►
From YouTube: IETF-TOOLS-20220510-1800
Description
TOOLS meeting session at IETF
2022/05/10 1800
https://datatracker.ietf.org/meeting//proceedings/
A
B
D
B
Now
the
joining
is
slowing
down,
so
I
think
we'll
go
ahead
and
start
everybody's
had
a
chance
to
see
the
agenda
for
a
couple
of
days,
I'll
ask
for
agenda
bashes,
but
remind
people
that
it
would
be
nice
to
have
the
agenda
bashes
in
the
notes.
Before
we
start
does
anybody
have
anything
that
they
want
to
add
here.
B
B
B
B
B
D
Well
then,
I
I
think
mark's
right.
I
think
we
should
call
it
or
something
whatever
and
that
then
provides
us
future
proofing
as
we
provide
in
other
formats,
because
we,
I
think
in
the
way
that
we
provide
these
shouldn't
just
be
thinking
about
ietf
authors.
Writing
ids.
We
should
also
consider
external
authors
writing
in
different
formats
that
are
because
we're
going
to
use
a
more
standard,
bibliography
format
than
viv
xml.
B
You
know
right
now:
the
data
tracker
serves
bibtex,
so
this
thing
might
serve
bib
tech
in
the
future.
So
casara,
let's
consider
in
fact,
let's
just
go
ahead
and
execute
on
running
the
service
as
bib.itf.org
we
can
cname
bibxml.itf.org
to
it
for
a
little
while
we
can
get
maddie
to
give
us
the
domain
name
well,
since
glenn's
out
for
another
another
day.
B
From
the
ietf
infrastructure
the
week
of
may
23rd,
I
did
find
this
morning
a
spot
in
ribose's
code
where
they
had
they
were
scraping
the
legacy
service
and
I'm
coordinating
with
them
to
make
sure
that
that
scrape
stops
so
that
we
don't
get
into
a
loop
when
the
legacy
service
was
replaced
with
with
our
own
and
depending
on
when
they
execute
on
that
it.
May
it
may
nudge
us
out
of
that
that
may
23rd
week,
but
it
won't.
It
won't
be
a
big
nudge.
B
B
We'll
still
have
all
those
channels
bridge
to
jabber
whoa
cindy
put
in
some
stuff
here
I
hadn't
seen
so
it's
just
the
getting
the
jabber
bridges
in
all
of
the
zoolop
streams
have
already
been
created.
Thanks
cindy
the
we
need
to
decide.
This
is
probably
a
conversation
that
should
be
taken
to
the
isg,
so
that
they're
not
surprised
whether
or
not
we
continue
to
keep
the
jabber
bridge
up
and
offer
any
jabber
services
at
all
post
114,
assuming
that
the
zulub
stuff
works.
B
B
B
B
G
B
Issues
we
had,
there
were
a
handful
of
issues.
The
largest
I
report
in
the
notes
was
a
issue
with
microsoft,
365
and
all
the
corporations
using
microsoft,
365
being
affected
by
our
new
ip
address,
not
being
on
their
list
of
addresses
to
accept
high
volume
mail
from
so
it
took
a
bit
of
a
scramble
took
three
or
four
days
to
resolve
to.
B
Any
questions
about
this-
oh,
I
did
mean
I
mentioned
in
the
notes.
You
can
look
at-
do
comparisons
against
now
versus
a
month
ago,
on
scout
apm,
you
can
see
that
the
data
tracker
is
running
about
three
times
as
fast
on
on
the
new
infrastructure.
It's
a
combination
of
the
new
infrastructure,
new
infrastructure
underneath
us
and
running
on
a
a
slightly
more
recent
version
of
mariah
db.
B
B
One
thing
in
the
conversation
that
I
I
just
pointed
out
to
the
isg
our
allow
list
has
many
addresses
in
it
at
this
point
that
I
think
we
should
purge
side
effect
of
adding
any
successfully
created.
Data
tracker
account
to
the
allow
list
in
our
super
low
barrier
of
entry
for
creating
data
tracker
accounts,
we've
had
conversations
among
this
group
in
the
past
about
tiers
of.
