►
From YouTube: IETF-UUIDREV-20220929-1800
Description
UUIDREV meeting session at IETF
2022/09/29 1800
https://datatracker.ietf.org/meeting//proceedings/
A
A
C
C
A
All
right
great!
Well,
let's,
let's
get
started
I,
think
I'm
Tim
Fenton
I'm.
One
of
the
co-chairs
of
the
group
Michael
Richardson,
who
my
turn
is
audio,
is
his
video
on
briefly
to
wave.
Although
it's
a
little
bit
fuzzy
and
dark
where
he
is,
is,
is
there
other
co-chair
and
so
welcome
to
the
first
meeting
in
the
first
virtual
meeting,
the
first
interim
meeting
of
uuid
Rev?
A
B
A
Okay
and
no
no
problem
in
any
case,
you
know
so
so
the
the
decision
was
made
to
not
only
extend
it,
but
given
the
fact
that
there
are
several
Errata
against
the
original
specification
that
it
was
probably
time
to
do
an
actual
revision
with
the
with
the
extensions
in
it.
So
that's
that's
the
direction
that
we're
going
in
and
what
we
were
chartered
to
do
we'll
have
the
so
as
as
his
tradition,
we'll
Michael.
Can
you
go
on
over
to
the
note
well
slide.
A
A
That
you're
presenting
they
need
that
needs
to
be
couched
in
certain
ways.
I
mean
we
do
have
ways
of
yeah.
That's
thank
you
for
highlighting
that
you.
You
must
disclose
that
that,
if
anything
is
is
covered
by
a
patent
and
and
we
have
associated
with
ietf
documents,
we
have
IPR
declarations
that
that
go
alongside
those
that
can
be
filed
if
appropriate.
A
A
And
so
the
the
agenda
is
fairly
simple,
basically
planning
for
for
the
next
ITF
meeting
in
London
in
a
few
weeks,
whether
we
want
to
do
virtual
interims
and
if
we
want
to
kind
of
go
over
any
any
issues
associated
with
with
adopting
the
you
know
what
what
needs
to
be
in
the
in
the
in
the
document
that
we
will
adopt
as
a
working
group
document
is.
Is
everyone
familiar
with
kind
of
the
this
notion
of
document
adoption.
A
The
the
the
concept
of
adoption
is
that
when
a
working
group
takes
on
a
document,
basically
the
the
original
authors-
seed
control
of
the
of
the
document
to
the
working
group,
so
it
goes
from
you
know,
draft
author
name,
something
something
to
draft
ietf
working
group
names
I
mean,
and
so
it's
it's
kind
of
a
a
joint
decision.
The
working
group
looks
at
the
Office
document
and
says
yes.
A
F
A
F
A
F
A
Notes-
and
so
there
isn't
a
whole
lot
there
yet,
but
really
it's
a
working
group
consensus
decision
to
to
use
GitHub
for
that,
and
you
know
I
put
out
a
a
call
for
consensus
and
basically
got
no
responses,
so
I'm
I'm,
hoping
that
that
people
are
happy
with
with
using
GitHub
for
that
are.
Is
there?
Are
there
any
objections?
Everybody
comfortable
in
that
direction?.
F
D
Okay,
okay,
great
and
it's
all
good
with
me,
sorry
I,
just
it's
all
good
with
me
too
I
just
didn't
know:
Kaiser
I,
had
you
know
this
this
previous
stuff,
these
issues
and
kind
of
discussions
that
was
on
a
different
GitHub
account
or
whatever
that
may
not
be
relevant
here
and
it's
fine
and
we
can
just
use
the
link
that
you
have
I
just
wanted
to
bring
up
that.
That
does
exist.
A
C
Mean
you're
welcome
to
migrate
all
those
issues
over
if,
if
they
have,
if
there's
value
there
that
keeps
a
history
and
that's
just
fine.
The
reason
for
this
particular
structure
of
things
is
that
it.
It
means
that
the
Secretariat
essentially
can
take
a
backup
in
the
event
that
GitHub
implodes
and
you
know
whatever
in
30
years.
Somebody
might
want
to
look
at
the
archive
and
they'll
have
them.
