►
From YouTube: Memory Safety Sig (April 13, 2023)
B
B
And
I
would
love
to
get
your
feedback
on
kind
of
talking
about
this
group's
efforts.
Sure
maybe
just
kind
of
give
a
little
couple
sentence
summary
and
then
kind
of
talk
about
how
you've
been
organizing
and
how
we're
moving
forward
to
get
the
updates
together.
Sure.
B
I
had
to
go
bug,
bressers
and
Sarah
for
s-bombs
and
then
find
out
who's
working
on
the
risk.
Dashboard.
B
Yay
yay
more
more
stuff
to
write
like
I'm
waiting
on
this
David
wheeler
jerk.
To
help
me
with
our
presentation,
yeah.
A
F
A
All
right,
I'm
gonna,
give
people
dude
who
got
a
thumbs
down.
We're
gonna
give
people
just
a
little
longer.
B
A
We've
got
it
right
here:
I
will
share
it
with
you.
Oh
not
that
one.
Let's
share
this
link
cool.
It
wasn't
anything
terrible.
Just
not
knowing
come
on.
You
wanted.
Okay,
yeah
yeah,
all
right
all
right
and
I'd
love.
If
someone
could
maybe
fix
the
permissions
on
the
dock,
so
it's
not
showing
up
as
suggestions
every
time.
I
add
something
or
other
people
add
stuff.
D
A
I
mean
it's
not
terrible.
Well,.
D
D
From
back
from
bitter
experience,
click
on
reload
of
the
of
the
tab,
oh
yeah,
it's
a
good
idea.
Your
access
I
mean
it's
a
quirk
of
Google
docs,
but,
and
that's
one
time
now
once
you've
done
that
it
will.
You
are
now
gotcha.
D
D
B
Used
to
have
I've
had
this
nickname
forever,
so
I
used
to
be
able
to
tell
where
people
knew
me
from
by
how
they
pronounced
it.
So
I
had
a
boss
that
mispronounced
and
called
me
c-rob
everywhere.
So
she
every
time
I
knew
somebody
I
just
find
I
knew
them
through
Mary
Rose
awesome.
A
All
right:
well,
let's
go
ahead
and
get
started
and
it
looks
like
everyone
is
returning,
so
I
don't
think
we
need
to
do
introductions
in
this
case,
but
hello,
everyone.
It's
wonderful
to
see
you
thank
you.
Thank
you
for
taking
some
time
out
of
your
morning
afternoon
or
evening,
depending
on
where
you
are
in
the
world
and
I
see.
People
have
filled
out
the
attendees
on
the
agenda.
A
Thank
you
and
let's
go
ahead
and
start
off
with
Abby
Shay
very,
very
helpfully
created
a
pull
request
to
our
shiny,
new
GitHub
repo,
which
is
adding
in
some
doc
some
skeleton
things.
Thank
you
to
Jay
for
her.
In
our
first
few
meetings,
getting
us
to
form
our
vision,
goal,
scope,
Etc,
I.
Imagine
that's
making
this
much
easier,
so
yeah
go
ahead.
E
Yeah
thanks
so
her
request.
Is
there
ready
for
review?
One
thing
you
will
notice
is
that
I
kept
a
the
correlate
empty
for
now
I
thought
we
can
discuss
it
here
with
the
community
to
decide
who
that
will
be.
D
That
said,
I
think
that's
important
to
fill
in
ASAP,
because
having
a
lead
to
talk
to
get
things,
moving
is
really
really
important.
Those
leads
can
are
not
trapped
forever.
They
can
leave,
they.
A
A
A
Does
anyone
else
want
to
be
considered
for
co-lead,
okay,.
D
G
D
Type
it
in
right
now,
that'd
be
great
and
same
for
you,
obviously,
because,
basically
that's
how
we're
gonna
handle
the
GitHub
permissions,
we
would
like
to
make
sure
we
get
the
right
ones,
cool.
A
Thanks
all
right,
cool
and
I'll
give
people
a
moment
to
review
that
pull
request,
I'm
just
wondering
if
there's
anything
we
should
discuss
now.
A
D
If
you
know,
since
that's
our
group
decision,
there
I
think
I
need
to
at
least
notify
operations,
because.
D
D
Okay,
okay,
very
good,
I
I.