B
Reputation,
I
guess,
for
data
tracker
accounts
where
you
would
have
to
have
done
something
a
little
bit
more
than
having
created
a
data
tracker
account
before
it
would
cause
that
thing
to
land
in
the
global.
Allow
list
or
not.
These
are
our
questions
and
conversations
that
are
are
going
to
come
to
the
surface
again
over
the
next
few
weeks.
So
I
wanted
to
plant
the
seed,
so
people
could
be
thinking
about
it.
G
B
B
I
believe
we
add
all
active
ones,
but
remember
that
the
white
list
is
also
a
monotonically
increasing
thing
at
the
moment.
Nothing
comes
out
unless
we
explicitly
take
it
out.
So
if
it
was
ever
active
even
after
it's
been
marked
inactive,
it
doesn't
get
removed
from
that
from
the
list
as
it
currently
exists.
B
Yeah,
the
all
addresses
thing
is
coming
up
in
lots
of
contexts
right
now.
There's
a
bug
on
the
data
tracker
for
the
api
submission
interface,
where
the
user
presents
themselves
as
an
identity.
That's
not
there.
It's
not
the
same
as
the
identity
that
they
present
in
the
draft
it.
It
does
the
wrong
thing,
and
we
just
need
to
change
that
one
to
accept
and
to
to
realize
that
any
address
known
to
the
data
tracker
for
that
user
should
match
the
address
that's
presented
inside
the
draft.
B
B
B
My
first
question
is
to
the
group
mainly
to
jay.
At
one
point
we
were
expecting
okay,
so
the
edit
here
is
that
that
we
were
just
gonna,
have
one
big
wiki,
the
wiki.itf.org,
and
that's
still,
what
we're
planning
to
use
for
these
groups
to
to
migrate
their
things
to
does
that
exist?
Yet?
Is
it
yeah
or
is
it
an
older
instance?
Do
we
need
to
do.
D
It
it
exists.
We
were
just
going
to
wait
till
we
were
ready
to
probably
turn
on
the
integration,
wiki
js
and
data
tracker,
and
when
that's
done,
then
we
can
start
all
content
across
well
greg
and
I
need
to
talk
about
a
plan
for
it.
It's
going
to
be
a
mix
of
of
staff
and
secretariat,
doing
it
and
working
groups
doing
it.
So
we
just
need
to
sit
down
and
look
at
it
to
to
plan
that
we're
going
to
do
that
now.
G
So
if
there
is
a
wiki
part
of
it
for
working
group,
what's
the
situation
when
the
working
group
is
closed,
I
guess
the
document
do
not
disappear
right,
but
how
can
we
can
we
can
keep
the
access
running?
The
point
is
basically
next
week
in
the
retreat.
We
talk
about
living
documents
like
implementation
status
or
deployment
status.
B
Sure
so
the
permissions
model
that
we
are
building
into
the
end
wiki
js
uses
data
tracker
roles,
so
the
access
for
a
groups
pages
could
be
limited
to
the
chairs.
The
chairs
of
the
groups
are
the
ads
right.
The
group
gets
closed.
We
could
have
the
past
chairs
continue
to
have
access.
If
we
want
to,
this
is
something
we
could
set
up.
The
ads
would
always
have
access.
The
secretariat
would
always
have
access
and
any
editors
specific
editors
that
we
designate
at
any
given
point
in
time.
B
B
D
Just
add
a
couple
of
things
I
mean,
I
think
we
should
treat
this
as
an
ordinary
wiki,
which
means
that
we
don't
really
have.
We
don't
use
those
restricted
permissions
unless
we
have
to.
We
leave
the
whole
thing
open.
Anybody
can
edit
and
it
is
self-policed
so
that
you
know
if
you
notice
something
that
looks
like
it's
bad,
then
you
revert
back
a
previous
revision
and
discuss
it
with
the
person
in
the
same
way
that
we
do
on
wikipedia
somebody
put
something
you
know
strange
about
the
building
famous
building
down
the
road.
G
D
B
It's
only
been
a
month
since
we've
talked,
but
we've
crammed
four
or
five
months
of
progress
in
on
the
data
tracker.