D
A
A
B
Yeah
and
just
from
my
perspective,
Brad
I
was
just
kind
of
freezing
that
in
time
and
just
leaving
it
be,
and
I
was
shifting
Focus
over
to
the
new
one
and
just
going
to
use
that
as
the
new
base
for
operations
for
all
the
topics
that
surround
converting
4122
into
maybe
a
new
table
contents
or
updating
and
modifying
erratas.
And
then,
when
the
time
comes,
I
have
to
merge
the
new
stuff
into
the
old
stuff.
Then
we
can
talk
about
how
to
do
that.
B
A
Okay
and
and
kind
of
an
anticipation
of
of
doing
this
and
kind
of
using
the
existing
4122
is
the
basis
for
the
new
document.
Michael
has
gotten
the
existing
text
converted
over
to
the
the
current
markdown
format
that
we're
using
for
creating
rfcs
and
so
that
exists
right
now
in
the
in
the
Repository.
B
B
C
C
B
Yep
yep,
because
the
only
thing
that's
been
that's
been
holding
me
back
from
starting
just
just
as
the
group
out
here
is:
hey
I
wanted
to
have
this
meeting
just
so
we
were
all
on
the
same
page.
We've
been
kind
of
holding
off
on
sitting
and
doing
anything
until
we
could
all
meet,
and
then
second
I
just
wanted
that
table
of
contents
from
the
old
document
and
the
RFC
4122,
both
of
them
in
cram
crammed
down
markdown,
so
that
I
can
then
start
kind
of
making
them
gel
and
once
I
have
that.
B
My
goal
is
to
really
put
my
head
down
at
next
week,
starting
Monday
and
attack
the
update
and
getting
a
draft
zero
ready
to
go
in
time
for
ietf
115s.
We
can
discuss
it
and
specifically,
my
goal
for
draft
zero
is
to
get
the
4122
text
into
the
new
better
table
of
contents.
So
we
have
a
good
idea
of
what
is
missing
and
then
for
new
text.
That
needs
to
be
added
for
some
of
these
things
that
folks
want,
and
then
we
can
also
attack
erratas
and
change
text
for
for
for
draft
01.
B
C
So
I
forwarded
the
message
it
was
sent
to
you.
Cisco
has
some
interesting
spam
filters
that
do
wrong
things
and
I
experience
this
regularly
and
Karsten
may
have
as
well.
So
you
want
to
check
your
inbox
and
see
if
you
did
get
something
from
me
and
if
not
well,
I'll
have
to
forward
it
through
Google
through.
B
F
B
You
so
planning
that
that
you
know
just
thinking
timelines,
just
talking
about
ietf
and
virtual
meetings,
I
think
our
deadline
or
our
goal
for
the
entire
document
was
March
2023
and
not
if
I
think
I
saw
was
the
the
group's
goal
for
getting
this
thing
to
a
point
where
it's
like.
Okay,
we
can
kind
of
see
if
there's
consensus
to
adopt
is
that
correct.
C
B
F
F
B
F
B
C
C
It
that
way
is
that
we're
going
to
submit
it
to
the
ISD
for
publication,
but
of
course
they
could
put
us
through
six
months
of
reviews
and
we
can't
control
that
it's
their
fault.
So
we
write
the
so
we
write
the
the
Milestone
as
the
thing
that
we
control,
rather
than
the
thing
that
they
is
out
of
our
control,
but
yeah.
B
Yeah,
so
if
I'm
looking
at
this,
then
November
5th
is
ietf.
115.
I
want
to
get
at
the
very
least
zero
zero
submitted
to
this
group
the
week
before.
So
we
can
at
least
discuss
it.
We
have
it
and
it's
in
people's
hands,
so
they
can
read
it.
So
we
come
to
the
meeting
with
a
nice
document.
We
come
out
of
ietf
115
with
some
goals
and
stuff
that
we
can
work
through.
So
it's
the
end
of
November
and
I,
like
monthly
releases.
B
Personally,
so
I'm
thinking
sure
we
you
know
at
the
end
of
each
month,
we
have
a
numbered
release.
So
October
is
zero.
Zero
November,
0-1
December
0-2,
and
we
get
all
the
way
up
to
March
and
if
we
don't
have
anything
to
change,
we'll
submit
it.
But
you
know
we
get
all
the
way
to
March
and
we've
got
a
good
between
now
and
then
what
four
versions,
possibly
that
we
could
iterate
on
really
major
versions
that
we've
been
editing
working
on
I
think
that
to.