Think
I
should
give
a
heads
up
just,
but
that's
great
that
sounds
awesome.
A
Cool
anything
to
discuss
on
the
GitHub
repo.
A
All
right
moving
on.
Thank
you.
Everyone
who
added
comments
Etc
to
the
current
language
of
stream,
four
I
just
put
a
link
to
the
Google
Doc
in
the
chat
here
here.
That
was
very,
very
helpful
and,
let's,
let's
talk
I,
think
through
some
of
the
comments
in
some
of
the
suggestions.
A
So
first
I
see
one
from
gab
Gabby
regarding
needing
to
acknowledge
that
these
languages
are
under
constant
evolution.
I
I
am
assuming
that's
referring
to
C
and
C,
plus
plus.
C
Memory
on
safe
but
I'm,
assuming
it
is
safe
too.
C
Yeah
I
guess
my
you
know.
The
point
I
was
trying
to
make
here
is
that
so
we
usually
build
software
by
picking
up
a
a
version,
no
and
then
riding
with
guys
and
then
a
couple
of
years
later
they
can
later
learn
a
bunch
of
stuff.
A
Moving
forward
I
do
think
this
is
a
good
thing.
Gabby's
suggested
first
Define
memory,
safe
languages
in
a
standalone
sentence.
Follow
that
up
with
examples
and
counter
examples.
So,
let's,
let's
discuss
the
definition
of
a
memory
safe
language
I,
personally
often
think
of
them
as
memory
safe
by
default
or
or
unsafe
by
default.
I
think
I
think
there's
two
two
category
two
meta
categories
in
there:
how
would
you?
How
would
the
group,
how
would
some
members
of
the
group
Define
a
memory
safe
language.
C
Another
precise
definition,
but
I
do
like
the
the
distinction
by
default
right.
E
C
A
B
D
A
memory
safe
language,
it
has
the
following
properties
and
I'm
assuming
somewhere,
there's
a
there's,
a
somebody's
worked
this
out
and
has
a
definition
and
I
would
always
be
careful
to
say
memory
safe,
not
just
safe,
I,.
D
But,
for
example,
I
mean
frankly
the
list
of
memory
unsafe
languages,
I,
would
say
most
languages,
are
memory,
safe
examples
of
memory
and
say
unsafe
language
of
languages
that
aren't
not
memory
safe.
You
don't
want
to
say
memory.
Unsafe
I
mean
it's
basically
C,
C
plus
plus
fourth
assembler
and
I,
think
is
it
Vala
the
the
list
is
short
and
you
get
increasingly
obscure,
Vala
and
Objective
C.
Oh.
D
F
A
And
while
we're
at
it,
I'm
just
realizing
I
did
not
ask
for
volunteer
to
take
notes,
though
that
might
be
copying
and
pasting
the
links
that
are
in
chat
and
the
definitions
that
people
are
putting
in
chat
to
preserve
it.
C
Yeah,
so
is
it
here,
I'm,
not
yeah
I,
actually
like
the
delay
but
I'm
not
seeing
that
important
distinction
that
they
made
earlier
it
is
default
versus
non
non-default.
I
think
we
have
to
find
a
way
to
factor
that
in.
A
F
Just
having
the
damnedest
time
copy
and
paste
and
Link
or
text,
oh.
A
D
D
A
No
worries,
I,
think
I,
hadn't
refreshed
or
the
page
had
it
refresh
for
me.
Yet
the
doc
and
so
I
hadn't
seen
the
the
things
you've
been
adding
to
it,
but
I
see
them
now.
So,
okay.
D
More
secure
is
a
little
more
dicey
because
it's
but
but
I
understand
the
goal.
I
mean
I
would
copy
at
least
the
first
part,
although
it's
very
vague,
you
know.
E
A
Yeah
I
think
I
think
that
would
be
useful.
That's
something
I
hear
a
lot
within
Microsoft
and
outside
of
Microsoft
is.
How
can
you
say
rust
is
memory
safe
when
you
can
do
unsafe
blocks,
which
larger
discussion
we're
not
going
to
have
here,
but
yeah
I
think
it
would
be
helpful
to
have
that
in
part,
because
I
know
people
who
read
the
stock
regardless
of
what
language
or
what
language
background
they're
from
I
don't
want
them
to
be
immediately
turned
off
by
the
definition
that
we
use.
