►
From YouTube: IETF-TOOLS-20230214-1900
Description
TOOLS meeting session at IETF
2023/02/14 1900
https://datatracker.ietf.org/meeting//proceedings/
A
C
A
B
D
Weren't
we
supposed
to
all
have
like
some
kind
of
Sun
Skylight
fiber
optic
skylights,
that
connect
our
basements
to
our
roofs
or
something
wasn't
that
supposed
to
be
a
thing
in
the
80s.
D
E
B
Not
not
really
trying
to
shut
it
down,
we
just
need
to
go
ahead
and
jump
in
so
doesn't
look
like
Jay
has
completely
joined,
yet
maybe
he
has
so.
The
notes
have
been
up
for
a
few
days.
B
The
outline
of
the
notes
set
the
agenda.
Does
anybody
have
any
agenda
bashes
before
we
start?
B
Thank
you
as
we
get
into
the
meeting.
If
you're
capable
of
sharing
video,
please
do
so
having
it
around
the
table.
Experience
is
our
goal.
The
do
remember,
though,
that
this
is
being
recorded
and
will
be
made
available
on
YouTube
at
the
end
of
the
meeting.
B
B
So
there's
been
a
big
push
on
clearing
bugs
that,
in
particular
affect
scheduling
and
then
affect
the
use
of
the
agenda
during
the
meeting.
So
we
think
that
we've
cleared
the
the
really
rough
parts
of
that
out
of
the
way
that
I
would
like
to
get
everyone
here
to
think
about
whether
or
not
there
is
something
that
really
got
in
the
road.
At
the
last
meeting
that
they
haven't
seen
evidence
that
we've
cleared
away
for
the
coming
one.
B
D
So
I'm
curious
to
know
what
percentage
will
come
in
with
the
new
versus
the
old
and
can
we
make
that?
Can
we
make
that
reduce
that
to
zero
for
the
old
for
the
for
117?
Once
we
know
what
the
trend
is,
I
don't
know
if
we
have
any
Trend
right
now,
data
right
now.
B
Yeah,
realistically,
the
users
of
the
apis,
the
you're
you're
a
pretty
rare
case
using
it
with
scripts
that
you
wrote
yourself
the
rest
of
it
is.
A
B
Heavily
dominated
by
the
Mt
right.
D
D
This
may
be
a
new
experience
for
some
people
that
they
have
the
asynchronous
stuff
and
because
some
people
only
submit
the
draft.
You
know
the
hour
before
the
deadline.
That's
the
only
time
they
do
it,
but
at
least
I
think
I
feel
like
we're
gonna.
It's
gonna
be
much
better
experience
for
overall
No
One's,
Gonna
Get
a
timeout
and
be
confused.
A
B
I'm
not
seeing
people
rush
to
the
microphone
for
hey.
What
about
this
issue
that
we
had
so
do
do
rock
your
brains.
If
you
can
think
of
anything
that
we
need
to
address
before
we
get
to
the
meeting,
it
would
be
better
to
know
about
it.
Early
all
right,
we'll
move
on.
B
We
still
are
planning
to
move
DNS
into
cloudflare.
We've
run
into
several
things
that
are
taking
longer
than
we
had
originally
anticipated.
They
would
one
of
the
ones
that
we're
working
through
right
now
are
the
consequences
of
who
our
current
registers
registrars
are.
We
won't
go
into
details
on
that
on
this
call.
B
Michael
asked
a
question
in
the
notes,
in
response
to
the
reminder
that
there
will
be
a
short
period
where
the
zone
is
not
signed.
Yes,
we
do
mean
that
the
DS
records
will
be
withdrawn.
B
Some
time
will
be
allowed
to
pass
to
let
caches
all
catch
up
than
a
new
DS
record
would
be
inserted
later
and
the
the
chain
would
be
built
on
that
doing.
A
proper
in
place
signed
rollover
is
really
utterly
impractical.
The
combination
of
what
the
registrars
provide
what
the
cloud
service
providers
provide.
B
We
you
know
we
would
run
into
trouble,
trying
to
do
that
at
cloudflare
because
of
some
policies
that
they
have.