C
There's
a
there's
a
blackout
period
for
the
two
weeks
before
the
ietf,
so
you
have
to
get
your
draft
in
by
Monday
October
24th
midnight,
UTC,
yep
and
and
then
you
can
and
then
the
next
opportunity
you
have
to
post
it.
Well,
it
used
to
be
the
Monday,
but
now
it's
become
Saturday,
so
Saturday
the
5th
of
November
becomes
the
next
opportunity
and
that's
the
idea
is
that
people
get
two
weeks
to
read
all
the
new
drafts
before
coming
to
the
meeting.
C
So
it
would
be
better
if
we
had
it
for
the
24th,
particularly
for
the
people
that
you
know
may
be
coming
to
our
in-person
meeting
and
and
may
not.
You
know
they
would
like
to
discuss
the
draft
that
were
they
would
like
to
read
the
draft
that
we're
actually
going
to
discuss.
So
that
would
be
better,
but
if
we
can't
make
that,
given
that
we're
a
new
working
group
and
we've
been
chartered
so
close
to
the
meeting,
I
think
that's
okay,
we've
posted
on
the
fifth.
B
I
think
it's
the
24th
is
is
fine
for
me,
I
can
I
got
it
written
down
and
I'm
95
sure
I'll
be
able
to
dedicate
the
time
to
get
0-0
by
the
24th
for
sure
that's
submitted
on
the
tool
up
there
and
act
active
it.
It
might
have
some
like
sections
that
say
like.
We
need
the
right
brand
new
text
for
this,
because
it
does
perfect
between
the
things
move,
but
we'll
at
least
have
something
we
can
come
together
and
then
maybe
hand
out
those
and
divvy
out
different
things.
Yeah.
C
B
C
F
C
B
Understood
and
I
I've
written
a
couple
books
and
you
can't
phone
it
in
on
the
table
of
contents,
because
if
you
don't
plan
that
out
perfectly
in
the
beginning,
you're
going
to
have
weird
things
that
you
can't
figure
out
where
to
put
in
later
so
yeah.
That's
right!
That's
that's
important
super
important,
Okay
cool,
so
I'm
I'm,
good
with
everything
I
wanted
to
kind
of
get
out
of
this
meeting.
So
I'll
turn
it
back
over.
You
guys,
sorry.
A
F
A
F
F
F
B
Just
anticipation
of
fixing
the
erratas
I'm
gonna
create
an
issue
tracker
on
on
the
on
the
working
groups
GitHub
for
each
Errata,
so
that
we
can
kind
of
just
start
discussing
what
will
need
to
be
changed
when
the
time
comes.
We
have
to
modify
text
great
idea.
That's
my
only
that's
my
only
thing
just
the
way
to
focus
discussions
on
those.
That's
that's
all.
C
You
know
things
just
fall
into
place:
bing
bing
bing
and
we
just
avoid
all
slide
where
it's
just
you
know
just
put
up
the
GitHub
issue,
tracker
click
on
it
and
you
know
you
know
if
you
get
to.
If
you
get
to
hit
the
merge
button
at
the
end
of
the
day
at
the
end
of
the
discussion,
then
you
that's
success
right.
B
Yep
and
that's
how
the
the
I've
been
look
following
the
quick
yeah
GitHub
trackers,
and
they
do
a
lot
on
you
know
they
do
a
pretty
much
everything
through
there,
so
they
have
pretty
good
results
with
it
as
well.
So
those
are
going
to
do
the
same
thing
and
we've
had
good
results
with
it.
On
the
previous
draft.
Just
getting
that
sort
of
understanding
in
the
beginning,
I
think
Brad
can
agree.
B
C
F
B
We
can
yeah
exactly
and
that's
that
brought
up
on
another
one.
We
can
always
edit
it
all
right.
There's
always
you
know,
there's
draft
four.
We
can
always
make
your
app
five.
We
can
always
go
back
and
make
draft
six.
I've
I've
seen
some
of
these
things.
Pretty
high
47
yeah
TLS
TLS
drafts.
They
go
pretty
pretty
long,
but
yep
I
agree.
C
Are
we
going
to
plan
a
series
of
virtual
interim
meetings
after
one
one
five
115
is
you
know,
second
week
of
November
I'd
like
to
think
that
we
have
time
to
have
one
more
before
Christmas
and
then
I
have
a
have
one
in
January
and
February?