F
It's
the
problem
is
on
my
side.
Does
anyone
else
have
access?
Well,
it's
it's
spectrum.I
triple
e.org
memory,
safe
programming,
language
programming,
languages
and
there's
an
article
that
comes
up
and
they'll
have
what
they
have
a
definition
for
there.
It's
on
my
side,
I'm,
not
able
to
copy
paste
control,
c
yeah.
F
E
D
F
D
F
A
You
need
a
reboot
okay
cool.
Well,
it
looks
like
that.
Abhisha
put
that
in
the
Google
Doc
the
IEEE
definition.
Thank
you
very
much.
Abby
Shay,
so
yeah
memory,
safety.
F
D
B
Just
a
brief
tangent
quickly,
you
mentioned
that
we
don't
want
readers
of
the
document
to
become
inadvertently
offended.
B
David
and
I
had
the
opportunity
to
brief
sisa
on
the
foundation
activities
yesterday,
and
they
were
super
interested
in
this
little
team.
Oh.
A
A
So
what
do
people
think
of
the
IEEE
definition
I'll
paste
that
into
Zoom
chat
as
well.
D
B
D
Yeah
and-
and
there
are
edge
cases
too-
I
mean
I
I
mean
obviously
there's
the
you
declared
as
as
unsafe,
as
one
I
actually
have
done.
A
lot
of
work
with
Ada
Ada's
free
actually
doesn't
guarantee
safety
either.
It's
it's
explicitly
marked
as
unsafe,
but
you're
less
likely
to
use
it.
So
it's
more
like
the
unsafe
marker
in
Rust,
so
correct,
yeah.
C
A
Question
for
you,
Gabby,
with
regard
to
C,
plus
plus
so
I,
know,
there's
a
specification
that
is
allows
use
of
C
or
when
you
use
that
use
of
C
plus
plus
is
more
memory
safe.
At
the
very
least,
could
you
tell
us
just
a
little
bit
about
that?
Oh.
C
Yeah,
so
you
yeah,
so
you
have
the
full
standard,
and
then
you
have
the
C
plus
plus
curl
guidelines,
which
is
a
a
set
of
rules
that
talk
to
a
smaller
subset
of
the
language,
assisted
with
tools
to
enforce
these
rules,
and
when
you
use
those,
then
you
actually
reduce
significantly
the
the
instances
of
memory
on
safety.
So
there
is
no
new
Malak
or
sorry.
C
New
delete
exist
memory
management
you
have
to
use
helper
type
like
spark
pointers,
unique
pointers,
shared
pointers
that
have
well-defined
protocols
and
you
have
to
use
the
standard
guideline,
support
Library,
which
guarant
against
out
of
bound
memory.
You
don't
have
you
just
have
to
free
those
sort
of
stuff
so
and
and
that
so
do
you
have
the
role
the
implementation
is.
You
know
various
you
have
some
in
the
visual
C
plus
plus
product.
You
have
some
in
clunk
tidy,
but
what
you
do
is
they
do
respect
you
to
to
six.
C
D
C
So
the
these
are
not
so
usually
when
people
think
of
subset,
you
know,
including
myself,
is
you
have
the
Azure
Center
and
then
you
can
only
use
these
set
of
things
like
misra,
for
example.
Do
that
like
very
constrained
substrates?
What
the
core
guidelines
have
done
is
that
you
say
well,
this
Library
gsle
a
guideline
support
Library.
It
is
very
special.
It's
going
to
provide
you
with
some
basic
abstractions
that
you
need
to
use
and
that
library
has
the
exception
of
using
these
forbidden
constructs
right.
C
So
it
is
not
a
strict
subset
if
you
look
only
at
the
let's
say
compiler,
but
it
is
a
subset
when
you
look
at
it
from
programming
perspective
like
I'm,
just
a
ordinary
devs
I'm
not
allowed
to
use
new
delete,
I
can
only
use
unique
pointer,
make
sure
or
make
unique,
but
as
a
library
implementer
of
the
GSL
I
have
full
access
to
all
those
facilities.
Then
now
comes
in
the
tool,
the
checkers
that
comes
like
your
GSL.