Since
then,
we
had
the
deployment
of
version,
eight,
it's
our
first
deployment
from
github
the
had
several
deployments
since
then
we're
starting
to
work
on
some
pretty
major
infrastructure
changes,
the
velocity
that
we
had
hoped
we
would
start
to
experience
making
the
change
into
github
is
appearing.
B
B
It's
only
3.9
and
not
a
more
recent
version
of
python
than
that,
because
3.9
is
what
we
can
get
on
the
bare
os
at
the
moment,
but
we're
making
steps
eventually
towards
having
that
part
of
the
data
tracker
in
a
in
a
ultimately
in
a
in
a
container.
So
what
the
barrel
s
provides
will
will
no
longer
matter.
B
B
B
B
I'd
like
to
say
towards
the
end
of
the
week,
but
more
likely
into
the
beginning
of
next
week,
and
then
we
need
to
start
the
work
on
the
time
zone
where
timestamp
transitions
again
and
that's
a
really
really
big
piece
of
work.
It's
been
holding
for
the
server
transition
and
the
move
to
github
and
it's
time
for
us
to
go
ahead
and
bite
that
off
and
do
it.
That's
going
to
have
several
of
us
distracted
for
quite
some
time.
I
think,
from
a
programming
perspective.
This
thing
is
going
to
take
us.
B
It's
a
big
bite,
two
or
two
or
three
weeks
of
of
pretty
hard
work
and
we're
hoping
that
we
can
re-architect
the
way
that
that
it's
being
approached
from
what
we
had
planned
the
last
run.
We
made
it
this
so
that
we
can
avoid
the
four-hour
outage
that
the
last
planned
attempt
at
this
entailed
I'm
trying
to
see
if
we
can
find
a
way
that
we
can
actually
make
the.
B
B
B
Are
operating
it
as
a
web
service,
we
are
allowing
our
workers
to
run
for
a
much
longer
period
of
time
before
killing
them
the
number
of
of
requests
that
they
can
run
before.
We
we
tell
them
that
they
must
restart,
because
restart
is
expensive.
People
have
noticed
when
we
restart
the
data
tracker.
It
is
slow
on
that
very
first
request
that
hits
any
given
worker
as
python
brings
everything
into
memory
right.
So
we've
we've
moved
things
to
where
we
have
on
any
given
thread.
B
B
Jennifer's
started
digging
in
deeply
to
identify
these
places
and
can
start
working
with
it
where
it
loads
in
its
notion
of
what
today
is
as
it's
loading
and
and
just
writes
it
down
hard.
So
now
that
we
have
threads
now,
we
can
have
workers
that
run
for
days,
we're
having
things
like
people
submitting
a
draft
today
in
xml
that
doesn't
have
an
explicit
date
inside
of
it.
So
it
creates
the
text
version
with
the
date
that
it
woke
up,
as
which
was
two
days
ago.
B
We
have
another
draft
that
is
somebody's
having
trouble
with
that
excellent,
that
the
data
tracker
is
complaining
that
the
date's,
not
okay,
because
it
appears
to
be
in
the
future,
and
it's
not
so
it's
an-
and
this
is
actually
an
architectural
issue.
We
may
work
around
it
in
the
short
term
by
doing
a
daily
restart
of
the
data
tracker
and
crying
because
we're
having
to
do
so.
B
B
B
B
But
the
there
are
a
lot
of
drafts
out
there
right
now,
where
the
referenced
and
referenced
by
graphs
are
absolutely
incorrect.
We
need
to
fix
the
code
and
rerun
all
of
the
just
rebuild
the
reference
relationships
for
the
draft.
Since
we
started
parsing
the
xml
for
for
building
these
graphs,
it's
not
going
to
be
hard.
It's
not
going
to
be
a
big
deal,
but
but
we've
got
to
go:
do
it?
B
B
B
B
B
All
right
yeah
it
it
if
anybody
has
any
thoughts
around
this
after
they've
had
a
chance
to
go.
Think
about
it
again.
It
is
a
fairly
big
thing.