That's
in
my
mind,
the
right
thing
does
this
time
slot
work
well
for
everyone.
C
D
F
C
F
C
F
C
Okay,
so
maybe
we
do
need
to
put
a
due
to
pull
up
and
and
what
I
would
what
I
would
love
to
have
is
a
view
as
to
which
are
good
days
and
bad
days
and
then
which
are
good
time
slots
and
bad
time
slots
within.
You
know,
basically
this
hour
and
the
two
or
three
hours
beforehand,
and
then
you
know
we'll
pick
a
a
time
and
we'll
probably
just
probably
just
try
to
stick
stick
with
that.
Put
the
virtual
interims
into
the
data.
B
B
F
C
C
F
C
A
E
I
I
have
done
a
review
on
of
the
uuid
4
format.
Document
I,
don't
know
if
this
is
the
the
right
venue
to
bring
that
up.
I'm
happy
to
post
to
the
list
as
well
and
I
I
drew
out
a
couple
of
issues
that
I
see
in
the
draft
as
it
stands,
not
the
original,
the
best
document
to
be
clear,
but
the
one
that
introduces
version
six,
seven
and
eight.
A
E
What
used
to
be
arms,
iot,
Services
Group,
where
individual
devices
had
their
own
uuids
and
they
generated
you
know,
events
with
uuids
and,
of
course,
those
those
uuids
weren't
rfc4122
uuids.
They
were
these
crazy
things
that
somebody
made
up
on
the
back
of
a
paper,
napkin
and,
and
so
I
I've
been
working
in
and
around
this
space,
a
bunch
and
so
I
care
a
lot
about
it,
which
is
strange
to
a
lot
of
people
that
I
care
about
unique
identifiers.
But
there
you
go
I,
guess
that's!
E
Why
we're
all
here
yeah!
So
so!
That's!
That's
sort
of
my
my
background
here
and
I
I
as
I
think
I
mentioned.
I
do
have
a
fair
bit
of
a
security
hat,
and,
and
so
some
of
these,
these
the
comments
that
I
have
are
going
to
end
up
veering
into
that
direction
a
bit
so
with
that
intro
shall
I
just
lead
into
it.
E
E
From
that
perspective,
the
problem
that
we
have
going
forward
is
we're
explicitly
removing
the
Fiat
and
we're
replacing
it
with
Randomness
and
what
was
not
clear
to
me
in
the
draft
when
I
read
it
through
the
first
time
is
the
lifetime
of
those
random
numbers,
because
if
they're
randomly
generated
for
your
device
once
and
stored,
then
I
fear.
We
run
into
the
birthday
problem
awfully
quickly
with
trillions
of
devices
and
48
bits.
E
Yeah
or
are
we
generating
new
random
for
every
monotonic
clock
increase,
although
for
uuid
6,
that's
scary,
because
you're
talking
about
100
nanosecond
intervals,
which
means
that
the
data
rate
out
of
your
random
number
source
is
going
to
be
measured
in?
Is
that
gigabits?
It's
going
to
be
pretty
fast
anyway,.
D
I
I
was
just
I,
was
just
gonna
kind
of
roll
that
into
I.
I,
definitely
think,
there's
a
lot
to
unpack
there.
I
have
a
bunch
of
things,
I
could
say,
but
I
think
it's
going
to
be
better
to
do
that
in
writing
in
a
GitHub
issue,
where
we
can
kind
of
take
each
each
thing
and
say:
okay,
good.
This
is
the
analysis
of
that,
because
I
mean
it's
really
good.
That's
a
really
good
points.
D
It
is
only
as
unique
as
that
system,
so
I
think
it's
I
think
it's
kind
of
like
it's
more
like
we
should
list
all
those
things
out
and
in
the
issue
and
then
kind
of
decide
what
wording
represents,
because
from
my
perspective,
I
think
there
you
get
into
this
application
specific
thing
where
some
people
might
have
an
application.
They're
like
look.
This
doesn't
work
for
me
because
I
need
the
guaranteedness,
so
I
want
the
Fiat
system
and
other
people
might
be
like
well
I.
D
That's
great,
that
you
have
that,
but
for
my
application,
that
Fiat
system
does
doesn't
physically
even
work,
so
I'd
rather
have
the
probability
and
I.