C
A
Cool
all
right,
I
lost
my
tab
there.
It
is
so
it
sounds
like
we
can.
You
know,
maybe
asynchronously
do
some
words
missing
on
having
that
that
definition
of
what
we're
talking
about
when
we
talk
about
a
memory
safe
language,
yeah.
C
Yeah
yeah
I
would
definitely
work
on
some
things.
The
the
comment.
C
So
in
terms
of
logistics,
do
I
set
up
a
request
or
do
I
use
the
mailing
list
or
the
chat.
You
know
just
I
I
like
to
have
some
draft
and
early
feedback,
but
not
a
random
June
internet
interfering
with
I.
A
Understand
yeah
we're
not
going
to
ask
you
to
put
it
on
Twitter,
so
I
mean
in
the
slack
Channel
works
well
for
me,
but
I'm
open
to
other
other
opinions
as
well.
Go
go
ahead!.
B
G
B
D
F
E
A
Right
moving,
oh.
D
Go
ahead
so
gdr
you're
the
one
who
agreed
to
do
that.
So
thank
you.
A
All
right
so
I
see
art,
Manion
added,
an
additional
link
regarding
did
you
what
science
can
tell
us
about
C
and
C,
plus
plus
security?
Oh
okay,
I
have
not
read
this
yet,
but
I'll
definitely
give
this
a
read.
E
The
parts
that
was
also
mentioned
in
the
slack
channel
that
the
stats
were
relevant
only
to
software
built
with
unsafe
code
right.
So
it's
not
70
right
of
all
vulnerabilities,
but
it's
70
of
vulnerabilities
in
memory
and
safe
languages
in
the
same
edit
was
done
for
Android
and
later
for
Google
project.
Zero.
D
A
All
right
and
then
going
down
the
comments,
there's
a
comment
for
me
on
the
under
proposed
approach.
We
know
how
to
entirely
rid
ourselves
a
memory
safety
vulnerabilities
just
that
that
language
is
too
strong
and
I
I
do
feel
that's
misleading
we
can
attend.
Maybe
we
can
rid
ourselves
of
known
memory.
Safety
vulnerabilities,
but
I
I
I'm,
wary
of
giving
people's
false
sense
of
security
that
this
will
be
eliminated
for
all
time.
C
Yeah
I
was
looking
at
the
the
chatter
and
like
what
we
really
want
to
do
is
make
sure
that
we
remove
or
significantly
reduce
memory
on
safety
issues
in
in
software
that
we
use
and
for
me,
I
think
that
would
be
a
a
goal
that
we
can
can
push
forward
and
invite
more.
You
know,
brain
brain
powers
to
help
us
achieve
that
goal
time.
If
we
just
went
in
from
link
programming,
language
advocacy
perspective
doesn't
mean
we
don't
talk
about.
C
We
don't
talk
about
COC,
plus
plus
just
this
means
that
the
goalie
really
is
to
reduce
memory.
Safety,
vulnerabilities.
A
D
I
I,
don't
think
that's
so
terrible
I
mean
people
often
refer
to
these
as
foot
guns
and
it's
the
old
joke
about.
When
you
shoot
you
know,
C
plus,
plus
you
shoot
your
whole
leg
off
so
yep.
A
Any
other
thoughts
on
this
this
section
or
this
particular
comment
in
the
doc.
A
All
right
moving
forward,
so
yep
I
added
in
for
large
Legacy
code
bases
not
able
to
move
away.
There
are
tools
and
specifications
which
improve
memory
safety,
even
if
it's
not
the
default
for
those
languages.
B
A
Hence
why
I
developer
best
practices
working
group
was
such
a
good
fit.
A
Yeah
I
I,
like
the
approach
of
you,
know
product
of
this
grouping,
something
along
the
lines
of.
Let
us
help
you
I.
Never,
like
the
approach
of
you
know
some
people,
some
loud
people
have
around
security
vulnerabilities
of
well.
If
you
have
these
vulnerabilities
you're
stupid,
and
why
don't
you
just
do
x
and
x
and
x?
No,
it's
it's!
We
all
are
facing
these
problems.
Let
us
help
you
get
to
a
better
place.
It
doesn't
have
to
be
this
painful.
A
All
right,
cool
and
all
right,
I
see
Gabby.