It's
likely
to
have
very
subtle
unintended
cons
that
the
opportunity
for
unintended
consequences
is
high,
so
do
spend
time
thinking
about
it
and
and
bring
stuff
to
the
list.
If,
if
you
have.
B
B
Yeah,
it
looks
like
he's
gonna
bounce
and
try
again
we'll
skip
ahead
and
when
he
gets
back
we'll
we'll
come
back
to
his
sections
on
the
common
bootstrap
thing.
No
changes
since
the
last
meeting
there's
been
some
discussion
of
the
new
data.
Tracker
coloration
mark
nottingham
opened
an
issue
that
was
titled,
it's
so
blue,
but
he
had
in
the
thread
some
legitimate
complaints
about
contrast
and
accessibility
that
we
might
want
to
take
a
look
at.
B
We
have
not
deployed
the
common
bootstrap
theme
that
springload
built
for
us
with
the
data
tracker,
yet
I'm
still
expecting
to
have
that
in
in
an
upcoming
release,
but
that
set
of
changes
is
not
going
to
address
several
of
the
comments
that
that
mark
made
in
the
ticket
that
we've
linked
to
here.
I
am
going
to
agree
that
we
would
close
that
ticket
and
just
take
it
as
input
into
a
bigger
conversation.
B
B
Have
spring
load
take
to
run
the
data
tracker
as
it
currently
exists,
or
after
we
apply
the
the
common
bootstrap
theme
to
it
through
the
test
tools
that
they
have
for
accessibility
and
just
give
us
a
report
on
the
findings
that
it
it
produces
good
idea,
yeah
all
right,
while
we're
waiting
for
casara
to
well
rain
for
casaro
to
return
I'll,
go
ahead
and
skip
over
xml
rfc
ryan.
I
believe
yeah
you're
here.
Do
you
want?
Do
you
want
to
briefly
talk
through
the
points
that
you've
added.
I
B
How
they're
interacting
and
the
registration
system,
I
think,
is
going
to
give
us
a
way
to
see
the
gap
for
how
many
participants
are
actually
interacting
with
the
registration
system
versus
how
many
of
them
are
letting
the
analytics
tool
gather
anything
whether
through
their
own
choice
or
because
of
browser
configuration
like
if
they've
got
something
like
ghostery,
because
ghostry
does
block
analytics.
I
noted
this
when
I
registered
that
my
my
browser
configuration
caused
the
matomo
javascript
to
not
load.
B
We
also
deployed
analytics,
I'm
really
skipping
ahead
to
the
web.
Analytics
section
of
this
thing.
We
deployed
this
with
the
data
tracker
as
well,
and
it
ran
for
a
while
before
I
noticed
that
the
configuration
at
the
data
tracker
would
cause
every
major
modern
browser
to
not
load
the
moto
mo
javascript,
because
I
did
not
include
analytics.itf.org
in
the
data
trackers
configuration
for.
B
G
B
Email
on
the
call
not
gone
and
tried
to
do
things
with
author
tools,
yet
I
mean
if,
if
you've
not
touched
it
and
if
you've
not
really
tried
to
make
it
do
something
useful.
Please
do
so
you
know
in
the
next
week
because
we
are
going
to
start
redirecting
the
xml
to
rc
tools.iatf.org,
and
you
know
the
slash
page
in
this
slash
experimental
page
towards
it
very
soon.
A
So
moving
to
xml
rfc,
there
were
a
couple
of
releases
since
the
last
call
and
now
authors
can
put
now
they
now
accept.
Rfc
renders
non-latin
characters
properly
on
on
the
rfc
front
and
there's
another
fix
coming
on
to
fix
a
fixer
defect
on
the
text,
format
for
references
with
non-electing
characters.
A
I
have
added
a
notes
page
about
what
data
create
about
xml
direct
settings-
things
non-letting,
letting
name
be
lighting
character
means
so
there's
a
bunch
of
unicode
code
points
there
on
the
link.
If
you
go
to
that
link,
you
can
have
a
look
at
that
and
we
might
try
to
come
back
to
that
in
the
field
sometime
in
the
future,
to
update
that
list
or
make
it
a
better
of
representing
what
letting
nailed
characters
are
on
upcoming
work
for
xml
to
rfc.