Don't
think
that
is
a
right
or
wrong.
There
I
think
it's
a
matter
of
just.
What's
the
wording,
that's
gonna,
you
know
make
each
of
these
systems
able
to
coexist.
So
that's
what
that's
kind
of
how.
E
What
does
this
stuff
all
mean
and
that
gets
that
gets
a
little
bit
concerning
right,
because
if
you
have
two
people,
you're
committing
entries
to
the
same
database
and
one
of
them
is
using
a
Fiat
system
and
the
other
is
using
a
random
number
you're
in
for
a
headache,
so
there's
an
argument
to
be
made
that
these
actually
have
to
be
different
revisions,
because
otherwise
you're
going
to
get
collisions.
There's
no
there's!
No,
if
about
it,
the
the
other
side
of
it.
B
F
B
B
E
B
Yes,
okay
and
then
second
piece
to
this
is
obviously
when
we
talk
about
internet
standards
and
everybody
being
compatible
and
and
things
that
makes
a
lot
of
sense
when
we're
talking
about
protocols
and
Communications
and
streams
and
people
are
talking
to
each
other.
But
are
we
if
somebody
doesn't
know
how
uuid
Works
between
two
different
systems?
How
often
are
we
sending
uuids
amongst
systems?
A
E
F
E
E
It's
kind
of
half
hidden
from
you,
that's
okay,
but
I
mean
it
should
never
show
up
in
a
user
interface
and
I've
seen
it
happen,
yeah
so
that
that
was
another
one
of
them.
My
comments
about
the
lifetime
of
of
uuid
six
random
data
also
applies
to
uuid
7
from
what
I
understood
of
the
of
the
handling
rollover
section
and
the
monotonicity
section
it
sounded
to
me
like
the
unique
like
the
random
data,
was
generated
once
per
timestamp
and
again
there's
this
question
of.
E
Can
you
handle
74
bits
of
random
data
every
millisecond?
Do
you
have
a
a
random
number
generator
that
is
capable
of
doing
that
and
I
guess
that
depends
on
whether
you're
using
a
prng
or
whether
you're
using
a
trng,
but
bear
in
mind
that
many
people,
when
they
see
this,
will
just
use
whatever
is
exposed
by
the
OS?
And
if
that
means,
you're
exhausting
the
random
number
generator
in
your
processor.
That
becomes
a
problem.
B
Uuid
generation
is,
is
very,
it
was
a
very
large
topic
right.
Let's
say
we
have
to
generate
a
thousand
uids.
How
do
we
make
sure
they're
unique
and
that
they're
in
a
very
specific
order
laid
down
in
the
database,
because
the
timestamps
are
there
and
there's
a
lot
of
different
ways
that
came
about
about
one
was
create
one
random
and
then
increment
it.
B
E
Neither
of
those
deal
with
the
issue
of
exhausting
your
trng's
entropy
right.
So
the
point
here
is
that
there's
a
physical
limitation
in
true
random
number
generators.
They
have
a
specific
amount
of
Randomness.
They
can
generate
in
a
given
period
of
time
and
to
give
you
some
idea
about
how
critical
this
actually
is.
There
are
proofs
of
concept
out
there
for
creating
a
covert
Channel
based
on
random
number
generator
exhaustion,
where
you
communicate
between
Pros
two
processes
which
aren't
supposed
to
know
about
each
other.
E
By
measuring
how
long
it
takes
to
create
a
random
number
on
one
of
them
and
generating
lots
of
random
number
land.
Sorry
lots
of
random
numbers
on
the
other
to
send
the
data,
and
then
you
can
modulate
data
through
the
timing
channel
that
you
create
there
okay.
So
this
is
a
big
deal
and
if
we
are
using
random
numbers
and
we're
using
them
at
a
high
data
rate
and
people
don't
use
prngs
if
they
actually
end
up
using
the
true
random
number
generator
from
their
core.
They
are.
You
are
constructing
a
performance
problem.
E
E
Yeah
yeah
I'm,
getting
there
I've
got
a
couple
more.
The
the
counter,
the
discussion
of
counter
rollover
and
specifically
with
this
centralized
registry
that
terrifies
me.
The
idea
of
handling
batched
uuids
with
a
centralized
registry
is
just
that's
horrifying
right
because
that
just
create-
and
you
mentioned
that
this
creates
a
single
point
of
failure,
but
I
I
I'm.