You
had
a
comment
on
the
move:
the
internet's
most
critical
software
away
from
unsafe
languages.
C
Yeah,
so
when
I
read
that
what
comes
to
mind,
firstly,
is
either
rewrite
or
replace
with
something
else,
and
you
know
it
it
become.
You
know,
cost
risk
analysis
like
move
to
what.
C
We
haven't
actually
tested
yet,
and
so
is
there
an
assumption
that
we
can't
actually
do
any
improvements
to
the
existing
code.
Like
let's
say
something
is
written
in
C
we
know
C
by
default
or
C
cos
plus
lets
you
write
mistakes,
things
that
cost
billions
of
dollars.
C
Are
we
making
an
assumption
that
there
is
no
way
we
can
deploy
any
realistic
technique,
so
those
tools
to
bring
them
to
a
place
where
we
don't
have
those
concerns
anymore?
That
could
very
well
be
the
case,
but
I
think
we
need
to
spell
those
assumptions
explicitly,
so
we
can
discuss
them
and
say
well,
given
the
evidence
here
is
our
conclusion.
This
is
why
we're
doing
this.
B
C
Oh
yeah,
like
definitely
that
would
be
you
know,
I'll
think
of
it
as
mitigation,
but
maybe
David
has
a
better
way
to
correct
for
Iceland.
D
Yeah
I
mean
I,
certainly
if
there's
help
by
all
means
interact.
I
I
do
think
that
the
if
you
think,
you're
thinking
of
the
CNC
plus
plus
compiler
options,
work
right.
Okay,
yeah,
yeah,
I
I.
In
my
mind,
I
mean
both
are
broadly
trying
to
improve
security.
D
I
I
do
think,
there's
a
difference.
Nobody
believes
that
the
compiler
options
are
going
to
suddenly
prevent
memory
safety
problems.
Oh
I,
know
Crow
you're,
you're,
so
shocked.
You
know,
but
I
I
view
these
as
both
reasonable
approaches.
You're,
we
are
not
going
to
rewrite
trillions
of
dollars
worth
of
CNC
plus
plus
today.
Tomorrow
is
not
looking
good
either,
so
we
are
going
to
have
to
find
ways
to
improve
where
there
is
existing
C.
G
D
I
don't
know
if
this
is
a
good
time,
but
I
actually
have
had
some
very
interesting
conversations
with
the
lineage
Colonel
folks
about
their
Integrations
with
rust.
I,
don't
know
if
this
is
a
good
time
for
that,
but
I.
D
E
D
All
seriousness,
I
do
welcome
you
it's
just.
It
was
funny
that
we
were
about
to
switch
topics
on
you
and
you
appeared
so
yeah.
So
we've
been
doing
several
kind
of
you
know
getting
started
things
with
this
with
this
sig
in
fact,
I'm
about
to
send
an
email
to
our
operations
team
and
we're
getting
the
the
repo
set
up.
We
want
to
Define
our
terms
and
why,
but
and
I
think
we're
we're
through
that.
D
So
I
mentioned
that
there
are
some
interesting
things
to
be
learned
about
memory
safety
from
the
Linux
kernel
work,
so
I,
don't
think
it'll.
Take
me
too
long,
but
I
think
everybody
here
knows
that
the
Linux
kernel
folks
are
taking
steps
to
integrate
rust
so
that
it
rust
can
be
used
for
writing
drivers
in
the
Linux
kernel.
Now
a
quick,
quick
expectation
set
right
now,
there's
no
expectation
of
rewriting
all
the
Linux
kernel.
It's
rewriting
the
drivers.
D
However,
drivers
are
huge
proportion
of
a
kernel
and
in
many
ways,
they're
the
most
problematic,
because
most
drivers
are
not
are
you
know,
a
driver
only
works
for
particular
Hardware
typically,
and
therefore
they
don't
get
the
level
of
review
because
most
other
people
don't
have
that
device.
They
can't
necessarily
help
so
sell
help.
So
focusing
on
the
drivers
seem,
like
the
you
know,
let's
focus
on
where
the
biggest
pain
points
are
all
right.
D
Some
interesting
lessons
that
have
been
learned
first
of
all
rust,
has
this
annoying
property
of
requiring
that
certain
things
be
true
to
make
easy
calls.