That
would
not
let
our
old
key
in
nicely
in
upgrading.
You
know
row
track
and
roll
our
keys
in
place
before
we
went
to
the
move,
has
different
problems
and
really
we've
we've
mined
this
pretty
hard
and
the
path
that
makes
the
most
sense
to
go
through
is
the
one
that
cloudflare
is
recommending,
which
is
to
just
go
through
a
short
period
where
we're
not
signed.
D
So
it
would
be
interesting
to
have
someone
debrief
you
guys
as
to
what
your
hassle
was,
because
it
sounds
like
this
would
be
actually
a
good
BCP,
yeah
yeah,
maybe
over
beer.
B
B
B
B
B
We've
got
just
a
few
left
on
wrestling
right
now
with
the
Django
user
username,
so
that
we
don't
end
up
with
the
potential
for
collisions
in
the
username
just
because
of
case
folding
life's
going
to
be
a
little
bit
better
for
us
under
postgres,
in
that
the
collation
that
we
have
been
using
under
my
sequel
to
date
did
some
rather
questionable
things
where
it
would
fold
accented
characters
back
into
unaccented
characters
for
equivalence
purposes.
B
E
B
Database,
we're
going
to
trial
a
migration
of
the
full
production
database
on
a
machine
that
is
the
equivalent
of
the
full
production
machine,
will
basically
be
hijacking
our
hot
standby
and
trying
that
work
there.
That
will
inform
us
give
us
a
better
sense
of
what
the
actual
downtime
we're
going
to
need
for
the
true
production
cut
over.
B
So
we
should
be
able
to
report
on
that
before
and
at
the
next
tools
call.
So
while
we
were
down
in
this
Discovery
period
focusing
on
collation,
we
noticed
that
the
postgres
install
on
our
main
production
machine
is
very
old,
was
probably
originally
installed.
A
B
Utf-8
was
the
ubiquitous
answer
for
what
your
default
character
set
and
galatian
rules
should
be
based
around
to
address
that
we're
planning
over
the
next
again
couple
of
weeks
to
rebuild
the
postgres
cluster
on
ietfa,
using
the
more
modern
sets
of
of
defaults.
B
There's
only
one
production
database
in
postgres
on
the
production
machine
right
now.
That's
the
database
that
backs
the
wagtail
portion
of
the
website.
That
database
is
fairly
small
and
updates
to
it
are
at
a
pace.
That's
under
our
very
much
under
our
control,
so
this
rebuild
of
the
cluster.
Doesn't
we
don't
anticipate
that
it
will
truly
inconvenience
anyone
and
will
be
in
a
better
place
once
that's
made.
B
Because
Sarah
has
been
pushing
pretty
hard
on
bringing
ID
diff
into
parity
with
RSC
def
to
get
the
last
pieces
where
ID
diff
was
not
doing
the
same
things.
That
rscdef
was
and
discovered
that
the
amount
of
effort
to
bridge
that
last
bit
of
Gap
was
probably
not
worth
the
investment
compared
to
switching
directions
and
moving
back
to
continuing
to
support
RSC
diff
itself.
B
What
we
had
been
trying
to
do
was
to
move
to
a
inside
a
programming
single
programming
language
framework
in
Python
was
where
ID
def
is
so
that
we
had
self-contained
tests
that
we
had
built
around
it
as
opposed
to
what
we
have
with
our
C
diff,
which
is
a
shell
script.
That's
calling
out
to
you
several
other
languages
like
awk,
but
again
the
lift
to
get
that
last
little
bit
tells
us
that
we
should
just
continue
to
focus
on
the
RSC
death
instead.
D
Does
that
mean
that
RFC
diff
will
become
the
default
again?
That
is
our
plan
are.
Are
the
I
don't
experience
changes
that
bother
me
I
know
there
are
a
few
who
are
quite
vocal
about
the
changes
and
I
a
little
bit
wonder
whether
a
button
on
the
ID
diff
output.
That
said,
give
me
RFC,
diff
and
I.
Don't
have
to
enter
all
the
details
again.
