that
yes,
Robert,
is
correct
throughout
all
of
our
conversations.
There's
very
Zero,
Zero
To
None
on
the
namespace,
based
other
than
how
it's
laid
out
in
the
previous
document
is
terrible
and
it
needs
work
there.
B
G
And
I'd
also
like
to
just
add
that
there
are
also
a
lot
of
implementations
that
just
choose
not
to
implement
those
versions
for,
for
that
exact
reason
and
they're.
Just
like
this
is
old
stuff.
We
don't
need
this.
We
can
just
leave
it
out.
Okay,.
C
There
are
there,
are
there
communities
where
a
specific
uuid
format
is
mandated
version
rather
is
mandated
foreign.
D
A
C
B
C
B
B
F
G
And
and
that's
on
the
generation
side,
I
think
that's
absolutely
right
and
that
typically
comes
up
of
whoever's.
Writing
the
code
to
generate
the
euid
has
to
choose
the
right
thing
for
that
application,
and
then
everybody
else
just
says
cool.
They
gave
me
a
uuid
and
that's
my
unique
identifier
and
doesn't
care
what
type
it
is.
A
Yeah
I
have
a
little
bit
of
concern
about
overloading
version
8.
For
that
reason,
just
because
it
it
doesn't
really
give
you
any
any
clue
about
the
you
know
what
any
characteristics
of
the
of
the
uuid
but
understand
that
it's
good
to
have
something
for
you
know
for
vendor-specific
one
or
for
future
use
or
whatever,
but
I'm
a
little
concerned
that
we're
overloading
it
with
a
lot
of
use
cases
here.
B
B
There's
some
scenarios
where
you're
sending
it
places,
but
then
it's
an
opaque
string
of
data
that
we're
not
doing
anything
with
right.
It's
not
like
I,
I,
informatted
or
I
did
not
have
the
correct
format
on
like
an
IPv6
address
or
something
right
where
it's
going
to
cause
Downstream
issues,
it's
a
it's
just
a
value
that
sits
there.
So
you
don't
really
have
some
of
those
problems,
overloading
version
8
in
my
opinion,
so
it
does
help
us
extend
the
spec
because
sure
we
make
a
version.
B
Nine
shop,
256
well
in
six
years,
shot
256
is
gone
that
just
as
an
example
and
shot
512
is
the
new
golden
standard
right
because
of
for
whatever
reason.
Now
we
have
to
make
a
version
10.
and
we
only
have
so
many
versions.
So
the
idea
behind
version
8
was
also
to
increase
this
Spec's
lifetime.
Make
sure
that
we
don't
have
to
keep
updating
and
creating
versions
with
a
very
finite
version
space.
B
C
Right
sure,
okay,
yeah
yeah,
that's
why
I
wanted
to
know
whether
you
could,
just
within
the
the
existing
namespace
version,
whether
you
could
just
use
sha-256
and
no
one
would
notice
or
care,
and
what
you're
telling
me
is
it's
that
sometimes
that's
okay,
but
that
well,
you
would
be
sure
you
should
just
use
version
8
or
something.
B
B
Doesn't
make
a
lot
of
sense
because
now
all
of
a
sudden
I
did
this?
They
don't
know
they
do
it.
Now.
It's
not
equal
and-
and
you
know,
I-
think
database.
Things
pop
up,
I
think
alerts
that
pop
up
weird
application
issues
that
don't
do
that.
But
if
you
just
use
version
eight
and
you
use
it
in
this
context-
and
everybody
knows
within
your
application-
you're
fine,
so
I
can
I
can
definitely
add
some.
So
so
the
notes
on
that
shot,
286
in
the
shaw
there
steer
towards
V8.
For
that
use
case.
C
A
Kirsten,
thanks
for
keeping
us
honest
on
on
ABN
F
appreciate
that
yeah.
B
I
did
have
one
last
thing:
Jim
lyos,
K
came
back
to
me.
I
did
email
him
and
he
said
that
lyos
k
for
the
acknowledgments
at
the
end
is
completely
fine.
That's
what
he's
been
going?
It's
okay,
Scandal!
Since
the
beginning
of
time
on
the
internet.
That's
everybody
everything
he
has
is
lyos
K.
Nobody
needs
to
know
his
last
name.
So
he's
good
with
that.
D
A
B
C
Folks,
previous
questions,
whether
or
not
we
would,
we
would
be
cleaning
out
this
acknowledgment
section
of
the
what
the
text
we
had
free
for.
B
C
G
C
Send
some
suggestion
to
this:
we
maybe
don't
need
as
much
detail
and
but
we
should
add
the
we
should
probably
retain
the
names.