In
particular,
you
know
if
you,
if
you
provide
something,
that's
read
only
rust
is
much
happier
as
soon
as
you
have
to
read
and
write
it.
You
have
to
make
sure
it's
okay
to
read
and
write,
which
is
actually
always
true,
regardless
of
the
programming
language,
but
rust
enforces
this
at
compiled
time
what
this
right.
So
so
what
this
has
has
led
to.
It
is
an
interesting
thing.
D
D
Now,
from
a
c
perspective,
syntactically,
it
doesn't
matter,
but
unsurprisingly,
if,
when
a
a
c
function
and
they're,
always
if
it's
callable
and
C,
it's
called
a
function,
a
c
function
that
only
gets
stuff
that
can't
that
can
only
that
can
read
but
can't
write
that
does
eliminate
some
problems
right
away.
D
D
So
there's
been
a
very
painful
long-term
effort
to
change
how
structs
are
are
handled
to
eliminate
a
lot
of
memory
safety
issues,
there's
going
to
be
a
presentation
that
OSS
North
America
this
year,
well,
they'll
talk
and,
and
also
the
Linux
security
Summit,
specifically
at
the
Linux
security
Summit,
full
disclosure
I'm
on
the
program
committee,
but
but
we're
basically
and
I,
can
go
into
details
of
what
this
is,
but
basically
they're
changing
the
way
they
use.
Structs
struts
Crux,
I
gotta,
pronounce
it
correctly
to
eliminate
some
challenges.
D
The
good
news
is
it's
going
to
eliminate
some
memory
problems
and
the
bad
news
is
it's
a
lot
of
work
to
get
there
and
I
I
think
in
their
but
they're
I
think
in
many
ways,
they're
inspired
by
weird
need,
in
general,
to
cut
down
memory
safety
problems.
One
approach
is
rusts
in
some
portions
and
the
other
is
changing.
How
we
do
c
I.
C
I,
like
that
story
and
I
guess
that
goes
into
the
other
comment.
I
was
trying
to
make,
which
is
we.
We
get
best
practices
that
evolve
over
time
and
having
rust
go
into
that
ecosystem
to
go,
get
exposed
to
new
ideas
like
oh,
oh,
by
just
changing
how
we
clean
our
dishes.
We
can
avoid
all
these
other
diseases
itself.
You
know
it
it
it.
It
evolves
the
best
practices
right.
D
Of
course
I
mean
to
be
fair,
just
changing,
is
you
know
it's
a
there's,
a
whole
lot
of
work
hidden
behind.
D
C
D
Right
well
in
this
particular
case,
it
really
is
a
combination
of
tooling,
but
also
lots
of
of
careful
code
changes
enabling
compiler
options
doing
other
things,
but
it
you
know
I,
don't
want
to
steal
the
Thunder
too
much.
You
know
if
you
want
to
see
all
the
details
come
from
the
Linux
security
Summit,
but
and
we
will
record
those
for
others,
but
you
know
there
is
a.
There
is
a
good
story
on
both
sides,
because
everybody
agrees.
Member
safety
problems,
bad
and
you
know
a
great
way
to
do.
D
B
It
since
we're
so
close
to
the
end
of
the
dock,
I
looked
ahead.
We
may
want
to
consider,
as
part
of
this
proposal,
to
add
in
notes
for
like
evangelism,
education
and
awareness,
because
we
know
we're
not
going
to
get
everybody
to
convert
to
a
memory
safe
language.
We
know
not
everybody's
going
to
use
a
tool
but,
like
you
know,
as
Gabby
mentioned,
but
just
by
observing
how
this
others,
these
other
methods
work.
B
You
know
that
it's
informing
other
Behavior
so
again,
if
we
can
get
have
some
tasks
in
the
plan
to
say
we
want
to
devote
time
on
a
training
class
as
part
of
the
education
Sig
or
we
want
to
go
to
conferences,
developer
conferences
and
talk
about
this
topic,
just
to
kind
of
raise
awareness
and
evangelism
to
help
move
the
ball
forward.
I
think
that
will
be
a
useful
add
to
the
plan.
G
Yeah
I
just
want
to
point
out
trying
to
channel
the
days
when
I
wrote,
device
drivers
that
I
still
there's
still
there's
a
proliferation.