E
E
F
B
I
just
want
to
say
that
that
section
wasn't
I
wasn't
trying
to
endorse.
Centralized
registry
I
was
saying
that
these
are
two
ways
you
can
do
this
and
this
one's,
not
the
one
that
we're
talking
about
and
if
outside
the
scope
of
this
document.
So
if
it
reads
as
if
we're
endorsing
it,
I
can
I
can
totally
change
that
verbiage,
because.
B
That
was
me
that
was
me,
saying
hey.
These
are
the
two
prevailing
ways
of
tackling
this
and
handling
it,
and
we're
not
talking
about
this
one
we're
talk,
we're
going
to
go
on
and
talk
about
the
next
one,
so
I
again,
I
can
add
verbage
around
this
saying
No.
This
is
the
bad
one.
So
let
me
know
if
it
just
reads.
E
Me
there
was
one
more
comment:
I
have
already
mentioned
the
the
question
about
uuids
and
human
readability,
so
there
was
one
other
thing,
actually
I'd
kind
of
like
some
text
on
that
added
to
the
bis
you
know
do
not
present
these
isn't
in
a
user
interface
they're
not
for
humans
seems
like
an
important
note,
and
there
is
a
comment
on
that
somewhere
about
288,
bytes
versus
128
bits
or
something
like
that.
I
I
probably
have
the
numbers
wrong,
but
I
remember
seeing
something
to
that
effect
and
yes,
absolutely.
D
E
D
Well,
I
agree
with
you
on
a
style
from,
but
it's
it's
you
know
probably
not
a
great
idea
for
a
lot
of
applications,
but
why
is
that
important
to
specify
like
why
is
it
I
like
I,
understand,
there's
people
might
do
it
without
thinking
about
it,
but
is
it
really
that
terrible
to
have
a
uuid
entered
by
a
user?
If
there's
a
case
that
warrants
it,
it.
E
Absolutely
is
a
terrible
way
to
do
things
because
they'll
type
it
wrong
if
they
ever
have
to
type
it
they'll
type
it
wrong.
If
they
have
to
remember
it
they'll.
Remember
it
wrong,
you
know
it's,
it's
just
not
something!
That's
built
for
a
human
and
sorry.
I
I
have
a
a
big
usability
chip
on
my
shoulder,
and
so
when
I
encounter
things
like
that,
it
really
it.
It
really
bothers
me
and
I
I'm,
a
big
believer
in
usable
security,
and
when
these
things
are
used
in
an
interface,
they
need
to
be
hidden.
E
A
A
At
the
same
time,
there
are
situations
where
you
know
maybe
an
account
recovery
or
something
like
that.
You
need
to
have
a
complex,
randomly
generated,
Secret
that
the
user
might
have
to
type
in
occasionally
in
order
to
recover
their
account
and
whether
it's
in
this
format
or
whether
it's
as
base64
I,
mean
there
just
isn't
a
good
way
to
do
that.
I
guess
maybe.
E
E
A
C
D
Yeah
I
mean
I
I,
yeah
I
really
would
like
the
the
opportunity
to
be
able
to
go
back
and
forth
this
morning,
really
bat
down
and
get
down
to
the
core.
I
I
agree
with
this
suggestion.
The
recommendation
I
just
don't
like
I
literally
I,
have
an
iot
device
that
is
used
as
a
uuid
as
its
identifier,
and
there
are
actually
a
terrible
person
has
to
type
it
in
it's
all
certain
it.
No
it's
it's
Unique,
that's
the
point,
and
somebody
has
to
type
it
in.
In
certain
cases,
I'm.
F
D
F
E
D
E
A
E
D
E
C
F
C
C
C
So
these
are
all
template
because
we
could
talk
about
this
one-on-one
if
you
like,
but
basically
I
pulled
in
this
Mark
Martin
Thompson
template
that
basically
does
everything
it
sings
it
dances
and
it
makes
you
breakfast
so
yeah.
Sometimes
it's
too
much,
but
it
does
work
when
it
works.
When
it
doesn't,
you
can't
debug
it
a
loss
but
basically
type
make
even
on
the
windows
for
Linux
machine
and
it
downloads
everything
you
need
and
compiles.
The
Spitzer
HTML
yeah.