D
Would
it
satisfy
what
I
suspect
are
small
minority
of
people
that
notice
the
difference
and
I
I
I
I
understand
that
RFC
diff
has
a
significantly
higher
CPU
load
and
is
much
much
less
flexible
in
terms
of
inputs,
so
I'm
surprised
about
this
direction,
but
I
understand
what
you're
saying.
B
Yeah,
so
the
biggest
problem
that
we
had
with
rfcdif
if
we
back
up
and
look
at
it
from
the
the
highest
level,
is
that
our
web
service
implementation
of
it
used
a
proprietary
attempt
to
build
a
python-based
rapid
deployment
thing
that
Henrik
had
written
this
pyhd.
B
The
security
issues
that
we
had
are
all
wrapped
up
in
the
pyhd
code.
The
deployment
of
RSC
desk
that
we
have
at
authortools
right
now
does
not
suffer
from
those
issues.
So,
however,
we
proceed
we're
going
to
deprecate.
The
pyht
wrapper
as
currently
exposed
under
our
GitHub
organization.
B
Put
appropriate
red
tape
around
it.
You
know
the
the
warning
tape
and
then
eventually
just
archive
that
that
particular
Repository.
B
A
And
an
RFC
div
on
another
tools:
RC
rfcd
will
be
a
drop
in
replacement,
so
you
will
not
lose
any
features
that
you
currently
enjoy
with
idd
like
providing
a
different
format
or
anything
like
that.
Those
will
be
in
place.
B
So,
let's
kick
around
because
our
Michael's
suggestion
for
at
least
while
we
have
both
of
them
in
place
a
really
easy
way
to
bop
back
and
forth
from
the
UI.
B
And
then
Michael
think
a
little
bit
more
about
the
the
issues
that
we're
bringing
up
here
about
long-term
maintenance
and
see,
if
you're,
if
you
think
that
we're
doing
the
right
thing
it
that
we
would
be
doing
the
right
thing
by
moving
to
just
focusing
on
RSC
diff
going
forward.
If.
D
If
you
say
that
it's
like
this
is
the
you
know
short
term
next
year,
yeah
we're
just
going
to
defer
some
of
this,
then
I
think
that's
great
I.
I.
Think
the
part
of
the
issue
going
forward
is
that
the
number
of
people
that
have
the
awk
plus
shell
plus
this
plus
that
knowledge,
deep
knowledge
is,
is
very
rare.
It's
already,
seemingly
rare,
and
you
know
awkward,
just
aqua's
terrible
25
years
ago,
right,
yeah
Pearl.
D
For
that
reason-
and
you
know,
we've
literally
had
an
entire
generation
born
that
since
who
probably
you
know
so,
that's
just
you
know
even
just
getting
contributions
from
people.
You
know
it's.
A
real
I
would
think
it's
a
real
pain
to
do
things.
If
someone
wants
to
set
it
up
themselves
and
have
it
locally
wow,
that's
probably
a
big
disaster
for
most
people,
but
anyway,
I
I'm,
not
I've,
I've
read
the
complaints
on
the
tools
list
and
I
was
like
yeah,
not
really
that
important
to
you.
D
You
can't
cope,
but
you
know
change
is
hard
for
some
people,
so.
B
Okay,
it's
very
been
very
good
input.
Thank
you,
I
think
we'll
we'll
take
and
look
at
it
again
for
the
the
long.
B
There
was
a
flurry
of
excitement
suggestion
that
we
set
something
up
to
give
it
a
try.
B
Nick
is
already
given
us
a
proof
of
concept
that
we
can
run
an
instance
of
Mastodon
that
is
fairly
well
controlled,
for
who
would
have
access
to
it?
Who
could
post
we're
wondering
if
the
right
thing
to
do
is
to
move
forward
is
to
try
running
with
this
site
such
that?
The
only
thing
that
can
post
to
it
is
either
automated
spots
in
the
data
tracker,
similar
to
the
places
that
people
had
been
feeding
off
of
and
see
sending
messages.
B
Twitter,
possibly
having
a
queue
similar
to
ietf,
announced
that
Gregory
team
could
moderate
so
that
things
that
were
outside
the
bounds
of
what
might
come
out
of
an
automated
processing
chain
inside
the
data
tracker
could
be
sent
to
this
instance.