Typically,
what
they're
called
sdks
for
a
given
space
so
I
think
that
creating
best
practices
or
trying
to
address
that
in
terms
of
sdks
or
different
things,
like
scanners
or
or
whatever
disk
things
and
oftentimes.
You
can
like
pick
like
work
with
companies
like
Intel
or
others
who,
like
the
architectural
level
things
they
appear.
That's
the
case
for
different
things
that
fit
their
architectures.
A
Oh
oh
good,
discussion
and
good
examples
and,
let's
see
here's
just
a
couple
more
comments,
I
see
another
one
on
say
be
you
know
declaring
whether
I
mean
safe
or
memory
safe.
It's
kind,
it's
it's
misleading
to
say,
safe
languages
when
we're
talking
about
a
specific
subset
of
safety.
I
think
that's
reasonable
and
definitely
something
we
can
spell
out
and
okay
and
then
just
one
about
saying,
C
and
C,
plus
plus
not
C,
slash,
C,
plus
plus,
which
happy
to
respect
and
thank
you
for
adding
that
perspective.
G
E
Yeah
I
think
that
the
plan
like
as
it
is
written
now
only
responds
to
how
it
was
originally
written.
However,
I
see
that
both
Ned
and
I
added
two
more
potential
goals
right
because
to
have
two,
but
now
there
is
three
and
four
do
we
want
to
maybe
go
over
this
and
then
that
will
maybe
expand
to
what
Kroger
is
looking
for.
A
A
A
Yeah
I
hadn't,
even
because
I
felt,
like
the
the
costs
in
timeline
and
such
were,
were
very
off
now
because
we're
about
a
year
out
from
when
this
this
plan
was
originally
proposed,
but
yeah
I
think
that's
a
good
idea
to
discuss
the
the
goals
of
this
working
group.
So
is
there
anything
else
on
the
document
other
than
that
that
we
should
discuss
before
we
move
on
to
goals.
B
I
put
a
stylistic
comment
towards
the
top.
The
English
major
in
me
is
flummoxed
by
the
word
memory,
space
safety,
languages.
B
A
D
D
A
All
right
so
goals
of
this
working
group-
let's
start
discussing
you,
know
tangible
things.
Let's
start
discussing
that.
B
B
Rust
TLS
the
ffi
some
grants.
Oh
three.
E
Items
I
was
actually
referring
to
the
one
move:
the
internet,
most
critical
software
to
investment
in
tools,
reinvest
in
tools
that
improve
memory
safety
and
for
investing
educating
social.
E
I
was
referring
to
it
to
sisters
director
easterly's
Talk,
where
she
mentioned
that
a
large
part
of
the
of
how
to
influence
the
industry
is
by
educating
right
about
memory
safety
and
about
security
as
part
a
fundamental
part
of
software
engineering
as
a
general
effort.
That
could
be
a
thing
that
our
working
group
Sig
can
also
align
to
and
I
think
it
scales
well,
as
Gabriel
also
noted.
A
A
A
You're
still
muted
there
we.
B
Go
all
right:
we
should
give
an
action
item
to
the
education
Sig
to
create
or
augment
the
coursework
there's
we
have
a
secure,
secure
development
fundamentals
class.
Maybe
we
add
a
module
there
or
we
do
conference
talk
I,
think
we
can
partner
with
them
to
create
some
education
material
that
should
be
an
action
item.
It's
very
low
cost.
A
Any
other
goals
we
should
consider
adding.
We
will
definitely
be
continuing
this
conversation
asynchronously
between
now
and
the
next
meeting
in
two
weeks.
B
I
liked
what
Gabby
suggested
around
talking
about
like
techniques
and
tools
to
constrain
the
use
of
the
unsafe
languages
today
that
way,
you
know
it's
kind
of
a
a
mitigation,
an
alternate
control,
but
again,
if
somebody
can't
convert
or
won't
convert
at
least
gives
them
options.
I
think
that
should
be
on
something
we
should
work
on.
A
Would
that
fall
into
the
invest
in
tools
that
improve
possibly
okay,
cool.
B
E
E
D
I
I
don't
know
about
scorecard,
but
certainly
dashboard.