I
will
I
plan
to
go
through
the
current
issues.
A
B
B
Okay,
eric
you
have
an
item
here
on
yen
catalog,
where
you
mentioned
nothing
specific
that
you're
optimizing
the
montomo
analytics.
If
you
could-
and
I
know-
I'm
probably
catching
you
cold.
You
know
this
is
a
flat-footed
question.
I
apologize,
but
could
we
start
here
on
this
call
to
get
a
little
bit
more
insight
into
the
road
map
and
any
any
new
features
that
you
know
any
okay
milestones
in
the
roadmap
that
we've
that
we've
achieved?
You
know
we,
since
we
have
given.
B
G
Okay,
then,
the
next
big
step
is
basically
to
incorporate
the
sid,
not
the
service
exit,
but
more
or
as
less
in
the
young
catalogue,
and
we
got
already
a
couple
of
calls
from
this
with
carsten
and
the
other
people
in
young
catalog.
But
next
time
I
will
prepare
more
thing
on
this
stockton
and
that
to
be
done.
B
D
Like
to
say
how
impressed
I
am
at
the
level
of
progress
that
is
being
made
on
some
of
our
core
tours
of
core
tools
at
the
moment,
the
work
that
I
am
now
you
know
is
very
transparent
through
github
just
shows
a
remarkable
rate
of
progress,
and
it's
it's
very,
very
good
to
see
it's
good
to
see
the
many
people
in
the
community
contributing
to
all
of
the
github
issues.
It's
good
to
see
things
being
done
quickly.
B
B
H
B
H
B
H
A
common
one,
but
I
I
mean
we
did
talk
on
the
isg,
about
whether
we
want
to
define
our
own
flavor,
because
we
have
authors
that
are
using
markdown
at
the
moment,
they're
using
two
different
flavors
and
we're
also
using
markdown
for
the
wiki
and
we're
using
markdown.
We
want
to
use
more
market
on
the
data
tracker,
so
I
like
the
idea
of
trying
to
make
that
a
broader
effort
and
basically
define
you,
can't
call
it
a
common
marker.
You
can
call
it
like.
H
B
G
H
F
I
think
it's
important
to
notice
that
the
the
only
real
discontinuity
we
still
have
is
the
oauth
48
process,
and
you
saw
me
saying
oh
48
four
minutes
ago.
I
think
we
really
need
to
understand
how
to
make
the
the
auth48
process
working
better
and
doing
the
the
right
amount
of
github
and
whatever
is
one
part
of
it,
but
also
making
it
easier
for
people
who
also
in
in
markdown,
would
be
another
part
and
alex.
D
D
D
E
It's
going
to
say,
yeah
carson,
I
think,
published
the
xkcd
thing
about
standards.
Yeah
just
pick
the
one
that's
closest
to
us.
That
does
what
we
need.
It'd
be
great.
If
we
could
unify
the
two
that
we've
got,
it
might
be
a
few
years
out,
but
certainly
getting
unified
all
the
markdown
people.
You
know
step
two
will
be,
then
they
have
to
convert
all
their
existing
data
content
and
that's
not
gonna
ever
happen.
B
The
we're
going
to
have
markdown
support
for
buff
requests,
we'll
ultimately
go
and
have
these
things
for
charters,
we're
going
to
be
moving
into
places
where
we
have
we're
going
to
be
taking
advantage
of
things
that
markdown
extensions.
Even
that
haven't
been
available.
When
we
just
had
plain
text
objects,
they're
gonna
need
to
be
portable
cross
services.
B
B
D
I
agree
robert.
I
think
that
we
ought
to
be
looking
at
our
use
cases
and
solving
for
our
use
cases
and
you've.
Just
then
identified
some
more
so
you
know
we
need
to
make
sure
we
have
that
list
properly,
documented
and
understood,
but
I
don't
think
we
should
attempt
to
solve
beyond
that.
Initially,
that's
sufficient
for
us.
That's
hard
going
to
be
hard
enough
as
it
is.