B
They
have
them
all
be
of
an
announced
nature,
so
thoughts.
Anybody
think
that
this
is
a
good
thing
to
do.
Is
this
a
waste
of
our
time?
What's
our?
What
would
people
like
to
see
happen.
F
Okay,
there
we
go
too
many
too
many
different
uis
and
conferencing
systems,
so
I'd
be
in
favor
of
doing
some
announce
only
thing
where
it
was
actually
tied.
F
F
The
way
we
had
tried
with
Jabber,
perhaps
but
yeah
I,
it's
it's
fun,
it's
the
new
shiny
thing,
but
there
are
a
number
of
groups
who
are
using
Mark,
Nottingham's
or
other
tools
to
post
announcements,
including
HTTP
API,
because
we
have
a
lot
of
non-itf
contributors.
So
yeah
I
would
definitely
put
this
on
the
list
as
something
automated
announcements
as
something
we
should
do
and
including
that
things
like.
Oh
there's,
a
new
ietf,
blog
post,
there's
a
new.
You
know,
announcements
of
upcoming
meetings
and
things
like
that.
E
If
the
effort
is
low,
it
feels
like
a
worthwhile
experiment,
but
I
think
it
would
be
good
to
pose
it
as
an
ex
so
that
if
it
turns
out
that
people
don't
actually
use
it,
that
we
don't
get
stuck
maintaining
another
server
and
service
that
no
one
ends
up
using
actually
so
that
that
would
be
my
suggestion.
B
D
D
We've
already
reasonable
activity
right,
even
if
the
Mastodon
turns
out
to
be
disaster
and
no
one
wants
to
go
there,
then
we
would
still
have
that
outgoing
those,
outgoing,
Hooks
and
I
think
that's
a
interesting,
useful
thing
to
be
able
to
do
something
with.
B
B
B
B
Next
topic,
I
had
I,
don't
think
is
going
to
take
a
lot,
but
maybe
I'll
be
surprised.
We've
got
a
bunch
of
names
of
lists
that
we
direct
people
to
send
bug
reports
to
that
are
just
no
longer
self-descriptive
Django
project
data
tracker
projects
a
little
bit
better.
Django
project,
though,
doesn't
really
mean
anything
to
people
that
aren't
like
deeply
on
the
inside
I
would
like
to
do
a
cleanup
around
where
we
scrape
all
of
these
things
away
with
redirects
until
enough
time
has
gone
by
that
it
doesn't
matter.
B
But
what
we
start
advertising
is
just
simply
tools
help
at
so
we
could
bike
shed
the
name.
The
tools
team
spend
a
little
bit
of
time.
Kicking
around
names
and
tools.
Help
is
what
we
landed
on
as
the
thing
that
is
most
likely
to
cause
people
to
know
when
to
use
that
versus
just
normal
support
at
ietf.org.
G
Sure
so
this
item
is
just
to
kind
of
communicate
that
we're
doing
some
work
on
moving
some
Legacy
reports.
So
there's
some
Legacy
progress
reports
which
are
monthly
reports
that
live
at
the
address,
and
the
document
here
that
contain
information
about
draft
actions,
group
actions
and
some
mailing
list
information,
number
of
posts,
and
subscribers
and
these
reports
were
produced
with
some
rather
old
Pearl
and
Python
scripts.
G
That
would
periodically
fail
when
the
underlying
systems
were
updated.
So
there's
work
being
done
to
move
the
reports
into
the
appropriate
systems,
data,
tracker
and
mail
archive
where
they
can
be
easier
maintained,
so
that'll
be
done,
and
then
this
page
will
be
updated
to
link
to
the
new
locations
for
the
for
the
monthly
data.
B
Yeah,
okay
into
the
FYI
section.
B
Some
of
these
we
should
probably
have
promoted
up
into
the
hot
topics,
but
so
we'll
go
through
them
a
little
more
slowly
than
we
might
otherwise
have
the
infrastructure
strategy
topic
the
the
RFI
has
been
released.
It
closes
at.