They
are
already
planning
to
capture
the
languages
and
warning
which
ones
are
memory
safe.
Our
memory
unsafe
languages.
D
So
you
know
obviously,
if
you're
looking
if
you're
writing
in
C
and
you're
trying
to
get
a
C
library,
that's
what
you're
expecting,
but
that's
not
always
what
you're
expecting.
C
Yeah
I
think
I
think
the
a
very
useful
thing
to
do.
Yeah.
A
D
F
The
next
dashboard
means
on
the
21st.
If
you
want,
you
can
add
an
agenda
item,
you
know
I'm
there
as
well.
We
have
a
oh
I,
think
it's
a
proturgia
or
my
I
know.
Miter
come
in
the
talk
with
the
dashboard
around
a
hip
check.
Did
that
coming
out
21st
and
then
we
can
also
add
memory
safety
to
come
and
talk
on
21st
as
well,
so
have
both
of
those
and
talk
provide
at
least
20
minutes
to
each.
We
could
do
that.
A
Okay,
I
can
talk
at
the
scorecard
meeting,
but
I'm
also
open
to
someone
else
doing
it
as
well.
F
I
I
run
the
dashboard
meeting.
That's
that's
my
I
run,
it
I
mean,
but
you,
but
from
a
technical
standpoint,
you
guys
are
are
more
technically
inclined
here
than
I
am
I
can
simply
say:
hey.
We
did
a
whole
bunch
of
work
and
here's
a
duck.
You
guys
can
actually
add
some
flavor
flavor
to
the
meeting
about
it
in
and
all
that.
Well
rob
you
know
what
Raul
is
here
is
too
roll.
Is
here
too.
A
Said
he
was
driving,
so
he
was
listening
but
open
to
open
so
are
either
or
Jay
or
Raul
relin
understand
you
might
not
be
able
to
answer.
Are
you
comfortable,
leading
the
discussion
in
the
meeting
on
the
21st
I'm?
Not
even
gonna.
Try
at
the
moment
keep
track,
of
which
one
is
which,
but
on
the
meeting
in
the
on
the
21st.
A
Yeah
and
then,
if
we
can,
maybe
in
the
slack
Channel
or
something
get
a
report
back
on
or
a
link
to
the
the
meeting
minutes,
that
would
be
quite
helpful.
Matt
you
just
unmuted,
so
I'm
wondering
if
you
well
that's
something
to.
G
Say
I
have
people
involved
in
scorecard,
oh
cool,
well
and
I
met
with
the
team
when
I
talked
about
through
the
design
they
were
doing
for
for
probes
that
are
coming
up
in
scorecard.
So
my
concern
is
to
my
team:
is
that
we're
adding
greater
Grand?
So
we
have
checks
already,
but
they're
going
to
be
subdivided
into
lists
of
probes
and
how
do
we
get
better
dashboard
readouts?
So
my
concern
is:
is
that
you
go
to
a
metrics
dashboard
and
basically,
oh
you
see
you're
red.
So
how
do
we
get
better?
Granularity?
G
Is
my
concern
and
my
admonition
to
them
in
the
metrics
dashboard
is
say
to
have
some
type
of
thing
on
the
dashboard
that
says
you're
running
this
many
probes
and
you
pass
10
of
12
probes
type
of
things
you
can
drill
into
it.
So
that's
why
I'm
trying
to
get
my
team
to
volunteer
development
for
as
I
find
time
but
I
think
we
need
it.
We
need
to
work
towards
guidance
on
like
Beyond
saying.
Oh,
you
see,
you
fail
right.
A
B
This
very
sticky
button
today
are
we
aware,
or
would
it
be
a
task
to
potentially
identify
resources
to
convert
developers
to
these
memory
safe
languages?
So
are
we
aware
of
resources
for
coding
in
rust
or
Java?
Are
these
other
safer
languages,
and
maybe
that
would
be
a
step
for
us,
is
to
find
those
resources
and
highlight
them
to
the
community
as
well
as
another
alternative?
There's.
A
A
lot
of
material
on
learning
rust
through
there's,
not
a
ton
on
converting
from
one
language,
one
language
to
rust,
there's
some
stuff,
there's
more
stuff
coming
out
with
interfacing
between
the
two,
which
I
think
is
one
of
the
key
parts
foreign.