If
not
the
details.
B
Yeah,
that's
a
good
point
like
Ted
and
and
Professor
lartmouth
can
probably
say
and,
and
the
other
gentleman
and
folks
in
there
could
probably
just
stay
and
then
just
remove
the
context,
because
it's
not
even
going
to
make
sense
in
this
document.
Anyways.
C
Yeah
I
know
that's
what
I
was
going
to
do.
I
was
just
going
to
basically
leave
their
names
in
and
say
they
helped
with
4122
and
then
go
on
to
that
the
additional
things
okay.
So
we
have
I'm
going
to
close
the
meeting
here
now
we
have
another
interim
on
basically
in
a
month
and
oh
that
didn't
become
a
URL
either.
Did
it
so
February
16th
that
did
not
work.
C
I
know
that
did
not
work
there.
We
go.
No,
that's
the
wrong.
Still
this
wrong
window,
I
can't
get
it
in
the
right
window.
There
we
go
Thursday
February,
16th
same
time,
there's
a
different
meet
Echo,
URL
and
I'll
post
an
agenda
to
the
list
so
that
everyone's
sure,
and
so
the
plan
is
that
you're
going
to
post
this
version.
Oh
one
I!
Guess
it's
going
to
be
in
the
next
couple
days
and
we'll
try
to
basically
have
a
another
iteration.
C
One
thing:
that's
really
good
is
if
every
issue
does
have
a
some
proposed
text,
even
if
it's
controversial
and
then
we
can
argue
about
it-
and
maybe
the
goal
would
be
to
you
know
by
the
end
of
February-
have
a
a
stable
document
that
we
could
put
in
a
working
group.
Last
call,
let
me
just
ask
if
there
is
interest
in
meeting
at
ietf
116
in
Yokohama,
whether
you
think
this
work
would
benefit
by
having
a
wider
audience
or
really
it's
just.
B
I
will
comment
that
I
have
already
purchased
that
one
day
pass
in
the
anticipation
that
we
would
have
a
meeting
okay,
just
to
nail
down
any
last
items
before
we
submit
this
because
I
believe
we're
trying
to
submit
it
to
iesg
at
the
end
of
March
ietf
happens.
We
have
our
one
week
change
window
beforehand,
so
we
drop
whatever
draft
two
draft
three,
whatever
we're
on
there.
We
talked
about
it
one
last
time
and
say:
is
there
anything
left
before
we
submit
it
and
it's
out
of
our
hands?
B
C
Okay,
so
if
we
do
have
a
meeting,
many
of
us
will
probably
be
remote.
I
believe
I
will
be
remote,
but
I
haven't
I,
haven't
I,
haven't
purchased
any
I
haven't
committed.
Yet
to
that
so
I've.
A
I've
bought
tickets
and
everything
for
going
in
person,
but
I
may
be
backing
out
of
that
Jen
and
going
virtual
just
because
of
infection
concerns
and
stuff,
primarily.
C
F
Is
there
anything
that
the
community
can
do
before
the
next
meeting?
Just
like?
Is
there
anything
that
that
can
be
done
by
like
external
contributors.
C
Read
the
document
argue
with
the
data
tracker
and
read
the
mailing
list,
even
if,
and
also
even
if
all
you
do
is
read
the
document
and
agree
that
the
changes
are
good,
that's
valuable
feedback.
Otherwise
it's
like
an
echo
empty
Echo
chamber
and
we
don't
know
if
anyone's
the
difference
between
nobody
read
the
document
from
nobody
has
any
disagreement,
so
that
would
be
great
or
if
you
have
other
community,
you
know
of
other
communities
that
maybe
should
be
informed.
C
Even
if
all
they
say
again
is
everything
looks
good
to
me.
That's
also
valuable.
B
Yeah
I'll
I'll
comment.
One
last
thing
specifically
from
my
perspective:
Robert
I
would
like
version
three
and
version.
Four
are
net
new
extrapolated
from
what
was
in
the
previous
document.
Give
me
a
tech.
If
somebody
could
please
just
give
me
a
technical
check.
I
did
a
lot
of
reverse
engineering.
I
feel
pretty
good
about
the
way.
The
text
describes
the
technical
creation
of
those,
but
it
is
net
new,
so
there's
a
chance
that
I
did
something
wrong.
So
if
somebody
could
look
at
version,
three
version,
four
version
three
and
version
five.
B
D
B
Yeah,
it
just
looks
good
like
if
you
look
at
a
pseudo
code
or
something
I
got
something:
I
can
send
you
Robert,
but
you
just
I'll
I'll
email,
you
this.
What
I'm
looking
for
out
of
that
as
a
secondary
check?