E
B
The
end
of
the
month,
hopefully
everybody's
aware
of
it,
if
anybody
knows
of
somebody
that
they
think
should
be
sending
us
an
opinion
on
what
we
should
be
doing,
make
sure
that
they're
aware
of
it,
so
that
we
have
the
best
input.
We
have
that
we
can
get
available
to
us
at
the
end
of
the
RFI
and
we
start
moving
into
the
RFP
phase.
B
For
our
Wiki
migrations
from
track
into
Wiki,
JS
Greg
has
reported
that
we
are
nearing
the
end
of
the
content
migration
effort.
Jay
calls
out.
Victor
trust
is
raised,
a
question
about
how
we're
licensing
things
and
tracking
attributions
and
problems
of
that
nature.
B
B
For
the
data
tracker,
the
report
form
is
the
same
as
what
we've
had
before
quite
a
bit
of
activity.
B
Since
our
January
meeting
with
the
the
five
releases
focusing
on
the
the
three
points
that
are
listed
in
in
the
message
there,
a
few
extra
other
bug
fixes,
in
addition
to
those
with
those
really
framed
up
the
the
larger
parts
of
the
efforts
that
went
in
we're,
going
to
finish
out
the
media,
impacting
issues
and
then
spend
our
time
between
now
and
the
next
meeting
focusing
hard
on
finishing
the
postgres
work
setting
it
down
then
queuing
up
the
move
to
Django
4,
which
we
expect
to
happen
right
after
postgres.
B
Michael
already
brought
up
that
we've
got
the
API
draft
submission
going
through
the
asynchronous
tooling
that
we
built
we're
hoping
to
add
the
basically
update
the
the
GUI
submission
endpoint
to
use
that
async
back
end
as
well
and
hand
hold
the
people
through
the
submission
experience
change
so
that
there's
as
little
confusion
as
possible.
This
is
something
that
we're
hoping
to
have
deployed
before
we
get
to
the
heavy
crunch
of
submission
for
116.
B
Casara
lets
you
call
out
anything
from
the
next
five
blocks
that
you
want
to
bring
up
in
real
time.
C
So
I
added
a
note
issue:
339
for
bib
XML
Services
is
blocking
for
the
RPC
for
the
publication
of
rfcs.
It
looks
like
a
trivial
issue
and
for
for
the
bridge
here
it
the
target
URL
for
references
points
to
RFC
versus
info,
and
so
it
doesn't
point
to
the
info
page
and
for
us
to
redo.
All
the
references
to
get
something
published
would
take
a
whole
lot
of
time,
so
it
would
be
wonderful
if
that
could
be
fixed
quickly.
B
So
if
you
follow
the
there's
already
a
PR
against
that
issue,
I
think
it's
scoped
me
a
little
bit
a
little
bit
too
large
I've
got
a
question
back
out
to
ribos
about
whether
or
not
we
can
scope
the
change
initially
to
the
bib
XML
RFC
data
set
where
we
know
what
Target
should
be
we're
fairly
comfortable
with
what
Target
we
have
an
answer
for
the
short
term,
for
what
Target
should
point
to
for
some
of
the
other
data
sets.
B
It's
not
quite
so
obvious,
yet
and
I
suspect
that
the
Global
change
would
probably
bring
in
Target
values.
That
pointed
to
the
wrong
thing,
I
know
that
when
we
asked
them
to
take
the
target
attribute
away
earlier,
they
were,
for
instance,
pointing
to
the
archive
text
for
internet
drafts
instead
of
data
tracker,
so
so
yeah
we're
we're
going
to
continue
to
push
getting
this
resolved
as
quickly
as
we
can.
B
B
Ryan's
here,
if
anybody
has
anything,
they
want
to
ask
about
the
male
archive
service,
otherwise
you
can
take
the
report
he
has
in
there
as
red,
which
would
bring
us
to
any
other
business.
B
Once
again,
thanks
everyone
for
their
time,
the
the
conversations
on
these
calls
are
very
important
for
the
tools
team
as
we're
moving
forward
and
I
really
appreciate
the
level
of
Engagement
that
we
get
from
from
our
community
members
and
making
this
making
this
all
go
in
the
right
direction.